Computing in Civil Engineering 2021

The state-of-the-art in construction material classification has used various machine learning approaches using 2D color image data sets. However, construction materials are often discolored due to other foreign substances and lighting conditions, making it challenging to classify construction materials only with color information. To address this problem, this study presents a new material classification method employing laser intensity values as another parameter. A hierarchical 3D deep learning approach is adopted for the material classification in laser scanned-3D point cloud data. The deep learning model was trained with a self-developed point cloud data set, which includes laser intensity value as well as 3D coordinates and color codes. This study conducted a case study at an actual building construction site to validate the proposed classification method. As a result, when additionally using intensity values as an input feature for network training, the accuracy was from 5% to 14% higher than when the intensity was not used. Therefore, we identify that laser intensity can improve material classification results.

Predicting building occupancy is crucial to design and operation of energy systems in buildings. The common method to model building occupancy is using static occupancy schedules, recommended by energy standards. Data-driven approaches are used to provide a more dynamic and accurate prediction of occupant interactions with buildings. This study built upon previously monitored data for 44 full days in a public educational building, collected via sensors on room-level occupant counts combined with indoor environmental quality data. We developed an automated machine learning framework with pipeline for pre-processing, model selection, and prediction result analysis. The framework evaluates various machine learning algorithms, including K-nearest neighbors (KNN), support vector machine (SVM), classification and regression trees (CART), and ensemble methods. Different evaluation metrics were adopted to quantify the percentage of errors. The results of these occupancy predictions were compared with the conventional static schedules in the ASHRAE Standard 90.1 through the application of dynamic energy simulation. The output of this study informs the construction sector about the advances of data analytics for occupancy schedule prediction, facility operation, and building simulation.

Over 97% of the US airports are not equipped with control towers, which limits their ability to keep records of flight operations. This problem can be alleviated with implementation of an automated video-based air traffic surveillance system capable of detecting light aircraft operations, which comprise the vast majority of operations at non-towered airports. To this end, we propose a rapid automatic light-aircraft detection method based on Haar cascade classifiers. The video data are collected from three public-use Utah airports and used for detecting both arrival and departure operations of light aircraft. The proposed model is compared with two state-of-the-art deep neural networks—SSD and YOLO—which typically achieve high accuracy at the cost of decreased speed. The Haar cascade provides competitive accuracy at a much-reduced processing time, which makes the proposed method suitable for real-time applications.

As one of the main sources of information in construction projects, analyzing and exploiting text data emerged as a fast-growing body of literature in the construction domain recently. Various applications appear in a wide range of domains, specifically exploration of different natural language processing (NLP) techniques. Usage ranges from construction contracts, design requirements, risk registers, change orders, claims and litigation documents, and safety reports. In this paper, the authors present a systematic review of the current NLP body of knowledge in the construction research domain. Various machine learning and deep learning-based NLP techniques, as well as their applications in construction research, are documented. Further, the authors introduce potential knowledge gaps and future research directions. In particular, this paper compares the performance of these NLP techniques through a risk classification problem. Analysis reports that Bidirectional Encoder Representations from Transformers (BERT) model outperforms other NLP models and achieves 80% of accuracy in the risk classification task.

With the emergence of automation and robotics, construction robots have been increasingly introduced to construction projects to relieve human workers from physically demanding and hazardous tasks. The co-existence of and interaction between workers and robots on the unstructured and dynamic sites will pose new safety challenges due to the potential risk of human-robot collision. This study proposes a new method to model the risk of collision based on predicted trajectories and associated uncertainties of construction workers. First, the movements of construction workers are predicted using uncertainty-aware long short-term memory (LSTM) network. Second, at any given time, the probability of human-robot collision on any location of the site is computed considering the distributions of predicted trajectories for all workers, resulting in a probabilistic representation of the collision risk on the dynamic site. Construction videos are used to demonstrate the proposed framework, which achieves 9.3 pixels of average displacement error in trajectory prediction. The result also suggests that the proposed method can effectively capture the collision risk at any given location over a time period.

Despite the proven benefits of data sharing between different stakeholders participating in civil engineering projects, a systematic classification of data sharing barriers and strategies for promoting data sharing in the civil engineering domain is missing. The Data Sensing and Analysis Committee of the American Society of Civil Engineers (ASCE DSA Committee) initiated an investigation to explore the current practices, barriers, and future vision in civil engineering data sharing. The approach is to conduct a literature review and interviews with domain experts to identify and classify data sharing barriers and strategies for promoting data sharing in the civil engineering domain. An online survey then follows, asking the participants having different job functions in various organizations (e.g., for-profit corporations, academic institutions) to prioritize five categories of barriers and three categories of promotion strategies summarized by the authors. The survey results show that: (1) most responders regard data sharing policy and data standard regulation as urgent needs; (2) different stakeholders prefer different promotion strategies due to they face different aspects of the data-sharing problem encountered; and (3) non-academic stakeholders hardly perceive the value of incentives for data-sharing research. The synthesis of these findings can guide the systematic design of three road maps that help advance healthy data-sharing systems in civil engineering: (1) policy road map, (2) technology road map, and (3) business road map.

Digital methods such as building information modeling (BIM) offer significant potential in the operation phase of a building. A prerequisite for these methods is the availability of digital as-built models of the existing buildings that are not given for many of them. This makes a retro-digitization, i.e., the generation of a digital representation of an already existing building, necessary. Thereby, computer vision (CV) and machine learning (ML) are essential technologies to extract the information from the data sources of a building. An important data source is 2D construction plans from which the initially planned geometry of existing buildings can be extracted. The extraction requires a highly reliable detection of lines and texts. However, the performance of the CV and ML methods is highly dependent on the quality of the data. Anomalies like discolorations, stains, and fold lines on 2D plans can negatively affect the detection. This is especially the case for old hand-drawn paper plans that are often the only data source available for old buildings. To integrate old 2D plans better in the detection process by CV and ML, their quality recovery is necessary. To achieve this, we propose using the cycle-consistent generative adversarial network (CycleGAN) that enables style transformation with unpaired data. Hereby transformation means the removal of the stated anomalies. Our results show that both text and edge detection methods perform better on improved 2D plans than on the original 2D plans with anomalies.

Current studies combining deep learning techniques and computer vision for the detection of personal protective equipment (PPE) are mainly focused on the detection of hard hats and vests for construction safety management. This study implemented two deep learning-based solutions using You-Only-Look-Once (YOLO) and convolutional neural network (CNN) architecture to detect safety gloves to expand the construction safety applications. In the first case, the customized YOLO-v3 algorithm was applied to directly detect whether a person is wearing gloves or not. In the second case, the customized YOLO-v3 was used to first detect a person’s hands from the images, and then CNN architecture was used to classify it as either “wearing gloves” or “not wearing gloves.” A better performance was found in the second case where the customized YOLO-v3 achieved 89.46% mAP (mean average precision), and the CNN network (VGG-19 and ResNet-50) classified “wearing gloves” or “not wearing gloves” with an average of 100% precision. On the other hand, the first case achieved 78.48% mAP in detecting gloves. This study enhances the implementations of computer vision for PPE detection, and the labeled data sets can be used by future research in this area.

A worker repeatedly exposed to slip, trip, and fall (STF) hazards is at an increased risk of an STF event. To monitor an individual’s exposure to STF hazards, previous approaches have assessed the loss of body balance (LOB). However, these approaches were not fully validated in field settings, where dynamic workplace environments can impact workers’ bodily movements. This study aims to develop and evaluate a new approach to assess workers’ exposures to STF hazards by LOB assessment in a real construction site. A waist-worn inertial measurement unit sensor was used to extract features of waist movements, which were modified into a single value to measure LOB using the Mahalanobis distance (MD) metric. The MD value was calculated for each step a worker took, and exposure to the STF hazards was detected when the MD value was larger than a predetermined threshold. The results provided an average of an unweighted average recall of 89.13% (without exposures: 90.30%, and with exposures: 87.96%) that would strongly indicate generalizability and robustness for practical applications.

Ground penetrating radar (GPR) is a non-destructive technique that has been used to evaluate the quality of concrete bridge deck. However, the lack of automated GPR data processing and interpretation methods hinders the use of GPR in large-scale bridge deck assessment. The objective of this research is to automate bridge deck condition assessment from GPR data based on image processing methods. An automated workflow is proposed to generate bridge deck deterioration map using GPR scans, which consists of three steps. First, the Random Forest classification model is trained to detect rebar regions in GPR scans using Histogram of Oriented Gradients (HOG) features. Second, robust hyperbolic fitting is used to fit each hyperbolic signature in the rebar region, and thus localizing the rebar by finding the peak of the fitting hyperbola. Third, detected rebars and their locations and depth-corrected amplitudes are used to create deterioration map. Field experiments were conducted on two bridge decks, and the results indicate the efficiency of the developed method. The proposed method achieves 98.6% accuracy on real GPR scans from the two bridge decks, which consists of thousands of individual rebar features. The developed workflow will help automate the processing of GPR scans and thus expedite the bridge deck inspection and evaluation.

The advent of smart monitoring and cyber-physical systems in civil engineering has significantly propelled a broad wealth of smart city applications. Representing a vital technological basis of smart city applications, Internet of Things (IoT) frameworks have been an increasing topic of research in recent years. However, different definitions of the terms “smart city” and “IoT framework” are used without consensus, and the technical aspects of IoT frameworks for smart city applications have not been fully reviewed, causing ambiguities and redundant developments in the community. This paper aims to condense the definitions of the terms “smart city” and “IoT framework” by summarizing and comparing concepts. Besides critical IoT technologies, sensor node hardware utilized in IoT frameworks is summarized in a systematic review. As a result of this study, IoT framework trends for smart city applications are provided. It is expected that the findings of trends of IoT frameworks for smart city applications presented in this study may serve as a basis for future IoT framework implementations that advance smart city applications.

Faulty operation of heating, ventilation, and air conditioning (HVAC) systems in a building can lead to thermal discomfort, wasted energy, and shorter equipment life. Early fault diagnosis in HVAC systems is critical to maintaining indoor environmental comfort, saving energy, and preventing further deterioration of the system. Data-driven predictive models have emerged as a popular approach to fault diagnosis of HVAC systems. For an HVAC system, failures are a small probability event. Therefore, the prediction accuracy of these data-driven approaches is affected by imbalanced data sets. The normal data far outweigh the fault data. Predictive models trained with these imbalanced data sets will perform poorly in fault diagnosis of an HVAC system. To address this limitation, this study aims to examine the potential of the synthetic minority oversampling technique (SMOTE) for sampling fault data points to improve the accuracy of the fault predictive model. To that end, six months of operational data (for example, humidity and temperature) were collected from embedded sensors in the HVAC system. A SMOTE algorithm was applied to increase the ratio of failure to normal data from 0.02 to 0.3. Both the original and improved data sets were used to train the fault predictive model based on different supervised learning algorithms. Results indicated that data sets improved through the SMOTE algorithm increased the accuracy of the predictive model by about 20%. This improvement can lay the groundwork for increasingly proactive facility maintenance.

Finding effective countermeasures of work zone crashes requires identifying the causes of work zone crashes. Recently, several researchers have identified some evidence that the influence of factors determining crash severity change over time (year to year), which is referred as temporal instability. So far, this phenomenon has been identified for factors determining crash severity and not studied for occurrence of work zone crashes. This paper focuses on investigating possible temporal instability in factors causing work zone crash occurrences. Three research gaps are identified: (1) Current models focus on statistical associations rather than causal relations. But causal relations are required for implementing work zone countermeasures; (2) Crash records cannot cover all variables defining crash risk such as human behavior, which induces unobserved observation-specific variations (unobserved heterogeneity) on safety impacts of observed variables; and (3) Current studies usually use work zone crash records in low spatio-temporal granularity, which cannot capture effects of fast-changing factors (weather conditions, traffic speed) on crash risk. In this paper, the fast causal inference (FCI) model is applied to data including work zone crashes and environmental conditions (weather conditions, traffic speed, and roadway characteristics) in high granularity to identify temporal instability in factors causing work zone crash occurrence. The proposed method is tested on the Pennsylvania work zone crash data from 2015 to 2017. Among 45 pairs of factors, the proposed model identified four pairs of factors whose causal relations changed across 2015 and 2016, and 10 pairs of factors whose causal relations changed across 2016 and 2017.

HVAC systems are crucial for proper ventilation and air circulation in a building. Current HVAC maintenance methods focus on visualization tools or process monitoring techniques. These conventional techniques are triggered at the onset of system failures, leading to exorbitant maintenance costs. While researchers have utilized the operational data for predictive maintenance of HVAC systems by detecting early failures in HVAC systems, these studies have relied mainly on the data collected from advanced embedded sensors. Moreover, the data collection is mainly based on a lab experiment in a controlled environment. Towards this end, the study aims to assess the feasibility of operational data (e.g., zone humidity, temperature, and airflow) collected from common embedded sensors in the HVAC system for early detection of failures. In this regard, three months of the operational data were recorded from an HVAC system in the Human and Health Development Building at Penn State University. To assess the potential of operational data in detecting failures of HVAC systems, measurable metrics in time-domain and frequency-domain such as mean frequency and kurtosis were calculated both from failure and non-failure data. Pearson product-moment correlation (PPMC) test showed a strong correlation (correlation coefficient greater than 0.6) between the calculated metrics and the working condition of the system. Results demonstrated the feasibility of applying operational data from the embedded sensors to early detect failures in HVAC systems. This potential of the embedded sensors will set the stage for assistive mechanisms whereby the HVAC operators can prognose the faults before they emerge.

Digital forms of citizen communication with response organizations through social media are becoming more widespread during disasters. Public agencies often use this information to examine community discussions to assess, determine, and prioritize critical areas in need of assistance. However, limitations on harnessing high volumes of precise geolocation information from social media restricts their ability to locate and promptly delineate actionable insights. Here, we propose a semi-supervised machine learning framework that integrates natural language processing (NLP) and spatiotemporal analytics to augment data from social media (Twitter) with a community-driven application (Waze) to achieve further evidence on location and type of emergency events. The framework is illustrated through a case study on Tropical Storm Zeta. This fusion provides increased context and may enhance the speed of emergency response. This study establishes a foundation for real-time crisis event detection, expanding organizations’ response capacity in allocating resources and reducing harmful effects of disasters.

Numerous studies predict energy consumption using building technology or occupant behavior data, but few focus on occupant behavior through a lens of energy consumption. This study aims to predict occupant behavior using energy use data. We collected actual household energy consumption data for one year, and the sensor-measured data were used to predict the appliances in use that represented associated occupant behaviors. The features used for the prediction included frequency, timestamp, cooling degree days, heating degree days, power, day of the week, and appliance based on the occupant behavior prediction model. In data analytics, we used classification and clustering with machine learning algorithms. The results showed that the features derived from the model to predict appliances could reach 96% accuracy using the decision tree algorithm. The finding implies that daily appliance usage and associated occupant activities in a household are well patterned and predictable using our model. Additional clustering analysis provided further energy consumption characteristics of the household. Also, the minute-level daily energy consumption can be used for detailed energy strategies for households having similar conditions with this test bed.

Buildings account for 40% of the energy consumption and more than 50% of the greenhouse gas (GHG) emissions worldwide. Various building energy efficiency policies have been issued in many cities around the US, requiring the disclosure of buildings’ energy performance and encourage building owners to take this information into consideration for retrofit investment decisions. However, various types of building information at different scales are meant to be collected and disclosed with costs that have substantial variations accordingly. Hence, complying with these policies is expensive and time-consuming for both government agencies and building owners. This paper proposes a non-intrusive and scalable data-driven approach to automatically capture energy-critical building characteristics to declare energy benchmarking or retrofitting policies. This approach first integrated open city data sets (building-related) and identified principal building variables in relation to energy performance indexes (i.e., electricity usage), then utilized transfer learning to retrain deep learning models on building façade images (retrieved from Google Street View) to extract their corresponding principal building variables. To elaborate, given a fully captured façade image of any building, this approach provides the data values for the identified principal variables (e.g., for a façade image given as an input, the outputs will include buildingType = Office). The best accuracy of the extraction model exceeds 80% in extracting building types from images, which indicates that the model is capable of extracting data values for the principal building variables that are essential for energy performance of buildings.

A population growth of around 1,000 people per day is the main factor contributing to the increase in housing demands in Florida. Optimization of the energy performance of residential buildings is crucial for the reduction of greenhouse gas emissions and fossil fuel consumption. This optimization entails the designers to accurately predict the energy consumption of buildings right from the design stage. Traditional energy modeling techniques require a lot of expertise and tend to be time-consuming and lacking in terms of energy predictions. Artificial intelligence (AI) techniques have been recently used to predict the energy usage of buildings and to be an alternative solution to such engineering methods. This study developed an artificial neural network (ANN) model based on a large data set of more than 18,000 newly constructed single-family houses in Florida between the years 2009 and 2019 to predict the heating and cooling loads on single-family houses. An ANN model based on this data set is developed to predict the energy usage of detached residences. The ANN will help designers to make the right decisions during the conceptual design stage of residential projects, and to explore the different design options using generative design for energy optimization purposes.

Support structures of relocatable large-scale outdoor conveyor systems are prone to unintended movements over time due to vibration induced by their moving parts, as well as the usually non-solid foundation ground. These movements can lead to a misalignment of the system’s individual segments, which in turn can cause an increase in power consumption as well as a potential belt ripping and therefore a breakdown of the system. This article proposes a novel approach to autonomously monitor the alignment of large-scale conveyor systems utilizing aerial images captured via unmanned aerial vehicles, semantic image segmentation via deep learning, 2D-to-3D label transfer using photogrammetry, and 3D point cloud analysis. The feasibility of the method is demonstrated on a working prototype. One of the main challenges is to find a good balance between precision and reliability on the one hand and computational costs on the other.

Climate disasters cause significant loss of life and damage to properties and infrastructure. In a disaster aftermath, building inspectors conduct preliminary damage assessment (PDA), which, depending on the size and population of the affected areas, may be slow and resource-intensive. In this paper, we present an uncertainty-aware convolutional neural network (CNN) for PDA by incorporating a Bayesian framework and implementing Monte Carlo (MC) dropout sampling to efficiently approximate the Bayesian inferences. Using dropouts not only during training but also at the inference time allows the CNN model to represent softmax probabilities of the damage prediction as random variables, enabling a more rigorous and systematic characterization of uncertainty and confidence in model predictions. The model is trained and evaluated with non-deterministic ground-truth damage scores obtained through crowdsourcing from drone footage of a hurricane-affected area. Results indicate that the trained uncertainty-aware CNN model leads to more informed and reliable PDA outcomes.

Automated safety compliance checking aims to detect field violations to construction safety regulations. Recent research and system development efforts have made good progress on automated tracking of labor and equipment towards improved violation detection and safety compliance. However, extracting and modeling safety requirements for supporting automated violation detection or safety alert systems remains highly manual. Towards addressing this gap, information extraction provides an opportunity to automatically extract safety requirements from regulatory documents for comparisons with field information to detect violations. However, existing information extraction methods fall short in their scalability and/or accuracy. To address this need, this paper proposes a deep learning-based information extraction method for extracting entities that describe fall protection requirements from construction safety regulations for supporting automated field compliance checking. The proposed method uses a hybrid bidirectional long short-term memory (BiLSTM) and convolutional neural network (CNN) model for recognizing the entities. The proposed method was implemented and tested on four selected Occupational Safety and Health Administration (OSHA) sections related to fall protection. It has achieved an average precision, recall, and F-1 measure of 81.5%, 80.3%, and 80.9%, respectively, which indicates good named entity recognition performance. The paper discusses the proposed method and experimental results, and outlines directions for further performance improvement.

Many studies were conducted to predict project time, cost, and risks using historical data of similar past projects, assuming similar projects within the same group follow similar patterns. However, the accuracy of prediction models depends on the project grouping accuracy in categorizing similar projects together. Current practice in US highway agencies indicates categorizing projects based on similar activities into different work types, which can be challenging due to the lack of a solid definition of project work types and the variety of projects’ activities. Besides, the accuracy of categorization with this method has not been determined. This paper develops a K-means clustering model to group similar highway projects together. The Euclidian distance was used as a metric to detect similarity and evaluate the performance of the method. The historical data of 730 projects from a highway agency in the US was used to validate the model and compare the performance of the model with the current categorization method used in US highway agencies based on project work type. The comparison showed the superiority of K-means clustering in grouping similar projects together. The proposed reliable clustering model can be used in future studies to increase the accuracy of prediction models that use the project type as an influential factor.

A massive amount of data/information that is buried in unstructured data sources such as bridge rehabilitation reports offer opportunities to complement other data sources such as structural health monitoring data for improved prediction of future bridge conditions. Extracting these data/information from their sources is, however, challenging for two reasons: (1) utilizing textual data for analytics remains to be a challenge due to the inherently unstructured nature of these data; and (2) existing information extraction methods, which aim to extract structured data/information from text, are limited in their capabilities and generalizability (e.g., a method developed to extract certain information from one type of document may not perform well on another type). To address these gaps, this paper proposes a deep learning-based information extraction (IE) model that automatically recognizes and extracts bridge-related data/information (e.g., bridge deficiencies) from multiple types of unstructured textual sources and represents the extracted data/information in a structured format to support further data analytics. The proposed IE model utilizes deep learning-based BiLSTM; begin, inside, and outside (BIO) encoding for the phrase-level segmentation; and conditional random fields (CRF) for both word-level and phrase-level labeling for automatic sequence labeling. The paper discusses the proposed model and its performance results.

With the development of urbanization, the energy-intensive building environment in cities is becoming increasingly responsible for energy consumption and greenhouse emissions in the United States. As a result, great efforts have been put forth to develop tools and methodologies to forecast urban building energy consumption in a spatial and temporal dimension. However, existing physics-based and data-driven models are insufficient to consider the impacts of building dependencies and micro-climates efficiently, which can significantly affect model utility and accuracy. Due to configurations and characteristics of modern cities, the interdependencies among buildings, e.g., heat transfer between buildings and solar impacts, are most often non-linear, high-dimension, and highly dynamic, which increase the difficulties to model them. To address those challenges, a novel urban building energy model (UBEM) based on spatio-temporal graph convolutional network (STGCN) algorithm was proposed to predict temporal urban-level building energy consumption in cities and better understand the interactions between buildings. In particular, we took a campus in Atlanta, Georgia, as a case study to validate the accuracy of UBEM. Results indicate that the UBEM tool has significant improvement in simulation accuracy, and model explanation compared with physics-based models and pure data-driven models.

Households often have relatively large electrical loads that typically run only for relatively short periods and make a disproportionate contribution to peak demand than overall electricity consumption. Although appliances such as clothes washers and dryers have a medium impact on peak demand periods compared to HVAC and lighting, they have potentially flexible duty cycles and can be scheduled to operate outside of peak periods, shifting away electricity loads from peak hours. This study aims to determine laundry activities’ contribution to electricity demand peak hours as well as to investigate the potentials through shifting cycles to off-peak hours in laundry appliances. This study uses ATUS and RECS data sets to analyze laundry activity time duration and laundry electricity consumption, respectively. The results illustrated an average of 23% of the time used for laundry activity during demand peak hours, which is equal to 1.6% of the total electricity used in an average US house.

Buildings are the largest user of energy in the United States and account for more than 50% of carbon emissions in large cities. Adopting effective tools for encouraging more energy-efficient buildings is therefore paramount. Cities in the US widely use the EnergyStar score for grading the energy efficiency of the buildings. The EnergyStar score uses multiple linear regression (MLR) model to assign a 1–100 score to each building, and its accuracy has been questioned recently by experts. This study evaluates the EnergyStar grading methodology and compares it with more complex machine learning based methods. The work focuses on two major aspects of the grading: (1) model accuracy and (2) building attributes used in modeling. The results show that the two developed machine learning models, random forest and extreme gradient boosting, outperform the linear EnergyStar model. Therefore, considering the significant impact of grading labels on energy improvement actions, implementing more accurate grading methods is suggested.

Automatically monitoring construction equipment and their activities using sound data has been an active research area in recent years. Despite their contribution to the literature, the previous studies can only be applied to isolated single-machine scenarios. However, most construction jobsites consist of multiple machines performing activities simultaneously, and the generated sound signals are mixed. Therefore, there is a need to develop a robust approach that identifies multiple machines’ activities. To do so, we propose a novel software-based method using a single-channel microphone and convolutional neural network (CNN). This method eliminates the need for using microphone arrays that require specific setup and configurations. We also propose a data augmentation method for generating synthetic mixed data to increase the method’s accuracy. We test the method on one synthetic and one real-world equipment sound mixture. The results demonstrate the capability of the proposed method in identifying activities with an accuracy higher than 80%.

While the use of deep neural networks (DNN) for computer vision is increasing in the construction domain, the shortage of training data sets prevents such models from achieving their maximum potential. To address this issue, we investigate the potential of using synthetic data for vision model development. Specifically, we synthesize construction images and train a DNN model only with the synthetic data. We then evaluate the performance of the synthetic data-trained model on a worker detection task, and the results demonstrate the great potential of synthetic images: 97.3% of mean average precision. Given the benefits of synthetic data—it is possible to automatically create an unlimited number of images without manual labeling—this finding is promising. Moreover, this approach can be readily applied to other computer vision tasks, without requiring the manual labeling. This finding will enable the creation of more accurate and scalable DNN models for construction applications.

Incomplete bridge inspection records containing missing information of bridge defects are inevitable due to sensor failures or bridge engineers’ omission errors during field inspections. Such incomplete inspection data could bias the condition assessment and mislead the maintenance decisions that lead to structural safety risks and raise maintenance costs. Existing studies proposed data imputation methods for filling in the missing values based on their spatiotemporally adjacent data. Unfortunately, these studies spent little effort examining the impacts of missing and imputed data on the reliability of bridge condition assessments and maintenance decisions. Biases within condition assessment and maintenance decisions can either waste maintenance resources or miss the chances of preventing structural failures. This paper uses sensitivity analysis to comprehend and explain missing data and data imputation methods’ influences on bridge condition assessments’ reliability. Such sensitivity analysis results could provide a basis for choosing data imputation methods that would not cause condition assessment biases and guiding data collection efforts to focus on safety-critical data sets.

Although a shield tunneling machine should excavate a tunnel along its planned alignment, deviations occur between the planned alignment and the actual result. In this case, to decrease the deviation from the planned alignment gradually, a target alignment that should be achieved in construction is generated. However, because excavation managers and operators are aging, their skills may be lost in the near future. Artificial intelligence is expected to play an important role in automating the operation of shield tunneling machines, but there are several limitations to systems proposed in previous studies. Therefore, in this research, an autopilot model was developed to automatically compute operation parameters for excavating along the target alignment. The results showed that the deviation between the predicted position of the shield machine by the optimum operation parameters of the autopilot model and the target alignment position was smaller than the measured deviation at an existing tunnel that was previously constructed.

Various types of emergencies can occur in buildings, exposing human safety to great danger. During building emergencies, it is of paramount importance for people to reach safe destinations quickly, for which wayfinding behavior is a critical determinant. In this study, the impact of visual access and crowd flow on emergency wayfinding was examined using simulations, in which human behavior was modeled based on empirical investigations. To do so, evacuation experiments were conducted in a virtual metro station. A total of 317 participants were involved in the experiment, and each participant experienced two or three decision points during the evacuation process. Using the experimental data, a mixed logit model was employed to investigate the relationship among participants’ wayfinding choices, visual access, and crowd flow. The results showed that both visual access and crowd flow significantly affected participants’ wayfinding choices. Crowd simulation was further developed according to the estimated mixed logit model, and multiple building designs with varying levels of visual access were evaluated with regard to the impact on evacuation performance and exit choice. The results suggested that increasing visual access of a shorter evacuation route could result in more agents choosing that route and improve the evacuation performance. The main contribution of this study is that instead of using predefined rules for crowd simulation, we established an approach to developing crowd simulation from empirical data to improve the fidelity of agents’ behavior. We also quantitatively assessed the impact of visual access and crowd flow on human indoor emergency wayfinding.

Visual inspection is the predominant practice used to determine the physical and functional conditions of bridges. Inspection processes typically require significant cognitive effort. Multiple bridge defects with various severity levels and locations could increase the cognitive demands of tasks. As a result, inspection records obtained from visual inspection often showed unpredictable errors. On the other hand, inspectors’ abilities to correctly identify defects can vary drastically among individuals and lead to different error rates in inspection reports. Personalized task performance predictions for individual inspectors are critical for proactively guiding bridge inspectors to avoid unique issues related to individuals’ nature. This study presents a personalized coach for bridge inspector (PCBI) training system that provides personalized predictions of an inspector’s performance in given tasks based on their behavioral traces in similar tasks. The predicted personal inspection task performance can help maintain a quality control system for bridge inspection. The authors developed a computer-simulated environment for bridge inspection processes to track inspectors’ cognitive behaviors and task performance. The authors analyzed and predicted inspectors’ performance individually using the personalized multi-linear regression (PMLR) model based on recorded cognitive patterns. Evaluation of the proposed model with the mean square error (MSE) between the actual and predicted performance achieved an MSE of 0.077. Finally, the authors compared human performance with state-of-the-art automated bridge inspection systems and discussed the effect of work experience on eye movement patterns.

Civil construction projects have numerous labor-intensive repetitive activities such as manual material handling, tile laying, concreting, among others. A handful of safety personnel cannot monitor the dispersed workers performing these tasks on the construction site. Besides, manual monitoring can be tedious, error-prone, and challenging to analyze postural safety. For computing the “safe work posture” of a task, unique actions and the moment exerted on major joints for each movement frame needs to be identified. For this, the paper proposes to use the Kinect to get the skeletal postural data while performing lifting and setting down task. Also, the research implements different machine learning algorithms to predict the unique actions and the moment exerted on the lower back for each movement frame. The training data was collected using Kinect in a controlled lab setup. The unique actions were manually identified for each movement frame. Then the kinetic analysis was performed in OpenSim using a human musculoskeletal model to compute the moment exerted on the lower back. The identified unique actions were used to train the action classification model using a random forest classifier, and the moment data was used to train the moment prediction model using a random forest regressor. Additional postural data were collected to validate the applicability of the developed model. The proposed research can play a pivotal role in assisting safety personnel in real-time safety monitoring of such repetitive tasks.

AI-powered smart homes bring high-quality intelligent services to occupants with digital virtual assistants. Through interactions with occupants, the smart home assistants (SHAs) can develop occupants’ profiles using a number of personal characteristic features for tailored and smart interactions. Based on these profiles, smart home systems can proactively offer automation services while conserving occupants’ comfort and convenience. In this study, we have sought to investigate characteristic features that affect occupants’ perception of the proactive concept, as well as their preferences for modes of interactions through an application of automation for energy efficiency management. Upon a data collection through an online experiment on campus, 58 valid responses with personal characteristic features were utilized to develop predictive machine learning models. These models can predict participants’ general attitude towards proactive SHAs, as well as their preferences for interaction modes with good performance (accuracy between 0.67 and 0.82 and F-score between 0.66 and 0.74). Various features were identified to have considerable significance, including personal beliefs of taking actions and energy expenses, as well as environmental protection values. The findings of this study provide an insight into the design of learning processes for virtual assistants in smart home ecosystems and the effect of the individual characteristics on the users’ preferences for interactions with SHAs.

The purpose of this paper is to examine the opportunities and barriers of integrated human-machine intelligence (IHMI) in civil engineering. Integrating artificial intelligence’s high efficiency and repeatability with humans’ adaptability in various contexts can advance timely and reliable decision-making during civil engineering projects and emergencies. Successful cases in other domains, such as biomedical science, healthcare, and transportation, showed the potential of IHMI in data-driven, knowledge-based decision-making in numerous civil engineering applications. However, whether the industry and academia are ready to embrace the era of IHMI and maximize its benefit to the industry is still questionable due to several knowledge gaps. This paper thus calls for future studies in exploring the value, method, and challenges of applying IHMI in civil engineering. Our systematic review of the literature and motivating cases has identified four knowledge gaps in achieving effective IHMI in civil engineering. First, it is unknown what types of tasks in the civil engineering domain can be assisted by AI and to what extent. Second, the interface between human and AI in civil engineering-related tasks needs more precise and formal definition. Third, the barriers that impede collecting detailed behavioral data from humans and contextual environments deserve systematic classification and prototyping. Lastly, it is unknown what expected and unexpected impacts will IHMI have on the AEC industry and entrepreneurship. Analyzing these knowledge gaps led to a list of identified research questions. This paper will lay the foundation for identifying relevant studies to form a research road map to address the four knowledge gaps identified.

Lighting energy consumption has a significant share of energy consumed in US commercial buildings. An accurate prediction of lighting energy consumption offers energy-saving or energy-management potentials, especially in office buildings. Although office buildings’ energy performance has been studied extensively in the literature, mainly focusing on HVAC energy consumption, studies focusing on lighting energy consumption in US office buildings are limited. This study aims to develop a data-driven prediction model to estimate the office buildings’ lighting energy performance accurately. This study used Commercial Building Energy Consumption Survey (CBECS) 2012 data. First, the predictors regarding lighting energy consumption in office buildings are analyzed, and the most relevant predictors to predict lighting energy consumption are selected using LASSO regression method. Second, an artificial neural network (ANN) model is trained, using the selected predictors, to predict lighting energy consumption. This study identified 12 main variables as the most influential predictors in lighting energy prediction, including square footage of the building, percentage of light during building operation hours, number of employees working in the building, building owner type, whether there is daylight harvesting, lighting percentage during off-hours, number of businesses in the building, whether there is lighted parking area, building America climate region, percentage of occupancy, whether the building is open during the weekend, and census division. In addition, the results showed that the developed ANN model is able to predict lighting energy consumption in office buildings with an R-square value of 76% and an RMSE value of 223.17.

Engendering trust in technically acceptable and psychologically embraceable systems requires domain-specific research to capture unique characteristics of the field of application. The architecture, engineering, and construction (AEC) research community has been recently harnessing advanced solutions offered by artificial intelligence (AI) to improve project workflows. Despite the unique characteristics of work, workers, and workplaces in the AEC industry, the concept of trust in AI has received very little attention in the literature. This paper presents a comprehensive analysis of the academic literature in two main areas of “trust in AI” and “AI in the AEC,” to explore the interplay between AEC projects’ unique aspects and the sociotechnical concepts that lead to trust in AI. A total of 490 peer-reviewed scholarly articles are analyzed in this study. The main constituents of human trust in AI are identified from the literature and are characterized within the AEC project types, processes, and technologies.

Robots will soon become as commonplace as heavy-duty vehicles at construction sites. Their emergence will create safety challenges, as most of them will work very near humans in the dynamic and unstructured workplace of construction. Among them, being stuck by robots could be a major cause of accidents. Robotic control techniques can tackle these challenges by providing efficient and reliable channels for worker-robot interaction. However, traditional techniques, such as haptic- and gesture-based control, require workers to use hands and arms as middleware. Therefore, in scenarios where workers’ range of motion is restricted—such as moving materials or roofing—physically interactive control may not be adequate. To bridge the gap, this study examines the feasibility of a novel approach. The technique employs an adaptive neural network to translate workers’ mental images of their bodily movements into robotic commands via brainwave signals. These commands regulate robots’ actions through human kinematics. To examine its performance, four subjects were instructed to control a masonry robot lifting and placing concrete bricks to build a wall. The results demonstrated that the average success rate of picking up and laying bricks was 84.02%, with a 3.35% standard deviation. This demonstrates the feasibility of controlling robots under the physically complex conditions of construction sites. It opens the door to safer interaction between workers and construction robots.

Industrialized construction revolutionizes the construction industry by transferring the majority of onsite activities to controlled, offsite manufacturing facilities. However, the prefabrication processes are still labor-intensive. Construction workers perform repetitive and physically demanding tasks that can cause severe injuries and musculoskeletal disorders. Previous studies have proven the success of medical exoskeletons in physical rehabilitation. Nevertheless, there is a scarcity of research on their efficacy in industrial environments. This research introduces a criteria-based framework for assessing the use of exoskeletons for the rehabilitation of injured and disabled workers in industrialized construction. The proposed research reviews the state-of-the-art and state-of-the-practice in medical exoskeletons to evaluate their readiness for construction tasks. A criteria-based decision matrix captures the relationship between the identified exoskeletons and industrial construction tasks. The present research is expected to address the unprecedented skilled labor shortage facing the construction industry. It will provide construction professionals with preliminary guidelines for integrating medical exoskeletons in factory-based settings, thereby retaining and rehabilitating Gen X workforce.

Poor safety and low productivity have pushed the construction industry to increase the use of technology and automation. Consequently, robotic systems such as aerial and ground robots are becoming increasingly popular in modern construction sites. As construction is still a heavily human-driven industry, these robots need to collaborate, interact, and share their work environment with human workers. This integration raises novel occupational safety and health issues for jobsite personnel. In this conceptual paper, human-robot interactions and their potential collaboration areas will be discussed, focusing on robot types and their application areas. Next, the health and safety risks associated with human-robot interactions on construction sites will be discussed. This discussion will mainly focus on physical risks, attentional cost, and psychological impacts of such interactions on workers who work directly or indirectly with or around robots. Finally, a discussion will be provided on considerations and recommendations to ensure the safe integration of robots in construction.

The development of robotics technology has created lots of potential for the construction industry. Due to unique features of construction sites (e.g., unstructured and dynamically changing), robots require human assistance to perform various tasks. Thus, human robot collaboration (HRC) has become an essential question when applying robots to the construction industry. However, it is not clear about the current research status within the HRC context in the construction industry. This paper conducted a preliminary review of the existing studies focusing on HRC in the construction industry. The research objectives/environment, HRC methods applied, construction tasks and robots involved, technologies applied, and factors affecting adoption of robotics in the construction industry from the existing research studies were summarized. It is envisioned that the outcomes of the paper could assist both researchers and industrial practitioners with appreciating the research and practice frontier of HRC in the construction industry.

Although there have been many attempts to increase student engagement and interaction in class through virtual reality and augmented reality (VR/AR), structural engineering has remained as one of the disciplines that lacks interactive learning. There was an earlier attempt by the authors to improve student learning in structural analysis and dynamics through a student-centric cyber-physical system (SCPS) that reacts to student movement and outcomes structural behavior. This system used mobile devices as sensors and sent data to be processed by a server. As a result, it suffered from the inherent limitation of Transmission Control Protocol/Internet Protocol (TCP/IP)—lags and instability. Therefore, this paper presents the development of a new SCPS tool with a new system architecture that is much more stable and provide better user experience. This SCPS captures and uses student movement (i.e., walking, turning, and jumping) as point loads on a virtual bridge. As students move around, structural behavior (i.e., deflection) is simulated and visualized in real time.

A successful construction safety management process involves systematic collection, management, and distribution of a significant amount of data through various means including stakeholders’ safety records, inspection procedures, risk assessment analyses, daily reports, preventive control tasks, incidents reports, and post-incident investigations. Despite the tremendous efforts and research in construction safety, still little is known about a systematic method for safety data management that provides accessibility, transparency, reliability, consistency, and traceability. This study aims to address this gap in knowledge by developing a novel information management framework for construction safety processes using blockchain. The proposed framework utilizes distributed ledger technology (DLT) to offer a new architecture of trust for reliably sharing information among all entities involved in a construction project who may not trust one another. The proposed method is designed based on a private blockchain system that maintains a decentralized control of information through permissioned channels to provide a shared view of reliable safety data and records that are accessible, traceable, and immutable. The outcomes of this study contribute to the existing body of knowledge and current state of practice in integrated safety management (ISM) systems that aim to protect construction workers and public through providing timely, reliable, and complete information to all entities involved in a project.

In the last couple of decades, various industries have taken advantage of emerging advanced technologies, such as the digital twin, to find more effective solutions in their respective areas. In the transportation infrastructure sector, the implementation of digital twin technology is slowly gaining traction but lagging behind other major industries. To better understand the limitations, opportunities, and challenges for the adoption of digital twin technology in this sector, a survey questionnaire was distributed to collect information from industry professionals involved in transportation infrastructure projects. The purpose of this study is to understand how digital twin technology is being perceived by the industry. Based on the results of this survey, the current state of digital twins in the infrastructure sector, as well as the expected benefits, and potential challenges for the deployment of this technology are discussed here. This study aims to initiate discussion for future research in digital twin technology for transportation infrastructure applications.

Data play a significant role in the life cycle management of road infrastructure and bridges, and their integration is a crucial element of smart infrastructure systems. However, the current practice of managing infrastructure involves the use of an abundance of data produced by a variety of non-interoperable information systems. Thus, the lack of interoperability creates major challenges in deployment of a fully integrated and smart management system for the infrastructure. This research focuses on the development of a digital twin that could be used as an umbrella for integrating the live load traffic data on bridges with other bridge life cycle data to mirror the bridge behavior and offer a smart and integrated bridge management system. Intelligent transportation systems (ITS) are remarkable instances of advanced technology that reduced the human role in the operation and management of transportation infrastructure. Weigh-in-motion (WIM) is a type of ITS system provides a stream of valuable data about weight and other dynamic attributes of the fleeting traffic over road network and bridges. In this research, the real-time WIM data are integrated into a BrIM model by visual scripting to form the digital twin of the bridge. Through this approach, weight sensor data could be streamed in the digital twin of bridge, and the level of data utilization by different stakeholders can be improved. The outcomes of this study will help the preservation and sustainability of bridges and helping their resiliency through pro-active planning and enhancing the utilization of the available data.

Healthcare facilities are essential in providing health, and as a result, they require effective facilities management systems. Unfortunately, current approaches to healthcare facilities management (FM) are based on “if it breaks, we fix it” approach. This is due to the lack of effective real-time information update and bi-directional coordination to monitor, manage, and control critical healthcare facilities. Thus, there is a need in ensuring a preventive maintenance approach. This research investigates the extent to which healthcare facilities management can be enhanced using digital twin technology. Digital twin (DT) enables the virtual representation of a physical asset and all the instances using real-time information. DT collects data from physical assets in real-time and uses the data to create virtual models of the physical objects. This research employs a set of methods such as literature review and interviews to address its objectives. The study further identified issues with critical healthcare facilities amenable to digital twin approach, reviews digital twin enabling technologies, and presents a system architecture for DT application in healthcare facilities management. The research concludes with a discussion of the potential benefits of DT application in healthcare facilities management.

Building information modeling (BIM) is increasingly used in the construction industry, but existing studies often ignore embedded rebars. Ground penetrating radar (GPR) provides a potential solution to develop as-built BIM with surface elements and rebars. However, automatically translating rebars from GPR into BIM is challenging since GPR cannot provide any information about the scanned element. Thus, we propose an approach to link GPR data and BIM according to Faster R-CNN. A label is attached to each element scanned by GPR for capturing the labeled images, which are used with other images to build a 3D model. Meanwhile, Faster R-CNN is introduced to identify the labels, and the projection relationship between images and the model is used to localize the scanned elements in the 3D model. Two concrete buildings are selected to evaluate the proposed approach, and the results reveal that our method could accurately translate the rebars from GPR data into corresponding elements in BIM with correct distributions.

This paper presents a new approach for reconstructing the geometric properties of steel members in bridges, such as steel girders, cross-frame members, and steel piles, by exploiting point clouds captured by terrestrial laser scanners. The key innovation is focused on automatically addressing the occlusions in the point clouds by leveraging a 3D occlusion labeling algorithm and parametric models. Through the 3D occlusion labeling algorithm, the spaces that are not occupied by the acquired point clouds are automatically labeled as empty spaces and occluded spaces, depending on whether they have been visited by the laser beams. By integrating this information with parametric models, the complete geometric properties of the steel elements can be effectively and automatically retrieved, even with a high level of occlusion. The effectiveness of the proposed approach is validated through preliminary experiments conducted on the steel girders in a highway bridge, and the accuracy is evaluated by comparing the computed geometric properties with that from ground-truth measurements.

Architectural programming (AP) is the first essential step before starting the schematic design of a project. This process allows for the identification of the description and layout of an area based on specific project requirements while defining scope of work and crucial factors towards client satisfaction. One deliverable within the AP is the room data sheet (RDS), presenting information pertaining to the project rooms. Specifically, RDS information includes room names, intended uses, locations, numbers, description of the finishes, fixtures and fittings, as well as mechanical and electrical requirements within the space. Despite several attempts to develop multiple programming techniques and effectively meet clients’ project requirements, previous projects aimed at capturing architectural programming have not been adopted, and poor early-stage planning remains a barrier in the construction field, potentially resulting in project delays, cost inefficiency, and ultimately client dissatisfaction. This study proposes an RDS data representation with the aim of enhancing project planning efficiency. The contribution of this study is to help pre-design practitioners and researchers in reducing design changes, associated project costs and delays, and ultimately improving client satisfaction and construction efficiency.

It is estimated that colleges and universities in the United States spend an average of more than a dollar per square foot on electricity annually that corresponds to large recurring expenditures. As university campuses continue to grow, it causes an increase in the energy demand impacting the land surface temperature. Energy audits of campuses are performed annually; however, they are mostly manual and require physical inspections that are time and resource-intensive. The technological advancement in satellite imagery provides an environment that can acquire rich, meaningful information in a time- and cost-effective manner. This study utilizes a medium-size urban campus in the United States with a hot and dry climate as a case study to evaluate the relationship between energy consumption intensity and land surface temperature using the remote sensing and geographic information system (GIS) tools. The study aims to explore the variations in total energy and cooling energy demand on campus due to changes in land surface temperature over the past five years. The study will help in better understanding the micro-climate changes as well as the thermal behavior of the campus landscape to identify hotspots. The results of this study will provide the basis for decision-makers to better understand energy consumption patterns on campus and adopt effective mitigation measures to save energy and ultimately reduce the carbon footprint of the campus.

Building information modeling (BIM) and finite element analysis (FEA) are widely used technologies for underground structure design; however, they failed to coordinate work well. Currently, existing BIM-to-FEA model conversion methods have a coarse conversion effect and a narrow scope of application. They do not consider the variational working conditions in the construction of underground structures, and thus can hardly be used in the simulations. To address this problem, this research proposes a semi-automatic BIM-to-FEA conversion method for construction simulation of underground structures based on Revit Dynamo and OpenSees software. The method is realized in three steps. Firstly, typical subway stations are selected as the research objects, and the parameterized BIM models are established. Secondly, based on Dynamo visual programming and tool command language (TCL) scripts, a BIM-to-FEA semi-automatic conversion method is formed to carry out FEA calculation. Finally, the method is extended for construction simulation. Based on the BIM model of the station’s enveloping structure, an FEA model of diaphragm wall with multiple working conditions is automatically generated, and the structural simulation of the open-cut excavation process is conducted. This method provides an automatic BIM-to-FEA conversion way that can efficiently generate FEA construction simulation models with various working conditions and realizes a time-saving construction simulation framework based on BIM technology.

Indoor air conditioning schedule has a significant impact on building energy consumption and occupants’ comfort. In order to assure the comfort and health of the occupants, it is necessary to continuously monitor indoor air conditions and adjust the setpoint of the building systems according. Occupant-centric operation of building systems often results in saving energy and more satisfaction in the occupants. Recent advancement in the digital twin technology has created the potentials for real-time monitoring and assessment of indoor air condition, which is an essential need for occupant-centric building operations. This article discusses a case study on creating a digital twin of a building at Roger Williams University to investigate the potentials and challenges of implementing a digital twin for the occupant-centric operation of facilities.

The construction industry is generally regarded as information intensive. Building information modeling (BIM) has been developed to provide considerable information support for construction activities. As more data is aggregated in building information models, it is more difficult and time-consuming to extract useful building information for construction project team members with limited BIM experience. In Industry 4.0, deep learning-based conversational artificial intelligence (AI) technologies provide more opportunities to support human daily life. Similarly, for smart construction, a conversational AI information retrieval system is considered important to provide building information support. This research aims to develop a conversational intelligent query answering (QA) system for attribute information retrieval from building information models. The methodology of the QA system for BIM was developed based on the Google pre-trained deep learning language neural network (NN) algorithm—Bidirectional Encoder Representations from Transformers (BERT) and natural language generation (NLG) methods. A Python-based prototype program was developed based on the methodology. Building attribute information-related queries were generated to test the accuracy of the developed QA system for BIM. The preliminary results indicate that the developed prototype achieved an 80.0 accuracy score, meaning 80 predictions exactly matched their ground truths, for providing answers to building attribute information-related queries. The developed QA system methodology for BIM was also found to be valid and relatively accurate. The contribution of this research facilitates the development of conversational AI systems that provide convenient information support for onsite and offsite construction personnel, and the developed QA system for BIM can be extended to other information retrieval areas.

We propose a method to improve the understanding and visualization of the actual condition of a construction site by automatically developing an as-is BIM using site-appearance information from images and the as-planned BIM model. This is achieved by generating point clouds (PCs) from site images applying structure from motion (SfM). The corresponding elements between PCs and the 3D BIM were automatically determined using geometric and position assets to register PCs into the as-planned BIM accurately. Moreover, material condition classification can be done using information from the point cloud data. This way, the as-is BIM can be enriched with additional information such as the actual material conditions. The proposed method has been demonstrated using a construction environment where the as-is BIM was developed automatically from a set of 130 site images and the as-planned BIM. The as-is model has been used to identify deviations between as-planned and as-is conditions.

The US Department of Transportation aims to comprehensively incorporate social equity goals into its nationwide programs and policies. Accordingly, state highway agencies seek to integrate social equity goals in their asset management plans. To better accommodate social equity goals in their decision-making, it is essential to quantitatively assess the current level of social equity in the transportation network. However, the existing literature on social equity has been primarily qualitative in nature, making it difficult to assess social equity quantitatively. This paper aims to address this gap by exploring different numerical approaches and visualization techniques to systematically evaluate levels of social equity and other measures of network performance. The study expands on quantitative techniques developed by the authors in an earlier study and incorporates GIS-based visualization techniques to assess the level of social equity. Highway network data for New York City is studied in conjunction with the CDC’s Social Vulnerability Index (CDC SCI) and the community’s socio-economic demographics to gauge the current levels of social equity in the community. The representations of this study expand the use of quantitative metrics of social equity to support decision-making for infrastructure.

For road managers to properly operate roads, it is necessary to immediately confirm road damage, perform frequent periodic inspections to check for defects and damage, transfer business knowledge, and develop human resources. To realize appropriate road maintenance, it is important to share information among all stakeholders, and easily shared three-dimensional (3D) data with high visual fidelity are effective means toward this end. With the introduction of unmanned aerial vehicles (UAVs) and terrestrial laser scanners (TLS), there will be increased use of 3D data in the future. In this study, TLS, UAV, and camera measurements on road pavement, bridges, and slopes were performed with the aim of realizing 3D data for road maintenance. We also investigated methods for collectively managing the generated 3D data and visualizing results of inspecting the 3D data and proposed the information system using 3D data and images. Furthermore, a road maintenance management system was proposed using images from vehicle-mounted cameras and investigated its feasibility. Two system use cases were considered for the system: the case of receiving notification of an abnormality from a road user and passing on knowledge. The system includes functions for same-screen display of images and maps, and storing descriptions of abnormalities. It was inserted GNSS data, display abnormality descriptions, and store photos. As a result, we contributed the collection of video data for roads, the combination of two different angle video data, and the usage of video images for road maintenance management system.

Building information modeling (BIM) is increasingly being integrated with facilities management (FM) to support decision-making and optimize the business value of facilities. There is a plethora of publications on proposed process and technology improvements and innovations relating to BIM use in FM. Several studies have indicated improved efficiency when FM is supported by BIM. Yet, there is a lag in the acceptance of BIM for FM and a lack of clarity on why people accept or reject BIM for FM. In this study, we formulate a conceptual model through an adaptation of the unified theory of acceptance and use of technology (UTAUT), which identifies critical acceptance determinants of BIM for FM and establishes probable relationships between these determinants. The resultant model is intended to improve the understanding of the determinants that lead to acceptance or rejection of BIM for FM. Empirical validation and application of the model is reserved for future work.

Automated checking of the compliance of building information modeling (BIM)-based building designs with relevant codes and regulations requires bridging the semantic gap between the Industry Foundation Classes (IFC) schema and the natural language. In most of the existing automated compliance checking (ACC) systems, the integration of the IFC schema and natural language is realized through hardcoding or predefined rules, ontologies, or dictionaries. These methods require intensive manual engineering effort and are often rigid and difficult to generalize. There is, thus, a need for an automated and meanwhile flexible and generalizable information integration method. To address this need, this paper leverages transformer-based language models to learn the semantic representations of concepts in the building information models (BIMs) and regulatory documents. An automated IFC-regulatory information integration approach based on these learned semantic representations is proposed. The preliminary experimental results show that the proposed approach achieved promising performance—an accuracy of 80%—on integrating IFC and regulatory concepts.

As the number of infected people by COVID-19 in the US has broken the record, the federal and state emergency planners are looking for different methods to cope with the challenge of expanding the current healthcare system capacity to provide care for people with a compromised immune system and preexisting conditions. Given the current capacity of US hospitals is estimated at 920,000 beds in total, if the number of infected people keeps rising, the nationwide shortage of hospital beds is projected to be higher. This predictable shortage has led healthcare facilities managers and federal and state authorities to seek alternative locations outside of existing healthcare settings to accommodate the expected surge of COVID patients. As a result, the construction permit request to convert the existing building to temporary COVID care facilities has recently increased. This spike has put pressure on the building authorities to process the plan review and inspect these converted buildings to allow for immediate occupation and operation. The study identifies the critical shortcomings in the current permitting process and proposes a new virtual permitting approach. The paper introduces a BIM-based Virtual Permitting Framework (VPF) to expedite the existing building’s permitting to repurposing ACS, especially the airborne infections isolation (AII) rooms.

The objective of this paper is to propose a comprehensive resilience assessment of disaster-induced disruptions in transportation networks and use Harris County, Texas, in 2017 during Hurricane Harvey as a case study. In this research, we integrate network percolation theory and spatial analysis of household survey data to holistically examine the spatial patterns of risks caused by disaster-induced road closures. The overall risk of infrastructure disruptions is measured in terms of a community’s experienced hardship. Hardship is a collective influence of the exposure of transportation networks, represented by connectivity, the duration of disruption of commuted roads, and the social susceptibility of households (tolerance) to the road closures. Integrating the spatial analysis of tolerance and duration of disruption with network percolation theory reveals the hotspots or coldspots of hardship. The results of the comprehensive resilience assessment show that the spatial patterns of risk to the road closures are the function of network connectivity, duration of disruption, and tolerance. Moreover, these together could explain the emergence of the risk hotspots. The emerging hotspots of hardship in the east of Harris County are due to clusters of low tolerance and a low connectivity of roads, and not duration of disruption. The west of Harris County did not have any significant clusters of hardship due to their high tolerance despite experiencing high duration of disruption and having a low connectivity in the road network. These results highlight the importance of a comprehensive resilience assessment when assessing prolong disruptions in the transportation network, which leads to devising effective plans for hazard mitigation.

A digital twin involves technology including data analytics, artificial intelligence, and IoT that acquire and analyze heterogeneous data from multiple sources, which creates a living digital simulation model that supports the decision-making process. The adoption of building information models (BIM) is one of the core digital twin technologies that can be applied in architecture, engineering, construction, operation, and facility management (AECO/FM) sectors. However, current developments related to building DTs are at an early stage in the building life cycle, especially regarding integrating and analyzing asset and facility management technologies. This study presents a novel holistic system to construct a building-level DT integrating BIM data in the building life cycle from a multi-stakeholder perspective. First, current challenges faced by stakeholders are documented and analyzed by a requirement model. Then, a multilayered framework is proposed to underpin the complex requirements when tackling the challenges. This framework assists five major areas of asset and facility management: space, energy, operation, safety, and maintenance. A web-based system based on the proposed framework is developed and visualized through a game engine in a case study for a commercial building in China. This developed building-level DT system is comprehensive, robust, and can efficiently and effectively satisfy the requirements of different stakeholders, such as the estate management team and maintenance staff, who may not have BIM experience. Finally, insights and challenges for researchers and practitioners are concluded and discussed.

Despite many potential applications of building information modeling (BIM), the automated evaluation of existing as-is information in images and its transformation into purpose-oriented digital as-is BIM models is still under investigation. Although several previous studies have systematically investigated technologies for creating as-is BIM models for existing building elements, they do not include the latest approaches to object recognition in terms of machine learning and deep learning methods. Many researchers point out the need for more efficient and feasible algorithms to provide an automatic and information-rich BIM model generation system. By following a systematic assessment methodology, this review paper provides an overview of current object recognition and detection strategies in the architecture, engineering, construction, and operation (AECO) domain and discusses their potentials and limitations based on the richness of semantic information with respect to as-is BIM generation of existing buildings. The combined application of image-based object detection and the automated transfer to BIM will be investigated in future work.

Conceptual graph (CG) is a graph representation of logic derived from semantic knowledge. In the rule interpretation process in automated compliance checking (ACC), the quality of knowledge exchange between rule experts and rule implementers is difficult to ensure. CG is an effective communication tool to address this issue. However, quantifying the amount of knowledge transferred through CG is challenging considering the subjectiveness existing in people’s understanding of rules. This paper proposes an assessment framework including both subjective and objective criteria of a CG to measure its quality by comparing the knowledge captured by the CG and the knowledge in the original rules. Four groups of criteria are studied: correctness, completeness, complexity, and comprehensibility. The framework is validated through a case study using a building code related to gross floor area (GFA). The CG generated by the rule implementer is evaluated to test the rule implementer’s knowledge. The evaluation result shows that the rule implementer can capture most of the knowledge in the original building code. The mistakes mostly exist in notations of the CG, which could be due to his/her unfamiliarity with it. Overall, the CG can be easily implemented to represent the building code and check rule implementers’ knowledge.

Quality and defect management is a huge topic in the construction execution phase. By now, digital methods like BIM are mostly used in the planning phase. In the construction execution, especially in the management of qualities are analogous or document-oriented systems state of the art. Current quality management systems usually focus on the management and correction of defects, but the quality-related construction supervision contains more activities like planning quality inspections. The distribution of quality and defect-related data is still limited due to corporate solutions, missing project relation, and non-standardization. Ontologies can link heterogeneous data sources and can be used to plan and check quality inspection by standardized rule languages. The following paper will overview current research in using ontologies for quality management in the construction execution phase. The research will investigate and compare current methods of construction quality management by ontologies. We examine and compare methods for querying and rule checking of quality information based on ontologies. This review enables further research and improves research activities lasting.

Digitalized communication and information exchange are the bedrock of smart cities and infrastructure. Enhancement of neutral data exchange platforms and data serialization is an inevitable step toward improving smart city architecture. Data serialization is defined as a process for encoding or translating data into neutral file formats so they could be transferrable or retrievable by other applications. The semantic data need to be communicated with other systems and, therefore, need to be converted into a neutral format that could be globally communicable. The Industry Foundation Classes (IFC) are one of the examples of neutral data exchange format that benefited the building information modeling (BIM) in the built environment. The expansion of neutral data exchange schemas is receiving more attention every day; however, it is full of challenges. Review and aggregation of challenges experienced by other researchers and their approaches for overcoming the interoperability complications could help future researchers. Moreover, reviewing the key elements in semantic and syntactic validation procedures improves the early planning of the project and improves the research process’s reliability. This paper reviews the most common data serialization methods and provides a summary of the challenges and gaps. The findings of this research can inspire other researchers with their data exchange system architecture, verification, and validation methods to increase the effectiveness of their research.

Applications of building information modeling (BIM) have widely infiltrated various construction fields. BIM serves as a set of technologies to maximize productivity, increase collaboration, and improve work efficiency throughout the project life cycle. BIM also supports the best practices of other emerging technologies by creating a platform for integration and coordination. The process, approach, and strategy of implementing BIM can impact the entire project’s cost and schedule performance. A need exists to evaluate the effectiveness and practicability of the current BIM practices. First, this study reviewed 84 selected BIM projects from 21 countries based on their features, including project information, delivery methods, BIM approaches, level of development, software, and emerging technologies. Second, the challenges and limitations of adopting BIM in the construction project were evaluated. One significant outcome of the research was the identification of key attributes that facilitate the effective use of BIM on construction projects. The result of this study can assist practitioners in making better decisions on project planning and BIM adoption.

The exponential growth of on-site visual data and the advent of computer vision techniques have invigorated automated construction progress monitoring. However, the state-of-the-art automated progress monitoring methods are still insufficient for the daily monitoring of construction activities. As-built 3D point clouds at the construction site are captured on a specific time interval and require a considerable amount of processing time. Between the intervals, the daily site progress status is often not recorded. Additionally, existing methods report the progress status of any building element in binary form (i.e., completed or not completed) without reporting partial completion percentages. In the construction phase, although BIM models created can help progress monitoring, some schedule activities are usually not modeled in BIM. Automated progress monitoring of these activities is still challenging. To overcome these methodological challenges, this study proposes a novel framework for automated daily construction progress monitoring incorporating image localization and deep learning-based semantic segmentation. The preliminary results show that the framework can successfully detect daily progress with a partial completion percentage.

This paper devises, implements, and benchmarks a novel framework, named CLOI, that can accurately generate individual labeled point clusters of the most important shapes of existing industrial facilities with minimal manual effort in a generic point-level format. CLOI employs a combination of deep learning and geometric methods to segment the points into classes and individual instances. The current geometric digital twin generation from point cloud data in commercial software is a tedious, manual process. Experiments with our CLOI framework reveal that the method can reliably segment complex and incomplete point clouds of industrial facilities, yielding 82% class segmentation accuracy. Compared to the current state-of-practice, the proposed framework can realize estimated time-savings of 30% on average. CLOI is the first framework to have achieved geometric digital twinning for the most important objects of industrial factories and provides the foundation for the generation of semantically enriched industrial digital twins.

A digital twin (DT) can be defined as a multi-physics, multiscale model in which a digital model, such as a building information model (BIM), is updated based on data obtained from a physical system as well as results from probabilistic simulations and models. This study describes the critical steps toward the implementation of DTs to support structural health monitoring (SHM) of mass timber buildings. In particular, the study defines a methodological approach used to integrate as-built geometry of existing buildings, as well as their material properties and sensors, to link SHM parameters into a digital model to assist in assessing a building’s structural performance. A coupled pair of post-tensioned cross-laminated timber (CLT) self-centering shear walls at the George W. Peavy Forest Science Center (“Peavy Hall”) at Oregon State University were used as a case study to test the proposed approach. The BIM of the shear walls was developed using a scan-to-BIM approach by converting LiDAR point clouds into a BIM. Sensors in the building recorded environmental and structural parameters influencing the long-term performance of the shear walls. Measurands included relative humidity, air and wood temperature, wood moisture content, displacements and deformations of shear walls, and tensile force of post-tensioned rods. The precise placement of these sensors and the possibility to associate the measured parameters of these entities within a BIM is hypothesized to assist with data management by adding a spatial element to data and analysis results, which could lead to the prolonged service life of a building.

Traditional road safety assessment methodologies do not recognize the driving environment’s fast-changing dynamics that influence the contextual complexity and, ultimately, its risk. This paper proposes a method to use diverse open-source sensor data (LiDAR) collected by Waymo autonomous vehicles to estimate the road environment’s complexity considering dynamic traffic conditions. The proposed contextual risk factor (CRF) model estimates the driving scene’s complexity using the density and proximity of the objects around the vehicle. The data was analyzed frame-by-frame, and contextual risk categories of high, medium, and low were assigned. The results revealed the objects in the scene well represent the contextual complexity. However, what the driver sees in front of them and within their forward reaction space is not directly representative of the complexity of the scene and vice versa.

Building façades play a significant role in the thermal and energy performance of a building. The presence of thermal anomalies in façades can cause uncontrolled heat leakages, which impacts a façade’s thermal performance. Unmanned aerial vehicles (UAVs) equipped with infrared (IR) cameras are becoming increasingly popular as non-destructive methods for the inspection of thermal anomalies in building envelopes. However, the thermal anomalies identified in UAV-captured images are difficult to localize on a 3D building model, which impedes the assessment and documentation of the thermal anomalies with their properties. This paper proposes a computational workflow to automatically register the detected imagery thermal anomalies to a 3D building model. First, IR images and thermal anomalies are sequentially aligned with the corresponding visual RGB images and mapped into façade reference images (RIs) by using image matching techniques. They are then registered to the 3D building coordinate system through a predefined coordinate transformation process. A pilot case study is presented to demonstrate the processing of UAV-images using the proposed method. As a result of this study, the proposed method enables automated registration and data fusion of UAV-captured images, as well as thermal anomaly information into a 3D building model, supporting the assessment of anomalies and evaluation of existing façade conditions.

The integrative nature of BIM technology renders it an ideal platform for implementing sustainable strategies into new construction and renovation and/or retrofit of existing structures. Recently, Saudi Arabia’s gasoline prices have dropped dramatically, necessitating the kingdom to establish its 2030 vision. The 2030 vision aims to make the Kingdom of Saudi Arabia (KSA) an ideal sustainable society by reducing its reliance on petrol, creating more sustainable buildings, and infrastructure. There is an imperative need to translate the green aspirations of the KSA into enforceable regulations. In order to achieve this vision, consistency across jurisdictions is needed to allow robust and well-informed decision-making in adopting and enforcing sustainability. The number of high-performance buildings in Saudi Arabia is inadequate, and the rate of change is too slow. Due to Saudi Arabia’s need to enhance its built environment, it established a new green building rating named “Mostadam.” Currently, there is no metric to measure the road map to achieve the aspects of the 2030 vision of Saudi Arabia, and the number of green buildings in the KSA is very limited. This study aims to formulate a BIM-Mostadam model to help facilitate the green building certification process in the kingdom that supports its 2030 vision. The proposed BIM-Mostadam model seeks to address critical elements of the 2030 vision in the KSA based on integrating BIM and green building rating systems to aid in achieving the envisioned 2030 sustainability goals and implement sustainable design principles in the KSA. This research follows a mixed methodology that includes both qualitative and quantitative approaches. The limitations of the proposed model are presented along with future phases of the study.

Modular construction has the potential to increase the efficiency of on-site construction activities, as well as reduce capital costs, project duration, and construction waste. However, currently, critical issues in the design and pre-planning phases including longer lead-in time, lack of expertise, and flexibility of design hinder its extensive deployment in the AEC industries. To this end, this paper proposes a new methodology for automated generation and configuration of a modular building design by deploying generative adversarial network (GAN) and building information modeling (BIM) technologies. This paper focuses on spatial design in the early design phase. By incorporating the requirements and constraints of a modular building project needed for its life cycle into the automated design process, the proposed method can automatically provide feasible, constructible, and optimal modular design alternatives. The results of this study show that it is expected to significantly reduce the time and effort for designing a modular building and generating diverse design options that can efficiently fulfill the final user’s requirements, functional requirements, and manufacturing and assembly constraints.

The interdependencies among construction activities are a key factor to define the construction schedule. The application of robots in construction projects not only impacts a single task but also the sequence and interdependency between tasks. The difference between humans and robots in physical characteristics, work constraints, and work performance will change the means and methods of construction, and potentially the dependencies or sequence of construction activities. This study develops a method for analyzing the impacts of using robots as alternative methods within network schedules to understand the impacts of employing robots in construction tasks. We first summarize the typical construction sequence dependencies and their causes based on extant literature. Then, we analyze the impact of emerging or currently used site construction robots on these dependencies based upon robots’ characteristics and work methods, as well as the potential propagation of changes through the network using graph theory. A case study project is analyzed to compare the proposed use of robotic construction in lieu of the actual methods employed, understanding the changes with and without robotic methods to analyze the sequence differences. The outcome is a defined process and four types of schedule changes identified through the use of construction robotics in construction sequences and network schedule dependencies.

In bridge inspection, engineers should diagnose the observed bridge defects by identifying factors underlying those defects. Traditionally, engineers search and organize structural condition-related information based on visual inspections and multiple data sources. Even following the same qualitative inspection standards, experienced engineers tend to find the critical defects and predict the underlying reasons more reliably than less experienced ones. Unique bridge and site conditions, quality of available data, and personal skills and knowledge collectively influence such a “subjective” nature of data-driven bridge diagnosis. Unfortunately, the lack of detailed data about how experienced engineers observe bridge defects and identify failure modes from multi-source data makes it hard to comprehend what engineers’ behaviors form the best practice of producing reliable bridge inspection. Besides, even experienced engineers could sometimes fail in noticing critical defects, thereby producing inconsistent, conflicting condition assessments and biased maintenance plans. Therefore, detailed cognitive behavior analysis of bridge inspectors is critical for enabling a proactive inspector coaching system that uses many inspectors’ behavior histories to complement personal limitations. This paper presents a computational framework for revealing engineers’ observation and cognitive-behavioral processes to identify bridge defects and produce diagnosis conclusions based on observed defects. The authors designed a bridge inspection game consisting of FEM simulation data (stress and displacements) and inspection reports (basic bridge information and possible defect types) to capture and analyze experienced and inexperienced engineers’ diagnosis behaviors. Mining these behavioral logs has revealed reusable behavioral process patterns that map critical bridge defects and diagnosis conclusions. The results indicate that the proposed method can proactively share inspection experiences and improve inspection processes’ explainability and reliability.

The digital twin (DT) concept is increasingly employed as a powerful tool to support data-driven decision making in management of the built environment. Various open data models (DM), addressing specific scale and domain applications, can be found in the literature. However, many DT implementations are based on organization-specific information management processes and proprietary DM, which hinder interoperability. This paper aims to present the process and information management approaches adopted for developing a federated metamodel supporting DT applications. The information delivery manual (IDM) technique is applied adopting a cross-disciplinary and multi-scale methodology. The proposed framework is applied to the West Cambridge Campus Digital Twin research facility, identifying five main phases: information requirements (IR) definition, data schemes mapping, conceptual model definition, prototyping, testing, and validation. An insight into IR definition, employing the Industry Foundation Classes (IFC) schema is presented. Although additional testing and case studies are required for validation, the proposed methods are currently used to explore and develop DT applications at the building and built environment levels.

Tower crane optimization is pivotal to ensure proper lifting and handling of materials, and warrant conflict-free work zones. This research, therefore, aims to optimize the position of a tower crane by maximizing its lift ability. To achieve the goals, generative design, a paradigm that integrates the constructive features of mathematical and visual optimization techniques, is used to develop a relatively comprehensible prototype. The first part of the research, thus, utilized generative design on two construction sites and warranted a 40% improvement in the lift score. A pool of potential alternatives was explored and supplemented by the trade-off illustrations. On evaluating the usability aspects by interviewing 12 participants, it has been unanimously observed that the technique has extreme efficiency of usage and can evidently prevent the occurrence of errors. The study concludes by providing recommendations to augment the significance and usability of generative design for tower crane position optimization.

The lack of seamless interoperability between building information modeling (BIM) and building energy modeling (BEM) has been a long-lived problem, which causes two main obstacles: (1) significant time and effort required in building design and energy modeling iterations and (2) limited opportunities for architects to optimize their design informed by BEM. A smooth model information transfer from BIM to BEM can significantly accelerate the BEM process since BEM-related information can be imported from BIM, rather than start from scratch. In the reverse direction, mapping BEM to BIM will provide better opportunities for BEM-informed design optimization and integration with generative design tools, allowing architects more opportunities to improve building designs towards energy efficiency, such as by geometry modifications. In this paper, the authors proposed a new methodology of developing computing technologies to automatically map BEM to BIM following the industry foundation classes (IFC) standard. The authors concluded that the reverse mapping automation from BEM to IFC can be achieved through the state-of-the-art Data-Driven Reverse Engineering Algorithm Development (D-READ) method, based on the invariant signatures of building objects. The developed algorithm was tested on a high-rise residential building model, and it achieved seamless and error-free model transformation. This builds a solid step towards enabling smooth bi-directional modeling workflows between BIM and BEM, and helps move the whole architecture, engineering, and construction (AEC) industry beyond the fairly isolated design and BEM practices towards enabling more opportunities to benefit the buildings’ life cycle from BEM.

Due to the growing construction costs and duration of large capital projects, owners face increasing difficulty in collecting and analyzing cost information during construction. It requires enormous effort to analyze cost items in progress reports with explicit references. This study aims to propose and validate a data tracking tool to achieve smart cost reporting during the construction phase. Through examining the current practice, data consistency and system integration are found as the key challenges. To overcome them, this study proposes a data tracking system architecture and demonstrates the major functionalities, which is able to merge multiple sources and prepare cost information for reporting. In the case study for feasibility validation, the conceptual architecture is applied via the relational schema design of cost information repository, and the procedure to accomplish semi-automatic monthly cost reporting. Further application and evaluation of the tool’s timeliness, accuracy, reliability, and ease of use are reserved for future work.

Digital methods, in particular building information modeling (BIM), are being used more and more in the architecture, engineering, and construction (AEC) sector. However, building operation and maintenance processes mainly still follow conventional processes and are hardly supported by digital databases. Research at Leibniz Universität Hannover investigates a fundamental concept for digitally supported maintenance of railway infrastructure to enhance availability as well as reliability and provide a structured data basis for the involved project participants. The paper will answer the question on how a product model in the context of BIM can be used to evaluate the bridge health combining different data sources. Therefore, a new evaluation method based on a three-dimensional damage characterization was developed, which uses the data from the BIM model to quantify the grade of maintenance need.

To ensure high quality in the construction of industrial plants (e.g., power plants and chemical plants), precise planning and execution of logistics and assembly must be achieved. Therefore, detailed step-by-step assembly instructions are required. In some cases, the visualization of the assembly procedures is already represented with the help of digital models. Our previous work showed that a semi-automatic generation of step-by-step assembly sequences for large scale plant construction is possible. The basic information needed for this is already available in BIM models (building information modeling). This information includes information about components, connections, and construction details. On this basis, a methodology was developed that allows an assembly sequence to be created automatically using a defined set of rules that is stored in a database. The assembly sequence can then be made available to the executing staff in the form of lists and 4D animations. Currently, the developed methodology lacks of a mechanism for selecting specific construction strategies. In this paper, the previous work is extended by developing an approach for the semi-automatic generation of assembly sequences by different construction strategies. This allows the planner to select and adapt the most suitable assembly strategy for their project, so that the assembly sequence can be further optimized. The assembly sequences generated for a project can thus support the planning of the processes on the construction site (storage of construction elements, delivery to and transport on the construction site, as well as assembly).

Three-dimensional thermal mapping has been commonly used for building energy performance audits. Many researchers have focused on single building audits and modeling multiple buildings in a large district through aerial images. To solve accuracy and efficiency trade-offs, researchers have used data fusion methods to project information from thermal images onto a 3D RGB model reconstructed by high-resolution RGB images to generate a 3D thermal model. This approach overcomes issues related to the direct use of thermal images to reconstruct a 3D model that might otherwise fail due to low resolution thermal images. It has been demonstrated that the accuracy of thermal images is influenced by camera shooting angles, the distance between the objects and cameras, and drone flight speed. Researchers have also demonstrated that edge pixels usually introduce more errors than center pixels of a thermal image. These facts emphasize the need for different pixel locations in thermal images to be assigned with distinct importance when generating a 3D thermal model. This paper introduces a novel framework that was developed to support the improved performance of the thermal and RGB data fusion process.

Keeping track of issues and their documentation in architecture, engineering, and construction (AEC) projects demand significant amounts of time, budget, and effort. While various types of documents and software aid coordination in AEC projects, project team meeting minutes, developed as a follow-up to periodic project team meetings, continue to be the most common and prominent type of documentation across project types for recording team communications, tasks, and assignments. Presently, due to its dynamic nature, identifying unique project issues and tracking their progress from meeting minutes is a manual process that is time-consuming and susceptible to error. This study aims to automate the identification of project issues and track resolution timelines using project team meeting minute documents via the Jaccard similarity method. In this study, over 50 AEC project team meeting minutes documents of varying formats from three different projects of various sizes were collected, automatically converted, and coded to train the Jaccard similarity model for detecting new and continuing issues. Accuracy, precision, recall, and F1 parameters were tested, and the accuracy rates of 81.86%–94.18% were obtained. The study provides the groundwork to automate the analysis of issue complexity, detection of bottlenecks, and analysis of expertise assignments for issue resolution.

Since the pandemic, most work environments have changed to home-based settings, which mainly serve for living purposes other than working purposes. There is no lack of studies in the impact of indoor environmental quality (IEQ) factors on occupant productivity in regular work environments; however, limited studies are conducted in home-based work environments, not to mention in the context of the COVID-19 pandemic. Therefore, an online survey was developed to explore the impact of IEQ factors on productivity between office- and home-based work environments among occupants with different genders and ages. A comprehensive list of key indicators was first developed. Then a survey was developed, distributed, and received 204 complete responses. The descriptive analysis and t-test are performed to evaluate the impact difference of all the IEQ factors on productivity of different occupants. The findings indicate that the visual factors’ impact on productivity decreases for both genders, the impact of all factors on productivity for younger occupants increases when work from home (WFH), and the impact of acoustic quality is the highest among all five IEQ factors. Because of WFH as the future of work, this study can provide insights for future built environment design.

Hotel guests’ energy-related behavior is one of the main factors that affects energy consumption in hotel buildings. However, there are few studies that provide practical and efficient energy-use reduction strategies based on guests’ energy-related behavior in hotel buildings. To address this research gap, this study aims to develop an integrated energy-use model in four steps: (1) incorporating four energy-related behavior models (i.e., motivation-opportunity-ability, norm activation model, theory of planned behavior, and pro-environmental behavior); (2) developing a set of hypotheses and their relevant measures to examine the relationship among the energy-related behavior models and hotel guests’ energy-use behavior; (3) conducting an energy-use survey to analyze the effect of each determined measure on hotel guests’ energy-use behavior; and (4) analyzing the energy behavior data to identify energy-use profiles (i.e., prone, indifferent, or resistant to change) of hotel guests. The findings can provide decision-makers in hospitality management with a better understanding of their guests’ energy-related behavior and accordingly develop effective strategies to reduce energy consumption in hotel buildings.

HVAC systems account for 50% of buildings’ energy use and could play a critical role in energy management in buildings both for energy-saving, and demand-side management to reduce peak energy use. According to contextual demands of occupants, efficient control of HVAC systems could result in peak energy saving and decreased energy costs in demand response programs. Accordingly, in this paper, we have introduced an agent-based model consisting of three agents: human agent, thermostat agent, and utility agent. In this model, the thermostat agent receives the real-time electricity price from the utility agent and aggregates thermal comfort profiles of the occupants from human agents. By considering these inputs, the thermostat agent employs a predictive model of a house and calculates the next set point of the HVAC system on energy cost and occupants’ comfort. Then, the thermostat agent signals the suggested set point to the thermostat of the building at each time step. For evaluating the proposed controller’s performance, the electricity price profile from ERCOT, which supplies the state of Texas, is used as a signal from the utility agent. Realistic thermal comfort data were used to simulate the thermal preference of occupants represented as human agents. The evaluation was carried out in a co-simulation using a Python and EnergyPlus model of a residential unit. Our results show that the proposed controller reduces the peak energy by 5.5% to 10% and increases occupants’ thermal satisfaction up to 12%. The main contribution of this paper is developing an agent-based model that humans, as the main stakeholders of the buildings, play a role in controlling HVAC systems for peak reduction and energy saving.

Dynamic models of occupancy patterns have shown to be effective in optimizing building-systems operations. Previous research has relied on CO₂ sensors and vision-based techniques to determine occupancy patterns. Vision-based techniques provide highly accurate information; however, they are very intrusive. Therefore, motion or CO₂ sensors are more widely adopted worldwide. Volatile organic compounds (VOCs) are another pollutant originating from the occupants. However, a limited number of studies have evaluated the impact of occupants on VOC level. In this paper, continuous measurements of CO₂, VOC, light, temperature, and humidity were recorded in a 17,000 sq ft open office space for around four months. Using different statistical models (e.g., SVM, K-nearest neighbors, and random forest) we evaluated which combination of environmental factors provide more accurate insights on occupant presence. Our preliminary results indicate that VOC is a good indicator of occupancy detection in some cases. It is also concluded that a proper feature selection and developing appropriate global occupancy detection models can reduce the cost and energy of data collection without a significant impact on the accuracy.

Bridge deterioration prediction plays a critical role in supporting bridge maintenance planning and prioritization. However, existing bridge deterioration prediction approaches are either model-driven or data-driven—missing the opportunity of integratively learning from both domain knowledge and data for enhanced deterioration prediction. To capitalize on this opportunity, this paper proposes a novel adaptive intelligence-based bridge deterioration prediction framework. It integrates model-driven and data-driven prediction approaches by (1) formalizing the domain knowledge captured by mechanistic structural deterioration models as constraints and (2) using the constraints to regulate and enhance data-driven learning of bridge deterioration patterns. Nevertheless, formalizing domain knowledge is challenging. One specific challenge is to explicitly represent the complex mathematical relationships between the concepts defined in mechanistic structural deterioration models in a machine-interpretable way. To address this challenge, this paper proposes a structural deterioration knowledge ontology (SDK-Onto), which leverages the arithmetic-logical expression to capture and represent the complex mathematical relationships. This paper provides an overview of the proposed framework and focuses on presenting the proposed ontology. The presented research contributes to the body of knowledge by offering a means to formally and explicitly represent domain knowledge captured by mechanistic structural deterioration models in a machine-interpretable way for supporting adaptive intelligence-based bridge deterioration prediction.

The need for efficient information exchange (IE) has been critical more than ever due to the paradigm shift in information technologies of the built environment. Breakdown structures are a driving tool that can result in automated IE among different participating entities. Many authors have acknowledged the strategic benefit of a comprehensive structure for a seamless IE. Nevertheless, existing breakdown structures have not been popular enough to maximize IE’s efficiency due to several limitations. In this context, this paper introduces a promising concept named unified breakdown structures (UBS) that facilitate an effective IE in a comprehensive and shared manner. Firstly, an overview of existing breakdown structures was summarized, and the UBS concept is introduced. Secondly, key variables to examine the feasibility of the proposed concept were identified. Finally, the viability of UBS was examined by investigating hierarchal elements of different literatures. Findings and suggestions for future directions were also outlined.

Pavement management requires up-to-date data from several sources. For a network of roads, the necessary data include: (1) pavement condition survey data such as pavement distress and surface friction; (2) pavement attribute data such as the number of lanes, length, width, surface type, functional class, and shoulder type; (3) traffic data such as volume, vehicle type, and load; (4) pavement construction and rehabilitation history data; and (5) a digital model of the road network. Currently, these data are mostly kept in heterogeneous databases, which makes the integration of data from sources distributed across the Internet difficult. This paper presents the results of an ongoing research project on a semantic approach to pavement management data modeling. The semantic approach creates a vocabulary to define the pavement management data and ontologies that represent contextual relationships behind the vocabulary. To facilitate automatic data integration from various knowledge domains such as the pavement inventory and the pavement conditions knowledge bases, this study uses a shared ontology that defines the vocabulary common to all pavement management knowledge domains; each knowledge domain must extend the shared ontology for developing its own domain ontology. The scope of this paper will be limited to a discussion of the shared ontology and pavement management domain ontology.

The architecture, engineering, and construction industry involves multiple disciplines and activities throughout the building life cycle stages (BLS). To enable collaboration amongst these disciplines iterative and coordinated exchange of information is required. This improves the design process over multiple BLS. Since the last decade, BIM is a well-known approach to achieve collaboration through semantic representation and exchange of domain data. Despite the improvement, there is a lack of efficient implementation and management of building life cycle functionalities in existing BIM solutions, because of their fundamental heterogeneity, complexity, and adaptability. This research focuses on these issues and addresses a clear perception through analysis of BLS from various standards and norms. The paper concentrates on the demonstration of efficient representation of various BLS through the ontological approach and their effective involvement in BIM data management. With the validation and evaluation through SPARQL queries, this paper presents an ontological framework for building life cycle data management.

Industrialized construction is gaining popularity in the industry due to its advantages of reducing pollution, shortening construction time, and increasing safety. However, the design process of industrialized construction projects is different from traditional design approaches, thus should consider not only the architectural and performance requirements but also the manufacturing and assembling requirements. Specifically, during the pre-construction design process, the designers have to consider the constraints of material, shapes, or mechanical property of the components from the assembly and manufacture technicians. In order to reduce design alteration times and improve the design efficiency, it is necessary to strengthen the communication and information interoperability between the designers, manufacturers, and contractors. The use of an ontology, an explicit formalization of a conceptualization that provides an abstract schema consisting of formal definitions of concepts and their relationships, is often applied to achieve semantic interoperability among different systems. The main objective of this paper is to present an ontology to support the manufacturability and constructability checking of prefabricated components. The methodology used in this research for developing the ontology includes four main steps: (1) purpose and scope definition, (2) taxonomy building, (3) relation modeling, and (4) ontology coding. A case study using Revit and a prototype application validates the exchange of information. It is expected that this developed ontology can provide the foundation for the establishment of a comprehensive prefabricated component manufacturability and constructability checking system.

Future-proofed transportation infrastructure planning is a complex process that requires consideration and integration of various external and internal factors. Although existing studies have explored the different factors for transportation infrastructure planning, few studies have investigated the inter-relationships among them. Therefore, this study proposed an association rule mining (ARM) based methodology to study the inter-relationships among key factors for transportation infrastructure planning. First, a list of factors was identified from 48 published documents on future-proofed transportation infrastructure planning via two topic modelling techniques: latent Dirichlet allocation and non-negative matrix factorization. ARM was then used for discovering relationships among these factors. Specifically, two quantitative association rule mining metrics—confidence (frequency of association) and lift (strength of association)—were used. Results showed that stronger associations existed between certain factors; e.g., a significant association was found between societal trends with environmental performance. It implies that in order to achieve better environmental performance of transportation infrastructure, capturing and taking advantage of societal trends could be useful, since societal trends such as less dependency on personal vehicles can significantly reduce the environmental impacts of transportation infrastructures (e.g., less emissions). Such results demonstrate the potential of using ARM to discover inter-relationships among a list of factors based on large text data. It could help transportation planners to better understand the inter-relationships among different factors and use them towards a more integrated transportation infrastructure planning.

Many developments and research have led in last decade to enhanced building automation systems embedding data analytics algorithms and performing energy-optimized building systems control. The presented approach aims at providing a complementary analysis layer compared to such systems by the means of semantic modeling and knowledge reuse. For that purpose, it relies on a semantic description of a building energy system from which energy conservation measures can be derived. Starting from an ontological description of the building energy system, an expert system is proposed that provides a monitoring plan and operation advices to building users and managers. This system consumes for a part real-time building data gathered during its operation. For another part, it relies on information about the built-in technical systems, which is usually available in initial design models. The resulting software application might be used in the future as an add-on to existing building management systems and for easing their configuration.

This paper presents a concept for the integration and application of BIM models for the condition assessment of transport infrastructure. For this purpose, necessary information for condition assessment is exchanged digitally and comprehensibly with the help of information containers in accordance with ISO 21597-1. The application of ISO 21597-1 requires a precise definition of the information to be exchanged, the specification of templates to be used, and formal logic for data links. Another very important aspect is the distinction between information that is contained directly in BIM models and information that is to be linked to external documents. Furthermore, it will be shown how semantic information is handled on one hand and geometric representation on the other hand. The concepts are presented and discussed based on real case studies.

The North Carolina Department of Transportation (NCDOT) created a new knowledge repository called Communicate Lessons, Exchange Advice, Record (CLEAR) as an official platform for end-users to store and retrieve knowledge. Through the CLEAR program, end-users can enter lessons learned and best practices gained in their workplace in addition to soliciting solutions to any ongoing issue. This paper briefly reviews the development of CLEAR and proposes an intelligent knowledge transference process of information on NCDOT projects using natural language processing and knowledge graphs based on neural language models developed by the CLEAR project team. The CLEAR project includes a collection of documented lessons learned and best practices. The AI model learns an inference model of the domain vocabulary from various sources such as contract documents, textbooks, and specifications. This model allows the system to make meaningful connections between lessons learned and best practices within CLEAR and the project-specific domain knowledge. The model output will initially be shown to NCDOT team members belonging to various project life cycle phases such as design, construction, and maintenance to certify the usefulness of the generated keywords and thereby the AI model in an iterative manner until the model has been appropriately fine-tuned. Necessary modifications will be made to the model based on the feedback obtained from project personnel to ensure high-quality output. In the long run, this automation in information retrieval will encourage NCDOT personnel to use the CLEAR program as a part of their routine work to improve project workflow processes.

The interest in smart, emerging technologies and offsite construction methods continues to grow. While previous studies investigated different offsite construction aspects, there is still lack of research work that explored the current and future use of smart and emerging technologies, especially those used in offsite construction. This study addresses this knowledge gap through a multi-step research methodology. First, a list of 24 technologies was identified. Second, a survey was distributed to study the current use and the experts’ future perceptions of the identified technologies in offsite construction. Third, descriptive statistics and computational clustering analysis were used to quantify and classify the current and expected future use of offsite construction technologies. The findings reflected that the industry will experience an increased reliance on all offsite construction technologies in the future. The outcomes also indicated that the top 10 technologies with the highest potential in the future include: (1) drones and remote monitoring; (2) smart sensors; (3) artificial intelligence, cognitive learning, and computer/machine vision; (4) extended reality (i.e., virtualization); (5) integrated real-time project management information systems; (6) wireless technology and nG networks; (7) big data, data analytics, and data ecosystem; (8) robotics; (9) internet of things (IoT); and (10) nD printing and additive manufacturing. This study adds to the body of knowledge by helping practitioners identify strategic technologies that will shape the future of the construction industry and offsite construction.

Understanding the complex and unique requirements of modular construction methods has become crucial given the ongoing increase in popularity of such construction methods. In fact, studying the various modular construction factors and their interconnections to attain successful project performance is perceived to help practitioners choose the modularization strategies that are more suitable for their specific context. To this end, there is a need to better understand the interconnectivities among the various modular construction decision-making factors. As such, the paper aims to enhance the knowledge of the interactions and interdependencies among the various modular construction decision-making factors. To achieve that, the authors identified 50 decision-making factors affecting the use of modular construction in the industry based on an extensive analysis of the literature. Second, a reference matrix reflecting the co-occurrence of the decision-factors in the literature is constructed. Third, spectral clustering algorithms are used to aggregate the decision-factors based on their interconnectivities. Fourth, social network analysis is used to determine the key interconnectivities and combinations of decision-factors in each cluster. The findings show that the 50 identified decision-factors can be categorized into four clusters where each cluster reflects particular interconnected aspects. This paper provides professionals with a proactive approach that enables them to identify interaction and dependencies between the different modular construction aspects. Ultimately, the outcomes of this paper shall enhance a better understanding of the unique and complex requirements of modular construction projects.

Adoption of new technology is a relatively slow process in the construction industry. The growing number of cyber threats has raised a serious concern about the data transmissions. Blockchain technology (BCT) has been introduced to curb some of the privacy and security challenges. It is a technology that decentralizes authority and makes it almost impossible for external threats to alter the transmitted content. Thus, this paper aims to provide an in-depth review of the BCT articles in the construction industry, specifically for its applications in construction management. Five categories were identified for BCT in the construction industry after reviewing the existing literature: (1) secure data transmission and data ownership, (2) IT in construction, (3) quality information management, (4) stakeholders’ collaboration, and (5) construction payments and trust. The findings indicated that BCT presented high promises to improve the overall efficiency of the construction industry. Finally, insights for future research about BCT in the construction industry are provided.

Construction site layout planning is a crucial job with significant impact on project time, cost, and productivity. It pertains to determining the locations and specifications of equipment units and material storage spots. Despite considerable efforts on location optimization of material storage spots, insufficient attention has been given to their specifications. Most conventional approaches used in previous research are limited, as they do not necessarily optimize factors, such as material transportation time, equipment utilization rate, and task execution flexibility, simultaneously, especially for large construction sites. To overcome these limitations, this study applies agent-based modeling (ABM) to develop a material storage management platform. This platform adopts a non-homogeneous storage system to allow for storing multiple material types in each single storage spot. This enables equipment units to feed from multiple storage spots, which in turn allows for further flexibility in task execution; however, this necessitates an in-depth plan to remove potential collisions between operating units and check material availability at storage spots. The platform is tested against numerous task execution scenarios to eventually find the most efficient material allocations during operation. The simulation’s results using a case study of multiple tower cranes and supply locations indicate significant improvement in total schedule time.

Construction has long been notorious for its high accident rates. Topics related to construction safety planning have thus been widely studied worldwide. With the rapid development of various tools and technology, there have been increasing studies investigating automation in safety planning in construction for around three decades. It is interesting that researchers tend to approach construction safety planning with a slightly different focus using different methods and tools. This study aims to review research on automation in safety planning in construction from 2002 to 2021. It covers the topics of automated safety planning for permanent (88%) and temporary structures (12%), using fully (56%) and semi- (44%) automatic approaches. The results not only highlight the need for further research in the areas of automatic temporary structure planning in construction but also reveal major challenges facing automated safety planning, including the lack of construction schedule for integrating work sequence; the labor-intensive, manual rule interpretation process; requirements of explicit model details; and lack of emphases on temporary work safety. In the future, more studies shall be conducted to investigate the application of artificial intelligence in construction safety planning.

The lowest bid approach is the most used technique in the US construction procurement environment since competitive bidding is a legal requirement in the public sector. Despite that many research studies were directed to help contractors in taking different bidding decisions, there is still insufficient research efforts that equipped contractors with decision-support tools to decide on optimal bid values. Hence, this paper takes previous research works a step further by developing a simulation model that can be adopted to determine the bid decision that would result in winning more projects. To this end, a methodology composed of multiple steps was utilized. First, a simulation model was developed to model the competitive bidding environment. Second, the modified Roth-Erev learning algorithm was used to determine optimal bidding decisions. Finally, the applicability and usefulness of the proposed simulation model was validated through a real-world set-up by relying on data from 503 actual public transportation projects in the US. The results reflected that the developed simulation framework can increase the winning percentage of contractors from 36.98% to 68.59%. In addition, the results showed that projects’ owners would also profit when contractors use the developed simulation model. Ultimately, this research adds to the body of knowledge by equipping contractors with a pragmatic decision-support tool that helps in winning more projects by determining optimal bid values.

Health and safety problems are essential for the construction industry, and such problems are more pronounced in small and medium enterprises (SMEs) due to the lack of financial resources and skilled personnel. Researchers have explored the feasibility and viability of addressing such constraints using artificial intelligence-enhanced, low-cost sensor systems. Our previous studies have investigated both conventional machine learning and deep neural network models for recognizing workers’ postures from low-cost wearable sensors and assessing the ergonomics risks for injury prevention. In the next steps for this research, we are investigating adoption drivers and diffusion barriers for the scaled deployment of AI-enhanced sensor networks and other emerging digital technologies for construction health and safety in a real-work setting. Although the COVID-19 pandemic has brought unprecedented challenges, it has also sped up the digital technology adoption. The discussion in this paper is directed at building on this momentum to advance the use of emerging digital technologies at construction SMEs. The authors conducted a systematic review of literature on digital technologies at construction SMEs and how COVID-19 affected the digital transformation at SMEs. After an initial screening of a total of 170 articles, the key publications based on the research questions were selected for a more in-depth analysis. It emerged that although construction SMEs have embraced the use of several digital technologies during the current pandemic, there is still a large digital divide between these companies and larger companies. The research discussed in this paper contributes to efforts directed at addressing this problem through the design and deployment of SME-centric digital technologies for construction health and safety.

Profit is one of the most important factors in managing companies. As construction companies run their business on a project basis, their cash flow differs from that of companies in other industrial sectors. This study empirically analyzes the financial factors and explains the profit of construction companies based on the financial statement data for construction companies in South Korea. This paper presents two approaches for extracting financial independent variables that presumably affect the return on equity (ROE) representing the profitability of a company. The first approach involves conducting an in-depth literature review and extracting variables frequently used in prior research studies. The second approach uses the stepwise logistic regression. In the experiment, the dependent variable, ROE, was predicted by a series of neural network analyses. The experimental results show that the model selecting the independent variables through stepwise logistic regression is more predictive than the model that selects independent variables from the literature review. This study adds empirical evidence of financial factors affecting financial performance in the construction industry.

Several studies have been conducted on various methodologies and techniques to improve construction labor production, but where studies have been lacking is investigating how to measure and improve the efficiency of project managers. It is not uncommon to have major project delays due to unforeseen conditions; how would one objectively compute a productivity factor for that project management team? Too many uncertainties exist in construction to directly link the efficiency of a team with overall project performance. Such issues led to the development of this new methodology, which offers a different perspective to productivity measurement in the architecture, engineering, and construction (AEC) industry. Productivity is how well the project tasks are realized and prioritized while incorporating the factors as given with the data presented at the project planning stage. Whereas in this study considered was the tangible outputs that construction delivers, applied weighted priority values, and factored back to measure a value of efficiency for the project managers, architects, quantity surveyors, and others involved in the construction process. Two numerical equations were derived through which a methodology that could theoretically be applied to any construction project regardless of location. Previous trials of the productivity method were conducted in controlled environments. In this study, the methodology is being explored to adopt productivity measurement at the company level, not just for individual projects. Advancements in the application of project controls need to be realized to allow for this adjustment. Delineated would be the benefits of having efficiency data available to employees and employers alike rendering the cost of adjusting the methodology to project management. Thus, in summary, proposed is a methodology for adopting a different perspective in the AEC industry, examining the data of how well the execution of the work is conducted and exactly where there is need for improvements.

Building information modeling (BIM) has offered substantial improvements in planning and scheduling in the construction industry. However, it is still time consuming to plan and coordinate between team members the detailed requirements and information for the generation of practical installation schedules. During the installation stage, there can be delays or interference that could lead to the failure of the initial schedule plan. To effectively improve the installation schedule, continual learning of the installation site situation is important, with the learnt knowledge being used to generate realistic schedules. This paper proposes a new artificial intelligence (AI) framework with machine learning (ML) and heuristic optimization techniques for automating practical pipe system installation schedule generation and optimization using 4DBIM. A BIM model is used as reference to provide pipe system component information. ML is adopted to map extracted components with the installation activities and their requirements for installation based on knowledge obtained from industry experts and piping codes. In addition, a heuristic algorithm is adopted to optimize the installation schedule. Finally, an optimal installation schedule that minimizes overlapping activities and time is suggested.

Building technologies’ advancement demonstrates the promising potential to utilize information and communication technologies (ICT) and the Internet of Things (IoT) for more efficient and effective building design, construction, and facility management. However, several accompanying issues arise regarding data quality, management, and utilization. Sustainable development introduces another layer of complexity in this data exponential growth era. Therefore, the research first identifies the main issues of current data use and management that hinder sustainable development through a literature review. Besides the commonly found issues of data quality, consistency, and interoperability issues, the lack of domain knowledge, data connection, standardized information requirements, and data correlations are the main impeding factors to support effective, sustainable design and operation. Consequently, the research proposes and demonstrates a Sustainable Development Data Management System (SDDMS) framework that can potentially bridge the gaps and serve as a solid foundation to support advanced data utilization and analysis.

Choosing a proper delivery system is one of the key decisions for the success of infrastructure projects because it affects the duration, costs, risks, and sustainability of the projects. Although plenty of research has been conducted in this area, there are few studies that systematically characterized and analyzed delivery systems research. Therefore, this research aims to conduct a systematic state-of-the-art review of the related research in delivery systems in the horizontal construction industry. First, research scope as to alternative delivery systems is defined. Then, the preferred reporting items for systematic review and meta-analysis (PRISMA) method and scientometric analysis were used to search for the literature on delivery systems in horizontal construction. Finally, quantitative and qualitative analyses were performed on the literature data. This study reveals the growing trend of the number of papers on the intersection of horizontal construction and alternative delivery systems from 2011 to 2021 and finds the main delivery systems and research trends, the heated debates, and future development trends.

In the long-term operation of highway infrastructure, timely monitoring and performance verification of maintenance tasks are integral. Some examples of these maintenance tasks are mowing the landscape of highway areas, detecting debris, patching unfilled holes in the pavement, and identifying and repairing road signs. This study will develop a framework that integrates unmanned aerial vehicle (UAV) image acquisition with image segmentation methods to automate the tasks needed to effectively maintain highway infrastructure. Existing research for UAV-based environment monitoring has limitations in the number of data sets relevant for highway monitoring and did not comprehensively analyze the effect of changing flight parameters. To overcome these limitations, the proposed research investigates the effect of flight parameters on UAV semantic segmentation performance by considering images taken from varying UAV heights and both vertical and oblique camera angles. This research uses a deep neural network based on U-Net to automatically processes the images and segments them into different regions. Efficient training data annotation is also carried out by performing large-scale ground truth annotation through automatic co-labeling of images and point cloud data. Validation experiments were performed on a real highway data set, showing that while the segmentation performance varies by 3%–25% depending on the flight height, the performance only varies by 0.5% depending on the camera angle.

Highway maintenance and infrastructure monitoring tasks often involve labor-intensive activities and long inspection times. Examples of these maintenance tasks include landscaping and lawn care, detecting damaged road segments, and identifying missing road signs. While several infrastructure monitoring methods have been proposed, many of them are only carried out using 2D images and do not fully utilize the 3D geometric information in the scene. Additionally, most methods often downscale the data and did not consider the fine resolution needed for inspection tasks. To efficiently automate the maintenance inspection tasks, this research proposes a new approach that combines a data collection framework using unmanned aerial vehicle (UAV) with artificial intelligence (AI)-driven data processing techniques. Structure from motion (SfM) is used to create dense 3D point clouds from image data, and deep learning techniques are used to segment and classify different highway assets. Point cloud-based temporal change detection is carried out with a focus on grass height estimation for monitoring highway mowing operations. A field highway data set is collected to evaluate the proposed method. Experimental results show that the method achieved 93% semantic segmentation accuracy and 6.31 cm root mean square error (RMSE) in grass height estimation.

In the United States, hurricanes are the most devastating natural disasters causing billions of dollars worth of damage every year. More importantly, construction jobsites are classified among the most vulnerable environments to severe wind events. During hurricanes, unsecured and incomplete elements of construction sites, such as scaffolding, plywood, and metal rods, will become the potential wind-borne debris, causing cascading damages to the construction projects and the neighboring communities. Thus, it is no wonder that construction firms implement jobsite emergency plans to enforce preparedness responses before extreme weather events. However, relying on checklist-based emergency action plans to carry out a thorough hurricane preparedness is challenging in large-scale and complex site environments. For enabling systematic responses for hurricane preparedness, we have proposed a vision-based technique to identify and analyze the potential wind-borne debris in construction jobsites. Building on this, this paper demonstrates the fidelity of a new machine vision-based method to support construction site hurricane preparedness and further discuss its implications. The outcomes indicate that the convenience of visual data collection and the advantages of the machine vision-based frameworks enable rapid scene understanding and thus provide critical heads up for practitioners to recognize and localize the potential wind-borne debris in construction jobsites and effectively implement hurricane preparedness.

In the context of prefabrication and modularization, termination points are defined as local coordinate systems where assemblies are either connected or constrained. These points are typically points of connection between assemblies, sub-assemblies, or modules. As such, it is critical to ensure that termination points are measured accurately. In this study, the impact of point cloud filtering as a pre-processing step for improving the accuracy of detecting termination points in point clouds is investigated. An industrial-scale experiment was conducted where 3D scans of 40 piping components were collected and analysed while being fabricated. For data collection, each piping object was scanned using a laser scanner as well as a SLAM (simultaneous localization and mapping) scanner (80 point clouds were collected in total). The components vary in their design geometry. Using a guided Hough transform, a circle fitting method was developed to find the termination points in the scanned point clouds. It was then shown that applying noise removal as a pre-processing step for the termination point calculation can substantially improve the accuracy irrespective of the source of acquisition.

Hand signals, as one of the common ways to communicate, are widely used on construction sites due to their simple but effective nature. However, they may not always be captured timely or interpreted correctly in construction fields, which easily leads to worker injuries/fatalities, work interruption, and stoppage, etc. This paper investigated whether construction hand signals could be captured and interpreted automatically with deep learning networks. A new data set containing 11 classes of hand signals for instructing tower crane operations is created under different scenes. The created data set is employed to compare two state-of-the-art 3D convolutional neural networks (CNNs), namely ResNeXt-101 and Res3D+ConvLSTM+MobileNet, and measure their hand signal recognition performance. The comparison results indicate that a high classification accuracy (99.0%) and a short inference time (0.21 s/gesture) could be achieved, which illustrates the feasibility of using deep learning networks to achieve hand signal recognition in construction.

The theoretical accuracy of photogrammetry-based reconstruction has been studied intensively in literature, but the problem remains in evaluating the accuracy of the generated point cloud in real practice. Typically, checking the coordinates of ground control points (GCPs) using total station is considered a promising approach. However, the GCPs are artificial objects with clear and identifiable features and have consistent normal vectors or less roughness than natural objects, which cannot be considered as typical samples to evaluate the point cloud accuracy. Therefore, the present paper proposes an innovative method based on cloud-to-cloud comparison to evaluate the accuracy of point clouds generated from the SfM-MVS algorithm. The proposed method utilizes close-range photogrammetry with high accuracy as the referencing point cloud. A case study is carried out on a piece of the bare soil surface, which is about 1,000 m² with about a 1.5 m elevation variance. Compared to the GCPs-based method, the proposed method is more reliable in evaluating the accuracy of the 3D point cloud model.

Scene understanding, such as object recognition and semantic segmentation on images, plays an important role in many existing information management workflows such as progress monitoring and facility management. Current studies on architecture/engineering/construction (AEC) scene understanding often focus on using visual data captured for a particular building at project closeout, which has two primary limitations: (1) A scene understanding model trained based on static images collected at project closeout data often is not useful for understanding an ever-changing scene where facility components could have intermediate and incomplete states, such as construction sites; and (2) many facility components are occluded at project closeout, which creates many challenges when labeling or detecting them. By leveraging as-designed information present in building information models, this paper proposed an approach to generate synthetic data for training semantic understanding models reflecting the changes in site conditions through 4D-BIM and Unreal Engine. The paper contains two primary contributions: (1) The proposed workflow addresses issues with changing scenes by generating synthetic images with ground truth semantic segmentations during any stage of construction based on given schedule information and (2) the proposed method reduced the labeling effort by utilizing the semantically rich as-designed information that exists in a BIM. The proposed workflow was tested on an academic building in its ability to create a useful synthetic data set using the Uniformat (2010) as semantic taxonomy. The experiment results showed that the proposed data augmentation can improve the ability of scene understanding for images captured in changing scenes.

Management and maintenance of existing buildings remains a major problem due to the lack of existing three-dimensional (3D) models and accurate as-built representation. In this paper the authors propose to use 3D scanner technology to capture the as-built and existing conditions of the buildings combined with building information modeling (BIM) as the underlying technology, which is a 3D semantic representation of all the life cycle phases of a building. This paper presents the results from creating as-built BIM models of existing buildings, using point cloud (a set of points in 3D space) and machine learning as an intermediate medium. Machine learning methodologies are used to speed up the computation of segmentation and classification of point clouds from a 3D virtual indoor environment using procedural modeling, which focused on two attributes, point density and the level of random errors. In this paper we will present findings on the evaluation of the performance of machine learning segmentation and classification algorithm based on the comparison of ten different point cloud data sets. Different sets of segmentation and classification models with comparison between models and within themselves were provided, which included the mean loss and accuracy between models with different point density.

Conventional methods of traffic monitoring primarily rely on sensors, on-board GPS, and human observation, which are generally resource-intensive, slow, and subject to implementation limits. Computer vision (particularly convolutional neural networks) is a more flexible, convenient, and cost-effective alternative to extract traffic information from traffic videos. This paper presents a methodology to extract traffic volume counts from cameras on Texas A&M University (TAMU) campus. Particularly, an object detector model, namely YOLOv5 (You Only Look Once), is integrated with an object tracking algorithm (Deep-SORT) to track vehicles in camera views. These tracks are projected using homography transformation from the local coordinate system of the camera onto an orthogonal map (world coordinates) along with information such as unique vehicle ID, type, position, and time stamp. This data is used to count vehicles that cross the pre-defined study zones (lanes) on different directions. A total of 60 samples (each 1 minute long) recorded during a 24-h period on TAMU campus are manually annotated and used to benchmark model predictions, resulting in a traffic volume count accuracy ranging from 68% to 93%. Moreover, the best input pixel resolution is determined to be 1,280 × 1,280 by measuring the average precision for each class of interest.

Modular construction is increasingly being seen as an attractive method for delivering building projects due to advantages in safety, quality, and lead-time. Despite these benefits, this method still relies heavily on human labor, which causes variability in factory assembly-line performance that can erode performance benefits of modular construction. Continuous improvement methods can alleviate some of these issues, but they also require continuous monitoring of human workers’ performance. Due to limitations of manual time study and automated sensor-based monitoring methods, recently computer vision-based methods have gained momentum in identifying the activities of construction workers from the videos of onsite construction. Therefore, this paper explores the use of computer vision-based human activity recognition techniques to identify and classify worker activities in modular construction videos. Computer vision-based tracking method has been used to track the human workers in each frame, and Resnet-50 network has been used to classify the activity of tracked workers. Evaluation of this framework has achieved higher than 90% accuracy and recall in testing.

With the growing prevalence of site surveillance cameras and rapid developments in computer vision techniques, much effort has been made to develop vision-based safety monitoring methods. An excavator is one of the most used pieces of construction equipment, yet its revolving mechanical arm poses major risks for struck-by accidents. Previous works that leverage computer vision techniques have attempted to monitor such risks by capturing a 3D representation of the excavator arm using either stereo vision or deep learning models. However, stereo vision requires rigid camera setups, while deep learning models require vast amounts of annotated 3D excavator pose, which are cumbersome to collect from real excavators. These drawbacks limit the practicality of such methods in actual construction applications. Therefore, we propose an optimization-based algorithm capable of estimating 3D excavator poses using monocular camera images with no dependency on annotated 3D training data. Specifically, the proposed algorithm attempts to find the optimal 3D excavator pose by imposing rigid kinematic constraints (e.g., arm length and bending directions) while also minimizing re-projection joint errors in 2D. The kinematic constraints help to resolve the inherent depth ambiguity from 2D images, while the re-projection 2D joint errors serve as the objective function for the optimization process. Tests using synthetically generated data sets showed a mean 3D location error of 0.16 m for key excavator joints, demonstrating the capabilities of this proposed optimization-based method. Moreover, the proposed method does not require the demanding preparation works typically required for collecting 3D excavator poses, making it a practical tool to facilitate excavator safety monitoring in actual construction environments.

It is well accepted that data normalization is an essential and imperative step in using the direct linear transformation (DLT) method for homography estimation. Nevertheless, the existing data normalization methods are either point or line-based, thus making them inapplicable in scenarios where both point and line correspondences are available. To fill this gap and further widen the applicability of the DLT method, this study proposed a unified data normalization method for homography estimation using combined point and line correspondences. In this method, the existing point-based normalization method was exploited and then extended to normalize the line coefficients by conducting extra data processing created by the authors. Finally, a unified normalization method was developed by fusing the normalization principles implemented on the point coordinates and line coefficients alone, which allows for choices of using different combinations of point and line correspondences for homography estimation. The field results showed that the proposed method can significantly improve the accuracy and offer higher reliability for homography estimation.

The power infrastructure in the US is facing many challenges concerning capacity, reliability, and sustainability. Some of those challenges are associated with the integration of distributed energy resources (DER) into the conventional grid system. DER may offer reliable and cost-effective distributed small-scale generation near or at end-consumers. However, it also creates new challenges for utilities and generating companies due to the uncertainties in estimating demands. Accordingly, the goal of this research is to investigate dynamic pricing in electric power markets considering the effect of the increasing penetration of DER. To achieve that goal, a complex system-of-systems model that combines agent-based modeling and machine learning is presented in this research. The model relies on reinforcement learning to enable dynamic pricing of electrical power in response to the shifts in demand resulting from the adoption of DER. Results of the emergent behavior of the complex SoS verify the interrelated causality between prices of electrical power and the penetration rate of DER, which is also called the death spiral. Ultimately, the presented framework should assist researchers and practitioners in the fields of electric power infrastructure and DER in investigating the pricing dynamics of agents in the electric power markets considering the effect of the penetration of DER.

Wildfire caused by faults in power distribution infrastructure is a typical threat in the wildland-urban interface (WUI) area. With advanced sensing technologies and deep learning approaches, several machine vision-based surveillance methods have been developed to determine the contact point of power distribution infrastructure and vegetation. However, there is still room for improvement, especially in terms of dealing with uncertainty associated with the risk of failures by a sudden gust of wind that does not reflect historical patterns. In this regard, this research aims to assess the impact of wind on wildfire ignition probability. This research evaluates the likelihood of fire ignition that could be induced by tree encroachment into the minimum vegetation clearance distance zone surrounding power lines. The result would be expressed as a set of coefficients that describe the effect of wind on reducing the ambient air temperature and arc duration and enlarging the interface area. As a proof of concept, the wildfire risk variance following wind speed is analyzed based on the point cloud data of power grids. Possible applications of this research include identifying the potential fire ignition sources near power lines, and ultimately, this research can be used to advance probabilistic analysis for supporting decision-making to explore how vulnerable environment interacts with the associated urban infrastructure.

The pandemic of COVID-19 has caused severe disruptions in urban lives. Understanding and quantifying these disruptions is important to inform the development of targeted and effective measures to control the pandemic and its impact. One way of achieving this object is to measure the urban mobility perturbation caused by the pandemic. In this study, we built mobility-based networks for seven major metropolitan statistical areas (MSAs) across the United States in the years of 2019 and 2020, respectively. We quantified the disruptions of urban mobility by computing and comparing a set of network-based metrics before and during the pandemic. The proposed approach is able to uncover the impact of COVID-19 in cities and provides new insights into the resilience of cities when facing large-scale disasters.

Housing resilience planning is a dynamic process that involves multisector stakeholders, including public agencies, private industries, nongovernment organizations (NGOs), academia, and community residents. Despite the importance of multisector stakeholder collaboration, there is limited understanding of stakeholder collaboration in housing resilience planning. To address this gap, this study analyzes how multisector stakeholders collaborate in producing housing resilience-focused plans, reports, and guidelines utilizing social network analysis (SNA). A two-mode, stakeholder-document, SNA model was built based on secondary data collected from 39 documents on housing resilience in three regions, including the City of Miami, the City of Miami Beach, and Miami-Dade County. The network analysis shows that there are significant differences in network measures across different stakeholder sectors. The findings from this study could offer insights on how to facilitate more effective and collaborative housing resilience planning.

Healthcare facilities are one of the most energy-intensive building categories due to their advanced medical equipment and continuous energy usage patterns. This is more crucial in large hospital campuses, which are consisted of multiple buildings providing a wide range of services with various functional units. Previous studies have shown that the building energy performance can be affected by its spatial layout in terms of spaces’ heat gain and solar exposure. However, the relationship between the spatial layouts of healthcare facilities and their energy consumption at the campus scale was rarely explored. To address that, in this study, the authors propose a layout planning framework for healthcare facilities to optimize their energy performance. The geometric models of healthcare facilities are developed and imported into energy simulation software. Reasonable ranges of values for layout parameters can then be extracted from similar healthcare facilities which serve as the basis for parameter optimization. Finally, the optimal layout parameter configurations can be identified by comparing the energy performance of healthcare facilities using different parameters. To validate the proposed framework, three municipal hospitals located in Shanghai were chosen as the case study, in which the building use ratio (BUR) was mainly investigated. The results showed that better energy performance can be achieved by decreasing the BUR of outpatient, emergency patient, and medical technology units within an acceptable range. It is expected that this framework can serve as a tool to assist the layout planning decisions for hospital managers and designers to improve their facilities’ energy performance.

Subsea engineering operations heavily rely on remotely operated vehicles (ROV), and the performance depends on the seamless interaction between ROV and the human operator. Due to the dynamics of subsea environments such as the uncertainty of turbulence, affected visibility, and the interference with subsea ecosystems, control of subsea ROV is challenging for human operators who have never exposed to such an environment. Even with the tremendous amount of practice, the human operator can easily misjudge the situation and make wrong decisions in subsea operations. This research proposes an intuitive human-robot collaboration method based on mixed reality system (VR/AR). To explore the potentials of utilizing a mixed reality system to accomplish ROV control, a haptic suit is combined with VR goggles to bring an immersive subsea environment to the operator. A comprehensive physics model is first transferred to VR equipment, depicting a detailed underwater environment including turbulence and interaction with ambient objects. Then the haptic suit executes the corresponding pressure signals to the operator’s body in real-time. The feedback signals from haptic suit provide the human operator with realistic sensation of subsea situations peripheral to the robot. This immersive virtual environment ensures ROV operators of real-time awareness of the proximity conditions and prediction of changes. As a result, a less-trained human operator can pilot the ROV based on his/her intuition and experience to maximize the performance and avoid potential mistakes.

Construction professionals have started exploring several UAS configurations and techniques to ideally enhance the quality of the UAS-derived point clouds. This study aims at evaluating the factors [i.e., single-frequency GPS (L1), dual-frequency GPS (L1/L2), post-processing kinematic (PPK) georeferencing technique] affecting the accuracy of UAS-derived point clouds. Results showed that the use of a customized UAS that is equipped with dual-frequency GPS (L1/L2) and subject to PPK coordinate correction improved the overall point cloud accuracy. The outcome of this study is expected to benefit both practitioners and researchers in better understanding the effects of the system on the resulting point cloud accuracy.

Studies suggest that fall accidents are one of the leading causes of fatalities in the construction industry. To mitigate the risk of fall hazards, an automated site inspection system is crucial, considering that conventional methods of monitoring construction sites are labor-intensive, time-consuming, and error-prone. Affordable unmanned ground vehicles (UGVs) can provide several advantages, as they are cheap and can operate long hours in a congested indoor construction setting. Toward that end, this research study aims to assess the feasibility of an affordable UGV for automated detection and localization of fall hazards in indoor construction environments. Notably, all the slipping, tripping, and falling hazards in the indoor construction environment were considered as potential fall hazards for the purpose of this research. The proposed robotic system is assembled by four low-cost hardware modules, which improves upon the affordability of the available automated robotic system. The objective of this study was achieved by leveraging a deep learning algorithm to identify the potential fall hazards. Further, hazard locations were marked onto the map created using Hector SLAM. The proposed system can contribute to the safety management of construction workplaces by effectively detecting potential fall hazards.

The use of robotics has demonstrated numerous advantages in civil engineering applications. Telerobotics is expected to play a key role in achieving a successful civil engineering project in humans’ hard-to-reach environment, such as scenes of nuclear leakage and mining disasters. Due to the high accuracy and flexibility, the robotic arm has gained popularity in telerobotics. Most existing approaches have focused on teleoperate robotic arms based on the pose (position and orientation) of the end-effector, leading to the risk of the pose of other joints moving in undesired directions and cause collisions. This paper proposes an innovative control method that teleoperates the robotic arm based on the pose of the end-effector and another joint between the start joint and the end-effector. The Baxter robot (Rethink Robotics) was developed as the testbed to be controlled based on a VR controller (end-effector) and a VR tracker (middle joint) in a pipe skid maintenance task. The human operator held the VR controller in his hand, wore the VR tracker on his elbow joint, and operates Baxter to complete the task. The inverse kinematics (IK) was used to calculate the rotations of the rest of the joints based on the poses of the start joint, middle joint, and end-effector. The proposed method can control the robot to move according to the pose of the human arm and reduce the risk of collision.

The rapid development of advanced robotic scanning and navigation technologies such as simultaneous localization and mapping (SLAM) enables a more automated building scanning process. To enable a more efficient robot-based scanning, more evidence on advantages and limitations of different SLAM methods is needed. In this paper, we present the findings of a comparison study on the performance of two SLAM methods, including the depth camera-based SLAM and 3D LiDAR-based SLAM, on a quadrupedal ground robot. The SLAM navigational effectiveness and computing cost were compared and analyzed. The result shows that 3D LiDAR SLAM outperforms depth camera-based SLAM in point cloud modeling quality. More detailed spatial information is provided by the 3D point cloud model. On the other hand, depth camera SLAM completes modeling faster with relatively lower quality, which fits more in time-urgent tasks with lower model quality requirements.

Architectural building technologies have been routinely adopted from the production of consumer goods, ship building, and automotive industries. One example of an automated assembly machine created by architects for architecture is the Location Orientation Manipulator (LOM) developed by architect/educator Konrad Wachsmann with his Ph.D. students John Bollinger and Xavier Mendoza (1969–1971). This visionary machine, designed to study building kinematics and architectural building assembly, anticipated today’s multi-axis and gantry-based construction technologies. In this paper, the LOM is virtually reconstructed using Autodesk Fusion software, resulting in a comprehensive parts catalog, an inventory of technologies employed, procedures of use, motion capabilities, and machine design principles. As part of a larger research project, this focused study of the LOM as case study facilitated through an Autodesk Fusion model—essentially a lost architectural building machine reconstructed—allowed the researchers to study the LOM’s capabilities and limits. The reconstruction critically extending Wachsmann, Bollinger, and Mendoza’s machine to the present day, where it can be more accurately studied and articulated, to inform contemporary automated building technologies specific to architecture. As part of further work, the team will adapt replicate and study the kinematics of the LOM in relation to a contemporary gantry based Universal Building Machine (UBM).

Efficient implementation of robots in dynamic, human-populated construction sites will require the assistance and supervision of workers. Several hazards exist when workers are within the robots’ working envelope, including high mental stress induced by the robot as well as the risk of physical collision. To mitigate these risks, this study proposes a cyber-physical system (CPS) for safe, worker-centered human-robot collaboration (HRC). The physical components of the CPS include a wearable biosensor for assessing workers’ mental conditions and a video camera for evaluating the risk of collision accidents. The cyber component incorporates a machine learning layer that employs a kernelized logistic regression to translate worker bio-signals into distinct physiological states. Simultaneously, the system monitors the relative distance between workers and robots by applying computer vision techniques (i.e., video frame differencing, dilation, and contour detection). To test the feasibility of the proposed CPS, six subjects were asked to perform a series of bricklaying tasks jointly with a vehicle robot with different risks of collision (various distances from the robot) and at different levels of cognitive load (various tasks’ complexity). The findings revealed that the proposed CPS allows the co-bot to adjust its speed relative to the subject’s physiological states with 93.70% accuracy. In addition, the proposed system shows promising performance in collision avoidance with the accurate monitoring of position between robots and humans. The results unfold the feasibility of the proposed CPS in enhancing safe and intelligent HRC in construction under the trend of Industry 4.0.

Operation and maintenance (O&M) of building indoor spaces involves various tasks, including occupancy monitoring. Occupancy monitoring is an important task in O&M of large buildings that impacts other tasks, such as heating and cooling control and security control within the building. Despite the importance, existing approaches for occupancy detection using static sensors have several limitations, such as the need to install many sensors to allow a thorough observation of the entire building and poor performances of the sensors not adaptive to dynamically changing conditions caused by dynamic placements and movements of objects. On the other hand, mobile robots with autonomous navigation, sensing, and perception capabilities have the potential to overcome these limitations. This study presents a method of using mobile robots for automated occupancy monitoring as an example of O&M tasks. In this study, a mobile robot autonomously navigates to target locations to detect and count the number of occupants based on the sensor data collected from a camera. The proposed approach was tested in a simulated environment of an educational facility with predefined occupancy requirements for classrooms and offices. The results demonstrate the potential of mobile robots in dramatically reducing the number of static sensors and further enhancing the performance of occupancy monitoring in large buildings. Even though this study simply checks the compliance to occupancy limits, accurate occupancy monitoring can provide crucial input for other O&M tasks like air-conditioning and lighting control.

Construction progress monitoring is a critical inspection activity. It is mostly a manual process performed through in-person visual inspection, which is time-consuming and prone to human error. The key in progress monitoring is accurate, timely, and regular data collection and analysis of the construction process. Autonomous data collection with mobile robots can potentially improve the accuracy of data acquisition for construction progress monitoring. This study uses a four-legged robot to automate data collection in construction progress monitoring. The goal is to study the navigation accuracy of the legged robot in the on-site data acquisition process of progress monitoring. The research studies human operator positioning in the data collection process and undertakes a set of experimental investigations with the legged robot to retrace the path recorded on its controller and analyze the accuracy of the robot's autonomous navigation in multiple missions. Findings reveal an Autowalk position error of 0.283 m to 0.633 m from the recorded path.

Traditional construction monitoring relies on specialized personnel to compare the prescribed steps in the construction procedural documents with executed procedure on the jobsites. This requires demanding manual efforts and may introduce human errors. To reduce manual efforts on collecting and integrating information for construction monitoring, the authors propose an integrated framework to process both workflow information from construction procedural documents and operation information from construction jobsites, in parallel and then comparatively. The proposed framework helps convert the traditional labor-intensive construction monitoring and checking process to an automated one, through: (1) integrating natural language processing (NLP) and sensing (i.e., computer vision) techniques to automatically extract, analyze, and process information both in the procedural documents and on construction jobsite and (2) automating the comparative analysis of these two types of information and the generation of a monitoring report for users regarding sequence correctness of execution steps. To evaluate the proposed framework, an experiment was conducted where execution steps from open-source specifications and the corresponding synthesized sensing data of a construction jobsite were automatically processed and comparatively analyzed. The results were compared with a manually developed gold standard and achieved 83.33% recall and 71.43% precision in the construction procedure compliance checking.

The in-pipe inspection is one of the high frequent and challenging operations for facilities maintenance. Because the complex inner geometry and hazardous environments of pipeline systems are less accessible to humans, the in-pipe snake robots have been developed to facilitate the inspection of defects, corrosions, and the sampling of sludges. However, snake robot control is still nontrivial due to unique movements (such as helical motion) and environmental constraints like confined areas. The unique locomotion mechanisms of snake robots make it harder for the operator to perceive the remote environment and navigate the robot. This paper introduces an upper-body haptic simulator to support the snake robot controls with the haptic assistant. To prove the concept, a virtual model is built to simulate the in-pipe environments and the locomotion of a snake robot with a helical corkscrew motion. The proposed method collects signals relative direction of the gravity to the snake robot and convert the gravity direction information as vibrations of different magnitudes via an upper-body haptic suit with 40 vibrators. To demonstrate the potential of our simulator, a pilot study was performed, and the results showed that the operator could easily control the snake robot with the desired velocity in different shapes of the pipeline (straight pipes, T joints, and L joints) both horizontally and vertically.

Power system restoration is a complex process due to: (1) the dynamic restoration needs arising from completed restoration tasks and new outage events and (2) complex interactions among various stakeholders (e.g., customers, restoration managers, and crew). This study proposed an agent-based modeling (ABM) framework that can simulate the complex restoration process in dynamic environments by capturing the micro-behaviors and interactions of multiple restoration agents. Four types of agents were considered in the proposed ABM, i.e., restoration manager, crew, utility poles, and critical facilities. Each agent had their own attributes (e.g., utility pole failure probability) and interactions with one another (e.g., crew assignment made by restoration managers) which resulted in higher-level restoration outcomes. The proposed ABM was tested in three real-world storm-induced power outage events with different levels of severity in Connecticut. Two sets of simulation experiments were conducted to quantify the effects of: (1) decision-making frequency and (2) risk attitude of restoration managers. The goal was to identify effective restoration strategies that could reduce the negative social impacts associated with power outages. Results showed that more frequent decision making and risk averse restoration managers could substantially reduce the social impacts of power outages. However, the magnitude of the effects varied across power outage events with different severity levels. The proposed ABM provided a useful tool for utility companies to quantitatively assess the potential restoration outcomes and make informed restoration decisions.

The planning of tower crane layout and operation is a difficult task. Improper planning can pose significant problems at large construction sites, especially those in which multiple cranes operate with complex interactions. While past research has made significant progress toward the layout planning of multiple tower cranes, relatively insufficient focus has been placed on crane operation planning. In fact, many projects involving multiple cranes suffer from limited flexibility imposed by a site’s spatial constraints, requiring further investigation on the operation planning of cranes beyond layout development. In response, this research develops a method of daily operation planning with multiple cranes, by integrating two simulation technologies, i.e., agent-based simulation (ABS) and building information modeling (BIM). This method explores the concurrent and alternative operational options of multiple cranes, which might sacrifice execution time of individual tasks but improves overall multi-crane operational performance. Using an actual construction project, computer simulations are conducted, and the results are compared with a commonly used method based on minimized individual task execution times. The results of the case study indicate that the developed method is able to improve the overall schedule time of multi-crane operations, with an average of 9.63% improvement in overall operation time.

Stacking plans of prefabricated exterior wall panels determine how the panels should be stacked and transported, and to generate these plans, practitioners must carefully consider parameters that affect the logistics processes of stacking, transportation, and installation. In the previous study, the authors developed the  panel stacking plan generator (PSPG), a methodology that allows practitioners to generate stacking plans for prefabricated exterior wall panels based on input parameters such as panel dimensions, trailer dimensions, installation sequence, reshuffling effort cutoff, and stability cutoff. The methodology can inform practitioners about how the generated stacking plans impact logistics processes and how much cost, schedule, and safety benefits they can realize by using prefabricated wall panels. However, the methodology does not inform them about how the generated plans affect the environmental sustainability, especially CO₂ emissions from transportation. In this research, the authors conducted a sensitivity analysis using stacking plan data from a real construction project to study how fuel economy (FE), distance, payload weight, and gross combined vehicle weight (GCVW) each impacts the CO₂ emissions from transportation. The results show that emissions were more sensitive to FE (–0.83 to –1.25 relative marginal rate) and distance (directly proportional) than payload weight (0.024–0.027 relative marginal rate) and GCVW (0.10–0.09 relative marginal rate). These insights can inform practitioners about how to realize environmental sustainability benefits when determining stacking plans and logistics of prefabricated wall panels.

In the United States alone, annual energy use has soared to over 100 quadrillion BTUs in recent years. The electrical power sector is responsible for an increasing portion of this consumption, and there is a need for increased research to discover opportunities for reducing these statistics globally. Nonintrusive load monitoring, also known as energy disaggregation, has been touted as an inexpensive mechanism to obtain higher visibility into electrical demand patterns. Fundamentally, energy disaggregation is a signal unmixing problem that attempts to extract appliance-level patterns from aggregate electrical measurements. State-of-the-art approaches for disaggregation today are dominated by deep learning models. For the most part, these models cast the problem as a single task (e.g., regression, classification, clustering). To add to this body of work, we explore techniques from a sub-field of deep learning called multi-task learning for its potential in estimating electric consumption on multiple appliances at once through parameter sharing. We developed and evaluated architectures for modeling conditional source separation across multiple appliances targets simultaneously using simulated high frequency data. This solution is desirable due to its high flexibility and scalable properties as compared to traditional single network appliance-based disaggregation. Benchmarks for performance were measured using MSE, MAE, MedAE, and sequence-wise RMSE to look at effective grouping of tasks. Our preliminary results suggest that combining disaggregation tasks of appliance groups in particular configurations may outperform up to 90% (RMSE) to single-task methods. We expect our results to open the door to more multi-task studies in this field.

Building energy consumption comprises a highly complex set of dynamics that critically depend on the behavior of building occupants. A key area of ongoing research is the simulation of energy consumption for evaluation of occupant-centric building design or management interventions (e.g., building layouts, suggestions to occupants). Researchers have shown that standard physics-based simulation tools like EnergyPlus exhibit large errors when compared to measured data, which has largely been attributed to difficulties in handling occupancy data. To address this performance gap, scholars have noted the promise of data-driven simulation tools. However, a key remaining challenge is the integration of granular (e.g., zone- or room-level) occupancy data for prediction of energy consumption at these granular scales. Such analysis can enable rapid simulation of the effects of new building states (e.g., occupancy patterns) for evaluation of potential building interventions (e.g., schedule-shifting). In this paper, we use real-time data from ambient sensors in a test bed and train a data-driven model to predict HVAC energy consumption using occupancy, temperature, humidity, and temporal features, each collected with high spatial and temporal granularity. We find that our data-driven approach outperforms EnergyPlus, even when adjusted with calibration tools [CV(RMSE) improvement of 49% to 84%]. We also find that our trained data-driven model is able to make energy predictions in near real-time (sub-second), which enables rapid evaluation of building interventions. Overall, this sensor-based and data-driven approach to building energy simulation can enable new occupant-centric building management strategies that save energy and promote well-being in the built environment.

This research presents a simulation-based genetic optimization approach to optimize earthmoving operation by selecting the optimal fleet configuration. The paper underlines the relation between the profit and loss (P&L) function and the utilization of each equipment type within the fleet. The proposed methodology investigates the impact of the optimized fleet configuration on the emissions cost. The model monitors the emissions cost and the P&L function, against the total cost of ownership (TCO) of each of the loading and hauling machine types, as the number of haulers in the fleet varies. The methodology is applied on a real case study that took place in Lebanon. The proposed approach is shown to yield significant improvements in the objective function of the optimized fleet selection, and investigates the relationship between the operational efficiency and the total pollutant emissions.

Construction planning is a fundamental task in the management of construction projects. The most common modeling methods for construction planning are linear scheduling (LS), the critical path method (CPM), and discrete-event simulation (DES). While LS and CPM have advantages of visual feedback and simplicity, respectively, DES as a technique for simulating the behavior and output of a real-world process, facility, or system is the most versatile of these methods. However, when the constraints and dimensionality of a planning problem increase, DES methods are cumbersome and struggle to accurately reflect decision options. As an alternative to DES, deep learning artificial intelligence (AI) methods can, through the use of reinforcement learning, more rapidly review and recommend more planning options for scheduling complex construction projects. The goal of this paper is to examine the feasibility of deep learning AI as an alternative to the DES in an existing project planning method, named Foresight. Foresight is a graphical constraint-based method of planning manufacturing and construction processes that uses DES to simulate possible outcomes. This paper presents the process changes necessary in Foresight to accommodate deep learning algorithms, composed of multiple network layers. In addition, challenges of implementing deep learning AI in construction planning are discussed. Early project planning leads to improved performance in terms of cost, schedule, and operations, balancing the competing needs of a project. A new modeling paradigm is proposed that is better suited to the needs of contemporary construction project planning.

Hurricane is one of the costliest natural disasters affecting the United States. Facing the threats of hurricanes, homeowners usually have to rely on general information and instructions to understand their level of risks. The knowledge of structural resilience to hurricanes at individual residential building level is often limited, which hinders personalized disaster preparation and mitigation decision-making. To address this limitation, a new method of customized and interactive structural analysis is proposed. The proposed method includes three steps: input collection, modeling and analysis, and result presentation. First, various structure information (e.g., type of building frames and material) will be collected from individual users/homeowners based on their localized knowledge through a list of interactive questions. Second, customized finite element models (FEM) of residential buildings will be automatically built based on the collected information using Ansys Parametric Design Language (APDL). FEM analysis will be conducted under hypothetical hurricane-induced loads with randomized wind speed and direction. Finally, analysis results such as the percentage of failed sheathing components and nailed connections will be presented to the user indicating the risk level of a specific residential building under the impact of hurricanes. The proposed methodology provides a useful tool for individual homeowners to enhance their knowledge and awareness of disaster risks. In addition, the method could facilitate a better understanding of the resilience of building structures at the community level through distributed computing and crowdsourcing with less centralized effort.

Simulation of indoor airflow in heating, ventilation, and air conditioning (HVAC) system in construction design is critical in order to control spread of volatile organic compounds (VOC) or germs and viruses in the case of a pandemic, as we witnessed during 2019–2021. This paper presents the findings from comparing two types of fluid simulations: grid-based and particle-based methods. The former is widely used for scientific computation due to its precision; however, it is time consuming and requires designers to do pretreatment of the building model for airflow simulations in construction design using grid-based methods. The particle-based method is used in visual effects, games, and other applications requiring real-time simulation. This paper presents a review of the literature on different methods of fluid simulation like finite volume method (FVM), smoothed-particle hydrodynamics (SPH), position-based dynamics (PBD), and provides a comparison of heat transfer between particle-based and grid-based methods in indoor airflow simulations with HVAC system.

Recently, researchers have paid increased attention to the social and cognitive aspects of workers’ safety behaviors. In the same vein, researchers have begun modeling the socio-cognitive process of workers’ safety behaviors. However, an inquiry into the impact of workgroup changes on social influence remains limited. Workgroups in a construction project dynamically change because workers enter and leave according to the project schedule. Considering that changes in workgroups affect the social network of the project, and the transmission of the social norms occurs through the social network, workgroup changes should be considered. In this paper, an agent-based model is developed to investigate how the dynamicity of workgroup changes affects the socio-cognitive process of workers’ safety behavior. Three experiments examining the effect of the dynamicity of workgroup changes [i.e., static (no turnover), mildly dynamic (modest turnover), and highly dynamic (high turnover) change] in three site risk conditions (i.e., low, modest, and high site risk conditions) were conducted. The results indicate that while the incident rates for the static and mildly dynamic changes are significantly higher than the highly dynamic changes in the modest and high-risk condition, the differences in the incident rates for the three dynamicity changes are not significant in the low-risk condition. The findings provide construction practitioners with insight into the development of safety management interventions by exploring the interaction between the workgroup changes and site risk conditions. Also, this study contributes to the body of knowledge of construction safety by testing the importance of the workgroup changes in the socio-cognitive process of workers’ safety behaviors.

The construction industry has traditionally been a labor-intensive industry. Typically, labor cost takes a significant portion of the total project cost. In spite of the good pay, there was a big gap recently between demand and supply in construction trades position. A survey shows that more than 80% of construction companies in the Midwest of US are facing workforce shortage and suffering in finding enough skilled trades people to hire. This workforce shortage is also nationwide or even worldwide in many places. Construction automation provides a potential solution to mitigate this problem by seeking to replace some of the demanding, repetitive, and/or dangerous construction operations with robotic automation. Currently, robots have been used in bricklaying or heavy-lifting operations in the industry, and other uses remain to be explored. In this paper, the authors proposed a feasibility breakdown structure (FBS)-based robotic system method that can be used to test the feasibility of performing target construction operations with specific robotic systems, including a top-down work breakdown structure and a bottom-up set of feasibility analysis components based on literature search and/or simulation. The proposed method was demonstrated in testing the use of a KUKA robot and a Fetch robot to perform rebar mesh construction. Results showed that the overall workflow is feasible, whereas certain limitations presented in path planning. In addition, a smooth and timely information flow from the Fetch robot sensor and computer vision-based control to the two robots for a coordinated path planning and cooperation is critical for such constructability.

With the advent of personal comfort models in predicting thermal state, investigating individual differences, including gender, age, and race, has been widely conducted to predict personal thermal comfort. The personal comfort model is a good complement to the conventional predicted mean vote (PMV) model; however, the efforts to understand individual differences to date have been quite fragmented in terms of intra-individual differences. The underlying assumption of the current personal comfort models is that the generalized one personal model is universally applicable to any sub-personal variants. However, this assumption does not hold if intra-individual variant subgroups exist, and these subgroups differ in their thermal comfort state. Given that a person can have different thermal preferences from day-to-day and time-to-time under the same environment, interpreting one’s thermal comfort with one single model is not enough to understand intra-individual variances. To address such research gap and better understand human thermal comfort, this study aims to investigate underlying sub-personal thermal states with differential thermal comfort responses building upon the model-based (MOB) recursive partitioning model. To validate the proposed approach, building occupants’ thermal responses and their corresponding physiological data were collected through field experiments, and the performance was compared with current personal comfort modeling approach. It is worthwhile to revisit the overlooked intra-individual differences and explore appropriate modeling methods for better understanding of building occupants’ thermal comfort.

One of the central goals of affordable housing is to provide a decent and suitable living environment, which includes ensuring the housing is of good quality and is economically affordable. It is also vital for housing to continue to maintain these qualities throughout the building’s life cycle. Incorporating sustainable building practices can improve energy efficiency, reduce the waste of resources, and provide healthier living environments for the occupants. With the increase in intelligent technologies, opportunities exist to enhance sustainability and affordability in residential communities by utilizing smart technologies. This paper involves a critical review of the literature involving sustainable buildings and affordable housing and seeks to identify opportunities to implement smart and connected technologies while enabling those communities to achieve sustainability goals. In addition, this study highlights the current trends such as building-to-grid connections, distributed energy resources, and the potential impacts of these technologies on connected communities. The study highlights several opportunities to improve sustainability and connectedness in affordable housing. However, several factors should be addressed in implementing new technology, such as utilizing proper data management, defining beneficial information exchanges, and addressing complex human-building interactions.

Accessory dwelling units (ADUs) have had some exponential growth in recent years. ADUs are independent living spaces built on the same land as a residence or an addition. ADUs are typically designed for one or two people, and they incorporate some essential home functions. As people desire to live independently or downsize their space, quickly delivering ADUs using modular construction becomes critically important to meet the various needs of people at different ages. This paper investigates the potentials and challenges of ADUs through literature review and data collection from open sources. The results revealed some potentials, such as using these spaces for affordable housing, generating additional income through rent, living close to family while still being independent, and having a space that is just big enough for the occupant’s needs. ADUs increase housing affordability, mobility, intelligence, and resilience. The results also uncovered some challenges such as city zoning requirements, building permits, delivery difficulties, and legal issues.

One of the crucial tools to control and fight epidemic disease outbreaks (such as the fast-growing COVID-19 pandemic) is contact tracing in public buildings (such as universities). A contact tracing system can identify and alarm potential individuals who had close contacts with confirmed cases so that they can voluntarily self-quarantine. The current automated contact tracing systems, which mainly use smartphone sensors (e.g., GPS, Bluetooth), have two main challenges: (1) protecting the privacy of the users and (2) relying on GPS sensor, which does not work well indoors and in many urban settings. On the other hand, Wi-Fi positioning systems have been considered one of the most used technologies for creating a real-time indoor positioning system (IPS), especially in university campuses where the required infrastructure usually exists. This study aims to study the feasibility of using Wi-Fi location tracking technology to develop a conceptual privacy-preserving contact tracing system in university campuses. Such a contact tracing system relies on smartphones connected to the central Wi-Fi system’s access points and the connected devices Mac addresses to inform at-risk users. This study performs a comprehensive literature review to study the applicability, current limitations, and future research directions of such technologies for contact tracing. Such technology could enhance the current automated contact tracing system in universities by illuminating the need to use cellphones’ applications while protecting users’ privacy.

BIM can be complex to teach during a regular class schedule, and the COVID-19 pandemic made it even more challenging. However, the need for social distancing brought to attention students’ needs other than academic: It was possible notice the importance of students’ emotional well-being. This research demonstrates this through three main stages: first by looking into the structure of the classes. Then, a classroom activity elicited from the students revealed important points that either they perceived as challenges our positive points of the lessons. Some of their suggestions were applied and then in a final state, the points elicited by the students were turned into a survey. The survey identified difficulties related to technical problems such as to Wi-Fi network or slow computers, which can be quite challenging for BIM lessons. The paper also describes the actions taken to undermine these issues. One important point that the research showed was that students appreciated the processes of peer-evaluation. This is important since peer evaluation can be quite valuable for BIM learners, since the skill to correct other professionals’ models is highly important on a BIM oriented engineering processes. Finally, students showed that empathy and a closer relationship with the professor and colleague helped them cope with the feelings of isolation during the pandemic.

Increasingly, online safety training has been used in the construction industry to reduce costs, save time, and offer workers flexible learning opportunities. This type of training utilizes safety videos to demonstrate unsafe behaviors and increase worker awareness of hazards. However, directing trainee attention to hazardous conditions within videos remains a challenge in remote instruction. This study evaluates animation variations within videos to direct trainee attention towards hazards. Three degrees of animations were compared: (1) static images: regular fixed pictures, (2) partially animated images: cinemagraphs, and (3) fully animated images: traditional videos. A case study was employed to understand the perspective of construction domain experts that trained using the different animation approaches to recognize a select set of struck-by hazards. A between-subjects experimental design was used to collect measures regarding hazard identification index (HII), attitudes, cognitive absorption, and transportability from 19 study participants. The results of this investigation did not detect statistical differences in the average hazard identification index scores using narrated static images (average = 69.1%; STD = 27.9%), cinemagraphs (average = 80.5%; STD = 16.4%), or videos (average = 66.7%; STD = 10.5%). Moreover, the survey responses suggested all the experimental conditions had equally positive attitudes, engagement, and sense of being transported to a jobsite location.

Online safety training is an increasingly utilized tool to deliver flexible learning opportunities for workers in the construction industry. Traditionally, online safety training relies mainly on text and images for delivering trainee learning and assessment methodologies. This study investigates the use of online virtual reality (VR) environments powered by virtual humans and omnidirectional images to offer an alternative safety training method for fall hazard identification. A pilot study using a between-subjects experiment design was completed with construction management students to compare online traditional and VR training approaches. Construction students trained using one of two online methods—traditional or VR—to identify fall hazards. Subsequently, student learning was assessed in an online virtual environment to determine their hazard identification capabilities. Metrics regarding hazard identification index (HII) and time of completion were collected from the students. The results of the training did not detect statistical differences between traditional and VR student scores for HII or the time required to complete the training and assessment activities. It was concluded that the VR delivery was equivalent in terms of hazard identification knowledge with traditional training.

Laser scanning and virtual reality (VR) are two advanced trending technologies for better project monitoring and visualization in modern construction projects. Their uses, however, were mostly separated, and the connection between them has not been much studied in previous literature. This paper investigated the workflow of using VR to view laser scans through two different approaches, namely by point clouds and 360° photos, and then compared the benefits between the two approaches. The study utilized a Native American earth lodge as an example of the built environment for laser scanning and applied two different types of VR devices. In the point cloud approach, the scans were processed by the scanner software and converted directly into VR mode for convenient viewing with a PC-based VR headset. In the 360° photo approach, colored photos taken by the scanner at various angles were first stitched together by the scanner software to create 360° photos, and these photos were then imported into a VR scene in Unity and compiled into an executable app for photo-realistic viewing with a standalone VR headset. The benefits of each approach were discussed, including the software and hardware requirements, efforts needed, and viewing experience. This paper provides valuable insights on the workflow of using laser scans with VR based on a real-life example.

Among the technological advancements of Industry 4.0, virtual reality (VR), augmented reality (AR), and mixed reality (MR) have been used in many industries to visualize and interact with digital twins in real-time. These technologies have also been used with building information modeling (BIM) and geographic information systems (GIS) throughout the life cycle of construction projects. Moreover, the integration of BIM and GIS has been proven to improve collaboration among stakeholders by allowing them to visualize information seamlessly between buildings and their surroundings. Following the COVID-19 pandemic, many undergraduate and graduate courses were moved online, which posed many challenges on instructors to deliver lectures in accessible and informative ways. Furthermore, many courses and curricula have been originally online and have not yet benefited from the full potential of MR in virtual classrooms. Therefore, the aim of this research is to evaluate how MR applications of integrated BIM-GIS can improve the online learning experience of construction management students. An experimental study was conducted with students to evaluate the effect of using AR on handheld devices and MR on head-mounted devices on their spatio-temporal reasoning and understanding skills.

Firefighters often need to digest complex spatial information within a short period of time for search and rescue. Previous wayfinding literature has documented evidence about how the general population in normal situations leverages different forms of spatial information to develop spatial knowledge and guide wayfinding. However, little is known about how the arbitrarily given spatial information affects firefighter wayfinding behavior and performance when the time is limited and there is no privilege for them to develop complete spatial knowledge in an evolving manner. To narrow the knowledge gap, we conducted a wayfinding experiment with firefighters (n = 40) using virtual reality (VR). In the experiment, firefighters were required to find three victims in a simulated office maze. Each firefighter randomly experienced different spatial knowledge conditions. The results indicated that the route and survey spatial information was more efficient in facilitating firefighters to memorize the layout, leading to a better wayfinding performance.

Spatial memory of building layout is critical to the design, construction, and management of built facilities. With the advancement of new visualization technologies, the building layout and spatial configurations can be illustrated by multiple means, for instance, 2D drawings, 3D models, and immersive virtual reality (VR). It is noted that different formats of spatial information can influence the memory retention of space in different ways. Yet the relationship between the spatial information format and spatial memory pattern is less clear. This paper introduces the findings of a human-subject experiment (n = 63). Participants were required to review a building design in the form of 2D, 3D, and VR and then to hand sketch the memorized layout. A quantitative comparison was performed to assess the discrepancies between the original design and the memorized layout. The result indicated that 2D drawings, in general, led to the largest discrepancy, while emerging methods such as 3D and VR indeed led to a better memory. Besides, it was found that the memory quality of using 2D drawings seemed to diverge, showing an obvious performance polarization. The findings are expected to provide more evidence on the effectiveness of engineering visualization.

Augmented reality (AR) presents virtual content in its physical context. This technology can enable construction professionals on-site and off-site to communicate using virtual annotations to highlight contextually-relevant details. Theoretically, AR has the potential to reduce the amount of site visits and the time spent physically traveling to and from projects. However, the majority of studies exploring AR commercially available devices present results from controlled implementations or laboratory environments. While those help for initial experimentation, they do not involve all the critical factors that can be present on active construction sites. Therefore, this work explores the technological performance of a commercially available AR device in an uncontrolled construction setting. The factors that enable or hinder AR performance are documented. The findings help construction practitioners strategically target the contexts where the technology thrives, while also helping researchers define more robust technology and environment workflows to enhance the functionality of current-generation AR devices.

The construction industry has a slow rate to adopt new technologies (e.g., virtual reality) with resistance to change and a lack of research and development within construction companies. Virtual reality (VR) and augmented reality (AR) have shown some practical applications in construction such as educational training, clash detection, project coordination, immersive visualization, and virtual project walk-through. Few construction companies use VR and AR in project preconstruction and construction phases. This is because of unknown benefits and different perceptions from project stakeholders. This paper investigates the perception of VR and AR in the construction industry in Washington State. The authors collected survey data from general contractors that are members of the Associated General Contractors of Washington. The survey included 20 quantitative questions using the five-point Likert scale. The collected data revealed the perception of VR and AR in project planning, development, construction, operation, and maintenance phases. The results informed current industrial practices and future research developments on virtual and augmented realities. The results also indicated a lack of training and education on virtual and augmented realities in the construction industry.

Recent reports indicate that cyclist fatalities are rising. Unlike automobile driver crash and safety studies, there is very limited information and data on how different environmental or design features impact cyclists’ behaviors, attention, and awareness. Real world studies evaluating cyclist behavior are limited due to their inherent safety risk; therefore, there is a need for alternate data to better inform the planning and design of roadways for all users. Immersive virtual environments (IVE) have shown to provide a realistic representation of real-world conditions; however, these tools have not been evaluated and validated for vulnerable road users, such as cyclists. The purpose of this study is to assess the use of an IVE bike simulator to study the impact of design and environmental conditions on cyclists’ perceived safety and behavioral changes. By benchmarking cyclists’ behaviors and perceived safety in real-life settings compared to its representative IVE bike simulation, we can validate whether these IVE simulators are realistic representations of real-world conditions. Furthermore, by connecting these environments with the latest low-cost human sensing devices, we have built a multimodal human sensing data collection system to track participants’ gaze, heart rate, and head movement. The preliminary results from a six-participant pilot study indicate that our simulators are capable of replicating cyclists’ speed profile, heart rate changes, and most of the head and gaze behaviors and that these measurements are sensitive to environmental changes.

The model of human behavior concerning urban plaza design emerges that designers mostly approach it with their intuitions. However, the built environment and how humans interact with it have various hidden consequences on users’ behavior. Besides immediate movement patterns and physical interaction, social interactions can be dictated by the built environment’s design. In recent years, with the expansion of design research, human behavior, and technology advancement, machine learning (ML) and multi-agent modeling/system (MAS) have become the driving factors in the evidence design field. Conventional urban designs were driven just by the practicality of the design and in order to satisfy the needs of the program of the space. Moreover, the data collection methods have been done through traditional methods of manual observation, user survey, and cellphone tracking, which are either obsolete, have user errors, or are not suitable for this design purpose. This paper is the first phase of a study investigating the potential of implementation of artificial intelligence as a means of intelligent data-driven design approach for open urban spaces. The aim of the research is to promote the user experience’s positive affordance and enhancing urban open spaces’ livelihood through ML and MAS. This paper aims to investigate the use of visual object recognition algorithms of histogram of oriented gradients (HOG) and edge histogram descriptors (EHD) for human detection in urban plazas. We investigate the impact of each stage of the computation on performance results and conclude that EHD is inaccurate due to the absence of adequate descriptors. HOG results were proven accuracy of 89% on a preprocessed image of MPII Human Pose Dataset containing over 500 annotated human images with a wide range of pose variations and backgrounds.

The use of drones in construction grew exponentially over the last few years, making the construction industry one of the fastest commercial adopters of this technology. Drones are widely used in the construction industry for a wide variety of applications (e.g., site surveying/mapping, earthwork volumetrics, safety inspection, progress monitoring). Nevertheless, the integration of drones in construction raises novel occupational safety and health issues for onsite construction workers. This study presents the technical design and development of DroneSim, a virtual reality (VR) safety training environment for human-drone interaction on construction jobsites. Specifically, DroneSim aims at training construction workers on how to safely collaborate and work with drones on jobsites. The paper discusses the complete workflow involved in designing and developing the VR environment, which consisted of (1) identifying current and future drone-mediated construction applications; (2) identifying the technical and functional requirements pertaining to each of the applications; (3) proposing different drone types based on the identified applications; and (4) developing the VR-based environment, which included the incorporation of static objects, sound effects, animations, text-to-speech, as well as audio, lip-sync, and gesture synchronization.

Building information modeling (BIM) adoption in architecture, engineering, and construction (AEC) has proven to improve the efficiency and effectiveness of delivering the project from inception to operation. Nevertheless, even with the current advancement in technologies, the AEC industry is still falling behind other industries in digitizing and automating its operations, particularly in using BIM with multi-dimensions (nD) and standardizing which application should be allocated for a particular dimension. Acceptance of information such as time and cost as the fourth dimension (4D) and fifth dimension (5D) is consistent in AEC. However, there is no uniformity of applications for dimensions beyond 5D. In this study, a survey of BIM professionals worldwide has been conducted to get these answers. One finding from the survey is the identification of barriers to not using nD BIM. Another discovering is the software used for BIM nD worldwide. An additional result is uncovering the current applications of BIM beyond 5D. For instance, energy, sustainability, safety, and facility management were frequently interchanged between the sixth dimension (6D), seventh dimension (7D), and eighth dimension (8D) of BIM. Since the collected facts were from current professionals in the industry, this study can be considered a reliable foundation for enhancing the nD BIM execution and normalizing the nD BIM applications.

Virtual reality (VR) is one of the building information modeling (BIM) use cases that can benefit a project from the early stages of design to the construction and operation phases. VR enables designers, contractors, and owners to experience interacting with a facility before it is constructed. Currently, interaction with VR environments takes place through controllers. The equipment helps users to navigate virtual facilities and interact with the surrounding objects. Although controllers address the need for creating an interactive virtual environment, they make the experience less realistic, since real-life interactions take place through body movements, not by/through pressing buttons. This paper reports a project in which vision-based artificial intelligence (AI) technology is used to create a contactless connection between users and virtual environments. In this project, a series of body movements and gestures are mapped to the input commands of a BIM-based VR software. The paper overviews the state of the art in remote human-computer interaction (HCI) technologies and discusses both the potential and challenges of applying HCI technologies to immersive VR environments. It then introduces the architecture of the developed, contactless virtual environment and reports a case study of using this technology in an immersive setting.

We describe an interactive augmented and virtual reality (AR/VR) system that supports bridge inspectors during in-office preparation, on-site inspection, and follow-up work back in the office using building information modeling (BIM) data. Originating from the process steps required to perform a digitally supported bridge inspection, we focus on our prototypes’ implementation, a combination of VR and AR methods in a prototypical development. Our goal is thus to present one possible realization of these processes on a technological level. Therefore, the collaborative aspect between on-site inspection and in-office preparation and debriefing plays an important role. Several issues of data synchronization and communication techniques need consideration and proper integration into the entire system. It is essential to consider the bridge inspectors’ needs, requirements, and working environment to avoid losing track of the target group of such an application. Low initial hurdles and user-friendliness require seamless integration of different tools and additional resources. For that reason, we aim to link engineering tools and game engine environments in an operationally and reliable application. We believe that our application may serve as a starting point for further discussions on user-friendly digital support and broader usage of innovative technologies bearing in mind the evolution and tradition of structural inspections.

The introduction of emerging technologies in the capital projects industry has led to an exponential increase in the volume of information collected during the life cycle of capital projects. Despite this, only few of these technologies have gone through commercialization and been standardized across different sectors of the industry. To investigate the root causes of this problem, the Visualization, Information Modeling, and Simulation (VIMS) Committee of the ASCE Computing Division launched an initiative to document research to practice (R2P) challenges faced by both academia and the industry. This paper provides an overview of R2P practices from related fields, followed by a structured study aimed to solicit input from experts in VIMS-related domains, and a discussion of drivers and barriers identified from our initial findings. Findings are sought to ultimately contribute to successful knowledge transfer, commercialization, and entrepreneurship in the capital projects industry.

Public construction owner organizations often depend on various funding sources to meet their ongoing capital improvement needs, including federal, state, and local grant funding. These funding sources typically provide usage guidelines and reporting criteria as part of the conditions owners are required to abide by for grant disbursement and continued eligibility for future funding. This presents challenges to the owner in ensuring the appropriate allocation of funding during initial project planning, and continued adherence to funding regulations throughout the construction phase, particularly during change management. This research proposed a methodology for the visualization of cost information in building information models (BIM) using BIM tools such as Autodesk Revit and Autodesk Assemble. Project parameters within the BIM relational database were created to account for the funding source information of the model components, enabling the visual representation and analysis of embedded project information at later stages of model development. This approach was designed to foster effective and consistent communication among decision-makers as to the apportionment and tracking of facility costs to specific funds. This methodology will be applicable to other non-geometric attributes an owner may wish to track during a facility’s life cycle.

Human visual attention to natural objects could change their stress levels in indoor environments. However, the impact of objects’ properties as visual stimuli on people in virtual environments has been not understood yet. Consequently, the lack of understanding may prevent researchers from using virtual environments as a reliable tool to support the design of indoor environments with a combination of natural and non-natural objects. We hypothesize that attention to the natural elements differs among objects with different visual properties. Accordingly, we conducted an experimental study to investigate the influences of visual properties (type of natural objects, size, and location). Twenty participants were assigned to experience one of the two comparable office settings (real versus virtual), including three natural objects. The fixation time was measured by the eye-tracking method for the natural objects in conjunction with non-natural objects. The initial analysis showed that the fixation time of natural objects in VR was not different from the real condition. Additionally, different fixation times of objects varied in visual properties suggested that visual properties affect human attention to various objects.

The need to increase the number of science, technology, engineering, and mathematics (STEM) students has been stressed, yet the growing number of students still fails the national demand for the STEM workforce. This mismatch between supply and demand is also found in the construction field. Storytelling can be a potential solution to this problem, as stories are believable, rememberable, and entertaining. Storytelling can allow students to experience the career planning and decision-making processes and embrace the possibility of themselves joining the construction field. This paper introduced the creation and development process of the immersive stories and assessed the influence of immersive storytelling on students’ attitudes towards the construction field based on gender groups. The results can enhance our understanding of immersive storytelling’s potential influence in motivating students and ultimately attracting them to the construction field. It was found that immersive storytelling had a positive influence on both male and female students. Moreover, it was observed that females’ aspirations to achieve a career in the construction field before and after the immersive stories were both significantly lower than males’.

As drone-mediated building inspection gains interest among the inspection disciplines and becomes more standardized in the industry, current construction, engineering, and management (CEM) students are encouraged to develop and enhance their drone flight operational skills. While integrating hands-on building inspection training using drones in the CEM education curriculum seems favorable, the high costs of the UAS hardware components, together with the liability and safety concerns associated with flying in proximity to buildings and people, hinder such opportunities. One potential alternative to real-world drone flights is the use of virtual reality (VR) simulation technologies. This study aims at exploring the technical development and effectiveness of utilizing a VR training environment—DroneSim—specifically designed to enhance students’ drone piloting skills while accomplishing building inspection tasks. Using a head-mounted display (HMD), construction students were required to navigate their aerial platform using typical control commands (e.g., takeoff, landing) and perform building inspection tasks. Participants were asked to complete post-experiment questionnaires with the aim of evaluating their feedback on the usability and sense of presence of DroneSim. Post-experiment questionnaire results (N = 18) revealed that students’ perception on using DroneSim was positive, finding the overall experience stimulating. The contribution of this research is to: (1) understand the steps involved in the technical development of a VR-based drone-mediated building inspection training environment and (2) provide students with the opportunity to improve their drone piloting and building inspection skills in a safe and controlled environment.

This study proposes a human-in-the-loop robot-augmented intelligent system for emergency reconnaissance. The system improves situational awareness of first responders by providing robot-collected information through an intelligent and interactive interface. Unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs) equipped with RGB-D camera, thermal camera, GPR, and LiDAR are deployed to acquire information on disaster sites. The collected sensor data are further processed to present actionable information to first responders, such as victim location, buried void shape, and accessible hole through an augmented reality interface. A virtual reality platform is used to validate and evaluate the proposed system. Participants are invited to complete a search and rescue mission with robot-augmented intelligent system to find survival victims under rubbles in the virtual environment. The experiments demonstrated that the proposed system reduces the searching time and false alarms, and increases the accuracy of detecting victims, which will change the search and rescue paradigm from an experience-based practice to an information-based one.

Construction workers need to be adequately trained to foster the required skills of collaborating with robots in construction jobsites. In this regard, immersive virtual environments have been employed in many fields as cost-effective and safe alternatives to conventional training practices. However, there are still concerns regarding the transfer of knowledge from a virtual setting to real-world situations. To address such concerns, this study investigates the feasibility of virtual avatar-based simulation as an enhanced training test bed for human-robot collaboration in construction. To that end, a comparative experiment was conducted in which subjects performed a masonry task in collaboration with a virtual robot using the conventional virtual reality and the virtual avatar-based simulation. Electroencephalography (EEG) signals were used to assess subjects’ sustained attention during the training process. The results demonstrated the outperformance of the virtual avatar-based training in provoking individuals’ sustained attention, indicating its feasibility as an enhanced training test bed for human-robot collaboration in construction.

The budget for the Airport Improvement Program (AIP) in FY21 is $3.35 billion. AIP’s contractual guidelines and policies are outlined in Federal Aviation Administration (FAA) 5100.38D. Substantial contractual changes within AIP projects are likely to create risks that can greatly impact cost, time, and quality. To implement improved change management, identification and evaluation of contractual change are essential. Thus, there is a need to create an automated model to identify and evaluate contractual changes in AIP projects. This paper fills this knowledge gap. Using 876 contractual changes made to FAA 5100.38D, the authors utilized an interrelated multi-step methodology. Firstly, the data is prepared, and the changes in the FAA 5100.38D are manually adapted. Secondly, the data is manipulated using NLP techniques. Thirdly, before building the machine learning (ML) models, the data is preprocessed. Finally, hyperparameter tuned random forest (RF) and support vector machine (SVM) ML models are developed to predict contractual substantial changes, and the optimal model is selected and evaluated. Results show that RF presented the optimal hyperparameter tuned model with an AUC value of 0.928. Ultimately, this research presents an AIP decision support tool that predicts substantial contractual changes effectively and efficiently, and thus proactively provides risk assessment for AIP projects.

Having well-maintained transportation infrastructure systems, including tunnels, is essential for the functioning and growth of the US economy. As such, monitoring the condition of tunnels through structural evaluation and inspections is essential to maintain their level of service. However, there is a lack of research that utilizes data from previous formal inspections to model and understand the condition of tunnels in the US. Being the case, this paper tackles this research need by developing a tunnel condition rating model that identifies the most critical factors affecting maintenance of tunnels. To this end, the authors utilize random forest (RF), which is a well-proven machine learning technique. RF is able to both identify the importance of features and utilize the selected features to accurately predict the condition rating of tunnels. The model is trained using data for tunnels in the US, from the Federal Highway Administration Agency (FHWA). Validation, testing, and hyperparameter tuning of the model are performed through a k-fold cross validation combined with a grid search algorithm. Results highlight the most critical features affecting maintenance of tunnels. In addition, the RF prediction model yields a satisfactory accuracy of more than 85%. Ultimately, the developed tunnel condition rating model can benefit transportation entities in understanding and predicting the condition rating of tunnels in the US, hence optimizing and prioritizing their maintenance plans.

In recent years, automated fault detection and diagnosis (FDD) has been investigated as a possible solution to the energy waste problem from faulty and inefficient heating, ventilation, and air conditioning (HVAC) systems. During the operation of HVAC systems, a fault in a component is often detected when building occupants feel the effects of its downstream symptoms. Although current FDD methods can identify the presence of anomalous conditions, they provide little insight into how we can address the fault. Root cause analysis (RCA) is a complex diagnosis task that identifies a set of cause-fault-symptom chains and advises corrective actions that a system manager can employ to eliminate the fault and its symptoms. In this paper, we will explore the challenges of autonomous RCA of hardware faults specifically in the air handling unit (AHU) with variable air volume (VAV) terminals and recommend potential ways to support various challenges associated with autonomizing RCA.

Image and video data can be valuable for supporting facilities’ operations and maintenance tasks. However, this data should be captured and cataloged during the construction phase of a project to enable use during the operations phase. With advancements in technology, there are many approaches to capture, store, and retrieve image- and video-based data. This research aims to review the existing methods to capture, tag, store, and retrieve such data along with how this data can be retrieved and updated to support operations and maintenance. First, a literature review was completed on the existing methods and technologies for image capture, storage, tagging, and retrieval. Afterward, a taxonomy was identified for image-based technology. Image capture can be performed by either fixed or mobile devices. Image tagging was defined as assigning information regarding the location and time of the image capture along with identified content within the image. This content can be identified either manually or automatically using computer vision techniques, such as object detection and image classification. A content analysis was conducted on common data environments for image storage. The findings of this research can be used by practitioners to support high-value use case implementation. Results can also be used to highlight areas where image-based technologies can be expanded through research and development initiatives.

Nuclear power plant (NPP) operators need to have sufficient real-time cognitive ability to ensure their adherence to operational procedures and solve unexpected problems. The heavy cognitive load will degrade such ability and lead to operation errors. Previous studies show that the NPP workflows’ (a network of activities) properties, such as workflow structure types, activity types, and recurring activities, can significantly impact the operators’ cognitive load. However, which workflow properties could lead to a higher cognitive load are unclear. This study examines what workflow properties pose cognitive challenges to operators and quantitatively evaluates these properties’ impact on the operators’ cognitive load. The authors designed a simulation experiment on a nuclear reactor operation simulator called “Rancor” and recruited four students as pseudo-operators to perform reactor startup tasks. In the experiment, we collected the students’ eye movement data using a mobile eye tracker. Participants’ pupil diameter changes and eye movement types can indicate cognitive load and information searching strategies. The correlation analysis results indicate that workflow structure (i.e., sequential and parallel structures) and activity types correlate with participants’ pupil diameter changes. Such correlation indicates that parallel structure and recurring activities demand relatively high cognitive efforts. Moreover, eye movement analysis reveals that recurring activities require more fixed attention from operators. Overall, a quantitative evaluation of the relationship between workflow properties and the cognitive process advances understanding about NPP operators’ cognitive load and information searching strategies while executing different types of workflows.

Despite the benefits and suitability of mixed reality as a learning tool for spatial and haptic delivery of educational content, the adaptability of mixed reality for teaching sensing technologies in construction education remains ambiguous. Sensing technologies have potentials to advance construction performance as highlighted by several studies and construction industry reports. To equip future workforce with the required skills, it is important they experientially deploying these technologies on the jobsite. However, institutions are experiencing different constraints such as weather, schedule, and COVID-19 restrictions to provide future workforce with required hands-on learning. The study argues the pedagogical value of a mixed reality learning environment for teaching sensing technologies through a usability study with construction engineering and management students who are familiar with sensing technologies. The study employed quantitative and subjective users’ evaluation through a cognitive walkthrough of the learning environment using a mixed reality head-mounted display. Subjective data were obtained using think-aloud protocol and questionnaires on users’ experience. Eye tracking device inbuilt in the mixed reality head-mounted display was employed to monitor the eye gaze movements and provide quantitative data of users’ cognitive workload. Results from eye tracking, think-aloud protocol and questionnaires were triangulated to validate findings from the study. The study reveals the ease of use and effectiveness of the environment for learning sensing technologies. The characteristics, benefits, and deficiencies of the learning environment as a pedagogical tool in construction education are further highlighted.