Computing in Civil Engineering 2019: Data, Sensing, and Analytics

Automated monitoring of construction operations, especially operations of equipment and machines, is an essential step toward cost-estimating, and planning of construction projects. In recent years, a number of methods were suggested for recognizing activities of construction equipment. These methods are based on processing single types of data (audio, visual, or kinematic data). Considering the complexity of construction jobsites, using one source of data is not reliable enough to cover all conditions and scenarios. To address the issue, we utilized a data fusion approach: this approach is based on collecting audio and kinematic data, and includes the following steps: 1) recording audio and kinematic data generated by machines, 2) preprocessing data, 3) extracting time- and frequency-domain-features, 4) feature-fusion, and 5) categorizing activities using a machine-learning algorithm. The proposed approach was implemented on multiple machines and the experiments show that it is possible to get up to 25% more-accurate results compared to cases of using single-data-sources.

The rise of location positioning technologies has generated enormous volumes of digital footprints. Translating this big data into understandable trip patterns plays a crucial role in estimating infrastructure demands. Previous studies were unable to correctly represent commuting patterns on smaller urban scales due to insufficient spatial accuracy. In this study, we investigated if, and to what extent, estimated commuting patterns identified from GPS data can replicate the results from transportation surveys and to what degree these estimates improve the estimates of trips distribution pattern on census tract level using higher resolution data. We inferred average daily home-to-work trips by analyzing phone GPS data and compared these patterns with U.S. Census summary tables. We found that trips detected by GPS data highly correlate with census trips. Furthermore, GPS data is a better proxy for Census tract-pairs with larger numbers of trips.

Measuring built environment users’ response using wearable biosensors could provide a new opportunity for understanding their experience in built environment. Electrodermal activity (EDA) sensors are especially useful in detecting people’s stressful interaction with the built environment. Despite this potential advancement, the detection accuracy is still limited because of noises in EDA collected from uncontrolled settings. Alleviating respiration noise is most challenging due to the similarity in signal characteristics between the respiration noise and EDA response to distress. The authors propose an adaptive denoising method that references photoplethysmogram (PPG) to detect and remove respiration noise in EDA. Quality of denoising and quality improvement in stress measurement were measured for validation. The results showed that the proposed method brought better quality of respiration noise removal than previous methods, and therefore improved stress measurement quality. The finding can contribute to improve quality of EDA from the field, which is essential to accurately understand people’s stressful interaction with built environment.

In the last decade, many programs have been developed to help decrease or eliminate the wage inequality in the United States; however, identifying employees who might be at risk of wage inequality remains challenging. This paper presents a framework to identify such employees in an organization, using a machine learning approach. This paper utilized the U.S. Department of Transportation (DOT) workforce demographic information to train and test the model. First, a prediction model is developed to estimate the salary range of employees based on historical data, using supervised machine learning techniques. Then a minority score is defined to determine the employees who might be in the risk of inequality, based on three factors: gender, ethnicity, and disability type. Finally, a framework is developed to identify the employees at risk of wage inequality, using the prediction salary range and minority index. The proposed framework enables employers to establish a fair wage, resulting in reduction and/or elimination of inequality challenges and their consequences in their organizations.

Construction entities interact with each other to accomplish assigned tasks, constituting working groups. Recognizing working groups is important as it enables the correct comprehension of jobsite context, which in turn facilitates the interpretation of entities’ intentions, and the prediction of their movements. Aiming at identifying working groups formed by interacting workers and/or equipment, this study devises a machine learning approach to integrate positional and attentional cues. Methods are created to represent the spatial and attentional states of individual entities and compute positional and attentional cues between two entities. A long short-term memory network is adopted to identify the working groups. The proposed method is validated using construction videos that are available online and taken by the authors. The results suggest that by integrating positional and attentional cues, the working groups can be identified with an accuracy of over 95%, much higher than that obtained using positional cues only.

Concrete is a very popular construction material utilized in the construction industry. There are several methods for determining the quantities of its components, being the concrete mixture design method by American Concrete Institute (ACI) the most common procedure among others. Construction laboratories have conducted innumerable concrete mixture designs applying different procedures through time, acquiring valuable information that is not being exploited completely. This study used experimental historical information of a construction laboratory in order to develop Sugeno type fuzzy inference systems (FISs) to estimate material proportions for fabricating normal concrete. Fuzzy modeling was accomplished by using subtractive clustering and adaptive neuro fuzzy inference systems (ANFIS) techniques. A Sugeno type FIS was developed for estimating the proportion of each component in a concrete mixture; namely, water-cement ratio, fine and coarse aggregates, where concrete compressive strength and aggregate properties including fineness moduli and abrasion resistance were the input variables. The resulting fuzzy models were able to estimate concrete constituents very well since computed coefficients of determination (i.e., R-squares) were greater than 90% when validating the models. All FISs can be used as mixture design tools to compute material proportions based on past experience when fabricating concrete on the jobsite. Also, the proposed framework illustrated in this research for concrete mixture design could be extended for any particular data set regardless of concrete component characteristics.

Construction tasks involve various activities composed of one or more body motions. It is essential to understand the dynamically changing behavior and state of construction workers to manage construction workers effectively with regards to their safety and productivity. While several research efforts have shown promising results in activity recognition, further research is still necessary to identify the best locations of motion sensors on a worker’s body by analyzing the recognition results for improving the performance and reducing the implementation cost. This study proposes a simulation-based evaluation of multiple motion sensors attached to workers performing typical construction tasks. A set of 17 inertial measurement unit (IMU) sensors is utilized to collect motion sensor data from an entire body. Multiple machine learning algorithms are utilized to classify the motions of the workers by simulating several scenarios with different combinations and features of the sensors. Through the simulations, each IMU sensor placed in different locations of a body is tested to evaluate its recognition accuracy toward the worker’s different activity types. Then, the effectiveness of sensor locations is measured regarding activity recognition performance to determine relative advantage of each location. Based on the results, the required number of sensors can be reduced maintaining the recognition performance. The findings of this study can contribute to the practical implementation of activity recognition using simple motion sensors to enhance the safety and productivity of individual workers.

In water distribution systems (WDSs), pressure transients reflect the responses of WDSs to normal and abnormal changes. The behavior of pressure transients is largely unknown and cannot be fully assessed by numerical simulation or modeling. This study proposed a two-step time series data mining (TSDM) approach for detecting and clustering pressure transients to reveal recurrent and consistent patterns based on high-resolution pressure data collected by a network of sensors. First, the modified cumulative sum control chart (CUSUM) approach is performed on the pressure data preprocessed by piecewise aggregative approximation (PAA) to detect and extract pressure transient events. Second, a novel clustering technique is applied to cluster the extracted pressure transients based on their similarity as measured by dynamic time warping (DTW). An example application of this method has been implemented to confirm that the proposed approach provides a fast and efficient way to reveal consistent and unique patterns in WDSs.

Deep learning (DL) algorithms such as convolutional neural networks (CNNs) can assist in tasks such as content search and retrieval, image tagging and captioning, scene description, motion prediction, and language processing. This paper presents research that aims at designing and validating DL models for automated content-based retrieval of daily construction images and videos. Information retrieval from visual data is key to labor-intensive tasks such as safety inspection, crew activity logging, and work progress documentation. In order to train deep neural networks (DNNs), large repositories of high-quality annotated visual data are needed. However, generating such labeled datasets in construction is non-trivial and resource intensive, and requires specific skillset. To overcome this challenge, we present a methodology for fast object detection and tagging in visual data using DNNs trained with a relatively small dataset. Two state-of-the-art object detection algorithms, i.e., you-only-look-once (YOLO) and mask region-based CNN (a.k.a., Mask R-CNN) are investigated. Training data is obtained via web mining (the Internet) and crowdsourcing. Results show that training on data from both sources yields the best classification accuracy. Testing the model on new data reveals that the fully-tuned model can achieve a minimum mean average precision (mAP) of 79% when tested on different image subsets.

Existing automated code checking methods/tools are unable to automatically analyze and represent all types of requirements (e.g., requirements that are too complex or that require human judgement). Recent efforts in the area of augmented data analytics have proposed the use of templates to facilitate the analysis of text. However, most of these efforts have constructed such templates manually, which is labor-intensive. More importantly, it is difficult for manually-developed templates to capture the linguistic variations in building codes. More research is, thus, needed to automate the generation of templates to support the tagging and extraction of information from building codes. To address this need, this paper proposes an unsupervised machine-learning based method to extract sentence templates that describe syntactic and semantic features and patterns from building codes. The proposed method is composed of four main steps: (1) data preprocessing; (2) identifying the different groups of sentence fragments using clustering; (3) identifying the fixed parts and the slots in the templates based on the syntactic and semantic patterns of the sentence fragment groups; and (4) evaluating the extracted templates. The proposed method was implemented and tested on a corpus of text from the International Building Code. An accuracy of 0.76 was achieved.

Interdependency of interconnected infrastructure is a critical and daunting issue. While existing research is still limited to single physical mechanism-based methods or data-driven statistics approaches; there is a predicament that certain target infrastructure may only have limited knowledge on physical operation mechanisms or lack enough associated data. To fill the gap, the study proposes an integrated data-driven (DD) and physics-based (PB) approach to explore the interdependency of interconnected infrastructure. The approach consists of three components: (i) “motivations”—to understand the targeted infrastructure, availability of data, and physical knowledge related to different infrastructure systems; (ii) “methods”—to select suitable DD or PB methods for individual infrastructure; and (iii) “modes”—to design the connection for bridging DD and PB methods. A case study is conducted to explore the interdependency of the water supply pipe system and road transport networks of a district in Hong Kong. The preliminary findings reveal that the framework can help identify the hot spots of water pipe bursts and predict their cascading effects on the road transport networks, despite certain limitations need to be overcome in future research.

The scanning electron microscopy (SEM) images are commonly used to understand the microstructure of the concrete. Many researchers have adopted the image processing techniques for the microstructure analysis, but little has been studied on how the magnification of the SEM images influence the accuracy of analysis. Therefore, this paper presents a machine learning (ML) based framework to study the effect of SEM image magnification on degree of hydration measurement. In this study, the authors looked into the impact of magnification of SEM images on the model training, accuracy, and degree of hydration measurement using two scenarios. First, the image segmentation was performed using a classifier of specific magnification, and then a common classifier is trained using the image of different magnification. The preliminary results show that there is no significant effect of magnification on model training and accuracy. However, it has a significant impact on the degree of hydration measurement.

3D modeling of the built environment has become common practice in the AEC/FM industry. Practitioners take advantage of the geometric and semantic information embedded in the 3D model to perform engineering analysis. Despite the benefits provide by the 3D model, the process is time-consuming, labor-intensive, and error-prone. In this paper, we propose a new neural network-based method for 3D point cloud semantic segmentation of building scenes using a hierarchical approach: first, we reason on the local and global contents of raw point cloud data to extract geometrical features. Second, the features are used as input to an artificial neural network that performs semantic segmentation on the points. These points are classified into: beam, ceiling, clutter, column, door, floor, pipe, wall, and window. We evaluated our approach on a dataset of several buildings and we obtained an accuracy of 73%. Our experiments produce robust results readily useful for practical Scan2BIM applications.

Construction safety has always been one of critical concerns in the construction industry with diverse approaches and technology for consistently managing safety hazard issues being examined and adopted. However, the current technology such as a vision-based approach does not appear to fully support consistent and robust safety monitoring because of its heavy-data processing and inherent restrictions including limited angle coverage and view detectability. A vulnerable construction environment requires an advanced safety surveillance and event detection approach. To provide a supplement safety monitoring method, this project has an objective to establish an audio-based autonomous safety surveillance system with alarming pre-notifications and map visualization based on identified work activities and potential safety issues. This proposed system adopts a schedule-based sound data training method and incorporates historical occupational injury and illness manual data with classified sources and events of accidents in each construction activity. The injury data are informed by a daily project schedule to link potential safety hazard to daily planned work activities. By applying a machine learning technique, the system accurately categorizes a sound type according to sound training data scope-downed by project schedule and safety data and provides pre-warnings in accordance with any detected irregular events. The system is expected to contribute to the body of knowledge for safety monitoring and provide the potential to be integrated into a robust automated safety surveillance system with occupational accident data and significantly improving construction activities classification accuracy.

Due to the characteristics of the construction projects, construction contractors are often vulnerable to business failure compared to those in other industries. Thus, predicting the potential business failure of construction contractors has been a crucial issue. This study proposes a model that predicts business failure using long short term memory recurrent neural network (LSTM RNN), which is one of the deep-learning algorithms. The proposed model uses not only a set of accounting data but also proxies for the construction market condition and the macroeconomic environment as input variables. The prediction performance of the proposed model is examined by varying the combination of input variable groups. The results showed that adding construction market and macroeconomic variables to accounting variables could increase the performance of business failure prediction. It was also found that macroeconomic variables had a slightly higher impact on the business failure prediction than construction market variables. The results of this study are expected to be useful references for both researchers and practitioners to develop business failure prediction models of construction contractors.

Statistical studies indicate that the design-build (DB) project delivery system has better overall project cost performance compared to the traditional design-bid-build (DBB) project delivery system. The project cost performance of DB projects could be further improved if project managers had a better understanding of the budget implications of their specific project attributes. Existing relevant studies are either inconclusive or only explore project characteristics in isolation. The overall objective of this research is to comprehensively utilize data mining and statistical analysis to identify common patterns associated with project attributes and respective interaction effects in terms of cost performance in design-build projects. Several control factors, such as evolving requirements from owners, unforeseen conditions, attributes of project team, and major project characteristics are taken into account. The analysis shows that owner-directed changes and project managers’ previous experience, particularly the lack thereof, are risk factors having significant negative impacts on cost performance. Progressive design-build and metropolitan locations are project attributes associated with best cost performance. Interaction effects of project attributes also have statistically significant impacts on cost performance. The findings of this research contribute to empirical knowledge related to cost performance and assist project managers in addressing project issues to limit their negative impact on cost performance.

This paper presents a method for the digital reconstruction of the geometry of a wind turbine blade from a point-cloud model to polysurface model. The digital reconstruction of the blade geometry is needed to develop computer models that can be used by architects and engineers to design and analyze blade parts for reuse and recycling of decommissioned wind turbine blades. Initial studies of wind-blade geometry led to the creation of an airfoil database that stores the normalized coordinates of publicly-available airfoil profiles. A workflow was developed in which these airfoil profiles are best-fitted to targeted cross-sections of point-cloud representations of a blade. The method for best-fitting airfoil curves is optimized by minimizing the distance between points sampled on the curve and point-cloud cross section. To demonstrate the workflow, a digitally-created point-cloud model of a 100 m blade developed by Sandia National Laboratory was used to test the reconstruction routine.

Falling objects accidents frequently occurs on construction site and cause a great deal of casualties every year. In addition to the reported injury accidents, there are still a large number of near miss incidents related to falling objects on construction site, since not all falling object incidents involves injuries because of sheer luck. Therefore, identifying the near miss incidents and their causes in time is an effective way to prevent a real injury accident. This paper proposes a deep learning-based video monitoring method to automatically detect and analyze the probable causes of the near miss incident (falling objects) on construction site, which forms an exhaustive case database from real-time monitoring for safety evaluation, training, and improvement. The proposed approach develops a comprehensive framework for near miss incident detection and cause analysis by advanced computer vision methods. The experiment results show that our approach achieves high performance in both falling object detection and probable causes classification.

Struck-by accidents are potential safety concerns on construction sites and require a robust machine pose estimation. The development of deep learning methods has enhanced the human pose estimation that can be adapted for articulated machines. These methods require abundant dataset for training, which is challenging and time-consuming to obtain on-site. This paper proposes a fast data collection approach to build the dataset for excavator pose estimation. It uses two industrial robot arms as the excavator and the camera monopod to collect different excavator pose data. The 3D annotation can be obtained from the robot’s embedded encoders. The 2D pose is annotated manually. For evaluation, 2,500 pose images were collected and trained with the stacked hourglass network. The results showed that the dataset is suitable for the excavator pose estimation network training in a controlled environment, which leads to the potential of the dataset augmenting with real construction site images.

In recent times, construction operations have reported enhanced performance as a result of utilizing reality capture technologies, robotics, and the Internet of things (IoT). However, recent studies indicate that these technologies could struggle to diffuse across the construction industry at an acceptable rate. One reason for the predicted stagnation is the lack of standardization and regulations for these technologies. Given the importance of standards and regulation to technology adoption and the need to keep construction stakeholders informed with industry trends, it is essential to synthesize and document the current state of industry standards and regulations associated with emerging construction technologies. Presently, a few studies have evaluated the current state of regulations affecting some promising construction technologies. However, these studies were limited to national regulations and focused on a specific technology without pointing out important trends, opportunities for interoperability standards, and synergies across multiple technologies. To fill this gap, a global search was conducted to identify and streamline the available standards and regulations governing four promising construction management technologies—exoskeleton, IoT, LiDAR, and UAVs. Technology implementation experts from six countries were interviewed to qualitatively analyze the intersection between standards and regulations and the diffusion of emerging technologies. Results from the study suggest that although there is a growing development and implementation of technology standards and regulations in Western Europe and North America, construction-specific standards and regulations are sparse. In closing, recommendations that could help accelerate the development of standards and practical regulations that drive adoption are provided.

The exponential increase in the volume of images and videos captured on construction sites and the growing availability of building information models (BIM) and schedules with production-level details has created a unique opportunity to automate how progress is monitored and reported on construction sites. However, the state-of-the-art methods of automated progress comparison are still in its infancy largely because of these methods either only leverage geometry of the 3D reconstructed scenes to reason about presence or detect and classify construction material from 2D images without considering geometrical characteristics. To the best of our knowledge, this paper is the first to offer a computer vision method that can jointly reason about geometry and appearance of observed BIM elements in site images and videos to monitor and report on their state of progress. The new method fuses structure-from-motion geometrical features together with directional and radial appearance features in a new deep convolutional neural network (CNN) architecture to detect and classify state of work-in-progress. Our experimental results show that using geometrical features reduces errors in appearance-based recognition methods and offers a new opportunity to scale the applicability of automated progress detection methods to real-world settings.

Prefabricated mechanical, electrical, and plumbing (MEP) module is one of the off-site construction technologies that can significantly improve the construction productivity. To successfully install MEP modules on-site, it is necessary to ensure all the MEP elements including pipes, ventilation ducts, and cable trays inside the module have accurate positions and dimensions before on-site installation. However, the inspection process is time-consuming using a traditional manual way. To address the problem, this study presents an automated dimensional quality inspection technology to estimate the geometric properties of MEP modules using 3D laser scanning. This study uses an automated technique to align the scan data with the as-designed BIM of the MEP module. Each MEP element is then extracted based on the scan data within the region of interest for each respective element. Each pipe is extracted by fitting the scan data within the region of interest to a cylinder, and each cable tray is obtained by detecting the side and bottom surfaces of the cable tray as planar surfaces and finding the intersection lines of the planes. Lastly, the checklist items are computed based on the extracted as-built objects. An experiment on a real MEP module was conducted to demonstrate the accuracy and efficiency of the proposed method.

Object detection has been widely used to extract visual information from resources and build components, enabling timely identification and evaluation of construction performance. However, it is difficult to evaluate the progress of a structure solely based on the detection approach. A segmentation approach using 2D images is proposed for finding discrepancies between an as-built and an as-planned structure. First, preprocessing is implemented to convert the drawing to a binary image, and a deep-learning based segmentation is conducted to extract structural components from as-built images. Then, the extracted structure and the converted drawing are compared using a 2D matrix of those images. An experiment using wood structure images was performed. The results demonstrate the accurate detection of discrepancies. Thus, the proposed approach can potentially be utilized for monitoring progress as well as inspection and maintenance, for which the identified discrepancies between as-planned and as-built images can be used.

In civil engineering, point cloud data are used for construction site monitoring. However, with such massive datasets, processing these data have barriers due to the size and computational intensity. This study presents the first scalable Hadoop based point cloud management and processing framework to overcome those barriers. The framework consists of 3 layers: (1) storage layer; (2) operation layer; (3) interactive layer. The storage layer is optimized by indexing techniques to accelerate the map-reduce applications. The operation layer is equipped not only with the common operations such as range query, kNN, or spatial join, but also with two powerful modules including “Change Detection” and “3D Geometric Model”, for efficient monitoring. The Interactive layer enables users to approach a parallel processing model on Hadoop without requiring deep related knowledge. In addition, this layer could also visualize massive point cloud data directly from Hadoop, which is useful for analytical processes. With such components, our framework is scalable, inexpensive, and full-fledged.

Inability to capture and resolve spatial changes (deviations and deformations) during the construction process can have severe impacts on the structural behavior of bridges during its operation and maintenance. The major challenge lies in understanding the origin of spatial changes of a bridge element, which develop either during the construction phase, post construction, or during its service life. The objective of this study is to explore the potential of leveraging 3D point cloud data to analyze the quality of bridge construction and identify the deformations caused during the accelerated bridge construction process. This paper presents a case study of an ongoing bridge construction project where a site is scanned multiple times to obtain several 3D point cloud datasets. The obtained 3D point cloud data were analyzed to discover any spatial changes in major bridge members (abutment, concrete bearing pad, and bridge girder) induced by either construction error or structural deformations. The authors believe that such detailed study will produce a novel method to differentiate between deformations caused in the bridge elements during the accelerated construction process and its service life. The findings will potentially pave new avenue of research by pinpointing the causes of long-term deformation and can play a key role in taking preventive actions during construction to rectify potential problems. The proposed research can also act as a tool for construction managers to evaluate the quality of work and help in saving time and money by identifying discrepancies as soon as they occur on site.

The cutting and packing (C&P) problem has been extensively studied as it has a wide variety of applications in many industries. Good packing solutions can effectively reduce manpower and production costs. However, approaches for packing 3D irregular shaped items common in construction are very limited. In this paper, a heuristic algorithm to pack a set of irregular shaped items into a box-shaped container with the objective to maximize the contact area between objects has been proposed as one step required for alternative packing solutions. A 3D scanner is employed to obtain the geometric information of items. The heuristic algorithm determines the rotation and translation of each item, moves the objects into close proximity, and fits the objects together automatically using point cloud representation. This is a new approach. Experiment results show that the proposed approach has the potential to support good packing solutions of realistic items in a reasonable time.

Visual inspection for bridges highly relies on inspectors’ experiences. Such inspections are time-consuming and dangerous due to the needs of putting inspectors to hardly accessible places. Terrestrial laser scanning has proved its capability to overcome the shortcomings of subjective and error-prone visual inspection. Unfortunately, only experienced engineers with profound bridge knowledge and 3D data analysis skills can reveal anomalous trends of structural deteriorations from images. Combined uses of inspection reports and imagery data could help address the lack of experienced inspectors, but manually analyzing observations from images against documented bridge defects is also tedious and subject to expert knowledge. This paper examines the practical value and technical feasibility of automating the integrated analysis of 3D imagery data and bridge inspection reports for supporting high-pier bridge inspections. The results show that the potential values, technical feasibility, and challenges of such automation for supporting timely and reliable high-pier bridge inspection.

Visual data used for progress tracking on job sites require the knowledge of where changes are happening and the as-built state of the construction site. Today’s common methods for capturing visual data only offer (a) manually designed and operated missions; and (2) automated collection using lawn mowing patterns. This paper presents a new method for automated flight planning and visual data collection that takes advantage from the availability of 4D BIM as a priori to create optimal flight missions at locations of expected change. The developed approach simulates the flight path to review flight execution safety and visibility of the camera frames to the in-progress elements while accounting for requirements of baseline among images for 3D reconstruction purposes. Finally, a case study conducted on a construction project reported improvements in visual coverage to the structure and reduction in 3D reconstruction errors obtained by improving the visual quality of the collected frames.

Adaptive reuse of buildings has been identified as a superior alternative to new construction, in terms of sustainability and circular economy. The programming of deconstruction works plays an important role in the adaptive reuse process in order to improve a project’s performance. Unfortunately, there is little evidence of efficient implementation and use of deconstruction programming for adaptive reuse building projects. The aim of this paper is to describe a selective deconstruction programming approach for adaptive reuse of buildings. First, we describe a new model for multiple-target selective disassembly sequence planning for buildings. Afterwards, we describe a method for programming the deconstruction works based on the disassembly sequences generated in the last step. The proposed approach helps to improve project performance through process automation that supports quantitative analysis, low cost exploration of alternatives, and an iterative design process for meeting project constraints, e.g. budget, schedule, environmental impact.

Studies suggest that the recent technological advancements in robotics can foster the automation capabilities of the built infrastructure by introducing service robots. Task-allocation and path-planning are two of the fundamental challenges faced by such task-oriented robots. Existing algorithms in this domain have been adapted from outdoor logistics based applications with context-specific assumptions. This study particularly addresses this issue, extends the authors’ previous efforts, and proposes a new methodology to optimize the task allocation and route planning in case of multiple starts and destination depots where each robot begin and end at the same depot. Scenario analysis is conducted to compare the performance (e.g., total distance) of the proposed multi-depot algorithm with the single-depot one. The developed methodology is generic and can be used for a wide range of indoor building environment applications. Finally, limitations of the current approach are identified, and future work directions that require further investigation are proposed.

Equipment activity recognition plays an important role in automating the analysis of onsite construction productivity, considering that human observation is always labor-intensive and time-consuming. Recently, several methods have been proposed to recognize the activity of construction equipment from videos. However, one of their common issues is the failure to conduct the recognition in a long video for classifying a sequence of equipment activities. This paper proposes a novel equipment recognition method with the support of a three-dimensional (3D) convolutional neural network. The network could extract both temporal and spatial information of the equipment to recognize its activities in a long video. The recognition results are further compiled and analyzed to identify the equipment productivity. The proposed method was tested to recognize the excavator activities (digging, swinging, and loading) on real construction sites. The results showed that the method outperformed existing equipment activity recognition methods in terms of accuracy and robustness.

The increasing labor shortage issue and the working safety awareness cause the urgent of the development of new construction methods. A new type of lab for new construction processes is required to expedite the innovation cycle. This paper presents an ongoing work of building a construction lab at University of Alberta. The goal of the construction lab is to provide a sandbox for developing new construction processes and machines. We designed the lab with four major systems: (1) sensors: to collect data from construction site for operation assistants and virtual reconstruction; (2) manipulators: to excavate the path planning algorithms and to develop the cooperation approaches between human and machines; (3) visualizers: to construct the digital twin of a real construction site for revealing the simulated results in virtual environment; (4) computers: to run machine learning algorithms for recognizing and tracking objects in construction environments. This laboratory allows the researchers in the construction engineering test and develop their tools in a controlled environment. Such scaled tests in the lab can bring significant benefits in finance, efficiency, and safety.

The Industry 4.0 wave has driven various industries to be increasingly digital. Under this trend, the construction industry is also seeking out new ways to keep pace with the rapid transformation. As one of the major technologies associated with Industrial 4.0, digital twins (DT)—a subset of cyber-physical systems (CPS)—shows enormous potential to benefit the construction industry. The aim of this paper is to present a comprehensive, up-to-date literature review and critical analysis of existing research into DT applications, with a view to identifying the opportunities for research and applications in the construction domain. Based on the results of the analysis, DT has the potential to gather and analyze information at different scales. In this context, DT would facilitate more efficient construction processes and allows for informed decision to be made, which will ultimately promote a safer environment at reduced costs/improved efficiency.

Construction robots have drawn increased attention as a potential means of improving construction safety and productivity. However, it is still challenging to ensure safe human-robot collaboration on dynamic and unstructured construction workspaces. On construction sites, multiple entities dynamically collaborate with each other and the situational context between them evolves continually. Construction robots must therefore be equipped to visually understand the scene’s contexts (i.e., semantic relations to surrounding entities), thereby safely collaborating with humans, as a human vision system does. Toward this end, this study builds a unique deep neural network architecture and develops a construction-specialized model by experimenting multiple fine-tuning scenarios. Also, this study evaluates its performance on real construction operations data in order to examine its potential toward real-world applications. The results showed the promising performance of the tuned model: the recall@5 on training and validation dataset reached 92% and 67%, respectively. The proposed method, which supports construction co-robots with the holistic scene understanding, is expected to contribute to promoting safer human-robot collaboration in construction.

A robotic teleoperation system of excavator has been developed to improve safety and remote-control methods at hazardous situations such as construction or disaster sites. This paper introduces an innovative remote-control method and a newly developed robotic hardware system, Phantom, to rapidly retrofit existing construction equipment for efficient robotic control from a remote place. The system allows one joystick to control one excavator. A remote operator with two joysticks at a remote workstation can intuitively control two excavators equipped with Phantom simultaneously. The proposed method was compared with typical excavator operation methods to evaluate its performance based on operation time and mechanical movement accuracy from given tasks. The tests show the promising results and the proposed robotic equipment operation approach can significantly improve the safety of the operators who need to work at dangerous site conditions.

Research indicates that workers often fail to recognize a significant proportion of safety hazards. To reduce injury likelihood, efforts have traditionally focused on developing and delivering training interventions. Despite such efforts, desirable levels of hazard recognition are rarely achieved. Therefore, augmenting human abilities with a technology-driven solution to improve overall hazard recognition can yield substantial benefits. Accordingly, the objective of this study is to develop a method for localizing workers with respect to pre-identified hazards in real-time. To achieve this objective, a 3D point cloud of a construction site as a global map is created and hazard locations are marked on this map. Workers are provided with a head-mounted camera that continuously records their first-person view (FPV) videos. The image frames from these videos are localized onto the global map using bag of word (BoW) localization. Apart from estimating the proximity to safety hazards, the system can also capture large-scale data that captures unsafe behaviors (e.g., entry to restricted areas) and near-miss incidents for training purposes.

Multiple object tracking using videos have gained interests in the construction industry to support real-time monitoring and control of construction sites for productivity and safety. State-of-the-art multiple object tracking algorithms could encounter various challenges in the field, such as inaccurate detection and identity switch because of occlusions. These field challenges cause various uncertainties in the information derived from tracking results (e.g., average waiting time of workers in a workspace). Studying the failures of multiple object tracking algorithms in various scenarios are essential for assessing the decision risks related to the use of these algorithms for deriving information used by field engineers. The authors have developed a multiple object tracking algorithm that outperforms state-of-the-art methods found by the authors in the literature and characterized the performance of the algorithm using video data sets collected during a nuclear power plant (NPP) outage for monitoring indoor check-in processes of multiple workers before entering a workspace. The results indicate that the video data collected on job sites involve complex interactions among human, equipment, and environmental objects, causing various challenges to the multiple object tracking algorithm. Specifically, the authors used videos collected from an indoor space of a NPP to quantify how reliable the developed algorithm could predict the average waiting times of workers in queues at different waiting areas of the check-in space. The authors categorized scenarios where multiple object tracking fails and found the major failures came from identity switching and false positive detection of workers in mirror or shiny surfaces. For the construction research community, this research will form a framework to assess the reliability of multiple object tracking algorithms in deriving information used by field engineers. For the computer science community, this research identified the scenarios where state-of-art visual tracking algorithms fail to motivate the development of new algorithms.

Excavator action recognition is a major task for cycle-time and productivity analysis. Many researchers have developed vision-based methods and showed promising results for automated action recognition. However, previous methods did not fully incorporate sequential working patterns of excavators although it is crucial to explain their natural operation procedure. To overcome the limitations, this paper proposes a vision-based excavator action recognition framework that considers the sequential patterns of visual features and operation cycles. The framework was validated with 1,340 images collected from actual earthmoving sites, and the average precisions and recall rates were 90.9% and 89.2% respectively. This research performed additional experiments with the other CNN model, and the precision and recall rate were decreased to 77.5% and 71.6% respectively. The experimental results showed the applicability of the developed framework and the positive impacts of sequential pattern learning. This study can contribute to more reliable automated cycle-time analysis and productivity measurement.

Excavators on construction sites can pose a serious hazard for both their operators and the laborers working alongside them because they can easily overturn when operating on uneven ground. Although there have been many suggestions for improving onsite safety, the ultimate solution is to avoid these hazards altogether by utilizing unmanned equipment. Unfortunately it would be both technically challenging and prohibitively expensive to undertake the complete redesign of an excavator to enable it to function autonomously, as this would require an integrated control algorithm capable of real time sensing, analysis, and control if the excavator is to react appropriately to its local environment. We propose an alternative solution, namely the use of a removable manipulator that can be installed in the cockpit to enable an operator at a remote location to control excavator operations for 6 types of movement. A prototype system was constructed and tested.

Detection of wall defects by autonomous robots enhances building inspection with reduced labor, improved productivity, and inspection accuracy, in comparison with manual inspection. Accordingly, various advanced technology-enabled techniques have been investigated to detect defects through collected images. However, rare effort has been put on developing an automatic data collection system to support the collected data with high quality (complete but not redundant). In this paper, an autonomous robot-enabled data collection system is developed for indoor wall condition inspection. The autonomous robot is equipped with sensors for navigation, map building, and obstacle avoidance. To generate safer, more reasonable collision-free wall-following trajectories, improved heuristic algorithms are used to optimize robot trajectory. The developed data collection system with navigation of an autonomous robot is evaluated by simulation. The obtained results indicate that the hybrid approach for the automatic data collection system is able to collect all available walls and corners without redundant trajectories.

Camera-equipped unmanned vehicles (UVs) have received a lot of attention in data collection for construction monitoring applications. To develop an autonomous platform, the UV should be able to process multiple modules on an embedded platform. Pixel-wise semantic segmentation provides a UV with the ability to be contextually aware of its surrounding environment. However, in the case of mobile robotic systems with limited computing resources, the large size of the segmentation model and high memory usage requires high computing resources, which a major challenge for mobile UVs. To overcome this challenge, this paper presents a light and efficient deep neural network architecture to run on an embedded platform in real-time. The proposed model segments navigable space on an image sequence (i.e., a video stream). The results demonstrate the performance efficiency of the proposed architecture compared to the existing models and suggest possible improvements that could make the model even more efficient.

The use of camera-equipped robotic platforms for data collection and visually monitoring applications is exponentially growing. Cluttered construction sites with many objects on the ground are challenging environments for a mobile unmanned ground vehicle (UGV) to navigate. This study presents a mobile UGV equipped with a stereo camera and a robotic arm that can remove obstacles along the UGV’s path. To achieve this, the surrounding environment is captured by the stereo camera and obstacles are detected. The obstacle’s relative location to the UGV is sent to the robotic arm module through robot operating system (ROS). Then, the robotic arm picks up and removes the obstacle. The proposed method will greatly enhance the degree of automation and frequency of data collection for construction monitoring. The results successfully demonstrate the detection and removal of obstacles, serving as one of the enabling factors for developing an autonomous UGV with various construction operating applications.

This research explores a means by which a construction robot can leverage its sensors and a building information model (BIM) to perceive and model the actual geometry of its workpieces, adapt its work plan, and execute work in a construction environment. The adaptive framework uses the generalized resolution correlative scan matching (GRCSM) search algorithm for model registration, a new formulation for fill plan adaptation, and new hardware for robotic material dispensing. Joint filling is used as a case study to demonstrate the formulation of an adaptive plan and to evaluate a robot’s ability to perform work when the actual locations and geometries of its workpieces deviate from their designed counterparts. The robot was found capable of identifying the true position and orientation of the joint’s center with a mean norm positioning error of 0.13 mm and orientation error of 1.2°. The adaptive framework offers significant promise for a range of construction activities, including those involving objects of complex geometry and precision work.

Existing RTLS solutions are mostly based on either wireless technologies, fiducial markers, or Lidar-based simultaneous localization and mapping (SLAM), and inevitably suffer from some drawbacks such as low accuracy, reliance on existing facilities, labor-intensive environment instrumentation, or high economic cost. This paper introduces an ORB2 RGB-D SLAM based indoor RTLS that can be readily adapted and applied to building or civil infrastructure applications while addressing the above limitations. Besides the original sparse feature map built by the visual SLAM (vSLAM), the proposed system builds and maintains an additional 2D occupancy grid map (OGM) and overlays it with real-time 2D camera pose and virtual laser scan for 2D localization. These designs not only allow users to interact with the system but also open up the possibilities of path planning and continuous navigation based on feature-based vSLAM. The localization accuracy of the system is evaluated with a marker-based method, which proves its feasibility and applicability in indoor building and construction applications.

In the recent years, industrialized construction (IC) is moving the construction industry toward integrating emerging manufacturing technologies into offsite construction and factory component assembly practices to improve productivity and efficiency. The key to success in the IC business is providing mass customization and mass production simultaneously, which is only possible in a highly flexible manufacturing system. The manufacturing industry is experiencing the fourth industrial revolution which is blurring the separation of the physical, cyber, and biological worlds. The primary objective of this study is to identify the new methods and techniques emerging in manufacturing industry and evaluate their possible application in IC. A questionnaire was designed to collect data from industry experts. Preliminary data shows a significant lag in the transition of the new trends in cyber physical systems and data analytics into the IC production process. The findings of this survey are precisely analyzed and discussed in this paper.

Digital fabrication is an emerging approach to transforming design to physical products. While a number of studies have been carried out to highlight the use of digital technologies in construction, there is still a dearth of studies focused on design-to-fabrication workflows in construction projects. In response, the overarching goal of this study is to develop a design-to-fabrication workflow using digital models and computer numerical control (CNC) machines with a focus on formwork fabrication. The specific objectives are to (1) present a detailed workflow from the design phase of formwork to the fabrication phase, and (2) develop automated procedures to support the workflow without extensive manual intervention. This workflow integrates virtual design and construction (VDC), trade coordination, parametric modeling, software customization, tool path development, and CNC routing in order to produce precise prefabricated formwork components. To test and validate the workflow, the authors conducted a case study for the prefabrication of concrete edge formworks in a 26-story post-tensioned cast-in-place concrete structure. The case study shows that the use of the workflow for CNC machines supports craftsmanship to improve labor productivity, safety, and fabrication quality. The results demonstrate the advantages of using this workflow over the traditional approach to support project teams’ productivity and enable them to make informed decisions for their implementation of digital fabrication.

Productivity in the U.S. construction industry has stagnated over the past 50 years, whereas manufacturing industries have about doubled productivity levels. Adoption of smart manufacturing with construction has challenges to achieving efficiency in a factory environment. Construction projects are one-off designs with little replication in the configuration of components. The ability to reconfigure factory production and network optimization performance help smart manufacturing systems. Artificial intelligence (AI) is well suited to this problem. This paper provides an in-depth review of AI methods and how the technology may be applied to automated construction manufacturing systems. This starts with a state-of-the-practice review of AI applications within construction manufacturing. This is followed by an identification of the AI needs of construction manufacturing systems. Lastly, the paper reviews the state-of-the-art of artificial neural networks (ANNs) (e.g. deep learning and transfer learning) from the domains of manufacturing and industrial engineering, and discusses the potential for application to construction manufacturing. The objective of the paper is to help identify the direction for future research and development in this field.

Modular construction has been proven to be more time-efficient comparing to stick-built construction. However, the lack of automation technologies adoption shows a missed opportunity in modular construction, where the time efficiency of the modular process can be further improved, along with quality and safety. To address this problem, the authors proposed a new simulation method for modular construction using BIM, game simulation, and robotics technology, to help analyze and promote automation in modular construction of wood structures in a controlled indoor environment. As a first stage, the authors presented their simulation methodology, definition of the parameters and constraints for the construction simulation, and creation of the interactive simulation model. This first stage will be the foundation for further development that will enable the assessment of construction productivity when integrating robotic systems into the modular construction workflow.

The primary advantages of utilizing a mobile robot, unmanned ground vehicle (UGV), are its autonomous navigation and data collection capabilities from a large-scale environment. However, the UGV is restricted by its limited perception ability in a cluttered environment. To address this limitation, an unmanned aerial vehicle (UAV) can cooperate with the UGV and build out a detailed and complete map. The objectives of this study are to operate aerial and ground robots cooperatively and to fuse point cloud data complementarily. This paper introduces an autonomous cooperation framework using UGV and UAV for 3D geometric data collection in a dynamic cluttered environment. First, UAV is deployed and collects 3D terrain data by using photo images to get an overall idea of the target site. Then, the path planning and stationary scanning locations for UGV are estimated using the gradient-based 3D model generated by UAV. This method was tested on a real-world construction site and obtained promising results. The proposed UAV-UGV cooperative robotic approach is expected to reduce human intervention in data collection and data processing time significantly. Furthermore, it enables frequent monitoring, updating, and analyzing a cluttered environment for timely decision making.

Cracks are the most common defect on the surface of tunnels, which potentially brings threaten to the safety of the tunnel and the running vehicles. Timely repairing of the crack is of critical importance. In the past two decades, various vehicle platforms have been developed on the purpose of efficient crack detection and maintenance. With these platforms, images can be captured in a traffic speed, and automatic methods can be developed for fast crack localization. However, for image-based crack detection, traditional methods often meet difficulties in handling cracks with low contrast and poor continuity. In this paper, deep learning based techniques are exploited for feature learning and representation for crack detection. A novel deep neural network is presented for pixel-level crack recognition. Hierarchical features in different stages of the convolution are fused together to overcome the influence of noise and a spatial constraint placed on the target pixels is used to guarantee the crack continuity. In the experiment, a tunnel crack dataset is constructed for performance evaluation. Experimental results demonstrate the effectiveness of proposed method.

Since construction specifications are normally over 1000 pages and are complicated and often inconsistent, reviewing them is a labor-intensive and time-consuming activity. Thus, the aim of this study was to automate the review process by comparing construction specifications with standard specifications using natural language processing. Standard specifications for road construction projects were collected from 43 different states in the U.S. and used as experimental data. Doc2Vec, cosine similarity, and named entity recognition (NER) were used to recognize construction objects, standard values, and execution conditions, which can be used to find specification errors. As an early stage of the research, most of related sentences were found from standard specifications with high relevancy, and the average F1 score of NER was 0.256. The research findings will contribute to enhancing the efficiency of checking for specification errors by automatically detecting abnormalities and the absence of specific standards.

Earthwork is one of the major components of construction projects. In recent years, unmanned aerial systems (UAS) as a data acquisition platform is becoming attractive for many surveying applications in construction. This work identifies the effects of various flight parameters and construction materials on the measurement accuracy and operational efficiency of UAS. Flight altitudes of 60, 90, 120, and 150 feet above ground level, two image overlapping rates (70% and 90%), and four common types of earthwork material with the use of 16 ground control points (GCPs), are chosen to analyze their influence on the positional errors through multiple comprehensive comparisons and statistical analysis. The results indicate that the use of GCPs is the most influential factor. However, it is necessary to obtain a balance among all factors because no single factor is able to improve the spatial quality of the model if others do not also perform well.

Sensors and computation increase the precision, efficiency, and agility of crane operations. This paper presents an ongoing work of developing computational methods to enhance the crane operations. This research focuses on three significant challenges for crane operators: (1) identifying construction equipment and their activities, (2) identifying and tracking personnel, and (3) tracking the rigging object. In the preliminary stage, we focus on the first challenge to identify construction equipment and activities. 5,000 images have been collected and manually labeled for training deep learning detection algorithms. In the future steps, we will employ inception-SSD method to locate personnel, trucks, and excavators. After that, we will propose a method that recognizes excavator activities from crane views. Once learning algorithm is the reliable, it will benefit the crane operators to operate the cranes with confidence. It can also reduce the difficulty of crane operation. Training time and safety concerns will be reduced simultaneously.

As a mature technology, ground penetration radar (GPR) is now widely employed in detecting rebar and other embedded elements in concrete structures. Manually recognizing rebar from GPR data is a time-consuming and error-prone procedure. Although there are several approaches to automatically detect rebar, it is still challenging to find a high resolution and efficient method for different rebar arrangements, especially for closely spaced rebar meshes. As an improved convolution neural network (CNN), AlexNet shows superiority over traditional methods in image recognition domain. Thus, this paper introduces AlexNet as an alternative solution for automatically detecting rebar within GPR data. In order to show the efficiency of the proposed approach, a traditional CNN is built as the comparative option. Moreover, this research evaluates the impacts of different rebar arrangements, and different window sizes on the accuracy of recognizing rebar. The results revealed that: (1) AlexNet outperforms the traditional CNN approach, and its superiority is more notable when the rebar meshes are densely distributed; (2) the detection accuracy significantly varies with changing the size of splitting window, and a proper window should contain enough information about rebar; (3) uniformly and sparsely distributed rebar meshes are more recognizable than densely or unevenly distributed items, due to lower chances of signal interferences.

The construction industry has been criticized for lower productivity than other sectors, which may be influenced in part by its historical resistance to change in adopting new technologies. Therefore, it is common for organizations to have dedicated change agents to support effective technological adoption. This study aimed to identify the key attributes of change agents that influence the success of adopting new technologies in construction companies. To achieve this objective, industry practitioners in the United States provided responses to open-ended questions and the answers were analyzed to identify common themes. The significant findings include: (1) change agents’ personality is the most frequently reported attribute that contributes to their effectiveness, and (2) human-related attributes have a larger role in successful technological adoptions compared to organizational- and technical-related attributes in change agents. This paper contributes to the body of knowledge by identifying the frequency of attributes that are reported to impact the effectiveness of change agents in industry.

Construction contractors often suffer from lack of accurate management of projects’ cash flow. Retainage, advance payments, and interest rates are among several financial terms that essentially influence the project’s cash flow. Many research efforts in finance-based scheduling target developing models to aid contractors in cash flow management, however, these efforts often concentrate only on the employer-contractor payment conditions. Providing a reliable management of contractor’s cash flow requires highlighting the contractor-subcontractor as well as contractor-supplier agreements as they have a considerable impact on the project’s cash flow. This paper proposes a user-friendly spreadsheet model that utilizes genetic algorithm to reach a schedule that minimizes the negative cash flow required by the contractor to finance the project and maximizes the internal rate of return. The model is not limited to number of activities nor to number of predecessors for each activity and is flexible to incorporate multiple payment conditions of different subcontractors as well as suppliers to reflect the real picture of general contractors’ cash flow. The proposed model is demonstrated through sample models incorporating different subcontractors’ arrangements. The model accounts for time value of money parameters, multiple subcontractors and suppliers with different payment agreements, contractor-employer payment conditions, as well as a certain credit limit among other user inputs. Comparing with other cash flow models, the proposed model proved to provide support to contractors in better cash flow planning, monitoring, and control through improved depiction of the construction industry’s payment relations.

The objective of this study is to identify a ranking method for assessing the potential risk of knee musculoskeletal disorders (MSD) among construction roofers. On a slope-adjustable wooden platform, nine subjects performed the shingle installation, comprising seven phases: 1) reaching for shingles, 2) placing shingles, 3) grasping the nail gun, 4) moving to the first nailing position, 5) nailing shingles, 6) replacing the nail gun, and 7) retuning to upright position. Knee flexion, abduction, adduction, internal, and external rotational angles were measured using an optical motion analysis system. To analyze the relative level of risk at each phase, these angles were combined using multiplication and aggregation-based scoring models that generated ranks of phases at different roof slopes. The ranking results provide useful information for identifying the postures that might pose greater MSD risk, and may facilitate effective interventions development to reduce extreme knee positions which is a MSD risk factor.

Many designers and researchers have grappled with the problem of optimally locating buildings and use types in a neighborhood-scale development. But little work has used data-driven optimization to aid in creating urban design schemes. The paradigm of single-use Euclidian zoning has heavily impacted the way our neighborhoods, cities, and suburbs are designed, resulting in the physical separation of uses. However, as we grapple with emerging issues of environmental and social sustainability in cities, there is a pressing need to consider alternative urban designs that require less dependence on personal automobiles and that foster healthier cities. In this paper, we develop a methodology for (1) automatically assessing the walkability of neighborhoods by adopting a common walkability metric and (2) optimizing the layout of buildings and amenities across a known grid in order to maximize the walkability metric. We apply this methodology to a case study of the Potrero Hill neighborhood in San Francisco, California. We find that, in comparison to the existing layout that can be characterized by Euclidian-style separation of uses, the optimized layout suggests distributing amenities across the street network, resulting in a two-fold increase in walkability. This tool and analysis have the potential to provide computational and data-driven support for urban designers and researchers hoping to understand and improve the walkability of urban spaces.

Even though HVAC systems consume around 40% of total building energy, they often fail to provide satisfactory thermal conditions to occupants in commercial buildings. Personalized environmental control systems (PECS) such as local fans and heaters have the potential to control the local environment around the occupant to improve occupant satisfaction. Existing methods for modeling thermal preferences primarily rely on temperature measurements and often neglect air speed. Current methods for monitoring air speeds are either too expensive or too bulky to be used in real office environments. We present our preliminary results on predicting air speeds using alternate environmental parameters such as air pressure, temperature, and humidity using low cost miniature sensors. The results show that we can accurately predict air speeds with a mean absolute error of 0.056 m/s for conditions investigated in this study. Although the results are promising, further studies are needed to improve the system before it can be used in real world environments.

Cyber-physical systems (CPS) are co-engineered interacting networks of physical and computational components. Most future technologies are believed to have much more human-aware. Accordingly, integrating human context in CPS instead of placing it outside the system boundary is becoming increasingly important. It is essential to have a systematic understanding of the principles that how to integrate human into CPS as human-in-the-loop cyber-physical systems (HiLCPS). However, in the literature, the roles that human should take in HiLCPS are not clearly identified. The HiLCPS are heterogeneous systems with high uncertainty and complexity associated with humans, thus specifying human roles in HiLCPS enables the optimum system performance by fully considering human perspectives. This study focuses on synthesizing the existing HiLCPS literature to gain the insight of human roles. Three HiLCPS configurations are identified. For each configuration, human roles with related studies are reviewed. Furthermore, the associated challenges and opportunities are identified and summarized.

Systematic evaluation of construction equipment activities is essential to efficient management of the fleet. Recent research has made significant progress in developing machine learning activity recognition frameworks. Deep learning in particular can circumvent the need for complex manually-designed feature extraction/selection procedures that contribute to lower accuracies in traditional and shallow models. The research presented in this paper develops and compares deep learning algorithms for construction equipment activity recognition in different levels of detail. Data are collected in a non-controlled environment from real-world activities. A convolutional neural network (CNN) called BaselineCNN and a hybrid network that contains both convolutional and recurrent long short-term memory (LSTM) layers called DeepConvLSTM are studied. In summary, DeepConvLSTM proved superior to BaselineCNN. In a six-class identification task, DeepConvLSTM achieved a validation accuracy of 77.1%. In the vibration-setting-only subproblem, DeepConvLSTM achieved a validation accuracy of 75.2%. In the direction-only subproblem, DeepConvLSTM achieved a validation accuracy of 96.2%.

The objective of this study is to propose a seed-search algorithm and develop seeding strategies in online social networks for disseminating credible situational information regarding built environment disruptions in disasters. Rapid and extensive dissemination of credible situational information is important for disaster preparedness, response, and recovery in communities. Online social networks such as Twitter have become popular media sources among the public to share information in disasters. Due to the directed relations and fragmentations in networks, however, little is known about the ways of selecting starting nodes (a.k.a., seeds) in order to broadly and rapidly spread information online. To address this gap, this study proposes a computational approach, which is an integration of greedy algorithm and graph analysis, to capture fragmentations, identify critical seeds, and develop seeding strategies to disseminate information in online social networks. A case study of an infrastructure disruption (water release from reservoirs in Houston) during 2017 Hurricane Harvey was used to illustrate the capabilities of the proposed approach. The results indicate that seeding top 10 users is effective for distributing information to more than 80% of nodes in a network. The findings inform about strategies to better report, transmit, and gather situational information on social media, which can further enhance situation awareness and community resilience in disasters.

The development of cyber-physical systems (CPSs) is an essential milestone in advancing operations in many industries including the architecture, engineering, and construction (AEC). Its development is supported by technological advancements that align with the CPSs goals of automation and intelligent interaction between cyber and physical spaces. This study focuses on the evaluation of the supporting technologies for CPSs in AEC projects. To this end, a holistic review of industry and academic literature is carried out and then, the technologies are evaluated in terms of their level of implementation and application thorough the project lifecycle for horizontal and vertical project scenarios. Finally, a categorization was made to present the interrelations between the technologies to provide new insight into future research into CPS applications in the AEC industry.

Since construction project rework is one of the most important causes of cost overruns, the best strategy is to identify the causes of rework at the right time. Therefore, the aim of this study is to investigate and analyze the rework indicators (RIs) belonging three categories of organization, project, and people. To fulfill the objectives of this study, a structured survey was developed in order to collect data associated with various construction project characteristics, change orders, cost, and schedule performance. Appropriate statistical tests were applied to the 39 collected survey responses. The result reveals that the experience of the project management team PMT in the design and/or construction phase, and number of PMTs who work in a construction project are the most important indicators in deriving rework. The findings of this study would assist project managers (PMs) in planning proactively to prevent construction project rework.

Stakeholder management plays a critical role in successfully delivering construction projects; thus, identifying causes of miscommunication within primary stakeholders (owner, designer, and contractor) would be beneficial. Previous studies identifying significant causes of miscommunication have not studied how different causes work together to influence communication within primary stakeholders. Therefore, different appropriate statistical methods were used to identify the effective communication indicators (ECIs) within three primary stakeholders. Factor analysis technique was then utilized to investigate main components associated with ECIs. Structural equation modeling (SEM) was then used to identify the pertinent relationship between the identified main components. For this purpose, 38 completed case study projects were collected. As information related to other aspects of case studies was needed, a structured survey was developed and 38 completed surveys were collected. Thirty ECIs associated with three primary stakeholders were determined significant, which belonging to 20 main components. The results revealed that project target, scope clarity, filed labor, objectives and restrictions, and resource availability were critical components. The outcomes of this study help project managers focus on main effective communication factors to mitigate the risk of project failure because of dispute and conflict.

An essential prerequisite for reducing natural disaster damage is to identify objects vulnerable to extreme weather events. However, it is not trivial to scrutinize large urban areas within a short period of time, using conventional data collection processes for disaster preparedness. To address this issue, we propose a novel geospatial localization method building on participatory sensing to localize vulnerable objects or areas in cities. The proposed method consists of sequential modules—a geographic coordinate conversion, mean-shift clustering, deep learning-based object detection, magnetic declination adjustment, line of sight equation formulation, and the Moore-Penrose generalized inverse method—to localize urban objects in crowdsourced data. The localization accuracy of the proposed method is evaluated in a case study of urban areas in Texas. The proposed method is expected to contribute to rapid data collection practice in disaster preparedness and enable practitioners to concentrate their limited resources on where focus is needed.

The high frequency of work-related injuries and fatalities experienced on construction sites makes the construction process a very hazardous endeavor. The collection and analysis of safety data is an important element in measurement and improvement strategy development. Wearable sensing devices (WSDs) and the internet of things (IoT) have been identified as emerging technologies with strong potential for a transformative change in many aspects of construction workers safety monitoring including tracking and transmitting workers safety information in real-time. This paper provides an evaluation of the potential applications of WSDs and IoT for the continuous collection, analysis, and monitoring of construction workers’ safety metrics to mitigate safety hazards and health risks on construction sites. Wearable sensors and systems that can be used for physiological monitoring, environmental sensing, proximity detection, and location tracking of a wide range of construction hazards and vital signals which can provide early warning signs of safety issues to construction workers are reviewed. A schematic model for integrating wearable sensors for interoperability and multi-parameter monitoring to capture and track several safety metrics is also presented. The challenges facing the widespread adoption of WSDs and IoT in construction are also identified and evaluated. Based on the outcomes of the review completed, recommendations are made on how WSDs and IoT can be effectively implemented to enhance safety performance on construction sites.

The effective layout of departments within a new hospital influences the efficiency and effectiveness of delivering healthcare services. This study explores a computational approach to generate and evaluate potential hospital layouts. Given a set of healthcare departments, areas, and structural bay sizes, a graph theoretical approach with a placement strategy was used to develop an initial set of optimal and near optimal layouts based on both adjacency ratings and distances. Input data of adjacency ratings was collected from experts involved in the design of a new hospital project. An optimal adjacency graph was calculated and a placement strategy with a discrete set of constraints was used. Each layout was given a distance weighted score based on the pair-wise distance weighted adjacency rating. Healthcare planning and design experts were surveyed for their input on the use of this approach. Comparison of their results to the layout scores indicate that planning and design experts select the best scoring layouts more consistently than the worst scoring layouts.

A classroom is a multioccupancy environment that typically involves a large number of students with varying comfort zones and comfort levels in one single space, making it challenging to provide a satisfactory environment to accommodate the comfort needs of this diverse population. Current building practices that rely on centralized heating, ventilation, and air conditioning (HVAC) systems result in poor occupant satisfaction with the indoor environment, especially when the occupant group is large and diverse. There is, thus, a need to integrate HVAC system with a cyber human system (CHS) that allows for adaptive HVAC operation based on an aggregated group comfort level that collectively accounts for the comfort levels of diverse student populations. At the cornerstone of the CHS system, this paper focuses on discussing our proposed group comfort aggregation model that quantifies the thermal comfort levels of a group of occupants based on their individual thermal comfort levels. The proposed model is theoretically grounded in social welfare theory and social welfare function. This research contributes to the body of knowledge by providing a more explicit understanding of how to measure group thermal comfort in the indoor environments. The proposed model serves as the foundation of adaptive and comfort-driven HVAC operation that offers more comfortable and healthier classroom environments to students.

The nature of construction activities limits the sensing capability of construction workers, which causes difficulty in perception of hazards threatening their lives. This difficulty is a continuous challenge for construction workers. In an attempt to overcome this challenge, this study proposes a tactile communication system as a means to deliver information about detected hazards to workers. The system is based on the sense of touch, which workers can perceive even when their senses of sight and hearing are impaired or limited. In the prototype system development, the researchers investigated three parameters of vibration signals: signal intensity, signal duration, and delay between signals, in order to categorize a set of distinct signals that would result in a minimum rate of signal misperception. This study implemented clustering analysis and categorized eight distinguishable signals. Such tactile signals, each with specific information, would enable workers to perceive surrounding information about hazards at construction sites, which otherwise would be difficult. This method demonstrates the potential of such a technique to seamlessly deliver hazard information to construction workers by enhancing their perception ability on a loud and dynamic construction site.

The design phase is one of the important phases in any construction project. Leading cost and schedule performance indicators of design phase have been studied by some of the researchers, however, these variables' robustness has been studied rarely. The goal of this research is to discern between the robust and fragile cost overrun and schedule delay indicators of the design phase. Both extreme bounds analysis (EBA) methods including Leamer’s and Sala-i-Martin were implemented in this research. The Leamer’s method only considers the extreme bounds of the indicator's distribution while Sala-i-Martin focuses on the whole indicator's distribution. Results of this study revealed robust cost overrun and schedule delay indicators in the design phase. Project managers can use the findings of this research to prioritize the more robust indicators as higher priority and reduce required design modifications within the project process.

Working on elevated surfaces such as scaffolds has high risk of work-related musculoskeletal disorders (WMSDs), injuries, and fatal accidents. In the past, researches have studied postural stability, rotational postural stability, and cardiovascular stress of manual works on scaffold. Most of these studies focused on erecting and dismantling scaffold end-frames. However, there is a necessity for biomechanical evaluation of manual material handling (MMH) tasks and study of the effect of independent factors such as scaffold height, task type, and weights on low-back compression, shear forces, and perceived task difficulty. This study focuses on evaluating the biomechanical stresses on lower back due to the three MMH tasks using three different weights at two levels of the scaffold height using 3DSSP models. The statistical analysis result shows that a significant increase in low-back compression and shear forces with the scaffold height and lifting weights.

Primary stakeholders (owner, consultant, and contractor) are commonly subject to challenges associated with internal miscommunication. Inaccurate transmission of data may cause major project delays and cost overruns. As it was found that few researchers have focused on prediction and analysis of designers’ effective communication indicators (DECI) in construction projects, this study aims at determination of DECIs in construction projects. For this purpose, a survey containing 52 questions corresponding to potential effective communication indicators was developed and distributed among experienced designers. Based on survey responses and implemented statistical tools, DECIs were identified. Confirmatory factor analysis technique was then utilized to determine major components of DECI. Results revealed that principal components of DECI are design and technology, scope clarity, technical and financial support, facility, experience issues, and decision-making issues. This study assists practitioners make proactive plan to utilize project resources properly in order to manage internal miscommunication.

Posture quantification is important in biomechanical analysis of loadings of workers’ trunk and upper extremities in studying back/neck/shoulder musculoskeletal disorders caused by labor-intensive construction activities. This study demonstrates the reliability and validity of a posture-matching method using inertial measurement unit (IMU) tracking system. A commercial IMU-based motion capture system was used to obtain upper extremity and trunk joint angles. Observers wore IMU sensors and mimicked prescribed distinct work postures shown on a screen; then joint angles were obtained by the IMU sensors. A reliability test and the root-mean-square difference indicated that the posture-matching method was reliable in estimating certain joint angles such as upper arm flexion/extension but unreliable in estimating upper arm rotation and wrist ulnar/radial deviation. The reliability and validity were affected by the body part, joint angle range, observers’ perception while mimicking and matching postures, angle of the camera captured the worker’s posture, and errors from IMU sensors.

The thermal environment has a great influence on individuals’ performance; however, factors such as one’s motivation to perform well under experimental conditions cause difficulties in assessing how room temperature affect subjects’ performance. One approach to overcome this problem is to understand the changes in individuals’ neurophysiological conditions. This paper reports on the results of an experiment where electroencephalogram (EEG) data were collected from 5 subjects while they performed four computerized cognitive tasks. Power spectral density of EEG signals in three different thermal environments, slightly cool, neutral, and slightly warm, was compared within subjects. In most cases, significant differences in PSD of the frontal theta (4–8 Hz) activity are observed, indicating individuals’ mental effort varies with room temperature. In the long run, the increased mental workload will reduce individuals’ performance and be detrimental to their productivity. The study indicates that the proposed method could be implemented on a larger scale for further studies.

Measurement of operator quality performance has been challenging in the construction fabrication industry. Among various causes, the learning effect is a significant factor, which needs to be incorporated in achieving a reliable operator quality performance analysis. This research aims to enhance a previously developed operator quality performance measurement approach by incorporating the learning effect (i.e., work experience). To achieve this goal, the plateau learning model is selected to quantitatively represent the relationship between quality performance and work experience through a beta-binomial regression approach. Based on this relationship, an informative prior determination approach, which incorporates operator work experience information, is developed to enhance the previous Bayesian-based operator quality performance measurement. Academically, this research provides a systematic approach to derive Bayesian informative priors through integrating multi-source information. Practically, the proposed approach reliably measures operator quality performance in fabrication quality control processes.

Technology plays an essential role in accelerating and improving various construction processes, so both academia and industry have interest in adopting and adapting technologies in their research or projects. In this respect, academic and industry collaboration has increasingly become essential in the construction industry. A critical step for academia-industry collaboration is to understand how academia and industry accept and reject technologies differently. Therefore, this study aims to examine the technology maturity gap between academia and industry. In a partnership with the Construction Industry Institute’s Horizon-360 team, we investigated the levels of acceptance for the emerging technologies by academic research and the architecture, engineering, construction, and facilities management (AEC/FM) industry through an onsite and online survey, and classified the identified technologies into relevance and maturity. The survey results show that there are several technologies viewed differently between academia and industry. The findings from this study will serve as the basis in closing the gap for more active emerging technology-oriented academia-industry cooperation.