Construction Research Congress 2024: Advanced Technologies, Automation, and Computer Applications in Construction

The construction industry is increasingly incorporating Artificial Intelligence (AI) and construction technologies into projects. However, it lags behind other industries in terms of digital transformation. One of the reasons for this disparity is the lack of evidence-based information on AI and technologies in construction projects. Hence, this research aims to comprehensively understand the benefits of AI techniques and construction technologies and their role in achieving objectives and improving functions. To this end, the authors followed a three-step research methodology. Firstly, an extensive literature review was conducted to identify 4 AI techniques, 13 construction technologies, and 28 benefits relevant to their implementation in construction projects. Secondly, a social network analysis (SNA) was performed to validate and confirm the findings from the literature. Finally, an industry-based survey was distributed among 52 construction practitioners to achieve two main objectives: (1) assess the benefits of AI and technology and their influence on different project objectives, and (2) quantify the potential improvement in project functions resulting from their adoption. Study results offer evidence-based information on the impact of AI techniques and construction technologies on construction project objectives and help construction practitioners in assessing the benefits that AI and construction technologies can bring to their projects.

This study aims to review and identify advanced technologies utilized in Facility Condition Assessment (FCA) practices together with Building Information Modeling (BIM). FCA typically identifies the physical properties and deficiencies of the building systems and components. The common practice of FCA comprises visual inspections and walk-through surveys while determining the condition within linguistic definitions. The traditional FCA practices are non-standardized, subjective, time-consuming, resource intensive, and imprecise. Thus, integrating advanced technology has a significant potential to address these challenges with increasing the benefit of FCA in sustaining the longevity of the built environment. Within limited studies, digital twin, Geographic Information Systems (GIS), Internet of Things (IoT), laser scanning and point cloud, and Unmanned Air Vehicles (UAVs) are the innovative technologies used in FCA compared to the traditional practice. This study also suggests future research directions for utilizing advanced technologies in FCA in light of identified challenges in the literature.

Pavement condition significantly impacts a region’s socioeconomic status by affecting safety, economic effectiveness, and environmental aspects. The condition of the pavement directly impacts crash rates, fuel consumption, and pollution levels, necessitating equitable access for all, regardless of socioeconomic characteristics. We measured social equity in pavement condition by comparing pavement condition, represented by International Roughness Index (IRI), with community demographics. Analyzing more than 8 million records from Highway Performance Monitoring System (HPMS) data across multiple years, we established links between pavement quality and socioeconomic factors. We found that areas with higher proportion of African American, linguistically isolated population, and disadvantaged neighborhoods—in terms of housing/transportation—have lower access to high-quality pavement regardless of controlling factors such as region, road type, and traffic. Furthermore, a predictive classifier confirmed the influence of sociodemographic factors on pavement quality classification (good, acceptable, poor), emphasizing the need for social equity integration in pavement maintenance planning.

Pipelines are the most widely used energy transportation infrastructure in the world. Pipeline failures caused by aging, corrosion, cracks, and damages may result in irreparable societal, economic, and environmental consequences. Therefore, pipeline safety and integrity are crucial to the robustness of modern societies. Robust failure models have the potential to effectively reduce the emergent risk and the probability of hazards in pipeline construction, and extend the service life of pipeline infrastructure. In this paper, we apply a data-driven approach, Gaussian Process Regression (GPR), to predict the failure of the oil pipeline. We train and test the models based on historical public data from a report on European cross-country oil pipelines. The best model type is selected with different variable combinations. Variables such as pipeline diameter, service, and gross spillage volume are considered in the models. The data-driven pipeline failure model will provide pipeline operators with an effective way to evaluate pipeline conditions and guide pipeline rehabilitation and maintenance. The model can also be used to support pipeline integrity management based on the ASME code, which can extend the service life of the pipeline infrastructure and bring benefits to the environment and the economy.

The user persona is a communication tool for designers to generate a mental model that describes the archetype of users. Developing building occupant personas has proven to be an effective method for human-centered smart building design, considering occupant comfort, behavior, and energy consumption. The current approaches to developing building occupant personas face a major obstacle of manual data processing and analysis. This study proposes a machine learning-based approach for occupant characteristics’ classification and prediction with a view toward partially automating the building occupant persona generation process. We investigate the 2015 Residential Energy Consumption Dataset using six machine learning techniques for predicting 16 occupant characteristics, such as age, education, and thermal comfort. The models achieved moderate accuracy in predicting most of the occupant characteristics and significantly higher accuracy (over 90%) for attributes including the number of occupants in the household, their age group, and preferred usage of primary cooling equipment. The results of the study show the feasibility of using machine learning techniques for occupant characteristic prediction and automating the development of building occupant persona to minimize human effort.

Construction equipment is subject to several types of breakdown throughout the project duration. The availability of operational data is crucial to understand the breakdown patterns and implement effective maintenance strategies. However, projects often operate with tight profit margins and limited resources; and therefore, access to such data is not readily available. The aim of this study is to establish a predictive maintenance framework based on machine learning (ML) that leverages historical breakdown data with the absence of information relating to the condition of the equipment and any output extracted from monitoring devices and sensors. The proposed approach entails developing and applying a multilayer perceptron (MLP) neural network to a real-life infrastructure project in the Middle East. The collected data include an equipment maintenance log database. The results show a significant improvement in accuracy compared to linear and non-linear models reported in the literature. The proposed framework helps enhancing the overall productivity of construction equipment by minimizing their breakdown rate, thereby reducing the associated operating costs.

The increasing complexity of construction sites and the growing need for enhanced safety, productivity, and project management necessitate improved methods for monitoring and tracking moving objects. This paper introduces a methodology for moving object segmentation and tracking in construction sites, with the objective of improving site safety, productivity, and project management. A deep learning-based model is utilized for precise object segmentation, while the advanced Observation-Centric SORT (OC-SORT) tracking algorithm is employed for tracking segmented objects. The methodology is assessed using a large-scale dataset of moving objects in construction environments. Segmentation results demonstrate a mean average precision (mAP) of 0.561 and an AP₅₀ of 0.809. Moreover, the approach effectively tracks construction workers and equipment across diverse and challenging construction site situations, including crowded conditions. The implementation of such a system can greatly enhance construction site management, minimizing workplace accidents, and optimizing resource distribution.

Heat-related hazards present significant challenges in construction. To effectively mitigate these safety risks, timely assessment of heat stress is vital for implementing heat mitigation strategies. Current field measurement approaches, however, encounter temporal and spatial limitations in monitoring heat stress. This study proposes a framework for near-real-time analysis of spatial heat hazards at construction sites. The framework incorporates 3D modeling to address geometric features and a customized weather station to monitor meteorological conditions. The collected meteorological data is transmitted to a server computer, where heat risks are simulated based on the geometric and meteorological conditions in construction sites. Additionally, the study presents preliminary results comparing the accuracy of this simulation-based approach to on-site measurements using a black globe thermometer and net radiometers. The proposed framework is expected to enhance our understanding of heat risks and enable more effective implementation of protective measures in time-sensitive situations relevant to heat risk.

Many fatal incidents in the construction project occur during the earthmoving phase. Teleoperation has been a promising solution for such dangerous tasks since this method can remove humans from hazardous workplaces. Although humans are not on the jobsite for safety reasons, teleoperation requires human-in-the-loop automation, so humans need to robustly operate the construction robot at a distance. Operating equipment from a distance often reduces task performance and safety performance since spatial awareness of human operators is restricted compared to onboarding control. It is an important issue since the decisions and actions taken by operators directly affect the performance of construction robots. This study proposes the use of a feedback system for the teleoperation of construction robots to improve proximity sensing of risks and improve the human operator’s understanding of obstacles in the workplace where robots are used. The proposed teleoperation feedback system is validated with a human subject experiment in the virtual environment. This research will contribute to the body of knowledge on improving teleoperation, human operators’ spatial awareness, and human-robot interaction task performance.

A robust teleoperation system has the potential to play a key role in mission tasks relevant to construction in a physically challenging environment. Here, the communication delay generated by long-distance and data transmission in teleoperation systems impacts human–robot interactions and work performances during the teleoperation. There is a need for more understanding of how time delay affects performance in construction activities and human operators’ abilities, such as situational awareness and mental workload. This exploratory study aims to provide knowledge on the latency impacts on work performance and task load by measuring and evaluating the relevant human factors. Participants conduct construction tasks under time-delayed conditions in virtual reality environments. We assess the performance (i.e., completion time, success rate) and task load (i.e., mental demand, frustration) and observe manipulation behavior changes to evaluate the relationship between time delay and human factors during construction work. This paper contributes to exploring the signal latency by long-distance teleoperation and assesses its impacts on work performance, task load, and behavior changes while performing construction tasks.

Artificial Intelligence (AI) and Machine Learning (ML) have been embraced techniques in various fields including construction processes and materials, yet rarely applied in Integrated Project Management (IPM). IPM using Earned Value Management (EVM) systems is the grouping of the project management processes to ensure they operate in sync for the success of the project. This paper focuses on exploring the potential use of Open AI’s powerful tool, Generative Pre-trained Transformer (ChatGPT) in IPM using earned value management systems. The authors survey industry practitioners to identify the capabilities, limitations, and implications of the use of ChatGPT and similar large generative AI models in these fields, specifically for project management field practitioners. The preliminary survey result shows that there are several considerations and limitations to address when applying it in the field. This paper contributes to researchers and professionals in assisting in the use of such a tool with caution aiming to improve EVM data analysis and the overall EVM profession.

Teleoperation is widely used in hazardous and uncertain site settings, allowing scheduled procedures to be carried out across long distances while workers are away from the sites. Teleoperators in off-sites collect both the site information and feedback from the interfaces which provide synthesized information that a robot collects. This interface mainly conveys visionary information for the operator’s intuitiveness such as the spatial awareness of objects and surroundings. To achieve a rich visual understanding of the site, the interface should fully contain and intuitively convey the associated contextual information. Excessive or unintuitive information not only makes it difficult for operators to exert their full potential but also increases their cognitive load. This study explores how different visual interface configurations affect operators’ work performance and their cognitive load during the teleoperation task. The findings from the experimental studies are expected to help develop human-centered interfaces for teleoperation in the context of construction tasks and provide the cornerstone for not only an intuitive but fruitfully informative interface in a provided task setting.

Extended reality is among the disruptive technologies that have been gaining interest in the construction industry, and for the past two decades, it has been the focus of many researchers in the field of construction. Many previous studies have investigated extended reality benefits and its potential uses in various application areas, including safety, constructability, design review, maintainability, and operability. However, there are mixed conclusions on how beneficial the technology is for design reviews, and there is no agreement among previous studies on whether extended reality has always had positive impacts on design and construction coordination performance. This systematic literature review paper aggregates the results of previous studies that have investigated extended reality for design and construction review considering different parameters such as the type of project, the experience of participants, and the type of extended reality tool used. Challenges associated with implementing the technology for design reviews that were reported previously are synthesized.

As construction projects become more sophisticated, model coordination is critical to mitigating risk. Even though clash detection is highly automated in existing software systems, reviewing clashes and making corrections are still manual and repetitive workflows. Previous researchers leveraged machine learning and data mining techniques to analyze model coordination data and streamline decision-making. Nonetheless, gaps still remain in the fact that existing studies used limited datasets and mostly focused on MEP systems; additionally, no previous study identified which clash attribute combination is necessary to accurately predict clash relevance. By applying an Artificial Neural Network multilayer perceptron algorithm with different combinations of clashes’ attributes in the dataset, the authors achieved a precision of over 80% in predicting clash relevance. Notably, this study contributes to the body of knowledge by identifying the BIM object attributes necessary to predicting clash relevance with high precision using all major disciplines of a construction project.

Concrete 3D printing (C3DP) is a new automated construction technology with a significant potential to reduce construction time and cost. The absence of formwork in C3DP exposes the freshly printed structures to ambient conditions and potentially excessive water evaporation. The freshly printed structures should be able to carry the service loads in a shorter time compared to conventionally cast concrete to make it possible to reduce the overall construction time. The destructive testing methods, which are usually time-consuming, are not ideal for this application, and nondestructive methods are preferred. This article provides a review of existing literature on the strength monitoring of concrete, with a focus on the early-age strength of 3D printed concrete. A new automated curing system was used to compare the effectiveness of three different curing conditions. Moreover, a new index for estimating the flexural strength of 3D printed concrete based on electrical resistivity and temperature history is proposed.

Locational information in building information models (BIMs) is essential for providing geographical context to a project as well as the relative spatial context to each and every individual building element that the project is composed of. From a construction automation perspective, one main application is the use of locational data as input for robot-assisted operations in the construction of building components. Nevertheless, obtaining locational information is a time-intensive, laborious, and error-susceptible process. To address this gap, the authors proposed a logic-based approach for examining BIMs and retrieving the positional data of building elements. A duplex apartment model was used to test the proposed method, which achieved 100% precision and 92.31% recall compared to a gold standard. Building elements, such as columns and beams, from the model were successfully extracted. Results show that logic representation and reasoning can be effectively used for extracting locational information in the context of construction automation.

Recent advances in physiological sensors and machine learning have led to the development of non-invasive heat stress monitoring frameworks that can continuously and objectively assess the heat stress levels of workers in the field by analyzing their physiological data. However, variations in the statistical distribution of physiological data due to individual differences in responses to stressors negatively impact the accuracy of the assessment. To address this issue, this study proposed a transfer learning-based framework to improve the performance of non-invasive heat stress monitoring. The framework utilizes autoencoder and domain adaptation-based transfer learning techniques to reduce the deviation of the statistical distributions of physiological data across different individuals, leading to a more robust assessment of workers’ heat stress levels. To evaluate the effectiveness of the framework, physiological data was collected from 14 subjects performing roofing tasks with different heat stress exposure levels (low, medium, and high). Results showed that the proposed framework had a more robust performance on physiological data with distributional shifts, achieving an accuracy of over 89.9% in assessing heat stress levels across different subjects, a 6.3% improvement compared to existing frameworks. This study contributes to the advancement of heat stress assessment for construction workers.

The installation of Mechanical, Electrical, and Plumbing (MEP) systems for building facilities accounts for up to 40% of the total construction, while Operation and Maintenance activities constitute about 65% of the Facilities Management costs where MEP works are part of the main contributors. In this vein, most companies generate and manage vast amounts of construction and operations data in diverse Information Systems, but they lack data-driven insights to better plan, analyze, and control costs and profit related to construction and maintenance work orders. To tackle this problem, this paper presents a method that helps general contractors and facility managers to automatically discover dominant MEP-related activities from construction data involving 99k work orders of Operation and Maintenance activities and 118k construction work orders executed in several projects across North America. The proposed method involves (1) data preprocessing and automated identification of frequent activities by applying text-mining and Natural Language Processing (NLP) techniques on free-form text activity descriptions; and (2) a quantitative analysis of significant activities in terms of cost and profit. Results include the identification of dominant activities and providing quantitative insights on their occurrence frequency. The presented data-driven approach has the capability to identify items in high demand and those carrying significant profitability or cost implications for general contractors and facility managers.

Hot thermal discomfort poses significant risks to construction workers in hot environments, necessitating effective measures to mitigate heat-related illnesses. This study validates the applicability of a thermal discomfort action recognition model in outdoor construction environments, addressing gaps in existing research. The methodology comprises four steps: data collection, model training, application, and evaluation. The data collection was implemented by the acquisition of video data depicting hot discomfort-related actions within indoor setting. Subsequently, this video dataset served as the foundation for training a thermal discomfort action recognition model using deep learning-based classifier. The model’s performance was rigorously evaluated in real-world outdoor scenarios, specifically in construction environments. Results indicate limited accuracy (0.5833) in predicting hot thermal discomfort-related actions outdoors, highlighting the need for further model refinement. However, a detailed analysis based on different angles and postures provides valuable insights for future improvement. The study emphasizes the importance of diverse datasets encompassing various angles and postures to develop a sophisticated model for accurately recognizing hot discomfort actions in outdoor construction environments. These findings will contribute to enhancing the existing hot thermal discomfort model’s accuracy when the model is applied to outdoor construction fields, ultimately improving worker safety and well-being.

The Occupant-Centric Control (OCC) strategy incorporates occupant information in the building facilities control to improve energy efficiency while maintaining an acceptable level of occupant comfort. Predictive control strategies are necessary to implement OCC in complex systems like HVAC, which pose a significant challenge given the stochasticity of occupant behavior in built environments. Nonetheless, the recent advancements in Machine Learning (ML) and the Internet of Things (IoT) have made data-driven strategies more feasible in OCC of building systems. In this context, Deep Learning (DL) and Reinforcement Learning (RL) techniques have gained significant attention due to their ability to handle large volumes of data and achieve high prediction accuracy. However, the current literature lacks systematic knowledge of algorithm selection in the different OCC contexts. To address this gap, this paper presents a systematic literature review of DL and RL algorithms applied to OCC and provides organized information on the choice of algorithms by classifying occupant information into four levels based on increasing personalization. Subsequently, it identifies the algorithms suitable for each level to establish a systematic foundation for selecting DL and RL algorithms based on the degree of personalization required. The paper also highlights areas for future research in this area.

While the increment in temperature and humidity can alleviate or exacerbate occupational heat strain, the correlations between temperature, humidity, and heat stress exposure level are not well defined across a wide range of hot environments in the construction sector. To this end, this study seeks to quantify the impact of varying levels of humidity and temperature on heat stress exposure levels at construction sites. For this purpose, 10 able-bodied workers were prompted to perform common construction tasks (roofing and material handling) while exposed to varying levels of temperature and humidity. During each task, three biosignals, namely photoplethysmography (PPG), electrodermal activity (EDA), and skin temperature (ST), were collected from the workers through wearable biosensors. To assess the workers’ heat stress level, various metrics were extracted from PPG, EDA, and ST signals. Correlation analysis shows that air temperature and humidity are highly correlated with the workers’ extracted physiological metrics. Analysis revealed compelling evidence for the temperature and humidity ranges that would harm construction workers. This study can inform new plans and policies for occupational health agencies, potentially maintaining health and productivity in a construction site.

The need for secure, dependable, and higher-quality infrastructure systems necessitates more intricate transportation construction and maintenance projects. This research aims to identify six crucial factors for effectively implementing new and emerging technologies in State Departments of Transportation (DOTs). There is a growing interest among DOTs to implement emerging and wireless technologies, as seen and supported in various Federal Highway Administration (FHWA) Every Day Counts (EDC) initiatives. The study employs a mixed-methods approach, using survey responses from DOT personnel to determine the relative importance of technology implementation factors and subfactors. This study investigates six factors for successful technology implementation: Organization Structure, IT Infrastructure, Data Security, Information Workflow, Personnel Training, and Stakeholder Engagement. The relative importance of People, Process, and Technology was also determined for each of the six factors of technology implementation. The study also evaluates different criteria within each technology implementation factor. The paper aims to provide guidelines to different agencies regarding successful technology deployment by providing an in-depth overview of key factors for successful implementation. It is important to understand the factors of technology implementation to successfully deploy emerging technologies, which eventually increases much-needed efficiency and productivity in highway construction and asset management.

There has been an increasing interest in smart cities and digital technology/information to enhance the quality of services, while reducing costs and resource consumption. Hence, the “shared economies” notion was introduced accordingly, which is a modern business concept highlighting people’s preference to share/rent goods and services (e.g., space, transportation, utilities) rather than to buy and/or own them. Smart cities enabled by peer-to-peer internet capabilities—e.g., IoT technology and devices—can fuel the growth of shared economies by improving information transparency, rationalizing resource allocation, and reducing waste. This paper proposes a dynamic pricing model developed from an innovative renter reliability index, which is measured and created by IoT-enabled and real-time data. The proposed intelligent dynamic IoT model was used and tested in a real-world long-term shared economy environment (i.e., a rental space) of a smart city in New Jersey, USA. The results showed that the proposed model can significantly enhance the transparency and legitimacy of rental pricing in a shared. This paper contributes to the literature by improving the efficiency of the shared products/services, the allocation of resources, and the pricing rationalization, which ultimately stimulates the shared economy concept in smart cities. The proposed model is useful for both the providers and users of shared economy systems, particularly those of shared properties (i.e., rental spaces).

Unmanned Aircraft Systems (UAS) have been increasingly used in the inspections of civil infrastructure due to their superior capabilities of fast deployment, comprehensive coverage, and labor-saving field operations. Depending on the use of the infrastructure, while some of them allow a visual check of the structural integrity, others require accurate inspection and verification of the structural members to resist natural hazards, such as flooding and hurricane. Photogrammetry and LIDAR are the two most commonly applied techniques among UAS inspection approaches. With the vast variety of available UAS airframes, navigation systems, and measurement sensors, the results of UAS inspections are not always consistent and comparable. This study presents the findings of measurements between a commercial photogrammetry UAS system and a customized LIDAR UAS system in bridge structure inspection. The measurements from two UAS systems were compared against the bridge plans and analyzed in terms of the span, deck, and foundation, and the results suggested that most UAS measurements are within 1% of the design values. The findings of the study demonstrate the workflow of using a UAS system for infrastructure inspection with both photogrammetry and LIDAR sensors and provide valuable insights into the accuracy of both UAS inspection techniques.

Evaluating the deterioration process of bridges is a fundamental part of a good bridge management program by providing cost-effective maintenance strategies, given agency-defined goals and constraints. The deterioration of bridge structures over their lifetimes depends on various factors, such as material types, design types, geographic locations, and operational and environmental conditions. Grouping bridges at the level of components or elements is necessary to reduce data dimensionality in data analysis and formalize deterioration models through a statistical analysis while producing the same analytical results (i.e., homogeneous deterioration characteristics). However, grouping factors by which bridge structures are believed to show similar deterioration characteristics over time are determined based on the improvised, heuristic classification of bridges. This study conducted a data-driven similarity analysis to statistically determine grouping factors for bridge components. The results of this study demonstrated the effectiveness of the similarity analysis approach used in this paper. This research makes noteworthy contributions by introducing a novel data-driven methodology for identifying factors that facilitate the grouping of bridges, an approach that has not been explored before. Additionally, it enhances homogeneity within bridge groups, improving the reliability and robustness of bridge deterioration models.

Bridge textual reports include massive amounts of data/information about bridge conditions. Extracting and analyzing this data offers opportunities to enhance bridge maintenance decision making. However, utilizing the data/information embedded in textual bridge reports for bridge data analytics remains to be a challenge due to the inherent coreference ambiguities in the text. There is a need for semantic analysis methods to disambiguate coreferences inside sentences to enhance the performance of data/information extraction from multiple types of reports. There is a lack of efforts that addressed coreference resolution in this domain. To address this gap, this paper proposes a deep learning-based coreference resolution model that automatically disambiguates the coreference ambiguities from multiple types of unstructured textual bridge reports and represents the extracted data/information in a structured format to support further data analytics. The proposed model utilizes a deep learning-based bidirectional long short-term memory (BiLSTM); a begin, inside, and outside (BIO) encoding schema; convolutional neural network (CNN) features; and span representations for coreference resolution. The model was tested on multiple types of textual bridge reports, including bridge repair manuals, bridge assessment reports, and bridge rehabilitation reports. The proposed model showed promising coreference resolution performance and generalizability across different types of reports.

The traditional bridge inspection method, which relies on manual visual inspection, is time-consuming, labor-intensive, and could be dangerous. Recent automated bridge inspection approaches aim to utilize unmanned aerial vehicles (UAVs) and computer vision techniques to collect and analyze images to improve the inspection process. A survey of existing literature and tools shows that defect detection/segmentation has been studied extensively. However, there has been little effort focused on segmenting and characterizing the bridge components that have the defects. The identification and characterization of bridge components is essential for bridge inspection, which can contextualize the defects to determine their importance in maintenance decision making. Moreover, existing bridge component recognition approaches lack generalizability in the presence of a variety of bridge types, complex background scenes, and varying shot sizes. To address these gaps, this paper proposes a convolutional neural network (CNN)-based image segmentation method to segment bridge components, which leverages DeepLabv3+ and pre-training from ImageNet to improve feature extraction and generalizability. The proposed method was trained and tested end to end on 13 classes based on the Federal Highway Administration (FHWA)’s Bridge Inspector’s Reference Manual. It achieved a mean precision, recall, F-1 measure, and Intersection over Union (IoU) of 86.7%, 78.2%, 81.4%, and 70.4%, respectively.

The architecture, engineering, and construction (AEC) industry is widely known for being fragmented. In this situation, blockchain technology has been introduced as a promising solution to bring trust and transparency to the industry. As blockchain technology continues to evolve, it has embraced new capabilities and features, including Fungible-Tokens (FTs) and Non-Fungible Tokens (NFTs), which have recently drawn significant attention in various industries. Tokenization can be seen as a next-generation solution to address challenges in the AEC. However, the application of tokenization for the AEC industry has remained undeveloped. To address this gap, this study provides an overview of tokenization, followed by an investigation of its applications in the AEC industry using a literature review method. By illustrating a clear outlook of the potential advantages and challenges, the study helps to set realistic expectations over the potential improvement that tokenization can bring. This can also serve as a fundamental source for further investigation by researchers and practitioners.

On-site sorting is a critical task to support construction circularity toward the increase in the diversion of construction, renovation, and demolition waste (CRD) from the landfill. Image recognition technologies have proved their capacity to facilitate on-site sorting activities by automatically identifying the type and location of the generated waste. However, the labeling of images, while a crucial step in training image processing models, is a time-consuming and error-prone process task. Consequently, finding a dataset of high quality with a large number of CRD waste images is challenging. Thus, with this in mind, this paper aims to assess the performance of the Segment Anything Model (SAM) in generating masks for CRD waste by assessing the labeling process and performance of the model trained with the SAM labeled dataset. Results of this study have shown that even though the use of SAM slightly damages the level of accuracy of the trained model, the use of SAM dramatically eases the labeling process.

This study addresses the gap in research concerning the segmentation of smaller infrastructure, such as fire hydrants, within large-scale point cloud models. The RandLA-Net model, an efficient and robust deep learning algorithm known for its ability to process vast quantities of data while maintaining high accuracy rapidly, was evaluated for this task. A comprehensive dataset was constructed using LiDAR technology to gather point cloud data from diverse urban scenes. The data was manually segmented to assign unique class identifiers to distinct objects within the point cloud. The model was trained using these annotated data, and its performance was assessed using Intersection over Union (IoU), a typical evaluation metric for segmentation tasks. The results demonstrated the model’s impressive capability in segmenting different classes within large-scale urban point cloud data. However, segmenting smaller, less distinctive objects like fire hydrants revealed room for improvement. Future work should focus on improving the recognition of complex or underrepresented classes and enhancing training stability. The study’s findings provide critical insights into the capabilities and limitations of the RandLA-Net model in the context of large-scale, real-world point cloud data segmentation tasks.

Digital Twins (DTs) have recently drawn considerable attention in the AEC industry and academia. Despite their applications, most models typically represent an isolated DT without considering how they can connect to other DTs, where they can share data. Blockchain-based solutions, and specifically Non-Fungible Tokens (NFTs), have provided various features that can be seen as a promising solution to build the lost bridge for data sharing between different stakeholders in DTs. This study capitalizes the advantages of NFTs to propose a framework for DT knowledge sharing. This framework is followed by providing future prospects of data management in the DT field. Our solution enables the ownership of various DT components, leading to the establishment of a secure data marketplace within the context of the AEC virtual world. This study can also serve as a fundamental source for further investigation by researchers and practitioners to develop an interoperable built Metaverse, where all DTs, in all Levels of Details, can exchange knowledge.

The complexity of construction projects and the multiplicity of stakeholders, organizations, and standards involved call for a unified and common language to support linked data and allow for semantic inference among multiple projects, processes, and actors. While several taxonomies are developed for knowledge management in construction, most of them either have remained at a high level of information exchange or are focused on adapting project data with standards in a specific domain and have served a very narrow and particular range of queries. Although taxonomies that are established at such levels create a broad and general framework for organizing knowledge, they will not enable “cross-analyses” because of incomplete coverage of a particular domain. Hence, this paper proposes a user-level taxonomy, focusing on critical administrative processes in construction projects, that is, change order management. Text mining was applied to extract key concepts from the data of past projects focusing on two main facets, namely, processes and actors. Taxonomy has been developed to serve as the foundation of a new ontology that aims to enable organizations to perform cross-facet and cross-phase semantic queries in large construction databases.

To assess the condition of infrastructure such as bridges, highways, and pipelines, trained inspectors typically perform visual inspections with site visits. However, this conventional method can interrupt the regular functioning of the infrastructure as well as be time-consuming, laborious, expensive, and dangerous. To address these problems, this study proposes an immersive metaverse-based infrastructure remote virtual inspection platform. Based on a 3D digital twin model created through digital photogrammetry, this study generates an immersive metaverse to support collaborations between stakeholders for remote inspections. In addition, this study synchronizes a virtual reality (VR) headset with an unmanned aerial vehicle to remotely inspect infrastructure conditions at the target location using augmented real-time images mapped onto the digital twin model. The conducted experiments provided evidence that users equipped with VR headsets were able to see the 3D digital twin model of the infrastructure as if they were physically present on the sites. Within the metaverse space, users could easily navigate to specific areas and conduct inspection tasks by viewing both the digital twin model and real-time images through the proposed platform. The outcomes of this study hold the potential to enhance the efficiency and safety of infrastructure inspection.

Onsite coordination with Building Information Modeling (BIM) could improve collaboration and prevent conflicts between stakeholders during the construction phase. Recent research shows that mixed reality technology can extend the usage of BIM and enhance coordination to the next level. However, current practices often use LOD300, which provides inadequate information for MEP system coordination and can lead to clashes and reworks. On the other hand, the MR interaction is often limited to single-user scenarios. This paper proposes an onsite collaboration framework for multiple MR device users to examine the MEP system from lower LOD models in the construction phase. This framework includes creating an MEP system prototyping, generating an MR interaction module, developing a multi-user collaboration module, and establishing a report model. The system allows managers to check the MEP system’s position, communicate issues, make changes, enable design change coordination, resolve clashes, and reduce reworks. We evaluate this framework through a case study, using scenarios such as moving components and measuring distances with multiple devices.

Recent advancements in wearable physiological sensing and artificial intelligence have made some remarkable progress in workers’ health monitoring in construction sites. However, the scalable application is still challenging. One of the major complications for deployment has been the distribution shift observed in the physiological data obtained through sensors. This study develops a deep adversarial domain adaptation framework to adapt to out-of-distribution data (ODD) in the wearable physiological device based on photoplethysmography (PPG). The domain adaptation framework is developed and validated with reference to the heart rate predictor based on PPG. A heart rate predictor module comprising feature generating encoder and predictor is initially trained with data from a given training domain. An unsupervised adversarial domain adaptation method is then implemented for the test domain. In the domain adaptation process, the encoder network is adapted to generate domain invariant features for the test domain using discriminator-based adversarial optimization. The results demonstrate that this approach can effectively accomplish domain adaptation, as evidenced by a 27.68% reduction in heart rate prediction error for the test domain. The proposed framework offers potential for scaled adaptation in the jobsite by addressing the ODD problem.

The COVID-19 pandemic had a massive impact on work-life, with most workers forced to work from home due to COVID-19 lockdown restrictions. This study aims to explore the trends in residential energy-use-related behavior patterns from 2019 to 2021. Agglomerative hierarchical clustering applied to respondents’ daily activity data extracted from the American Time Use Survey data revealed changes in respondents’ activity profiles over the period. A time series analysis of the daily energy-use activities patterns in the residential setting showed significant relocation of work-related tasks from the commercial to residential settings. The studies revealed a more than 5% increase in Work from Home (WFH) from 2019 to 2021 and an over 4% decrease in Work from Office (WFO) from 2019 to 2021. These findings provide information on residential energy-use-related behavior trends to energy policy decision-makers and can help make critical energy policy decisions that can help lessen the energy bill burden and promote energy equity and justice.

Departments of Transportation (DOTs) are constantly looking for more efficient ways to inspect culverts. This study proposes a data-driven approach based on federated learning for culvert inspection in Utah. As the Utah DOT (UDOT) had a limited dataset of culverts in Utah, we collected data from several other state DOTs. However, instead of using traditional centralized machine learning techniques, we used a federated learning approach in which data from other DOTs was not shared with UDOT, but rather developed local models based on those datasets were shared. This allows us to keep the data of the DOTs private while benefiting from their collective knowledge. As a result of using federated learning, the performance of UDOT’s culvert condition prediction model was improved by 16%. The study demonstrates that DOTs can mutually benefit in scenarios of data scarcity, while still adhering to their preference for not sharing data directly with one another.

Unmanned aerial systems (UASs) are now standard tools for 3D mapping across diverse terrains. Despite their ubiquity, achieving consistent and reliable mapping results across varied terrains remains a challenge due to differing site conditions. To ensure consistency and the highest accuracy across terrains, it is crucial to discern the optimal flight configuration (defined as the UAS’s predetermined flight pattern and altitude to capture overlapping survey images). This paper delves into four distinct terrains: construction, pit, preconstruction, and rock surface sites, collecting photogrammetry data from varied flight configurations. The research revealed that all sites achieved Class A horizontal accuracy, with the best accuracy levels falling within 1−3 times the Ground Sampling Distance (GSD). The findings of this study offer valuable insights to surveyors and other professionals by elucidating the level of accuracy achievable for UAV-based photogrammetric surveying across various sites, enabling them to make informed decisions about flight configurations and data collection methods for more precise and efficient surveying results.

With the purpose of facilitating process tracking such as inspection reports and progress monitoring, the AEC industry has adopted an as-built 3D model that is reconstructed using a 3D scanner during or after construction. In response to the laborious difficulties of converting a point cloud into a semantically rich model, for example, BIM, researchers are attempting to automate this process via machine learning, applying 3D semantic segmentation and parametric modeling. However, there are no publicly accessible 3D datasets that target construction sites, regarded as unstructured and cluttered scenes, thus yielding a barrier to construction scene segmentation development. To this end, this paper aims to generate a 3D construction dataset that can be utilized for machine learning models requiring ground truth and to suggest foundation processing for general scene segmentation on construction datasets. In addition, we identify and discuss several challenges pertaining to construction sites, in terms of 3D semantic segmentation.

Runways are among the most important assets for many airports due to their high life-cycle costs and functional importance. Runway inspections are mainly conducted through manual inspections requiring visual identification of distresses. These practices may suffer from potential inaccuracies and biases due to the time- and resource-consuming nature of these inspection procedures. Manual data collection procedures also hinder the potential for adopting more sophisticated approaches to asset management featuring artificial intelligence and digital twin technologies. In this paper, the potential of photogrammetry and laser scanning—two emerging methods to construct 3D models—in improving runway inspection procedures is examined. Results obtained from online surveys of airport authorities and industry leaders and semi-structured interviews are presented. A road map is presented to assist airport authorities with adopting photogrammetry and laser scanning technologies to improve their condition assessment procedures. While these technologies have found applications in other domains, their adoption in the aviation sector has been somewhat limited. It is anticipated that this paper and other publications in this area will foster further research and practical applications in the 3D modeling of airport infrastructure, resulting in considerable improvements to airport asset management practices.

Recognizing daily human activities offers a great promise to develop human-centered efficient, assistive, and healthy built environments. However, the state-of-the-art sensing methods for human activity recognition are mostly intrusive: they either rely on capturing private personal information or require humans to wear sensors. Such intrusive sensing often raises privacy concerns or suffers from adherence problems (i.e., people stop wearing the sensors with time). There is, thus, a need for a nonintrusive sensing method to better support daily activity recognition in buildings. To address this need, this paper proposes a novel nonintrusive sensing and analytics method. At the cornerstone of the proposed method is a new multi-purpose sensing system, which captures the composition changes of multiple indoor gases induced by daily activities, without capturing private occupant information and requiring sensor wearing, for supporting activity recognition. As a pilot study, this paper focuses on evaluating the feasibility of the proposed nonintrusive sensing method by testing the significance of the differences in air composition data collected under different daily activities (e.g., cooking, sleeping, and idling). The experimental results show the feasibility of the proposed method to recognize daily human activities in a nonintrusive way.

Methodologies for conducting highway infrastructure inspection activities are experiencing a substantial change with the proliferation of technology. Instead of relying on destructive testing methods to perform inspections, state departments of transportation (DOTs) use a wide range of non-destructive testing (NDT) methods for inspecting highway construction projects and assets. However, the varying approaches adopted among the state DOTs can lead to confusion and inconsistency. Therefore, this study aims to provide a comprehensive and systematic overview of several NDT methods available for inspection activities used across the nation. A national survey of 50 state DOTs was conducted to collect data. The results showed that the typical NDT methods used for highway construction inspection include nuclear density gauges, dynamic test loading for piles, cross-hole sonic logging for drilled shafts, surface profile measuring systems, ground-penetrating radar, and ultrasonic testing. The typical inspection activities during construction using NDT are in situ material characterization, foundation investigation, structural inspection, and bridge deck deterioration. The top inspection activities involving NDT technologies during asset maintenance include structural inspection, automated pavement distress measurements, and identification of fracture damage. This study contributes to the literature by summarizing and analyzing the NDT technologies for highway infrastructure inspection activities.

The integration of cutting-edge technologies has become prevalent, promoting state departments of transportation (DOTs) to progressively adopt an array of digital tools and advanced technologies to improve their highway construction projects. There is a wide range of remote sensing and monitoring technologies, including light detection and ranging (LIDAR)/3D laser scanning, radio frequency identification (RFID), remote sensors (RS), infrared sensors (IS), and unmanned aerial systems. However, the varying approaches adopted among the state DOTs can lead to unclearness and inconsistency. This study aims at investigating how state DOTs use sensing and monitoring technologies for highway construction inspection during construction and asset management. This study engaged a research methodology that included an extensive literature review and a survey questionnaire of 50 state DOTs. Results show that 28 state DOTs use remote sensing and monitoring technologies to measure material strength and temperature, capture site photos and videos, and measure construction progress. The main challenges to using sensing and monitoring technologies are a lack of training, device maintenance and user support, a lack of reliable internet connection, and cost issues. This research contributes to the body of knowledge by exploring the use of sensing and monitoring technologies for highway construction inspection.

The trend of remodeling living spaces has risen exponentially due to pandemic-led work-from-home culture. Consequently, a sharp rise in material costs, labor shortage, and material lead times were observed. Homeowners increasingly resorted to a DIY (do-it-yourself) approach to remodel their existing homes. However, they face several challenges such as lack of tools to visualize different designs and a centralized platform to compare products from various vendors. None of the existing applications provide essential information (material availability, costs from different suppliers) and AR features (projecting digital models in real world) to cater to the diverse needs of homeowners. This study aims to conduct user market research, determine value propositions, and identify multiple revenue streams of such an application using surveys collected from diverse customer segments (e.g., homeowners, contractors, realtors, and suppliers). A major outcome of this study is to gain insights into the unique value proposition of an AR-based application for each customer segment in the home remodeling market. In addition, this study will also try to examine the necessary and supplementary application feature(s) required by the respective customer segments. The findings are therefore believed to transform the home remodeling market.

The advancement of Artificial Intelligence (AI)-driven defect detection has already demonstrated promises to boost quality assurance and control, as well as condition assessment in the built environment. However, training defect detection models requires hefty amounts of reality capture data, and labeling is considered expensive. In most cases, such data may not cover all situations of defects. Synthetic data, most recently made with Building Information Models (BIM), is turbocharging model development for learning defect features. Nevertheless, few studies focused on characterizing defects to classify their severity, which is crucial to the condition assessment. To that end, this study explores the requirements for generating synthetic data. Parametric physics-based modeling approaches are carefully examined. Using the underlying geometric properties of such data, the condition of each defect can be determined. The feasibility of synthetic defect data is validated with a case study of crack segmentation using the transformer-based model, SegFormer. Examples of how different scenarios can be generated photo-realistically with the use of physics-based rendering for creating varying geometrical characteristics, appearance, and viewpoints of defects are presented. The generated synthetic crack datasets can successfully be used to train the SegFormer model and reach promising predictions on real crack images.

Transportation infrastructure systems are subject to deterioration because of several factors, such as traffic loads, weather conditions, construction quality, and age. Therefore, regular assessment and maintenance are required to keep these systems in acceptable condition to provide their intended functions. However, traditional assessment procedures are time, cost, and labor-intensive. Thus, several Departments of Transportation (DOTs) have moved toward more innovative assessment methods, such as Unmanned Aerial Vehicles (UAVs). However, the use of UAVs in infrastructure assessment is still developing, and many DOTs face limitations and challenges in fully adopting UAVs for infrastructure assessment. Thus, this paper aims to identify the DOTs’ current status of UAV applications and the challenges of these applications in infrastructure systems assessment from five perspectives: Social, Workability, Operational, Economic, and Environmental. A survey has been distributed to 50 state DOTs, and responses from 25 DOTs have been recorded. Based on the survey results, this paper identified and categorized 25 limitations from the five perspectives. Finally, this paper maps the future research needs based on existing limitations and previous research.

The advent of robotic systems has brought significant transformations across various industries, increasing the quality of products and services. However, due to construction projects’ intricacy and robotic manufacturing’s technological challenges, robotics in the construction sector is still in the nascent stages of development. The variability in construction materials presents a major challenge to the integration of robotics-based manufacturing. Lumber misalignments can cause costly reworks to the wood framing process due to lumber damage, structural deviations, and nail gun misfires. This paper seeks to address the critical quality control challenge for robotic-based manufacturing in industrialized construction. The proposed automated quality control system detects alignment issues using computer vision and deep learning technology. Detected misalignments are transmitted through a graphical user interface (GUI) to construction workers to allow them to determine whether corrective actions are required or not. The field experiments illustrated the significance of the proposed system in ensuring a proper framing process and enhancing the quality, safety, and productivity of robotic manufacturing in industrialized construction.

Mixed reality has achieved acceptance within higher education owing to its ability to provide experiential learning with enhanced engagement and attention. In construction education, mixed reality environments can serve as alternative learning platforms for jobsite visits and hands-on learning without exposure to jobsite hazards. However, the design of such learning environments is often oblivious to users’ cognitive states. As human working memory is limited in capacity, cognitive overload is inherent in such learning environments and inadvertently impacts learning efficiency and knowledge retention. This study presents the assessment of a mixed reality environment for learning sensing technologies in construction education and investigates the effects of cognitive load within the learning environment. The study shows the relationship between cognitive load and students’ learning outcomes and the impacts of students’ demographics on cognitive load. This study sets precedence for exploring intelligent learning environments that adapt to students’ cognitive load for improved learning experiences.

Indoor air temperature, which is one of the main factors affecting the thermal comfort of occupants, varies across locations/spaces. However, current occupant thermal comfort models rely on predefined formulas or data-driven approaches that often ignore the importance of the specific location in the room at which the sensor is placed. This research aims to study the impact of sensor location on occupant thermal comfort state prediction. A set of 90-min occupant experiments were conducted in a controlled environment. Multiple temperature and humidity sensors were placed at different locations in the room. During the experiments, the room temperature changed from 19°C to 29°C, and the humidity, mean radiant temperature, and wind speed were controlled. The subjects performed office duties and provided feedback about their thermal comfort periodically. Personal parameter data were also collected. For each sensor location, a thermal comfort state model was developed using the XGBoost algorithm. Each model was tested in predicting the occupant comfort state using temperature and humidity data from other room locations. The results showed that the location of indoor parameter data used for prediction could affect model performances by up to ±7.2% accuracy and ±8.0% F1-measure.

The US Department of Energy’s Industrial Assessment Centers (IACs) conduct energy assessments to increase energy efficiency of small- and medium-sized industrial facilities and commercial buildings. However, it is not always possible to have all the necessary experts on-site, leading to lost potential ideas and recommendations. In addition, new energy assessment team members should attend as many assessments as possible to gain experience, but scheduling conflicts may prevent them from attending all. Virtual Reality (VR) could potentially address these issues, but no studies have been conducted on its use for energy assessments. This research aims to develop a virtual energy assessment process for industrial and commercial buildings and evaluate, based on participants’ feedback, how well the use of the technology mimics a real-world environment. Virtual environments are created using an infrared scanning device, and participants are trained in how to use VR and conduct energy assessments. Results from participant surveys suggest that VR environments closely mimic real-world environments and can be used for learning and training in energy assessments.

Data-driven energy prediction models can help urban planners and policymakers evaluate urban energy consumption patterns and then make informed decisions on how to improve urban energy efficiency. Typically, energy use data and building characteristics data were used to train these data-driven models. Few research utilized water use data and socio-demographic data, which have nexus with the energy consumption of buildings. This research utilized energy use data, building characteristics, socio-demographic, and water use data of multi-family buildings in New York City to train Least Absolute Shrinkage and Selection Operator (LASSO), Ridge Regression (RR), Support Vector Regression (SVR), and Random Forest Regression (RFR) machining learning models. The effects of socio-demographic and water use features on the performance of energy prediction were analyzed. Results showed that water use feature had significant positive impacts on the performance of LASSO, RR, and RFR models. Socio-demographic features had obvious positive impacts on the performance of SVR and RFR models. RFR trained with the BW dataset (including building characteristic features and water use feature) performed the best.

This research proposed a Convolutional Neural Network (CNN) model for Asphalt Concrete Computed Tomography (CT) Scan image classification, which uses SoftMax activation for the output layer to yield the Air Void prediction with a testing accuracy of 0.9684, weighted average precision, recall, and f1-score are 0.97 in the early stopped model, and 0.99 in the best model (smallest validation loss). Moreover, by switching to Linear activation, the CNN model can generate numerical values for Air Void, which has the average percentage difference of 0.42%, −1.47%, 2.87%, 2.3%, and 2.27% for Air Void contents of 10.4%, 15.6%, 18.9%, 21.9%, and 25.6%, respectively. Furthermore, by integrating the CT image with Grad-CAM++ generated heatmap, instance segmentation of the Asphalt Concrete CT image as components of Air Void, Binder, and Aggregates is more effective than using image only. The application of using segmentation results for Asphalt Concrete 3D modeling is also presented.

Unmanned aerial vehicles (UAVs) have recently become popular in collecting positional data of road users. In comparison to other tools at ground level, UAVs have the advantages of low cost, wider view coverage, and significantly less occlusion. However, the depth relief of road users and the perspective distortion of the onboard camera induce nonnegligible error while applying UAVs for localization of road users. This study proposed a method for accurate road user localization in aerial images. First, the localization error induced by the depth relief and perspective distortion was examined and modeled. Then, a deep-learning-based method was applied for automatic road user detection and localization in the aerial images by leveraging oriented bounding boxes to achieve higher localization accuracy compared to applying horizontal bounding boxes. Finally, an error compensation strategy was proposed to eliminate the perspective- and depth-relief-induced localization error by rectifying the oriented bounding boxes obtained from the previous step. Field experiments were conducted to evaluate the method’s performance. The results demonstrated its promising accuracy for road user location estimation and its potential to improve the reliability of UAVs in traffic applications.

The integration of deep learning (DL) technologies into construction offers great potential for promoting the level of automation in construction. However, the implementation of the DL model requires the acquisition of substantial data, which is error-prone and time-consuming. Additionally, due to safety and privacy concerns, not all real-world data can be retrieved. To address these issues, synthetic simulated data have emerged as promising alternatives, and various methods have been developed to generate such data. However, currently there is neither a summary of synthetic simulated data generation methods nor unified metrics to evaluate these methods. In this paper, a comprehensive review of 129 scholarly articles from Web of Science is conducted. Based on the source of data assets and the techniques employed for their combination, we categorize synthetic simulated data generation methods into seven distinct categories. Furthermore, we summarize seven metrics for evaluating these methods and consolidate the evaluation results in a table. The provided table serves as a reference for practitioners in identifying and selecting suitable synthetic simulated data generation methods for their applications.

Digital Twin (DT) refers to a digital technology that virtually represents a physical asset and reflects its behavior in real time. As the fully implemented DT is largely unachievable, a maturity spectrum has been proposed representing different milestones for DT implementation. DT Enabling technologies vary depending on the maturity level of DT. This paper intends to review DT cases in construction to help DT adopters understand DT-enabling technologies based on the maturity level of their demanded DT. The outcome of this study provides DT adopters with practical insights into how to decide the DT maturity level that meets their demand and what are the enabling technologies that support their DT development. With maturity level, AEC practitioners gain forward-looking insights on how to expand existing DTs or how to strategically apply their technological skillset to explore the potential of creating DTs, rather than solely focusing on their current job scope and needs.

Decision-making in the construction industry involves a high level of uncertainty stemming from numerous external factors, which are outside of the control of decision makers but cause the industry’s complex and dynamic behavior. As such, informed decision-making requires the identification of various external factors and continuous monitoring and analysis of abundant information derived from them. However, the existing data-driven approaches focus on limited aspects of external factors to understand the industry, which is insufficient for long-term policymaking since more diverse external factors are associated with the long-term prospect of the industry. To address this gap, this study proposes a composite index framework to allow decision-makers to monitor and analyze factors across various aspects of the construction industry. The composite index framework creates a comprehensive hierarchy of all influencing factors, allowing decision-makers to synthesize the data and monitor the dynamic behavior of the industry.

Heavy equipment contributes to accidents in construction as a major category of causes. Various studies develop simulation models to investigate the correlation between equipment’s state variables (e.g., center of mass, orientation, velocity, and acceleration) during movement and the risk of accidents such as tip-over. However, the impact of material attached to equipment is rarely incorporated as the dynamic state variables of the material are not readily available onsite. To address this issue, this research streamlined the process of preparing the training data, which enables the implementation of existing vision-based 6-DoF object pose estimation models. An evaluation test is performed to demonstrate the potential of the proposed approach. This research also develops the method to convert the estimated pose from the image’s coordinate system to a world coordinate system, which enables estimating state variables in a dynamics simulation. Finally, a simple stability check using the stability triangle is demonstrated using the vision-based approach.

The cross slope is a critical geometric feature of pavement surfaces as it affects safety due to its relationship to the potential of hydroplaning during wet weather. Appropriate cross slopes provide adequate drainage, so water will run off the surface to a drainage system such as street gutters (urban streets) or side ditches (rural facilities). An inadequate cross slope could lead to several safety issues, including hydroplaning, loss of control, and run-off-road crashes. This research study compared two methods of data collection, namely a conventional survey and a LiDAR-based survey using a terrestrial laser scanner, to evaluate roadway surface’s cross slopes. Two existing rural road segments in San Luis Obispo County (California, USA) were selected for evaluation. A comparison between the results from the two methods showed that the difference follows a normal distribution, indicating no systematic errors during data collection. Also, the two-sided paired t-test between the traditional survey surveying and LiDAR shows no statistically significant differences between the slopes estimated using the two methods. Moreover, the results indicate that the difference between LiDAR-derived cross slopes and field surveying measurements is less than 0.2% at a 95% confidence level.

The green building (GB) sector contends with a significant energy performance gap. Building information modeling (BIM), Artificial Intelligence (AI), and Internet of Things (IoT) technologies can address this issue effectively by optimizing design and accurately predicting and monitoring energy consumption. However, research on integrating BIM and AI of Things (AIoT) for GB is nascent. Intelligent processing and analyzing heterogeneous data schema from various information systems is the main challenge faced by many researchers in GB domain. Thus, this study aims to systematically analyze the application of BIM and AIoT in GB and construct an integration framework for improving energy performance. In addition, this framework illustrates how to exchange, transmit, and process massive amounts of heterogeneous data from BIM and IoT platforms by leveraging AI and Semantic Web technologies. Results show that BIM and AIoT integration can assist in intelligent energy-saving decisions through effective data exchange, cloud/edge/fog computing, and user interface (UI). This research contributes to the creation of the BIM-AIoT integration framework. This framework lays a foundation for energy efficiency, facility management, and intelligent construction in the GB domain. Finally, this research highlights the challenges and recommendations related to BIM-AIoT applications in GB.

For successful long-duration missions and to establish a long-term presence of mankind on the Moon and Mars, the construction of supporting structures is imperative to protect humans and enable exploratory logistics in the extreme space environment. Extrusion-based Construction 3D Printing (C3DP) offers a large-scale automated construction approach that can be leveraged for planetary construction. Although Portland cement concrete (PCC) is the most common material for C3DP and traditional construction on Earth, it is difficult to obtain on the Moon and Mars. Sulfur concrete is an alternative waterless construction material that is viable for off-world construction due to the availability of its ingredients. This study investigates the behavior of sulfur-regolith concrete as a candidate material for planetary C3DP, by studying the printability and performance of 3D printed samples made with a Martian regolith simulant. Based on the experimental findings, the printability and mechanical performance of sulfur-regolith concrete can be impacted by process parameters such as the printing temperature and interlayer time gap. Furthermore, the performance data for 3D printed specimens under vacuum conditions and temperature variations showed weight loss and flexural strength reduction as a result of sulfur sublimation.

The state-of-the-art PDF2BIM algorithms enable semi-automatic creation of 3D geometric building information models (BIMs) of bridges based on 2D bridge plans. However, this 3D geometric model is represented as one entity instance with no semantic information associated with it (e.g., concrete strength, structure type). To pursue the full potential of Industry Foundation Classes (IFC) representations in BIM for bridges, in this paper, the authors proposed a framework to segment the bridges into different components based on their semantic features and further assign semantic information to them accordingly. The proposed framework was tested on four bridges, which shows promising results. The semantically enriched bridge models can enable more accurate analysis and evaluation of bridge performance, facilitate better asset management and maintenance strategies, and enhance communication among stakeholders, including designers, engineers, contractors, and asset managers.

Automation in construction is essential to improve the efficiency of this sector. There have been many advancements in this field, and the use of natural language processing (NLP) has attracted increasing interest in recent years. This systematic literature review considers the contributions made in the past decade in the specific domain of implementation of NLP models in the commercial construction industry to help improve efficiency in project management. The review focuses on identifying relevant research efforts and their contributions, analyzing the NLP techniques used and results achieved, and determining the gaps and limitations based on the literature. Reviewed articles were selected from a total of 565 records retrieved from the Scopus, ASCE, and ProQuest databases. The articles were then filtered, classified, summarized, and analyzed. The review resulted in the identification of 19 most interesting articles in which the developed technologies have the potential of immediate applications in different project phases—from documentation and bidding, project closeout, to post-construction facility management. It was concluded that improving and automating the processes in these project phases with NLP resulted in increased project efficiency. Some gaps and limitations in the literature review were identified, and recommendations for future research were provided.

This study aims to provide a feasible framework in achieving automation of construction contractual clause interpretation and analysis through literature review on recent publications from the turn of the century to date. The concept of “smart contracts” was investigated in depth regarding its correlation and interoperability with other automation technologies, namely building information modeling, sensing technologies, virtual reality and augmented reality, and construction robotics. As a result, the proposed framework integrates knowledge modeling for computer-interpretable contractual clause formulation, automated reasoning for contractual monitoring and tracking (e.g., for fulfillment of bonds and securities that bind the involved parties), and machine learning for predictive model development for action planning and decision support. Furthermore, the use of blockchain technology proves to be beneficial to support the payment process. Last but not least, the most cutting-edge deep learning technologies such as ChatGPT by OpenAI were analyzed in this context. Research and development needs were identified, and recommendations were made for future research. This framework opens the door to more extensive research in construction contract analysis and management leveraging cutting-edge automation technologies.

Current bridge projects mainly rely on PDF plans as the official deliverables and documents to be stored, communicated, and transferred among different stakeholders. With the Industry Foundation Classes’ (IFC) Building Information Modeling (BIM) standard adopted by the American Association of State Highway and Transportation Officials (AASHTO) as the national standard for modeling bridge and road infrastructure projects, upgrading the documentation of bridge projects to 3D BIM in compliance with the national standard has become an urgent need. In this research, the state-of-the-art PDF2BIM algorithms were leveraged to semi-automatically create 3D geometric models of bridges based on PDF drawings. To enrich the 3D geometric model with semantic information of bridges’ components (e.g., bridge name, structure type, concrete strength), information extraction algorithms based on optical character recognition (OCR) and natural language processing (NLP) were developed to extract data from the bridge plans automatically. It significantly increases the efficiency and productivity of information extraction and enrichment of IFC-based BIM instance models for bridges by leveraging the rich information that already resides in the PDF plans. The results show that it achieved 97.6% accuracy in the information extraction task and reduced the overall time consumption on processing bridge data by 96.3% compared to the manual approach.

Small collaborative ground robots that are efficiently capable of accomplishing a variety of tasks have become more ubiquitous on jobsites. With the advancement in technology and automation, it is expected that close collaborations between humans and such robots will drastically increase in the future. However, there is a need to understand how humans working with robots at different interaction levels may result in different human perceptions and cognitive workloads. In this study, a between-subject experiment was created to explore humans’ perception towards robots by measuring their cognitive workload at three different human-robot collaboration levels of coexistence, cooperation, and collaboration when accomplishing a real-world bricklaying construction task. Results showed that collaborating with small ground robots may lead to lower physical demand compared to cooperating or simply coexisting with it. However, mental demand, temporal demand, performance, effort, and frustration did not significantly differ between the three collaboration levels. The outcomes of this study provide a better understanding of the safest and most efficient human-robot collaboration practices on jobsites.

This paper presents a construction robot schema (CRS) for construction planners to facilitate decision-making and project planning in operating robotics. CRS is a database schema structure that was developed in our previous study, which can facilitate collecting and exchanging data of various construction robots based on the data requirements of the construction domain. We validated the applicability of the schema by the simulation of robotic construction operations. In addition, we conducted interviews with experts from the construction industry to validate the information in CRS. As a result, the schema was validated with minor revisions to some parameters. The characteristics of CRS compared to other types of robot schema are that its development and application are based on the perspective of the construction domain and are designed to cover different construction robots broadly. The conclusions highlight the contributions of the data schema use and applicability for the construction industry.

Exoskeletons, also known as wearable robots, are being studied as a potential solution to reduce the risk of work-related musculoskeletal disorders (WMSDs) in construction. The exoskeletons can help enhance workers’ postures and provide lift support, reducing the muscular demands on workers while executing construction tasks. Despite the potential of exoskeletons in reducing the risk of WMSDs, there is a lack of understanding about the potential effects of exoskeletons on workers’ psychological states. This lack of knowledge raises concerns that exoskeletons may lead to psychological risks, such as cognitive overload, among workers. To bridge this gap, this study aims to assess the impact of back-support exoskeletons (BSE) on workers’ cognitive load during material lifting tasks. To accomplish this, a physiologically based cognitive load assessment framework was developed. This framework used wearable biosensors to capture the physiological signals of workers and applied Autoencoder and Ensemble Learning techniques to train a machine learning classifier based on the signals to estimate cognitive load levels of workers while wearing the exoskeleton. Results showed that using BSE increased workers’ cognitive load by 33% compared to not using it during material handling tasks. The findings can aid in the design and implementation of exoskeletons in the construction industry.

3D excavator poses providing the motion information of the boom, arm, and bucket in 3D space support monitoring excavator activities for safety and productivity management in earthwork. Thus, previous studies have attempted to estimate 3D excavator poses using deep learning relying on the large data with high-quality annotations, which requires time-consuming and manual processes. To address this challenge, this study proposes cycle GAN to automatically create large generated real training images with 3D pose labels from synthetic images. The proposed model is trained on 800 pairs of synthetic and real images and evaluated through pre-trained ResNet50-based 3D pose estimations. The results reveal that 3D pose model trained on generated data, reaching 0.50 m key-point loss and 8.53-degree angle loss for testing on generated images, and 9.33-degree angle loss for testing on real images, yielded better results than model trained on synthetic data (i.e., 0.64 m, 15.18-degree, and 15.39-degree, respectively). This demonstrates the effectiveness of the proposed method for generating training images from synthetic images for 3D pose estimation. This 3D pose estimated from generated images enables construction managers to monitor excavator safety and productivity in the construction sites.

Human–robot co-working is a promising autonomous resolution for alleviating the sufferings from labor shortage, occupational hazards, and low productivity. Instructing robots to complete construction tasks using natural language (NL) is a hand-free and straightforward working method, requiring less expertise. Towards fluent intention communication, ontology serves as an important knowledge base for the robot to intelligently understand NL instructions. However, few studies utilized ontology to facilitate a natural-language-based human–robot interaction in construction applications before. This preliminary study designs an integrated ontology for natural language-instructed excavation tasks. Then, a validation framework with a use case illustration is presented for evaluating the ontology from a reasoning perspective. This paper is expected to demonstrate the feasibility of controlling robots to complete on-site construction tasks through NL instructions.

Advances in robotics represent a potential shift in the construction industry. Construction planning is planned based on craft work; it is necessary to emphasize external factors such as construction robotics. Improving constructability can enhance design-phase construction opportunities, thereby expanding the potential scope of robot operations. However, robotics are often neglected concerning constructability. Previous studies on constructability concentrated on human-based construction methods; hence, gaps remain in assessing constructability for robotics. To minimize the barriers in robotic construction, this paper presents a method for using a rule-based framework for robotic constructability assessment checks with the help of BIM. Focusing on CANVAS—a drywall finishing robot—this paper applies a BIM-based object-oriented model integrating with ROS to utilize constructability reasoning about robotic operations. A model of rule-checking for robotics in the case study is demonstrated and tested. The availability of design information in the model containing robotics is discussed, showing the need for assessing robotics-related constructability information to support an automated review of robotic constructability assessment. This paper applies a case study to validate use of the framework for robotic constructability assessment in the design phase, leading to an automated constructability assessment of construction robotics.

Building Information Modeling (BIM) integration with Unmanned Aerial Vehicles (UAVs) can enhance construction safety through proactive hazard identification to reduce and eliminate injuries and accidents on construction sites. Although numerous approaches exist to improve construction safety, only a handful explore combining BIM and UAVs. This research presents parts of a holistic framework of UAV-BIM integration for hazard identification on construction sites for safety performance improvement. A snowballing process was used to review the existing studies on the different applications of UAVs and BIM. Thereafter, a conceptual framework was developed using the findings of the review. This study contributes to the body of knowledge in a novel way by providing a systematic framework in the areas of technological capabilities, regulatory requirements, and their overlap to proactively determine hazard locations on construction sites. The research findings provide insights into how UAV-BIM integration can be utilized to enhance safety on construction sites.

Unmanned aerial vehicles (drones) are widely used for reality capture, data collection, and aerial logistics in multiple industries. Existing research in autonomous drones focuses much on path planning and automated obstacle avoidance. However, in complex environments such as urban areas, a safe and efficient drone navigation depends on many more dynamic constraints, such as layout of buildings, local aerodynamics (e.g., side winds), signal coverage, and malicious attacks. And many of these factors are interdependent, such as the wind field being affected by the geometric features of nearby buildings. This paper proposes a multi-objective reinforcement learning algorithm for the drone’s path planning while satisfying a variety of interdependent constraint conditions. The drone will be able to sense multiple environmental and physical conditions of the near field, and develop dynamic policies for prioritizing the navigational decision to optimize the path while minimizing the negative environmental impact. A policy network will be trained to set priority weights for specific environment and physical factors at a given point. A simulation study is performed to test the proposed algorithm in an urban aerial logistics case.

Construction robots have gained attention due to their potential for automating various construction tasks. Although recent research efforts have been made to develop robotic systems that assist manual work and reduce human intervention with different levels of automation, robots still need human assistance and collaboration for various tasks because of the unique characteristics of construction projects. To facilitate seamless human-robot collaboration in a dynamic and unstructured construction site, it is necessary that robots are able to understand not only the collaborator’s behavior in time but also its behavioral intention proactively to perform appropriate collaborative work. To achieve this goal, this study proposes a worker’s motion intention recognition method that captures the muscle activity of the workers using surface electromyography (sEMG) and detects motion intention from the muscle activity at the early stage of taking motions using a deep-learning approach. Using the sEMG signals, a deep-learning algorithm is used to predict the motion intention of the workers performing tasks for human-robot collaboration. With the proposed method, it is expected that robots can predict the motion that the worker in collaboration is going to take at the next moment, and ultimately this can improve the contextual awareness of robots for human-robot collaboration.

Robotized agents are becoming increasingly important in the construction industry as an extension of human capabilities. In many applications, the complexity of a teleoperation task could only be completed alongside the same physical environment to ensure synchronized situational awareness. With the recent advances in mixed reality (MR) and reality capture technologies, a new interface for teleoperating robots with a physical world reference is made possible. This paper introduces an intuitive MR control interface that could mitigate the gap between humans and robots by using a HoloLens2 optical see-through head mounted display (HMD), a Franka Panda 7-DOF robot arm, ranging sensors, and Unity game engine. With real-time acquired spatial information, a high-fidelity digital twin of the physical workspace is teleported to the user in mixed reality. Simultaneously, a virtual space around the user is also defined by the HMD with their hand gesture processed as input commands. The two virtualized spaces are then registered together and merged to be organized by the Unity game engine. As a result, the user takes control of the robot’s arm through direct manipulation of the virtual/real version of their hand. The system is tested with intricate pipe fitting tasks to validate its performance.

A hot bot apparatus is an enclosure system used to test the thermal performance of different items when exposed to controlled laboratory conditions for a period of time. Typically, testing building components are performed using hot box apparatus following the ASTM C1363 standard. However, performing such tests in accordance with the standards by an accredited laboratory can be extremely costly, and many laboratories have waiting lists of several months. This is a significant expenditure for small-scale business companies in the construction industry. Therefore, the goal of this study is to construct and test a smaller-scale hot box apparatus to measure the thermal performance of different building components at various climate conditions. The results indicated that the built hot box apparatus is capable of maintaining the steady-state temperature during the experiment within ±1℃ meeting the requirement of the ASTM C1363-19 standard and is capable of measuring the thermal performance of a building component within an accuracy of ±0.5 ft²·°F·h/BTU. This allows small companies to test several components at relatively low cost and minimum duration.

Unhealthy working environments, such as high heat stress, can significantly decrease worker productivity, reduce work performance, and lead to unsafe behavior. While embedded/wearable sensors/thermometers can reliably measure the heat stress on construction sites, they can only be implemented locally and are often needed to be deployed for multiple workers, which requires a massive sensor network. To address these issues, this paper proposes a contactless heat stress monitoring and water-spraying robotic system that can be deployed repeatedly using drone infrared thermography. Four main components are in the system, including (1) a prototyped quadrotor [i.e., an unmanned aerial vehicle (UAV)] to provide mobility; (2) an infrared camera to take thermal images, along with an environmental sensor to measure air temperature and relative humidity; (3) a spray tank to carry water; and (4) a single board computer to dynamically process the thermal images and environmental sensor readings for controlling the spray tank. Real-case validation results showed that the proposed system could accurately measure the heat stress and automatically spray water to improve thermal working conditions. Ultimately, this paper contributes to the body of knowledge by developing a drone thermography-enabled robotic system for enhancing worker comfort, health, and well-being at construction sites.

Although modular construction has become increasingly popular worldwide due to its advantages of offsite production, most construction managers are still inexperienced with such projects compared with onsite construction project management. To facilitate modular construction management, various Internet of Things (IoT)-enabled Building Information Modeling (BIM) platforms have been developed to provide better information visibility and a more collaborative working environment. Nevertheless, such existing platforms have typical shortcomings such as not being able to trace single components and significant work time lost in material missing and locating. This study presents a new logistic chain-enabled BIM system that is designed with a three-layer framework: a material packaging layer, an inventory management layer, and a material locating layer. The logistic chain-enabled framework allows any material modules of a modular construction project to be tracked on route, inventoried in a material management system, and visualized in the BIM model with a barcode scanner prior to installation. This paper provides significant insight on improving modular construction practices by developing a practical logistic chain-enabled BIM system.

This research explores the challenges of human-quadruped robot partnership in the process of construction inspection and monitoring. Quadruped robots have shown potential in construction inspection and monitoring because of their ability in traversing uneven terrains. Construction inspection and monitoring which is a crucial process in construction can benefit from the use of autonomous robots by collecting frequent as-built data for facilitating inspection by different stakeholders. However, the adoption of robotic technologies in construction is still low. To realize the full potential of quadruped robotic technologies in construction inspection and monitoring, it is important to understand fundamental challenges that can restrict its widespread adoption. Therefore, this study uses a qualitative approach to identify the challenges of the adoption of an inspector assistant quadruped robot in construction inspection and monitoring through on-site observations, interviews, and a focus group study. The results of this study identified 20 different challenges grouped into five categories, namely, economical, operational, organizational, social, and other challenges. The findings of the study can facilitate the development and implementation of quadruped robotic technology in construction inspection and monitoring.

Using robots to automate construction processes can enhance productivity and safety and reduce cost. However, the construction robots haven’t been widely accepted primarily due to two challenges: lack of data and computational resources to train robots for performing specific construction tasks, and lack of the capability for robots to deal with new tasks. This paper introduces a framework leveraging large language models (LLMs) to enable robots to perform assembly tasks based on in situ worker instructions without extensive prior robot training. By using predefined prompts tailored to specific tasks, the LLM can interpret worker instructions and generate corresponding action sequences and robot functions without additional training. Hence, this LLM-based approach can effectively reduce training costs and facilitate robot adaptability for new tasks and environments through simple prompt modification. To validate this framework, several interlocking block assembly tasks performed by LLM-based robots were evaluated. It is found simple instructions were accurately executed by the robots, but future research is needed to improve the model to handle complex instructions from humans. Overall, this study demonstrates the potential of LLM-based robots in performing assembly tasks in construction with human collaboration.

Material procurement comprises a considerable portion of costs in construction projects. In addition to increasing costs, inefficient material procurement can impact a project’s schedule and duration. Nevertheless, efficient procurement management use has been limited in the construction industry. A more efficient approach to information management is needed in the construction industry to meet the needs of a growing industry. Digitalization of the construction industry has enhanced the automation of certain processes related to project planning and design. Building Information Modeling (BIM) solutions are often associated with handling high volumes of information. Hence, a data transfer and processing approach must be developed that can be applied to various modeling software applications. Integrating Industry Foundation Classes (IFC) with blockchain provides a basis for facilitating active and secure procurement data exchange and material purchase. This study proposes a framework for effective procurement management by focusing on the automation of processes utilizing IFC and blockchain concepts. The outcomes of this study can be specifically valuable for project managers. This framework provides flexibility in decision-making by being tailored to each project’s specific financial conditions and facilitates and automates procurement-related payments.

Digital Twin (DT) technology provides a comprehensive virtual representation and dynamic mirror of the physical infrastructure, which is invaluable for Facility Management (FM) applications. However, its operations and sensitive data are managed by the centralized infrastructure, which is relatively vulnerable to being compromised. To address these problems, this paper proposes DT-DAO, a conceptual blockchain-based framework integrating Digital Twin (DT) and Decentralized Autonomous Organization (DAO). Given the immutable nature of blockchain technology, the DT-DAO framework can strengthen the security and resilience of the DT. Moreover, taking advantage of the DAO’s decentralized governance mechanism, task automation capabilities, and the real-time sensory data retrieved from the physical assets, DT-DAO can also be used to create an autonomous physical asset (e.g., smart buildings, infrastructure) that can have its operations run automatically and autonomously as predefined in the smart contract. In addition, this paper also provides potential use cases of DT-DAO in smart building FM.

This study presents a framework for human-aware robotic sensing and control for safe human-robot coordination in dynamic construction environments. The framework addresses the critical need for enhanced safety measures in the construction industry, where robots are increasingly being deployed for various tasks. The proposed framework incorporates task-based risk assessments, safety rule definitions, real-time human detection, distance estimation, and adaptive robot control mechanisms. It enables a robot to autonomously adjust its actions based on the estimated distance to the closest human worker, allowing for prompt stopping, speed reduction, or continued full-speed operation as required. The framework enhances safety and fosters trust between human workers and robots in collaborative environments. A prototype application was tested in a simulated residential construction site with a mobile robot with multiple depth cameras to provide 360-degree visibility. The result of the case study demonstrates the potential to safely control autonomous construction robots working around human workers.

The integration of robotics technology with Building Information Modelling (BIM) to plan robot operations is a promising approach for automating construction tasks. While 4D BIM-related technologies are arguably the most advanced method of planning for construction projects, they lack adequate elemental information required by a robot to plan its operations within the spatiotemporal context of the construction site. This poses a challenge to planning robot operations for construction tasks considering their spatial coordination with human workers and site conditions. This study identifies the information required in 4D BIM for planning and simulating construction robot operations. A Robot Operating System-based software program was developed to demonstrate the utilization of this information in the planning and simulation of construction robot operations. The program was evaluated in a simulated construction environment, employing a Husky mobile robot to install drywall boards using information from 4D BIM. The simulation demonstrates the importance of information modeling in the integration of 4D BIM with robotics technologies for the planning and execution of construction tasks.

State-of-the-art in construction document analytics and progress detection has experienced accelerated growth over the last decade. However, each area encountered isolated growth, not considering their interactions. Today, progress monitoring practices are often neglected due to requiring manual input of visible progress against schedules. Such a challenge can be attributed to (1) vision-based progress tracking lacking formal construction work templates applied in common construction workflows, and (2) research in automated schedule generation and analytics lacking focus on extracting fragnets from a body of existing schedules. This study brings together insights on research trends for automated schedule generation and analytics using Natural Language Processing (NLP) and detection of under-construction objects using Computer Vision. Finally, the AIConstruct system is presented to demonstrate, for the first time, how the integration of text and image can create seamless data synchronization for construction progress monitoring and automated schedule generation, unlocking a new research paradigm.

A building must meet various requirements during the design and construction process to ensure the benefits of stakeholders and well-being of construction workers and occupants. These requirements may come from different functional areas such as structure, electricity, and fire protection, and focus on different building materials, such as concrete, steel, and glass. They may overlap or even conflict with each other. In order to identify the sources and focus of building code requirements and further clarify the relationships between them, this paper presents some recent results on using graphic convolutional networks (GCN) to classify building code requirements. One hundred building code provisions were randomly selected from the International Building Code 2015 and labeled into six categories manually, and a cutting-edge GCN model was trained to classify them. Experimental results showed an average precision of 91.67% and an average recall of 94.44% when 10% of the data was used for testing, which is comparable to the 84.30% precision and 97.30% recall of the state-of-the-art machine learning-based approaches applied on construction document classification. The effect of the size of training data on testing accuracy was also discussed in this paper.

Heavy equipment is crucial for construction work and depends on skilled operators to ensure productivity and safety. The inclusion of robotics and its applications, particularly teleoperation, has the potential to revolutionize the construction industry. Via semi-structured interviews, this research explores the current practice of equipment automation technology adoption, potential needs, and benefits on equipment teleoperation, as well as its impacts on workforce diversity. The study highlights the significance of equipment teleoperation on increasing work safety and worker comfort, as well as its importance in improving workforce diversity and reducing labor shortage. With technology limitations being identified as a main barrier to practical adoption, the research also emphasizes the need for user-inspired and worker-centered teleoperation and automation technologies to facilitate implementation in practice. The study contributes to the development of future teleoperation technology by identifying the needs of construction professionals and the potential impacts on both construction work and workers.

Human-Robot Collaboration (HRC) is a promising approach to relieve workers from repetitive and physically demanding tasks and improve safety and productivity in construction. It is critical for robots to understand worker intention in order to adapt their motion to facilitate smooth HRC. Evidence has shown that visual attention reveals human intention. However, it is still unclear how visual attention is distributed in human-robot collaborative construction tasks. In this study, a pilot experiment was conducted to examine human visual attention in a wood assembly task with the assistance of a collaborative robot. A mobile eye tracker was used to collect participants’ gaze movements. Data were validated and processed in terms of various metrics to analyze visual attention patterns. It is found that construction workers’ visual attention is related to the detailed process of the task—around 30% of the eye gaze is located at the connector areas and the design drawing area, which is primarily relevant to their task. Furthermore, workers’ attention could be affected by the movement of the robot, with their gaze following the path of robot arm and gripper during the collaboration. The findings can stimulate further research into attention-aware HRC for intelligent construction.

Virtual design and construction (VDC) managers must assess the costs and benefits of implementing VR before deciding whether to use it during mechanical, electrical, and plumbing (MEP) coordination to facilitate early issue identification and resolution. We conducted a comparative case study of two facility buildings: one where a VR pilot study was conducted in addition to 3D clash detection, and one where only 3D clash detection was used during MEP coordination. The direct and indirect costs of VR use were determined through observations, interviews, and by comparing MEP coordination workflows with and without VR use. Our analysis of buildability and operability issues identified uniquely in VR revealed that more than 100 field-detected issues, requests for information, and change orders could have been avoided through VR use. The resulting cost savings exceed the cost of implementing VR. As such, our research provides a preliminary method to evaluate VR pilot studies towards deciding on VR adoption during preconstruction MEP coordination.

Welding robots are often employed to improve the welding efficiency and quality of steel structures. In the past few years, weld seam detection technologies based on structured light have undergone huge advances. However, due to the limited measuring range of structured light sensors, high accuracy of workpiece placement is required by most popular approaches in order to avoid invalid sampling. This fails to meet the actual needs of manufacturing. To overcome this burden, a two-step seam detection method is proposed in this study. Firstly, a symbol-assisted 2D digital image contour detection method is employed to obtain the projection of the workpiece in the horizontal plane. Meanwhile, the actual point cloud and the model point cloud of the workpiece are generated based on projected image and CAD model, further calculating the transformation matrix between the two point clouds. Secondly, a linear structured light vision sensor is used to detect weld seams, whose sampling poses are adjusted by the obtained transformation matrix. The robustness and accuracy of the proposed method are demonstrated by simulation experiment results.

Construction projects extensively use steel. Welding safety and quality concerns are vital since they are mainly used as structural components. By 2027, the welding industry will be short of 360,000 welders, despite its importance. The labor shortage and the need for consistent and efficient welding are strong drivers for developing an automated welding system. This paper presents the development of a mobile robotic welding system by integrating an unmanned ground vehicle (UGV) with visual sensors, a robotic arm, and a welding machine. This integrated robot can navigate to the welding location while avoiding a collision. It can detect welding joints automatically using a camera through deep learning algorithms. The robotic arm can move along the joint while scanning the joint in 3D using a line-laser scanner. Finally, it can perform welding along the 3D joint captured by the laser scanner. The results show that the integrated robot can reach the goal position within the margin of error and demonstrate the consistency and accuracy of welding.

For autonomous robots deployed in human-populated workplaces, adhering to people’s interpersonal space is a key prerequisite for safe and efficient navigation. Especially in safety-critical work environments such as construction jobsites, robots should be capable of not only avoiding collisions but also maintaining a socially acceptable distance from other workers to achieve safety and efficiency. However, what is deemed as appropriate distance can vary based on work contexts. To this end, this study aims to understand the preferred interpersonal safety distance in different construction work contexts such as the size of the workspaces. We developed simulated work environments and conducted proxemics experiments. We found that participants tend to keep a larger minimum distance when interacting with a robot and in a more constrained condition than the normal one. Our findings can be used to develop a more safe, socially acceptable robot path planning in workplaces.

The construction industry is a major contributor to the US economy but is constantly challenged with low productivity, workforce shortage, and health and safety issues. Construction robotics potentially mitigate labor shortages and productivity inefficiencies facing the construction industry. However, construction jobsites are dynamic and ill-structured, and robot integration in such complex workspaces requires a high level of safety for successful collaboration with construction workers. Therefore, understanding existing robotic engagements in construction activities and necessary human-robot interactions for promoting a safe collaborative workspace with construction robots becomes critical. As such, this paper presents a review of state-of-the-art research in human-robot interactions in the construction industry. The paper contributes to the body of knowledge on the different human-robot interactions to facilitate the successful implementation of robotics in the construction industry.

The rapid advancements in Artificial Intelligence (AI), driven by increased computational power, have revolutionized the engineering industry in recent years. A growing body of literature has emerged, focusing on the application of AI, advanced robotics, and the Internet of Things (IoT) in various sectors. However, the construction industry appears to lag in adopting these intelligent automation technologies for enhancing construction operations. This study aims to examine the current state of AI and robotics adoption within the construction industry for automation purposes. To achieve this, a systematic review of academic peer-reviewed articles on AI and robotic applications in the construction industry was conducted. Keyword-based semi-supervised machine learning was employed to classify the articles according to the phases of construction. Subsequently, an unsupervised machine learning algorithm was utilized to perform content analysis on the articles concerning different construction phases. This analysis enabled a better understanding of relevant AI applications, which could be integrated across construction phases to enhance efficiency. Additionally, this study investigates the various barriers and benefits of adopting AI aimed at improving productivity and safety. Finally, the article discusses the implications of AI, robotics, and automation on job opportunities within the construction sector.

Construction 4.0 (C4.0) technologies have numerous benefits to the architectural, engineering, and construction (AEC) industry. To achieve the full potential of C4.0 technologies, firms and organizations are required to invest in developing workforce hard and soft skills. The shortage of additional skilled human resources hinders the effective implementation of C4.0 technologies as these technologies demand advanced skills compared to traditional practices in the AEC industry. This study aims to explore workforce skills, hard and soft, to adopt C4.0 technologies effectively. For this purpose, a systematic literature review has been conducted. The results demonstrate that critical thinking, problem-solving, teamwork, communication, and self-management are essential soft skills to achieve a successful transition of C4.0 across the industry. The findings of this study will help firms and organizations in the AEC sector invest in training programs by collaborating with educational institutions to develop necessary skills in their workforce.

With unmanned aerial system (UAS) technologies’ rapid development in the last couple of decades, UASs have been widely used in different areas, including Architecture, Engineering, and Construction industry. Integrating with a high-resolution camera, LiDAR or other sensing system enables the UAS to conduct complex tasks at construction sites such as pile volume quantification. Understanding pile volume is critical to assist construction scheduling. Traditionally, pile volume measurement is mostly conducted with manual inspection, which is inefficient and expensive. Applying UAS with sensing components onboard provides the possibility of measuring pile volume in a more efficient and affordable approach. In this paper, we introduce a UAS-based pile volume measurement system and conduct experiments measuring the bulk pile volume through a series of field experiments with this system. The volume measurement data collected from the UAS is exported to 3D point cloud processing software for analysis. This software is also used to quantify the bulk volume and record the volume changes over time. The volume measurement results are compared with other records and demonstrate promising accuracy in measuring the pile volume. At last, the performance of the UAS-based pile volume measurement and ongoing investigation to improve the system performance are summarized.

The escalating challenges in the construction sector, such as comparative lower predictability and productivity, labor shortages, occupational safety, and hazardous working conditions, indicate the necessity and urgency of investing in automizing construction. Identifying the barriers to implementing construction automation and investigating ways to overcome them can pave the way for a greater use of these technologies in the construction industry, which, in turn, could increase its efficiency, safety, and productivity. In this study, a survey was conducted, which led to the identification of 11 major barriers to the widespread application of these technologies. They were (1) job loss, (2) initial investment cost, (3) complexity, (4) limited flexibility, (5) lack of skilled labor, (6) inadequate infrastructure, (7) resistance to change, (8) technical complexity, (9) safety concerns, (10) lack of human interaction, and (11) regulatory barriers. The paper is expected to shed light on how the application of automated techniques in construction projects should modify to better fit the concerns and demands of the industry.

Building information modeling (BIM) technology in construction has become increasingly prevalent in recent years, and integrating robotics is seen as a natural step to improve efficiency. To increase the level of development (LOD) of a BIM model to support a construction robot, parametric modeling can be used to create highly detailed models by supplementing and defining the geometric and physical properties of the construction elements, such as the components’ size, shape, and material parameters, which are used as inputs for designing robotic tasks. Component information and data are stored as extractable parameters within the BIM model, allowing a robot to perform highly precise and repeatable tasks. This study develops a framework for implementing computational parametric modeling for masonry wall systems with Dynamo. This study tested six wall configurations constructed of 8″ × 8″ × 16″ concrete masonry units (CMUs). Dynamo successfully interpreted most wall geometries placing full-sized CMUs into the correct design locations. Errors occurred when placing partial-sized CMUs, typically at wall intersections, revealing a need for future refinement. The study shows the careful planning and considerations needed to implement computational modeling to generate model content for creating robotic tasks.

Highway and road construction and maintenance involve operations of heavy construction equipment consistently exposing workers to injuries and fatalities. Frequent collisions between heavy equipment and workers on foot cause many deaths and disabilities. A large portion of these work zone fatalities, such as being struck by equipment and caught in/between equipment, can be effectively prevented by automatically detecting objects around the equipment, determining their poses and movements, and eventually alerting operators and workers about unsafe situations. This study attempts to embed a situational awareness to heavy construction equipment by integrating sensing technologies into heavy construction equipment. This study developed a Robots Operating System (ROS)-based software program to determine the placement of multiple sensors ensuring 360-degree visibility around the equipment and process the sensor data into an accurate 3D representation of the work zone environment to detect predefined unsafe situations. The prototype safety monitoring system evaluated in a simulated road construction environment successfully detected the presences and locations of human workers around the equipment.

Autonomous Unmanned Aerial Vehicles (UAVs) are rapidly transforming industries requiring inspection and surveillance. However, conventional UAV systems often require complex control schemes and lack adaptability, limiting their efficacy in variable environments such as indoor inspections. This paper introduces an innovative system integrating the cutting-edge Generative Pretrained Transformer (GPT) models and dense captioning models for autonomous navigation and fault detection in indoor environments. Our approach, displaying human-like flexibility, allows the drone to interpret and respond to natural language commands, vastly enhancing its accessibility and user-friendliness. Simultaneously, the drone utilizes object dictionaries derived from dense captioning of its captured images, facilitating an advanced understanding of its surroundings. These capabilities equip the drone to adapt its behavior and effectively handle unexpected scenarios, significantly enhancing the efficiency and accuracy of indoor inspections. This research contributes to revolutionizing building inspections, making the process more user-friendly, and localizable to a broader user base.

The reusability and recyclability of demolition waste are significantly affected by demolition operations, particularly material separation activities, which are largely driven by productivity considerations. As such, investigating the productivity of demolitions operations is key to understanding the decision-making processes affecting the reusability and recyclability of demolition waste. Traditional approaches for tracking the duration of demolition operations and thereby monitoring their productivity are costly, time consuming, and prone to human errors. To enable more effective and efficient demolition productivity monitoring, this study presents an automated approach for identifying demolition waste material separation activities using the motion data of demolition machinery. As proof of concept, small-scale heavy equipment is used to simulate demolition operations. Inertial measurement unit (IMU) sensors are attached to different moving members of the small-scale heavy equipment to collect angular and linear acceleration data. Collected time-stamped sensor data are preprocessed and subsequently used to train and test an activity recognition model using various supervised machine learning classification algorithms. The output of the developed model facilitates the delivery of actual productivity information, which can be used to optimize demolition planning and decision-making in a way that increases the recycling and reuse of demolition waste.

Construction robots are a powerful driving force to enable intelligent processes in construction. User-friendly interfaces to support human-robot work collaboration are critical for increasing adoption of robots. Among different interfaces, eye gaze and hand gesture are effective and reliable interaction cues in the noisy construction environment. This paper proposes a novel context-aware method, which integrates eye tracking and gesture recognition for human-robot collaboration in construction. The proposed method employs a two-stream network architecture comprising a first-person view-based stream and a motion sensory data-based stream. The first-person view-based stream models the user’s gaze using an attention module to generate an attention map, which helps the stream to focus on the relevant spatiotemporal regions for context extraction. The motion sensory data-based stream is used to process the motion sensory data to extract features related to hand motions. Finally, the extracted vision context and motion features are combined to achieve the gesture recognition for conveying a message between the worker and the robot. This method was tested using a dataset gathered on construction sites. The test results show the proposed method can achieve accuracy and mean class accuracy of 96.8% and 97.7%, illustrating its effectiveness for human-robot collaboration in construction.

Buildings deteriorate over their service life. The early detection of defects such as cracking, spalling, corrosion, and moisture can benefit the preventative maintenance of building systems. Autonomous robotic systems have enormous potential in automating indoor building defect inspections, along with challenges of inaccurate prediction and unorganized information. With the implementation of state-of-art multimodal learning methods and large language model (LLM) techniques, we present a cutting-edge workflow composed of image captioning, landmark documentation, and real-time on-site human–machine interactive path planning. Compared with previous vision and language navigation (VLN) algorithms, our workflow introduces defect prompts to improve indoor inspection captioning performance. These pivotal defect features are extracted by YOLO (You Only Look Once) v5, a PyTorch-based deep learning model. As the robotic system recognizes the environment clearly, inspectors are capable of providing target-oriented instructions to control the survey path. By implementing the large language model GPT-3, vocal and textual instructions are transferred to the robotic localization system and summarize a brief inspection report. In this way, with the assistance of GPT, numerous inspections that previously demanded substantial effort can be conducted efficiently and expeditiously.

The fill factor is a key metric used to evaluate the performance and efficiency of the excavator’s autonomous functions and operations. Optimizing the fill factor allows for an objective definition of a “poor bucket” and provides an opportunity for reducing the emissions from operating heavy equipment. This article aims to study the effect of fill factor (%) on fuel efficiency for a bench-loading exercise. Two cases were developed “Bucket Normal” where excavation is completed at a pre-selected fill factor, and “Bucket Return” where initial excavation is returned to achieve 100% fill factor. The task efficiency of different fill factors was reported, and the statistical analysis shows there is no difference in fuel efficiency at 70% fill factor and above. This indicates an optimal fill factor where it is no longer energy efficient for an automated excavator to return the poorly excavated load. The optimal fill factor and relationship with fuel efficiency can be incorporated into existing algorithms for teleoperated and automated excavators to improve the productivity and energy efficiency of soil excavation.

Flaggers work closely to open traffic lanes which makes them a high-risk profession. They might get hit by distracted, speeding, or intoxicated drivers, leading to severe or even fatal injuries. To address this safety issue, the concept of Automated Flagging System Devices (AFSDs) was proposed. In an AFSD, one of the essential components is to recognize the turn signals of the vehicles in the lanes. However, most existing studies for vehicle turn signal recognition focused on autonomous driving to help autonomous vehicles interpret the intents of the vehicles in front. Little work has been conducted for AFSDs to guide traffic. This paper presents a novel method for recognizing vehicle turn signals using the video camera on an AFSD at the roadside. The method first relies on object detection and tracking models to locate the vehicles and headlights in the video frames. Then, the detected headlights are matched to the vehicles. For each vehicle, the luminance features of its headlights are further extracted. Based on these luminance features over frames, the flashing patterns of the headlights could be recognized to determine the vehicle turn signal. The method was tested with real traffic videos, showing that its overall recognition accuracy could reach 80.32%.

Natural disasters such as wildfires, landslides, and earthquakes result in obstructions on roads due to fallen trees, landslides, and rocks. Such obstructions can cause significant mobility problems for both evacuees and first responders, especially in the immediate aftermath of disasters. Unmanned Aerial Vehicles (UAVs) provide an opportunity to perform rapid and remote reconnaissance of planned routes and thus provide decision-makers with information relating to a route’s feasibility. However, detecting obstacles on roads manually is a laborious and error-prone task, especially when attention is diverted to needs that are more urgent during disaster scenarios. This paper thus proposes a computer vision and machine-learning framework to detect obstacles on a road automatically to ensure its possibility in the aftermath of disasters. The framework implements the YOLO algorithm to detect and segment roads on images from UAVs and reference images from publicly available datasets. The images retrieved from UAVs and reference images are segmented and counted pixels of the roadway for comparison of the difference in pixels to identify the obstruction on the road. In addition, the method is proposed to automatically detect obstructions found in the region of interest (ROI) only on a roadway with images and videos from UAVs. Preliminary results from test runs are presented along with the future steps for implementing a real-time UAV-based road obstruction system.

The pursuit of construction projects with error-free design, streamlined manufacture and assembly, and rework-free construction is often dominated by project management best practices. Yet this paradigm can be increasingly supported by cyber-physical systems that include elements such as generative design, laser scanning, computational algorithms, and digital fabrication, thus closing the gap between design and reality. This paper presents a regenerative BIM system that adapts 3D geometry in real time to match field conditions, thus achieving the ability to digitally fabricate to field conditions (i.e., “digital fab-to-field”). Such a system is particularly useful in 2D exterior panelized construction (e.g., curtainwalls and composite metal cladding), which frequently undergo design iterations and onsite rework due to geometric inconsistencies between site interfaces and manufactured components. The proposed framework is demonstrated using a digital mock-up of an exterior cladding assembly, which is shown to adapt to changing site conditions as described by a 3D point cloud.

Augmented Reality (AR)-enabled wayfinding applications have been widely implemented in complex building environments. These applications require a precise building information model of the service area and, therefore, are not applicable to the majority of existing buildings in the US, such as grocery stores. This paper presents an AR-enabled indoor wayfinding assistant app that visualizes the locations of typical destinations in building environments. Instead of building models, the app utilized data on relative spatial relationships that nonprofessionals can collect with minimal training and easily accessible tools. Therefore, spatial maps of different building environments can be developed with a crowdsourcing approach, and standardized data will be maintained in a cloud database. The feasibility of the data collection method and AR app was validated by a case study conducted in a grocery store. This paper contributes to the body of knowledge by replacing models with relative spatial relationships in indoor AR wayfinding apps. The app enables economical indoor wayfinding solutions for existing buildings that do not have accessible and up-to-date models.

The United States (US) is the largest Greenhouse Gas (GHG) emitter globally, with its transportation sector being a major carbon dioxide contributor. This is a matter of significant concern, given the impact of these emissions on climate change. Prioritizing emissions-reducing transportation projects is crucial to address this problem. Currently, such policies are generally discussed at the national level. However, since individual states have varying infrastructure and priorities, this paper proposes studying carbon dioxide emissions at the state level of granularity. To achieve this, ensemble methods such as Random Forest, Extremely Randomized Trees, AdaBoost, and Gradient Boosted Decision Trees, along with a meta-estimator, are used to predict statewide carbon dioxide emissions using variables related to statewide transportation systems. Based on this analysis, we suggest that states should discourage car dependency in a way that is feasible for local characteristics, as well as promote public transit to reduce carbon dioxide emissions.

As robots advance, their application in facility management tasks such as cleaning, maintenance, security, surveillance, warehouse management, and customer service is expanding. As buildings become smarter, robots will likely either replace or assist humans in maintaining them, leading to more complex interactions between robots and buildings. To accurately simulate and predict these interactions, we propose a conceptual framework for a simulation platform. This framework comprises four main components: occupants, robots, building systems, and a digital twin. The digital twin receives data and requirements from the other three entities, issuing commands to meet these needs. This research underscores the importance of viewing robots as occupants and promoting the coordinated development of smart buildings and robots. Future studies could focus on designing various types of robots to meet occupants’ and building systems’ needs, as well as creating smart buildings that account for robots’ convenience and efficiency.

The risk attitude of contractors impacts their bid values in competitive bidding. However, the existing literature assumes a static risk behavior for the contractors, and that competitive bidding does not alter the risk attitude of contractors. To this effect, this paper aims to represent the influence of external (market competition; project information, etc.) and internal (financial performance; need for work, etc.) factors on the adaptive risk attitude of the contractors, their submitted bid values, and their survival in the market. The paper’s objective is to develop an agent-based model (ABM) that represents the competitive bidding environment and that relies on data gathered through the literature on the importance and impact of multiple factors that drive the risk behavior of contractors within the bidding context. The presented ABM is an important tool to help contractors understand and evaluate the importance of adapting their risk behavior to their internal and external bidding conditions in order to ensure their survival in the market.

Inefficient facility layout in construction manufacturing leads to process waste, thereby increasing production time. In particular, it results in unnecessary movement of workers to transport materials/tools during production. The research presented in this paper examines the impact of a precast concrete factory’s layout on three parameters in the production of precast concrete panels: labor travel time, activity cycle time, and production time. Recommendations are proposed to reduce process waste regarding each of these parameters, that is, relocating the material preparation mills and introducing a cart-based material/tool handling system. The benefits are quantified using a discrete-event simulation model of the process. It shows a 10% reduction in the production time of a single panel and a 39% reduction in the labor hours incurred in transportation activities when compared to the current state. It also shows that 6.9% of the labor hours in the current state is allocated to unnecessary non-value-added transportation activities.

Pavement maintenance management requires artificial intelligent techniques to evaluate pavement condition automatically, predict pavement deterioration, and optimize maintenance actions under restrictive budgets. This paper presents three techniques that address the three objectives raised. First, Convolutional Neural Networks (CNN) are used to analyze images obtained by a camera installed on a vehicle. Several CNNs are trained to detect, classify, and quantify the 3D distresses. Second, pavement deterioration is predicted by Feed-forward Neural Networks (FNN). Pavement Condition Index (PCI) throughout a planning horizon is estimated from the information obtained by the inspection and the traffic and climate conditions. Third, heuristic optimization algorithms are used to determine the optimal maintenance plan. This plan indicates which sections should be repaired each year of the planning horizon and the treatment that must be used to optimize the maintenance cost, the CO₂ emissions, the user cost, the network condition, and the accidents. These techniques are presented and discussed in this paper.

The increasing requirements of robots in construction have brought great challenges related to understanding the complex environment and completing the downstream tasks with timelines. Focusing on stacked object detection, which is a generally complicated scenario in construction sites, this paper proposes a novel framework by using both RGB and LiDAR for object clustering with low storage occupation and high detection speed that support real-time implementation. An RGB camera is first used to capture an image of the overall scene, and a pre-trained CNN network is applied to give the rough prediction of region-of-interests (ROIs) along with their confidence scores. The ROIs are linearly sorted based on their scores to select the potential stacked areas with low confidence. The center locations of ROIs are then transferred into the LiDAR system with the calibration matrix, and a Velodyne-16 scanner is used to perform adaptive scanning on the ROIs for detailed object clustering and detection. The result shows that given the pre-detected ROIs from RGB, the scanning time and computational time of clustering could be largely reduced. Furthermore, a confidence-based criterion is illustrated to linearly determine the required scanning frames to get desired detection results.

Building Information Modeling (BIM) technology streamlines the establishment of systematic methodologies for managing built asset information effectively throughout each asset’s entire life cycle. The recently introduced ISO 19650 series provides standardized processes and guidelines to facilitate the efficient management of such information. In this context, the ISO standards advocate for using a common data environment (CDE), which serves as a centralized repository for collecting and managing all graphical and non-graphical construction project data for the entire project team. Ideally, a single CDE should be utilized and maintained throughout the life cycle of each project. However, due to practical and technical challenges, multiple CDEs are frequently employed and managed concurrently in actual projects, leading to significant issues such as data inconsistencies, duplications, and missing information. This paper presents a framework for managing multiple CDEs, which enables the preservation of data integrity. The proposed framework is anticipated to foster the progression of BIM maturity levels in real-world projects.

In the last few years, Building Information Model (BIM) has become a construction industry standard for exchanging as-built information for construction assets. BIM can be integrated with data collection technologies, such as Uncrewed Aerial System (UAS), to reduce manual errors and improve the efficiency in information retrieval processes. Traditionally, BIM feature extraction from UAS-collected images is done manually. This process is time-consuming, costly, and tedious. There has been a growing interest in automating this process, utilizing tools such as Computer Vision and Machine Learning. The objective of this paper is to develop an as-built BIM from UAS data, using these tools. This paper describes a foundational methodology to collect, process, and analyze the UAS data, stored as raster Geo-TIFF format files, where digital elevation models (DEMs) are used to create an as-built3D model of the project. The three steps of the described methodology are: (1) Data Collection and Processing, (2) Data Analysis, and (3) As-Built Development in Civil 3D, illustrated through an MDOT highway construction project. The outcome of this research can be leveraged by project managers as a record of work completed, beyond the construction phase of the project.

Accurate construction workers’ activity detection distinguishes workers’ activity efficiency and identifies activities with high risks, which enhances construction productivity and safety management. Automated construction workers’ activity has become feasible in the construction industry since the emerging of computer vision techniques. However, few studies explored workers’ activity detection in transportation-related work zones (e.g., mobile work zone operations), which have their unique characteristics and different requirements than building construction jobsites. Previous studies in the construction domain usually used convolutional neural networks (CNN) for computer vision-related tasks. The transformer-based model achieved higher performance in computer vision tasks after it was first applied in 2020. However, few studies have applied the transformer model for worker activity identification. Therefore, this study aims to detect construction workers’ activities in mobile work zones using a pre-trained Vision Transformer (ViT) model. This study starts with the video data collection of construction workers. Then, a dataset containing different activities is developed by manual labeling. Next, the ViT model is fine-tuned using the dataset developed in this study. The results show the model has 94.17% overall accuracy and achieves 100%, 100%, and 84% precision in detecting “placing mix,” “shoveling,” and “walking,” respectively, and outperforms a CNN-based classification model.

Maintaining and operating building facilities constitutes a substantial portion of the overall life-cycle costs, which, if not befittingly handled, can yield immoderate expenditures and a shortfall of financial resources. The evidence implies that building facilities are degrading due to inadequate knowledge of building facilities and fallacious management procedures. This paper proposed a building information modeling (BIM) framework to enhance the effectiveness of BIM open standards for facility management operations. The study expands to creating a comprehensively integrated framework that provides a paradigm apropos to employing BIM capacities that manage and retrieve operation and maintenance (O&M) data to engender a semantic repository of multiple existing building facilities and their peculiarities information. The proposed framework incorporates six chronological stages to ensure prime findings for the O&M phase and unerringly exhibit the distinctive attributes of the maintenance data paradigm. Facilities managers can employ the presented framework as a blueprint for implementing the proposed strategy for incorporation, exploitation, and handling of building facilities-based knowledge throughout the O&M phase into practice.

Cranes play a significant role in construction cost and time management. Tower cranes are highly used in dense urban residential construction sites due to congestion and lack of enough space for the crane’s foundation. However, optimizing the location of tower cranes in construction sites with multiple cranes and multiple trailers in an integrated application core is still a gap in the literature. This study aims to fill this gap by developing an application programming interface (API) for Navisworks, a common software in building information modeling (BIM), to facilitate practical use. The focus of this paper is on prefabricated structural elements that are planned to be lifted by tower cranes. The research utilizes VB.net language to develop the API and connects an access database containing tower cranes’ information for the optimization process. A case study is presented to validate the outcomes of the API. The results show the best locations for tower cranes and the least possible cost and time based on the database factors and each crane’s pick-point location. This application has the potential to significantly save construction budget and time.

Economic market uncertainties, such as those experienced during the COVID-19 pandemic, can make determining accurate prices estimate for construction materials a challenging task. While previous research focused on the contractual aspect of this issue by studying price escalation clauses, there is still a gap in the literature when it comes to proposing an accurate pricing model. Thus, this study develops an accurate pricing-estimate game-theoretical model that can efficiently and competitively account for escalations in construction materials prices during uncertain market conditions. First, data on past Producer Price Indexes (PPIs) of different construction materials were collected. Second, the percentage changes in the prices of four common construction materials, including asphalt, aggregates, non-reinforced concrete, and steel reinforcement, were calculated. Third, an algorithmic game theory model that leverages learning from historical bid data was proposed. The findings provided insights on how to account for construction materials price escalation under uncertain market conditions. Overall, this study contributes to the growing body of research related to construction materials price escalation under uncertain market conditions by proposing a practical approach that combines predictive modeling with game theory models.

Investments in green infrastructure (GI) can greatly improve the well-being and resilience of urban areas. However, the implementation of GI also has different challenges, including social considerations. Therefore, GI solutions need to recognize those social challenges and plan spaces where uses are complementary rather than conflicting. Existing research has suggested potential strategies to address the social challenges of GIs; however, little-to-no studies have focused on modeling the vulnerability to GIs’ social challenges. This study aims to fill this research gap by proposing a GI social vulnerability index (SVI). A list of 35 indicators was identified. Then, the weight for each indicator was assigned using Principal Component Analysis. Finally, the weights of the 35 indicators were used to develop the SVI. The findings offered insights into the most vulnerable/susceptible Census tracts in NYC. The results also provided insights into the relative importance of different indicators that should be considered when implementing GI initiatives. This study adds to the body of knowledge by helping city planners and policymakers target areas for interventions and mitigate the social challenges of GI. The research also provides a better understanding of how GI implementation can impact urban areas and the residents.

Recently, intelligent realities (IRs)—integrated systems that combine advanced analytics technologies such as digital twins (DTs) and artificial or virtual reality technologies—have emerged as a value-adding enhancement of DTs for facilities management (FM). However, DT-based IRs for maintenance are yet to be studied to determine their fitness for purpose from a practice perspective. The lack of practitioners’ perspectives on the development of DT-based IR poses a potential disconnection between desired end user preferences for interaction with a DT for maximum application benefits and the theoretical development of this emerging knowledge field. This paper presents FM professionals’ ex-ante evaluation of a conceptual prototype for a DT-based IR system. Using an electronic survey among selected FM professionals from educational and transportation facilities, it was found that the proposed solution can improve the visibility of managed assets, enhance predictive maintenance (PdM) practice, and provide cognition and judgment aid for FM professionals when they are mobile. The study provides relevant insights for developers of DT solutions and researchers on the preferred end user interaction modalities.

Facility management systems aim to maintain an acceptable level of service for the facilities through timely inspections and proper rehabilitation decisions. While inspection reports and maintenance requests embody the key data for all decisions, the manual and subjective nature of this data introduces errors and inconsistencies that complicate rehabilitation decisions. This paper enhances the utilization of inspection and maintenance data for rehabilitation purposes using data mining. Text mining is used to prioritize the components in most need of rehabilitation and trace fault patterns across multiple facility components to facilitate work packaging and crew assignment. An example of a 600-villa housing compound with over 2,000 maintenance requests was used to demonstrate the developed system. Analyses’ results based on component or type of work are then provided, as well as means to schedule required repair work. The paper aims to streamline the rehabilitation work packaging and delivery processes for large facility owners.

This study evaluates Ground Penetrating Radar (GPR), a non-destructive method, for inspecting and assessing concrete bridge deck conditions. Despite GPR’s potential to detect diverse bridge deck defects, a comprehensive review of its configuration and scanning parameters for effective deployment is lacking. Moreover, the extent of GPR’s ability to detect various bridge deck defects remains unclear. This study focuses on identifying suitable GPR parameters and analyzing GPR signal characteristics associated with different bridge deck defects. The study first establishes suitable system configuration and scanning parameters, enabling efficient GPR data collection. Subsequently, GPR scans reflecting different defects (e.g., cracks and delamination) are analyzed to assess GPR’s capabilities. Computer simulations and laboratory experiments reveal signal patterns associated with rebar corrosion and concrete deterioration, as well as limitations in detecting certain types of delamination.

Scattered Repetitive Projects (SRPs) such as multi-school or multi-bridge rehabilitations are rising in numbers, complexity, and costs. To properly allocate resources for these complex projects, efficient planning and simulation become necessary. At the detailed level, Agent-Based Modeling and Simulation (ABMS) is among the powerful techniques that can be used to analyze the impact of crew movements and behaviors on task productivity. To support efficient allocation of crews to SRPs, this research developed an ABMS model using the AnyLogic software to simulate multi-crew allocation to scattered units, incorporating crews’ travel times among the units using GIS. The paper discusses the model and its implementation on a case study of scattered linear projects where the activities in each location are sequential. Model validation against a powerful schedule optimization model proved its flexibility and applicability. Future integration with a powerful repetitive scheduling engine is highlighted to consider more complex networks at different locations.

When disasters such as hurricanes approach, residents rely on transportation infrastructure to evacuate. Evaluating road utilization is essential for emergency evacuation as it facilitates timely traffic congestion identification and alternative route selection. The objective of this research is to evaluate road utilization during Hurricane Ian in 2022 in Lee County. In this study, traffic data, population distribution data, and sociodemographic data were collected. The utilization level of government-declared evacuation roads was evaluated using volume to capacity ratio and the road density. It was found that interstate highways were highly congested during the evacuation period, while other highways and county roads remained underutilized, which highlights the unbalanced use of roads during hurricane evacuation. Practical wise, the road utilization levels can be used by emergency planners and local authorities to design efficient evacuation plans and improve existing transportation infrastructure networks to adapt to higher evacuation needs.

Toll road concessions are instruments used by governments to transfer risks and finance infrastructure projects needed to sustain economic development. However, these projects are plagued by heavy uncertainties, in part because their free cash flows are affected by unreliable traffic demand forecasts. Accordingly, this study analyses the literature on the field of project evaluation in concessions to extract common practices when forecasting traffic demand. From this information, this research assesses the most common traffic forecasting methods and proposes a new approach that might address flaws in the estimation methods existing in the literature. Thus, the study compares Geometric Brownian Motion (GBM), Mean Reversion (MR), and Fourier series models (FSM). The comparison evaluates the best-fit patterns in traffic demand forecasts given the long-horizon feature of toll road concessions and uncertainty. The models use the historical traffic demand of a real concession initiative within Colombia’s Fourth Generation Roads Concession Program (4G). The results indicate that the Fourier series outperform GBM and MR when forecasting seasonal traffic. Because this finding was obtained from a real toll road with marked seasonality, the study opens a path for further research in traffic forecasting for seasonal behavior.

Diverse components and properties of the disaster resiliency-related features need to be explicitly defined and satisfied within a BIM model. Revealing building resilience rules and BIM validation processes, however, remains elusive. To ameliorate this challenge, this paper aims to unravel logical resilience features embedded in the building codes and building standards and define a formalized logic for representing the categorized rules. In addition, this study defines an XML schema for flood-based building regulations to broadly distribute the formalized resilience requirements. The methodology adopts a functional structure of flood disaster-resilience regulations for single-family houses into the validation process of BIM data that removes restrictions on predefined queries, transforming related flood features in BIM data into an open and quarriable database. This study also involves the development of an automated BIM-based validation system that is designed based on the identified rule logic. The expected outcomes and validation system provide automated resilience assessment and reporting of related building design data checking.

Smart cities provide intelligent services to enhance infrastructure systems’ performance and community well-being. The heterogeneity of the 3D model data formats in their infrastructure systems, however, frequently hinders seamless data exchange, sharing, and integration. Model View Definitions (MVD) and Information Delivery Manuals (IDM) have been broadly used to develop interoperable data schemas that can address this heterogeneity. For developing IDM/MVD of a smart city, generalized standards and streamlined approaches that can avoid inconsistencies and redundancies of complicated infrastructure systems and features remain elusive. Particularly, existing domain-specific standards primarily focus on data exchanges within individual infrastructure, overlooking underlying interactions and interrelated functional operations between multiple systems. Thus, a unified system that contains necessary relationships of each infrastructure system covering sub-domain standards is required to represent the interactions and exchanges between urban-level infrastructure of a smart city. In this regard, this study provides a framework that integrates various infrastructure domains based on modularized knowledge of infrastructure systems and their smart functionalities. The expected outcome also includes a unified ontological data structure for functional elements and properties. The framework connects functions, elements, and properties for each specific use case that enables the development of an interoperable environment at an urban level, and allows for defining new types of connections between different infrastructure systems.

Creating semantic segmentation datasets is more time-consuming than the preparation of image-level classification labels. This paper presents a cost-efficient pixel labeling method and experimental results of using Grad-CAM-generated heatmap for weakly supervised post-disaster site semantic segmentation. It starts with training CNN binary-classification models, then removing non-target objects, reducing searching regions according to the distributions of heatmap values, and assigning the pixel-level label to the highest heatmap value regions as segmentations. The proposed CNN model architecture has a 1 × 1 convolutional heatmap layer with a suitable number of channels to fit the target objects. Testing results showed the 16-channel heatmap layer reached the best building-background classification performance (F1-score 0.89), and the 128-channel had the best building-damage classification performance (F1-score 0.78) among the evaluated options of 4, 8, 16, 32, 64, and 128 channels. Furthermore, the proposed pixel labeling method can understand complicated post-disaster scenes by correctly locating target objects (buildings and damages) in aerial images via repetitions of searching the target objects in regions of higher heatmap values.

Recent studies have shown that using sensing technologies like wearable sensing devices (WSDs) can potentially mitigate the high rate of worker injuries in the construction environment. When warning alerts that assist workers in recognizing the hazards are provided by WSDs, workers respond based on their consciousness, pre-training, and experiences. However, some workers may not respond and/or react to the alerts because of a lack of knowledge about how to respond and/or react. Virtual reality (VR) can be utilized in educating and training workers to understand hazardous activities and their related sensor alerts to potentially enhance their safety behavior. This study uses a literature review to highlight WSD and VR features to identify their possible and practical integration for the aforementioned worker safety improvement application. The study presents a conceptual framework that can be validated and implemented to potentially enhance workers’ understanding of response and reaction to warning alerts from WSDs or sensing technologies for injury, illness, or accident prevention.

Accurately predicting stock prices in the construction industry is crucial as it can help reduce losses for construction companies and mitigate the impact of economic downturns on the industry. However, uncertainties often lead to inaccurate predictions in the Egyptian construction industry. To address this issue, researchers used mathematical models and machine learning models to predict stock prices and assist decision-makers in responding to sudden changes. The main challenge is identifying the factors that drive stock prices. This paper proposes two models: artificial neural networks (ANN) and support vector auto regression (SVAR) to provide predictions for construction companies’ stock prices, which can be useful tools for contractors and construction material companies to predict and understand fluctuations in major construction stock prices and take appropriate measures to avoid negative impacts on their stability. The aim of this study is to provide a comprehensive evaluation of different methodologies in predicting stock prices in the Egyptian construction industry.

The human society’s mobility tool is shifting the paradigm from combustion engine vehicles to electric vehicles (EV). The wide adoption of EVs will benefit the sustainability of society by improving fuel efficiency, lowering fuel and maintenance costs, and reducing carbon emissions. However, the rise of EVs also brings challenges, particularly in EV emergency responses, in which EV fires burn hotter and longer than traditional vehicle fires, and most first responders are not well prepared. This paper introduces a multimodality virtual reality (VR) training system designed to help firefighters understand EV fires. The proposed VR system allows the user to visualize different phases of the thermal runaway process during EV fires. The proposed system has four modules to enhance the fidelity and improve the presence for the user to understand the EV emergency response scenarios. The results from the preliminary evaluation case confirmed the usability and feasibility of the proposed system.

Effective project requirement management is one of the significant tasks for the success of project planning and execution in complicated and dynamic construction environments. For improved requirement management, it is recommended to identify and allocate project requirements at an early stage prior to construction for appropriate project progress monitoring and control with little effort and cost. Current requirement allocation practices rely heavily on time-consuming and error-prone manual processes due to a sheer volume of free-formatted project documents and the necessity of practitioners’ judgment based on their experience and expertise. This paper aims to propose a systematic approach to assigning and clarifying project requirements to appropriate project parties. The proposed approach employs natural language processing (NLP) technique to automatically assign project requirements to responsible project parties. The allocated requirements are then incorporated into project geographical information system to present associated project requirements along with geo-spatial information. By integrating the information system with classified project requirements through NLP techniques, pro-active requirement allocation will further facilitate both managing of a variety of requirements and the real-time dissemination of them to all project parties.

Off-site construction has been steadily gaining popularity due to significant potentials and benefits in repeatability and standardization of processes. Since most of the construction activities are manually executed by labor force, it decreases productivity and manufacturing capacity. To this end, this study proposes a Deep Reinforcement Learning (DRL)-based decision-making tool for optimizing labor allocation to each ongoing activity with the aim to improve production pipelines by reducing delays and cost overruns. The study includes a framework that models the interdependencies between workstations and dependence on productivity of available labor and material resources to determine the optimal number of workers at various workstations, thereby reducing the make-span. The model utilizes the Q-learning algorithm along with real-time data on labor and material availability, current progress status, and deadlines of modules in the queue for predicting best possible action regarding distribution of workers to different workstations in the facility. The developed framework is elucidated through a case study to demonstrate the capabilities of the proposed model in automated handling and management of resources. The proposed framework contributes to the body of knowledge by (1) considering all renewable and non-renewable resources involved, (2) accounting for uncertainties, and (3) providing critical information required for well-informed decision-making in real time.

Heating, Ventilation, and Air-Conditioning (HVAC) operations consume around 40% of buildings’ energy consumption (EC). To reduce this consumption, Phase change material (PCM) can be incorporated in buildings’ envelopes. However, the effectiveness of PCM varies depending on its thermophysical properties and climate conditions. Conventionally, finding the optimal PCM properties requires a computationally extensive parametric study of white-box model, which may not always provide accurate results due to input simplification. To overcome this shortcoming, machine learning (ML) algorithms can be developed to identify optimal PCM properties based on predicting HVAC loads and PCM state. The best performing models for predicting cooling and heating energy consumptions were achieved by Artificial Neural Network (ANN) with MAE around 8% and Random Forest (RF) with MAE around 45%, respectively. As for classification, RF with Polynomial kernel PCA showed an accuracy of 81.5%. These results demonstrate the potential of ML algorithms to optimize building energy efficiency.

Additive Manufacturing (AM) is rapidly transforming various industries. In the construction sector, 3D Concrete Printing (3DCP) is gaining traction as a solution to address problems like labor and housing shortages. While other industries successfully embrace digitalization, such as Digital Twins (DTs), to improve AM processes, its implementation in the construction industry is in its infancy. This paper aims to pave the way for future DT implementation in 3DCP by proposing DT architecture, including necessary components, and layers (physical, digital, sensing, computing/perception, communication, devices, and control/actuation) throughout project phases (design, preconstruction, construction, postconstruction). The findings revealed that the current literature focused solely on the preconstruction phase and the use of “Building Information Modeling” (BIM) for 3DCP. This paper provides insights into the DT architecture system for understanding the main components and tools essential for starting future implementations of DT for 3DCP and reducing the digitalization gap between construction and other industries.

Improving the reliability and efficiency of anomaly detection is important for establishing and maintaining a resilient built environment. Meanwhile, Digital Twins (DT) have attracted increasing attention in the built environment management. In this study, through a structured literature review, we compiled and categorized existing DT-assisted anomaly detection methods applied in built environment based on three anomaly detection criteria (scope, target, approach) and three key attributes of DT (integration level, fidelity level, and decision speed). In doing so, we presented the current research trends and analyzed how DTs contribute to the built environment management with different anomaly detection approaches. This study lays the foundation for establishing DT-assisted anomaly detection frameworks utilizing Digital Twins of different integration and fidelity levels to tackle anomaly detection problems with different approaches. Moreover, it shows the efficacy of Digital Shadows (one directional DTs) in built environment anomaly detection and calls for future research endeavors in high-integration-level DT applications and DT-assisted hybrid anomaly detection methods.