Computing in Civil Engineering 2023: Resilience, Safety, and Sustainability

In each year, flooding causes tremendous damages to communities across the United States. A common but difficult decision problem for impacted communities is how to rebuild the housing stocks so that they will be resilient to future flood events. Choices such as elevation and buyout are often intensely debated in local communities, yet communities often lack the adequate data and tools to make wise decisions. In this paper, we present an urban digital twin-based decision support approach to empower local communities to rebuild flood resilient communities. The study shows with an urban digital twin consisting of 3D urban data, high-resolution hydrodynamic models, and regulatory data, stakeholders can effectively reduce the uncertainty in making critical rebuilding decisions such as elevating homes and identifying clusters of improvement opportunities such that the most cost-effective rebuilding choices can be made. Throughout the paper, we will use Manville, New Jersey, a township that was catastrophically impacted by Hurricane Irene, and more recently Hurricane Ida, as a case study to demonstrate the utility of our proposed approach and its superiority over other current approaches.

Rapid structural health inspection process is essential for post-disaster reconnaissance. Nowadays, visual inspection still dominates the practice, which is time-consuming, expensive, and unsafe. To reduce labor costs and guarantee personnel safety, we develop a fully autonomous UAV-based structural damage exploration platform that explores and rebuilds post-disaster environment maps with limited prior information. As a result, environment maps can be used for further damage detection and risk evaluation. The proposed platform implements a closed-loop control strategy in the UAV with a 3D lidar sensor for navigating itself to explore the unknown post-catastrophe environment. The control system integrates a lidar-based simultaneous localization and mapping algorithm and a next best view planner. The UAV can automatically rebuild a partial environment map and determine appropriate viewpoints to visit next. We tested our autonomous UAV inspection system with constrained damaged infrastructure in a virtual reality environment, RSASDA. The proposed UAV-based platform can rebuild the targeted infrastructure map and guarantee the required accuracy and completeness.

Building disaster-resilient communities requires a better understanding of what community stakeholders value regarding resilience. Communities with various characteristics (e.g., geographic, demographic, social vulnerabilities) are expected to prioritize resilience strategies to align with the value systems of their stakeholders. Stakeholder values include the aspects that are of importance, merit, and provide utilities to them, and stakeholders from different communities may hold values with varying degrees of importance, forming their distinct value systems. Despite the importance of involving stakeholders in resilience planning, there is limited awareness of stakeholder value systems with respect to resilience planning. To address this gap, this study focused on identifying stakeholder value systems and analyzing how the value systems vary across different communities (i.e., coastal vs. inland; metropolitan vs. non-metropolitan; more vs. less socially vulnerable). A semi-supervised learning technique—anchored correlation explanation (anchored CorEx)—was used to analyze the social media data (i.e., Twitter) collected within the Florida communities. The results demonstrate that communities with various characteristics have significantly different priorities regarding several stakeholder values. The findings of this study could inform policymakers to better plan for community resilience by aligning their strategies with stakeholder value systems, thereby increasing the effectiveness and efficiency of implementing resilience practices.

In this pressing time of climate change, the nation’s civil infrastructure systems are challenged by more frequent and intensifying natural disasters. Natural disasters have a disruptive impact on human activities. In coping with ever-increasing disasters, research studies have strived to develop resilience strategies to build physically resilient infrastructure systems. However, there is a lack of research investigating the impact of disasters on human activities across different population groups. To address this limitation, this paper proposes a clustering-based method to analyze the changes in human activities during disasters in a human-sensitive way. The proposed method includes two components: (1) population group discovery to cluster atomic population units into distinctive population groups and (2) activity change quantification to quantify the changes in activity frequency for each identified population group. As a preliminary study, this paper focuses on presenting the proposed method and its implementation in analyzing the impact of Hurricane Ida on different population groups in Manhattan, NYC.

Extreme events have become more frequent, severe, and widespread in recent years. Typically, human suffering and financial loss in a disaster are unevenly distributed among different communities due to their social vulnerability levels, which weaken the capacity to respond to and recover from disasters. Therefore, understanding social vulnerability distribution is crucial to explain how communities experience the same hazard event differently and support community resilience management. However, current assessment methods (e.g., CDC/ATSDR SVI) often regard social vulnerability as a static concept and assume that residents do all activities within their home community, which is in contract with the fact that there are inter-communities interaction due to human mobility. To solve this problem, this paper proposed a framework integrating human vulnerability and mobility trajectories to create a spatial-temporal community social vulnerability assessment. Through the process, we will first use Markov Chain to simulate activity schedules of different vulnerability groups defined by the CDC. Then, we will track the location-based mobility of people with positioning data collected from mobile devices. After that, a connection between vulnerability-based activity schedules and location-based mobility trajectories will be established to integrate humans, activity, and location together for generating vulnerable people distribution. Finally, we will evaluate the community vulnerability level with the new vulnerable people distribution among different communities to get the spatial-temporal vulnerability assessment result. This can help improve the community’s resilience against hazardous events by making in-time response strategies for different vulnerable people. Besides that, this can also benefit daily energy planning with minimum economic and safety risks to improve the community’s resilience.

The floods in 2022 resulted in significant damage to infrastructure and livelihoods. Being vulnerable to climate change, Pakistan must convert this adversity into an opportunity by developing robust, efficient, and resilient post-disaster reconstruction (PDR) strategies. This paper aims to identify and develop GIS-enabled resilience approaches and strategies that Pakistan’s stakeholders can use in the PDR. The triangulation approach is adopted to identify GIS-enabled resilience strategies for PDR. This research involved a systematic review of current GIS and PDR resilience strategies, as well as an exploration of GIS-enabled resilient PDR strategies. The strategies were further refined through a questionnaire and interviews with stakeholders in Pakistan. The results from triangulation research show that GIS-enabled resilience strategies play a pivotal role in making PDR more resilient and sustainable. The research will help the Government of Pakistan and other developing countries to improve efficiency during PDR and priorities investments in climate-resilient infrastructure.

Due to global warming, wildfires are becoming increasingly frequent and destructive, threatening environmental, economic, and human well-being on a global scale. Recent advancements in remote sensing and advanced data analytics have spurred the development of fire occurrence prediction models (FOPMs) to tackle this challenge. Although a plethora of features have been employed in the development of FOPMs in prior studies, identification of the most relevant features and optimal feature subset remains a critical knowledge gap. Utilizing California as a case study, this study fills this knowledge gap by conducting a comprehensive investigation on 96 relevant features gathered from seven heterogeneous databases. Ten machine learning algorithms were tested and employed with four feature importance methods to derive an importance score for all the features. Eleven features were identified as the optimal feature subset, and XGBoost achieved the best prediction performance with F-score of 97.35%.

Debris removal is one of the most expensive and challenging aspects of recovery from disasters. Erroneous debris estimation and delayed cleanup operation can block or slow down access to affected areas, thus putting public health at risk, and seriously hinder effective post-disaster search-and-rescue (SAR) and resource allocation. Existing disaster debris estimation techniques merely produce rough estimates of debris volume with significant errors. Recent research has demonstrated the value of low-cost unmanned aerial vehicles (UAVs) and artificial intelligence (AI) to rapidly collect information and assist in performing damage assessment of the building stock. However, the utility of AI models in debris estimation and classification is still underexplored. To this end, this study aims at enabling the measurement of debris volume and composition in residential structures from the outcome of AI-based damage assessment. Results from scaled experiments indicate that the proposed approach can provide high-fidelity estimation of disaster debris volume and composition.

Out of the estimated 926 million Americans that use airports as their transportation systems, 27 million of them have a disability. People with disabilities face many barriers in the built environment, especially within public transportation (e.g., airports, train stations, bus terminals). This study applies unique text mining techniques to discover the topics that travelers with disabilities experienced. We collected and analyzed 553,430 reviews for 67 international airports across the United States to find the possible inequity issues involving passengers with disabilities. As online reviews are typically shorter in nature than other forms of social media, a unique short text clustering method (a collapsed Gibbs sampling algorithm for the Dirichlet multinomial mixture model, GSDMM) was used to identify 10 topics through 3,516 unique nouns. Our results showed that topics involving restrooms and intra-airport navigation score lower for travelers with disabilities than with all travelers. Our approach and findings can be used practically to identify concerns for designers and airport operational entities to find potential concerns to mitigate the inequity issues of passengers with disability.

Energy-related occupant behavior can significantly impact building energy consumption. Identifying occupant behavior and individual activity patterns is essential to modeling human-building interaction and buildings’ energy performance. American Time Use Survey (ATUS) has been used widely for analyzing occupancy and occupant patterns and developing occupant behavior models in the last few years. This study aims to comprehensively review occupant behavior models in residential sectors built based on ATUS data. First, a total of 31 peer-reviewed articles in which ATUS was used to develop occupant behavior models were collected. Then, a comprehensive analysis of the model inputs, outputs, and methods used in the collected studies, as well as the most practical occupant behavior methods, is presented. Finally, the future research direction in occupant behavior modeling and simulation using time-use data is discussed. The results showed that the Markov Chain method is the most common technique for modeling occupant behavior in building energy simulation.

Many studies have emphasized the link between the built environment and occupant well-being. However, it is not fully understood which aspects of the built environment impact which metrics of well-being, and further investigation is warranted. In-person laboratory experiments have provided significant insights, yet they are resource intensive. Immersive virtual environments offer a promising alternative method of conducting studies in a more cost- and time-effective manner while allowing for controlled manipulations of the variables of interest. In this study, we develop an immersive virtual environment (IVE) framework to investigate the impact of different built features on psychological, physiological, and behavioral well-being outcomes. Our framework includes multimodal data collection through a virtual reality (VR) headset, physiological sensing devices, recordings, and integrated in-VR questionnaires. To evaluate the effectiveness of the IVE framework, we conduct a pilot study (N = 20) modeled after a previously conducted study in a real environment. We hypothesize that the well-being measures will be sensitive to the manipulations to the built features that participants experience in the IVE. The preliminary validation of the proposed framework lays a foundation for further investigation on relations between the built environment and occupant well-being leveraging IVEs.

Humans have an innate need to connect with nature, and designing built environments that promote this connection can support well-being. While research has shown that natural materials, views of nature through windows, and presence of indoor plants, are design features that promote well-being, there is a lack of understanding of how much connection to nature inside buildings is needed to achieve desired well-being outcomes. To answer the question of “how much,” we first need tools to quantify indoor doses of nature. In this study, we compare two methods of measuring nature indoors to a baseline pixel analysis. We use a semantic segmentation computer vision technique and a parametric method. These methods are applied to photographs and 3D models of a conference room. While both methods face limitations relating to quantifying views of nature through windows, they perform well overall compared to our baseline. We discuss the advantages and limitations of each method to support studies of the role of connection to nature indoors on well-being.

By 2030, global water demand is expected to surpass supply by 40%, and only 81% of the global population will have access to clean drinking water, leaving 1.6 billion without water. In response to the global water crisis, the circular economy as a concept has been proposed and advocated over the past few decades, seeking to minimize waste, recover resources, and reduce consumption by sharing, reusing, and recycling water resources. However, there have been limitations in distributing the accumulated knowledge to the general public due to resource accessibility and interpretability of condensed information. To further promote the implementation of a circular economy, it is crucial to educate the public on its concepts and effects to encourage behavioral and policy changes. Consequently, this research seeks to bridge the gap by using cutting-edge technologies to increase public awareness. Mixed reality (MR) provides an immersive experience and integrates the physical and digital worlds as it allows users to interact with digital objects that are overlaid with the actual world, resulting in a more effective learning experience. This study presented a general framework for developing MR applications in educating the public on circular economy in urban water systems, which involves developing an MR application to be deployed on a mixed reality headset, namely Microsoft HoloLens2. 

Access to accurate and timely data by disaster response decision-makers is vital in the management of flood risk. The possibilities of using emerging and historical data have increased the need to develop systematic methods for evaluating data input requirements to ensure appropriate use of data-driven methods in disaster response. The goal of this work is to present current data formats and requirements in response and recovery decision models and identify potentials to include measures of equity to make them more flexible and equitable for a larger segment of the US population. This study evaluates existing decision models and presents a framework to aid researchers and modelers to incorporate social vulnerability and community data with other data sources used in disaster response and recovery. The framework contributes to the advancement of flood resilience management by providing an analysis process to facilitate inclusion of diverse data streams for use in decision-making.

The increasing availability of global positioning systems and smartphones has enabled the crowdsourcing of actual travel behavior. Using this data to measure the impacts of flooding on traffic functionality, however, remains a challenge, as different traffic variables can be used to quantify functionality, and it is not straightforward to determine the duration of the traffic disruption. This research analyzed five flood events to determine the traffic variables that better capture the impacts of flooding on traffic functionality. The study also proposes a methodology to determine the moment at which the impacts of flooding on traffic end. The results suggest that number of trips, vehicle miles traveled (VMT), and vehicle hours traveled (VHT) are the traffic variables that better reflect the impacts of flooding on traffic functionality and that changes in the standard deviation of these variables can be used to determine the duration of the traffic disruption.

Flooding is the second-most deadly natural disaster in the US, affecting a large population. However, the lower-income population often fails to undertake necessary measures against flood risks due to their economic incapacity and limited accessibility to information. Using at least 1 foot freeboard to elevate the first-floor level of a building above the base flood elevation is encouraged as a precautionary measure against flooding by National Flood Insurance Program without considering its implications at the micro-level. The generalized strategy prevents an optimal level of freeboard implementation, affecting the required initial investment and associated economic liabilities. This study proposes a new framework for analyzing the risks to benefits of using an optimal level of freeboard at the micro-level of the population and identifying disparity among different income levels. The framework helps identify factors that affect the ability to adapt to the new policy on freeboard and develops a method to measure the vulnerability of different communities empirically. One case study was conducted on Jefferson parish to elucidate the methodology. The research outcomes showed that lower-income households entail higher economic risks in adapting to the new freeboard policy; however, the benefits are also higher for lower-income groups than for higher-income groups. The study contributes to the existing body of knowledge by providing a prioritization framework that can help to identify and rank (1) the most vulnerable communities that face economic and other liabilities toward freeboard implementation and (2) the vulnerable areas that require further development for improving long-term resiliency.

Climate change is increasing the frequency and intensity of extreme precipitation events, resulting in more extreme flash flooding events. To increase flood resilience in urban areas, there is a need for local-scale flood information systems to guide evacuation and rescue operations. However, current guidance for inland areas prone to riverine flooding is often limited and ambiguous due to limited application of urban-scale hydrodynamic models and digital urban feature datasets. To address these challenges, this study proposes an operational framework that integrates a validated urban-scale flood forecasting model, high-resolution urban point cloud datasets, human vulnerability model, and transportation network to predict rescue and evacuation routes. Using Manville Township in New Jersey as a case study, the study reveals the time-varying nature of evacuation route for residents and rescuers during Hurricane Ida. The results highlight the importance of an operational framework with flood forecasting systems to increase the effectiveness of evacuation actions.

Access to building and environmental data is critical for firefighters when they respond to a building fire. Existing research mostly focuses on information gathering at fire emergency sites but not on information integration. This research proposes an information integration system for firefighters using a graph database to integrate relevant building and environmental data. An ontology that represents firefighters’ data requirements is used as the knowledge base for the proposed system. Multiple data sources are identified to provide both static and live building and environmental data. After the data was integrated, two sample applications were built based on the unified database. The first is an automated database assessment for the most critical data firefighters need, while the second application is on-demand access to required data. A case study on a real-world building is used to provide technical details and evaluate the method’s usefulness.

The architectural, engineering, and construction (AEC) industry has repeatedly shown that the adoption of technology can improve efficiency, quality, and productivity. Several barriers need to be overcome to enhance sustainability in the built environment through technology adoption. This study aims to highlight how technology adoption can contribute to achieving sustainable goals. A survey of 46 construction professionals in different roles within a large construction firm provided insight into the challenges and opportunities of using digital technologies, such as building information modeling (BIM), cloud computing, big data, and virtual reality, to improve sustainable construction practices. The key benefit mentioned includes a waste reduction in terms of time, cost, and personnel. However, there is an issue with the industry’s readiness to implement these technologies. The future of digital technology adoption within the construction industry is promising and provides an effective way to promote sustainable construction practices.

As the deadliest climate event over the past 30 years in the US, heat waves are projected to be more intense and dangerous under global warming. Cities, composed of unevenly distributed populations and diverse landscape, suffer from significant adverse impacts of heat waves with varying land surface temperatures across different areas. In addition to the climate, effective adaptation to heat waves also requires an understanding and evidence regarding how urban natural, physical, and social processes influence urban thermal variations. Prior studies that mostly rely on traditional regression methods fall short in interpreting the complex role of these factors in influencing fine-scale thermal variations within and even beyond the physical boundaries of these factors. To address this research gap, we propose an interpretable machine learning method based on graph neural network to disentangle the significant factors that influence local thermal conditions. Our model can explain 35.3% extra variations compared to the baseline ridge regression model. The results of the model can inform urban adaptation strategies, including nature-based solutions, such as evaluating the cooling efficiency of spatially variant greening interventions at fine-scale levels.

During the COVID-19 pandemic, many academic institutions have been facing the challenge of how to maintain face-to-face instruction while also reducing the chance of an outbreak within the student population. One potential solution that has proven to be efficient is wastewater surveillance of student residence halls. In a case study, BIM models of 18 residence halls on a university campus were developed, each of which included parametric components such as building envelopes and plumbing fixtures. These models also included the locations of COVID-19 wastewater autosamplers. A Dynamo visual programming script was developed to highlight a vertical riser once the connected autosampler showed the existence of the coronavirus. Supplementary 2D floor plans and 3D isometric views can also be generated to help the institute’s pandemic management team make informed COVID-19 management decisions. In the absence of someone who is knowledgeable in plumbing, this novel application of the BIM technology can significantly improve visual comprehension of the extent of COVID-19 spread within a facility and assist the management team in decision-making processes under tight time constraints.

Civil infrastructure is a key driver for growth, employment, and better quality of life, which leads to communities transitioning from the natural rural vegetation to urban infrastructure areas. Urbanization exacerbates worrying climate change trends due to man-made activities and increased anthropogenic heat production resulting from urban population growth. This contributes to numerous climate change-related challenges, one of which is the urban heat island (UHI) effect, which affects human health and welfare. While several states in US have experienced high number of heat-related illness cases in the past years, minor research efforts were conducted to determine the areas that are subject to the highest heat-related risks associated with UHI. In relation to that, this paper addresses this knowledge gap by assessing the vulnerability of 95 communities in the state of Tennessee that are exposed to the UHI effect by considering demographic, geographic, climatic, and health factors. To this end, this paper followed an analytical approach based on the integration of unsupervised machine learning algorithms with multiple criteria decision-making methods to cluster or group communities based on 11 UHI-vulnerability-related factors. The results showed that clustering communities based on their vulnerabilities to UHI-related considerations can reveal the most critical geographical areas that are in immediate need to implement strategies that reduce the UHI effect and enhance heat resiliency. Ultimately, this research adds to the body of knowledge by helping states prioritize the design and implementation of optimized urban planning and infrastructure management measures to address UHI and climate change consequences.

Innovation and technological advancements have enabled energy production to meet growing energy demands. The challenge, however, is to achieve this goal without generating significant greenhouse gas emissions. Leveraging electric vehicles (EVs) as a distributed power storage solution for solar PV-powered energy presents an opportunity to significantly reduce the carbon footprint of energy generation. This paper explores the potential of vehicle-to-grid (V2G) technology to reduce greenhouse gas emissions by utilizing solar-generated energy stored in EV batteries. This study presents the findings of a case study that evaluates the economic feasibility of integrating a single-family household with a V2G capable system. As a result of examining the findings of the case study, a better understanding of the feasibility and advantages of V2G technology can be gained, as well as its potential as an alternative to traditional energy storage systems.

The prevalence of natural hazards and extreme climatic events highlights the critical importance of disaster recovery for communities. In this study, we examine the impact of Hurricane Harvey on private well water in Texas and explore the relationship between the contamination of well water, stewardship behavior, and demographic and social capital characteristics. We develop a multi-level regression model that shows bonding and linking social capital index, median house income, owner-occupied houses’ percentage, and the percentage of households speaking only English correlate with people’s vulnerability to unsafe drinking water supplies. We validate the results through a partial dependence framework built upon a random forest model. Our study provides valuable insights to address issues of fairness and vulnerability in society and foster resilience and preparedness in the face of shocks and hazards.

Predicting crowd flow patterns in a physical space can be useful for infrastructure management and safety planning. A simple representation of individuals in Euclidean space is insufficient for representing people’s spatial distribution and movements over time. This paper describes a spatiotemporal graph formulation, namely crowd mobility graphs (CMGraphs), to represent the spatiotemporal data. The CMGraphs model employs dynamic node features that store temporal crowd flow information, while the time-invariant edges represent spatial connectivity of locations of interests in the surrounding space. The spatiotemporal formulation using the CMGraphs allows for crowd flow prediction. Specifically, graph neural network is used to aggregate neighborhood nodal information on CMGraphs to capture spatial connectivity. Subsequently, recurrent neural network is employed to generate future sequences of crowd flow. An experiment is conducted using a publicly available video dataset at a train station to demonstrate the effectiveness of the proposed CMGraph formulation for crowd flow forecasting.

The emerging interest in smart, sustainable cities has propelled an exploration race by adopting various technologies to improve business efficiency. Specifically, the smart technology has a considerable potential in a shared environment, which is defined as a socioeconomic system that allows an intermediated set of exchanges of goods and services between individuals and/or organizations to increase efficiency and optimize resource allocation. Smart technology is capable to track the utilization of shared goods or service in a continuous and real-time manner, which effectively strengthens the trust between the shareholders (i.e., users and providers); however, its application in the sharing economy is still immature. This paper proposes an approach to adopt smart technologies in shared properties in smart cities. The developed method includes (1) identifying critical indicators and developing various sensors to obtain real-time data to assess the user’s utilization of the shared property, (2) developing an IoT device using open-source embedded systems (i.e., Arduino) to integrate all the sensors, and (3) creating a cloud-based, visual interface accessible by property users and providers to monitor service utilization and the user’s compliance with the terms of service. The proposed IoT device was implemented and tested in a shared environment in New Jersey (NJ), USA. Ultimately, this study aims to pave the way to transforming traditional business practices toward more intelligent models whereby both the providers and the users have fewer, or ideally no, regrets in relation to their decisions about the sharing economy environment in smart cities.

This study investigates the rural-urban disparity regarding the impacts of severe weather-related power outages in the United States, an issue that has not received much attention despite its significant socioeconomic losses. Specifically, the research focuses on the effects of Tropical Storm Isaias in Connecticut and utilizes geographic information system to assess the storm’s impacts on 71 urban and 78 rural towns across three dimensions: trouble spots, impacted populations, and outage durations. The results indicated that while rural communities experienced fewer trouble spots and affected people in absolute quantities, they had longer recovery times compared to their urban counterparts. Furthermore, the disaster impacts were more severe in rural areas, with 60.5% of the rural population being affected as opposed to 37.5% in urban areas. These findings provide new insights into the geospatial dynamics of disaster impacts, which can assist emergency management professionals in identifying vulnerable areas for improved pre-disaster mitigation measures.

Incorporating building-integrated photovoltaic (BIPV) systems at the urban level can reduce emissions and increase clean electricity generation. However, these systems face challenges in matching electricity demand and supply, particularly during peak demand. A proposed solution is optimizing the orientation of BIPV panels, but this can be challenging due to limited space available for panels, self-shading, and difficulties integrating panels into the building façade. This paper proposes a framework that simultaneously optimizes both BIPV orientations and building forms while considering the energy demand of the urban block. The approach is demonstrated through a case study of six mid-rise apartment buildings in California. The process sought to minimize overgeneration of energy while minimizing the system payback period. Depending on designer priorities, quality designs can have a payback period ranging from 0 to 18 years and overgeneration between 0% and 30%. The methodology offers a performance-driven design optimization technique for BIPV systems at the urban level.

Hospitals across the world have been significantly impacted by the novel COVID-19 pandemic, which has required incident commanders (hospital decision-makers) to make changes to the hospital such as opening and closing wards, changing screening protocols, reducing or canceling elective procedures, using temporary treatment spaces, etc. These changes impacted the relationships between the spaces within the hospitals. This study used the graph theoretical approach (GTA) optimization technique to analyze the impact of these changes during a surge event such as the COVID-19 pandemic. In particular, the study is interested in how these changes impacted the relationships between the various departments and areas within the hospital facility layout. In both cases, the layout created with surge data outperformed the layout created with normal conditions. This research provides insight into the impact of layout on operations and provides recommendations on how hospital facilities can be designed/modeled for flexibility in future surge events.

With a warming climate and existing inequities in the built environment, it’s critical to examine pathways for reducing energy consumption and heat stress on city residents, especially in disadvantaged communities who bear the brunt of climate change. Modeling the urban context remains a challenge for accurate performance prediction in urban building energy models (UBEMs). We build upon existing research by leveraging open-access data to describe the urban context. First, we use socioeconomic data to develop archetypical UBEMs that describe disparities across a city. To capture features of the environment often ignored in UBEMs, we introduce an “urban context vector” created from satellite data. In doing so, microclimatic and urban heat island effects are captured in the weather file, providing insight on how the urban context affects building performance. This paper demonstrates a generalizable model to produce multi-scale predictions of retrofit impacts on energy consumption and heat stress. By leveraging the interpretability of UBEMs with open-access contextual data, cities will be better equipped to develop informed policies to reduce energy inequities and heat stress.

This study aims to provide material scientists and building designers with valuable insights regarding the properties of shape-stabilized phase change material (SSPCM)-cement composites to maximize building energy efficiency. To this end, the energy-saving potential of SSPCM composites is evaluated using the Department of Energy’s (DOE) reference commercial building for small office buildings, and energy simulations are conducted through EnergyPlus. A parametric study is conducted to investigate the impact of peak melting and solidification temperatures on the heating, cooling, and total heating; ventilation; and air conditioning (HVAC) energy consumption in two representative cities, Houston and Chicago. The results indicate that the optimal peak melting temperature and optimal solidification temperature in hot climates are higher than those in cold climates. Furthermore, the study reveals that heating dominant climates exhibit a higher sensitivity to the choice of peak melting temperature, while cooling dominant climates are more sensitive to peak solidification temperature. These findings have significant implications for building designers interested in maximizing energy efficiency. Specifically, the selection of peak melting temperature is crucial in cooling dominant climates, while the selection of peak solidification temperature is more critical in heating dominant climates. Overall, this study emphasizes the importance of considering regional climatic conditions when developing and implementing SSPCM composites in buildings for optimal energy performance.

Mitigating the impact of winter storms on essential lifeline infrastructure and services, such as groceries, water and wastewater, electricity, communication, and transportation, is crucial for effective disaster management and response. However, the joint impact of multiple infrastructure/service disruptions during a pandemic is a relatively unexplored aspect. To assess and predict the severe impact on residents’ well-being caused by winter storms, understanding experiential knowledge of the joint impact of infrastructure/service disruption is necessary. Although surveys and interviews can provide this information, they have limitations, such as the level of detail obtained, and the time and cost required for conducting interviews. To overcome these limitations, AI-driven interview chatbots, such as the Juji chatbot platform, can be used to collect experiential knowledge related to household resilience to winter storms. Chatbots can improve productivity and efficiency while eliciting detailed information about household challenges and responses to infrastructure/service disruptions. Analyzing the results of the chatbot interviews can help researchers and decision-makers better understand residents’ challenges and responses to the challenges during disasters in targeted neighborhoods.

Flooding is a frequent disaster in urban areas. It disrupts the urban transportation and results in longer response times from emergency public services. The primary approach of predicting the response time in prior studies is to simulate vehicles’ movement on disrupted road networks. However, this approach implicitly assumes free-flow transportation and lacks real incident data for prediction and validation. To test the effectiveness of the simulation-based approach, this study applies the approach to predict the ambulance response time change during flooding events. The result shows that the approach could predict the response time increase due to road disruptions and the routing change, but is weak in modeling more complicated mechanisms, such as congestion, demand surge, and service capacity limits, all of which play a significant role in the expected response time. As a promising future direction, a data-driven approach to overcome the weakness of the rule-based simulation approach is discussed.

The increasing demand for senior housing has led to efforts to transform senior housing design to enhance the quality of life of older adults. The sustainable design theory of biophilia posits that humans are biologically encoded to have an affinity to natural systems and processes. Biophilic design has shown the potential to provide benefits associated with emotional well-being, healing, and cognitive performance. However, it is challenging to identify the impact of biophilic design on older adults, mainly due to the overlooked and misinterpreted preferences, perceptions, and needs of older adults. To address this challenge, this research proposes to create a new framework to simulate daily activities in an augmented reality environment with specific biophilic design strategies and interventions. The proposed method also involves collecting real-time biofeedback data from older adults using wearable devices and accessing the cognitive processes and responses of older adults using cognitive tests.

Ensuring the security of coastal infrastructure lifelines is crucial for maintaining sustainable economic development and enhancing resilience against climate change. Flood is one of the most common natural hazards that cause damage to infrastructure in the Great Lakes region. Bridges are the most vulnerable elements of the infrastructure system during flood hazards. Faced with frequent flood impacts, it is imperative that we systematically investigate and assess the risk of infrastructure due to inappropriate infrastructure design. In this paper, we conduct a computational investigation of the flood impact on bridge infrastructure in the Great Lakes region using fluid-structural interaction (FSI) modeling. The FSI model integrates computational fluid dynamics (CFD) and finite element analysis (FEA) to simulate the flow behavior of flood and its loading on the bridge piers. Case studies are conducted to simulate different scenarios of the flood wave and its impact on bridge infrastructure. The results of this study will provide coastal communities and stakeholders with a quantitative framework to predict the performance and risk of bridge infrastructure subjected to flood. This study will enhance pre-flood preparations and mitigate the long-term risk to life and property from future flood events in the Great Lakes region.

Design intent is communicated using 2D plans or 3D rendered building models displayed on 2D monitors. Buildings are designed to comply with American Disability Act (ADA) and code requirements. It is impossible to visit buildings before they are built to check how they feel and perform for diverse occupants’ needs. How can we know that future buildings will not only comply with ADA requirements but will also be as welcoming for a person in a wheelchair as they will be for people without disabilities? This requires an empathic design mindset. How can virtual field trips in building models advance project-based learning towards developing an empathic design mindset of the next generation AEC workforce? The paper presents an empathic design mindset framework developed and tested to study and define the AEC learners’ transformative journey from emotional to cognitive empathy experiences gained through virtual field trips in buildings towards developing an empathic design mindset.

Urban decarbonization targets and building performance standards are two ambitious policy packages that are rapidly being adopted by cities around the world. In order to evaluate relative energy efficiency in buildings many cities use a process known as building energy benchmarking. The output of energy benchmarking is useful for cities to determine what buildings are most inefficient in addition to providing data on estimated building-level greenhouse gas emissions. However, energy benchmarking often uses annual data to assess building performance and therefore may fail to capture insights on time-varying energy use and associated emissions. Without a more comprehensive view of energy usage and associated real-time emissions, current benchmarking methods and building performance standards are unlikely to accurately evaluate the decarbonization potential of urban buildings. Although there has been a growing interest in time-granular benchmarking the applications are limited due to concerns surrounding the collection and processing of such data. We utilize annual hourly data from a proof-of-concept hourly benchmarking platform to demonstrate the feasibility of time-of-use energy benchmarking. This data allows us to evaluate the performance of various benchmarking methods and building performance standards across different levels of energy use data temporal granularity. This analysis examines data at two scales: (1) inclusive of the buildings enrolled in the proof-of-concept and (2) extrapolated to all commercial buildings in New York City. The results are expected to demonstrate the value of shifting policy toward time-of-use benchmarking and its potential to help accelerate decarbonization of the urban built environment.

Debris generated and left in the wake of a disaster is one of a community’s greatest obstacles in the ability to restore function. The objectives of this paper are (1) to provide systematic procedures to develop and visualize the debris vulnerability index based on the four dimensions and (2) to develop to prioritize debris removal by integrating all dimensions into GIS mapping. The proposed methodology for Social. Unmanned aerial vehicle was applied to collect the 3D volumetric information, and 11 metrics were selected and evaluated to establish vulnerability of a location. The proposed process incorporates all systems into a process giving decision makers the ability to quickly determine scope, complexity, potential impact, and recovery needs. Therefore, the outcome of this research will ensure maximum utilization of the limited resources to reach a larger number of people who might be facing life-threatening situations in a timely manner, which will enhance resilience of the community.

Recent advances in robotics and artificial intelligence have influenced the broader adoption of human-robot collaboration (HRC) in industries such as manufacturing and healthcare, but the same cannot be said about construction due to the dynamic nature of the work environment. To facilitate informed decision-making in HRC adoption, there is a need to evaluate the potential benefits of incorporating robots in construction activities. This study evaluates the impact of HRC on construction workers’ workload. Experiments involving human-robot collaborative wood assembly tasks were conducted, and workload levels of humans in different scenarios were evaluated using physiological data collected with a wearable sensing device. Thereafter, a survey was administered to participants to assess their mental workload. Findings show that HRC posed less workload on workers compared to human-human collaboration, which demonstrates the potential benefits of HRC in the aspect of workload or fatigue reduction among construction workers, and thus could help enhance productivity.

Work-related musculoskeletal disorders (WMSDs) are a leading cause of injury for workers who are performing physically demanding and repetitive construction tasks. With recent advances in robotics, wearable robots are introduced into the construction industry to mitigate the risk of WMSDs by correcting the workers’ postures and reducing the load exerted on their body joints. While wearable robots promise to reduce the muscular and physical demands on workers to perform tasks, there is a lack of understanding of the impact of wearable robots on worker ergonomics. This lack of understanding may lead to new ergonomic injuries for workers wearing exoskeletons. To bridge this gap, this study aims to assess the workers’ ergonomic risk when using a wearable robot (back-support exoskeleton) in one of the most common construction tasks, material handling. In this research, a vision-based pose estimation algorithm was developed to estimate the pose of the worker while wearing a back-support exoskeleton. As per the estimated pose, joint angles between connected body parts were calculated. Then, the worker’s ergonomic risk was assessed from the calculated angles based on the Rapid Entire Body Assessment (REBA) method. Results showed that using the back-support exoskeleton reduced workers’ ergonomic risk by 31.7% by correcting awkward postures of the trunk and knee during material handling tasks, compared to not using the back-support exoskeleton. The results are expected to facilitate the implementation of wearable robots in the construction industry.

Active exoskeletons are emerging as ergonomic solutions in the construction sector to reduce work-related musculoskeletal injuries. While the benefits of active exoskeletons are promising, they can also cause increased muscle activity, leading to local muscular fatigue. This study aimed to examine the impact of the active exoskeleton system on the muscular activity of construction workers during common construction activities. Ten subjects completed material handling tasks under two weight conditions (10 and 30 lbs) in a lab-controlled environment, with and without using an active exoskeleton. Portable electromyography (EMG) sensors were used to measure lumbar erector spinae (LES) muscle activity in each condition. Four descriptive statistics features in the time and frequency domains were extracted from the collected signals. Results of the t-test showed a significant difference in the physiological metrics extracted from the subjects’ EMG signals of the LES muscle. Findings demonstrated that using active exoskeletons reduces the internal muscle force in the lower back regions of construction workers.

Occupational exoskeletons and exosuits (collectively termed as EXOs) have the potential to reduce the risk of work-related musculoskeletal disorders (WMSDs) by reducing exertion and muscle fatigue caused by heavy material handling, and repetitive and prolonged tasks in construction. However, the implementation of EXOs in construction is in its infancy, and there is a knowledge gap in terms of accepting, adopting, and applying this technology. This study applied a sequential multistage Delphi approach to identify facilitators, barriers, and potential solutions to solve those barriers to adopting EXOs in construction workplaces. Eighteen experts from academia, industry, and government agency provided insights and exchanged opinions in a workshop on these three aspects from a holistic view with respect to business, technology, organization, policy/regulation, ergonomics/safety, and end users (trade people). The expert panel reached a consensus on all discussed items. The outcomes of this study would help the research community have a comprehensive understanding of the potential of EXO use in the construction industry. This will subsequently underpin the development of a viable roadmap to study the envisioned future technology, workers, and work in construction workplaces.

Exoskeleton has been introduced as one of the effective ergonomic intervention strategies in the construction industry for reducing the physical burden of workers often exposed to heavy lifting and repetitive work. While extensive research has investigated the potential of the exoskeleton in site for reducing muscle fatigue, little is known regarding its effect on cognitive performance such as attention. Considering that attention plays a critical role in recognizing hazards in construction sites and incident involvement, this study aims to address this knowledge gap by determining the impact of wearing an exoskeleton on the hazard identification performance of construction workers. To achieve this objective, the research team has designed an experiment in which participants identify construction hazards before and after conducting repetitive manual material handling with and without an exoskeleton. In addition, functional near-infrared spectroscopy (fNIRS) was used to measure participants’ visual and cognitive processes during the experiment. The data was collected from 12 participants and analyzed through a permutation analysis to determine the differences between hazard identification performance with and without the exoskeleton. The result indicates that exoskeletons tend to positively affect the attentional distribution of workers with lower accuracy for the number of hazards identified. The findings demonstrate the influence of wearing an exoskeleton on the cognitive performance of workers and highlight the need for further research to reduce this effect.

This study investigates the use of immersive storytelling videos that leverage virtual humans and 360-degree panoramas to deliver easy-to-access learning contents for construction safety training. A pilot study using a pretest-posttest design was completed with residential construction professionals to understand the safety learning, hazard identification, and usability from interacting with immersive training videos. A 1-h training session was held with eight participants with over five years of experience in residential construction trade work (e.g., laborer, carpenter, and plumber). Metrics regarding safety learning, hazard identification, and usability were collected from participants using surveys. The results showed significant statistical differences in the safety learning scores. However, no statistically significant differences were found for the hazard identification scores. Participants rated the immersive storytelling videos as highly usable. It was concluded that immersive safety storytelling provides an easy-to-access and effective method for construction professionals to learn safety concepts through using 360-degree interactive videos.

Falling from heights remains a major safety concern on construction sites worldwide. In 2020, falls accounted for approximately 34.8% of the deaths in construction in the US. Previous studies have proposed a range of construction technology applications for fall hazard identification and protection. However, few technologies are applicable in real-time scenarios on construction sites, and some of them create a certain level of interruption to operations, as physical attachments to the human body or personal protection equipment are required. The objective of this study is to propose a real-time proximity warning system for fall hazard detection and protection with limited interruptions to operations. The proposed system integrates a microprocessor board, multiple ultrasonic distance sensors, and an early fall alarm with a buzzer. When the sensor unit detects a worker’s movement in a pre-identified hazard zone, the early fall alarm will activate, warning the worker to stay alert and take precautionary measures to prevent fall-related injuries. The proposed system is intended to serve as a supplementary measure for fall protection and prevention, in addition to existing measures.

Drones have been widely used for various construction applications, and human-drone collaborations and interactions are expected to increase significantly on jobsites. However, the safety risks associated with the onsite presence of these aerial robots are yet to be explored. Such safety risks might have higher impacts on construction professionals working at heights, who are potentially more susceptible to higher risks of injuries and fatalities. In this user-centered study, construction individuals were recruited to participate in an experiment and accomplish a series of at-height tasks within a virtual environment. A between-subject design (with drone or without drone conditions) was adopted to evaluate the emotional and attitude impacts of these aerial robots on construction individuals at heights. The results showed no evidence supporting that drone presence caused significant changes in participants’ emotional states or attitudes towards robots. The contribution of this study is to advance the safety knowledge about human-drone interactions on construction jobsites and ensure safe usage of this technology in construction.

Due to the labor-intensive nature of construction tasks, workers are exposed to an increased risk of work-related musculoskeletal disorders (WMSDs). Recently, a wide range of wearable robots (i.e., exoskeletons) has emerged in construction to reduce the risk of WMSDs. Although there are promising prospects for wearable robots to minimize the physical demands on workers, there is limited knowledge about their fundamental impacts on workers. In particular, there is a lack of studies on the impact of wearable robots on workers’ gait stability in dangerous situations, such as heights, which puts workers at high risk. To fill this gap, this study evaluated the impact of powered back-support exoskeletons on workers’ gait stability at heights through kinematic analysis. The analysis revealed that working at height with a back-support exoskeleton does not diminish users’ gait stability. The findings can shed light on the widespread adoption of back-support exoskeletons in construction.

Over the years, falls have been the leading cause of severe injuries and fatal accidents in the construction industry. Various automated detection and prevention systems have been developed to improve worker safety conditions on the job. However, there have been limitations in the use and accuracy of these systems due to the peculiarities of the construction sites. Hence, this study was conducted by reviewing multiple studies on 10 sensing technologies, their modes of operation, and the factors that influence performance, strengths, and weaknesses based on working conditions at varying levels of fall risk. This study proposes a framework that can facilitate the integration of two or more systems to detect and prevent workers from falling. A comparative analysis was conducted to identify the usefulness and limitations of various sensing technologies, with the aim of developing a hybrid system that can be employed in various fall hazard levels.

The effect of struck-by hazards as one of the most lethal dangers has not been given enough attention due to its variable and complex contributing factors. This study aims to conduct an analysis of struck-by accidents to identify the key factors that contribute to struck-by accidents, such as the accident time, weather conditions, worker age and body part injured, equipment involved, and location of the accident. The random forest (RF) model was used to analyze the factors related to struck-by fatalities. Results showed that the equipment involved, the body part affected, and the location of the struck-by accident are significant predictors of fatality. The four most important factors influencing struck-by fatalities are ground, head, vehicle, and afternoon. Based on the research findings, more attention should be given to equipment safety. In addition, safety training should be strengthened on improving workers’ awareness of urgency in safety.

Understanding distracted pedestrian behaviors is critical to road user safety and preventing traffic-related injuries. Front-facing vehicle cameras (a.k.a., dashcams) have increasingly become popular for documenting driving behavior and patterns. However, a relatively underexplored application of dashcam footage is to automatically identify distracted road users that may pose a threat to pedestrians. Detecting distracted behaviors in dashcam-captured imagery can enable drivers to take preventative measures and avoid potential traffic accidents. To this end, computer vision techniques powered by the prediction capability of artificial intelligence (AI) can be leveraged to identify pedestrians’ distracted behaviors when crossing streets and intersections. In this paper, pedestrians’ phone-related distracted behaviors are detected and classified in dashcam footage by leveraging convolutional neural networks (CNNs) assembled in the form of a two-stage detection and classification architecture. In particular, we propose to first detect pedestrians (stage 1) followed by classifying the most prevalent distracted behavior visible in each detected instance (stage 2). This technique has been developed on an in-house video dataset collected from urban intersections around a major university campus. Results indicate that the pedestrian detection model achieves 76% average precision (AP), and the classification of distracted behavior reaches 72% precision, 98% recall, and 83% F1-score.

Transportation construction sites generally encompass unsecured materials, temporary facilities, various equipment, and structures that are incomplete which result in an unsafe area, especially during or after the occurrence of environmental disaster events such as floods or hurricanes. Improper and insufficient safety pre-assessment and measures planned before a project begins can cause not only physical and financial losses but also additional costs and delays, especially for a construction project in a disaster-prone area. To address the existing challenge, this study aims to predict potential safety hazards according to a jobsite condition and expected severe weather in roadway construction projects. The proposed framework adopts machine learning for safety assessment that is conducted based on a project schedule, occupational safety database, and flood zone areas. With the scope of flood disasters and roadway construction, the framework is expected to evaluate the impact of severe weather on roadway construction safety, identify the possible sources of work events, and predict the class of events for a project team to promote safe and resilient work environment against flood disasters.

In 2020, more than 102,000 work zone crashes occurred in the United States, resulting in over 45,000 injuries and more than 850 fatalities. These numbers are higher than 2019 records, despite lower traffic volumes due to the COVID-19 pandemic. Also, population growth and the increased load on infrastructure are expected to lead to more construction work zones and, consequently, more crashes and fatalities in the future. As such, understanding the factors contributing to work zone crashes and their severity can assist the Department of Transportation (DOT) in safety management and planning. Accordingly, this study investigated the effect of freeway construction work zones on the severity of crashes by employing machine learning algorithms. Moreover, the performance of three of the most common machine learning models, decision tree, random forest (RF), and XGBoost, were evaluated based on accuracy, precision, recall, and F1-score. The model was developed using crash data and work zone information from the state of Utah between 2017 and 2021, considering an extensive set of work zone attributes, such as road factors, environmental conditions, driver attributes, and work zone features. The study results showed that RF has the best performance in severity classification, achieving an accuracy of 88.6%. Moreover, feature importance analysis reveals that roadway surface conditions, crash type, motorcycle involvement, weather conditions, and roadway junction type are significant contributors to work zone crash severity. The findings of this study provide valuable insights for the analysis of construction work zone crashes and DOT’s safety management plans.

Analyzing historical construction accident narratives is widely implemented by researchers to understand critical causes of past accidents. Previous studies have identified accident causes and affected body parts to assist safety managers in making timely decisions to prevent accidents. However, the sequential relationship between workers’ actions and accidents, which can provide valuable insights, has not been investigated. This study addresses the gap by conducting a novel investigation into the most common sequential patterns between workers’ pre-accident actions and accidents. PrefixSpan sequential pattern mining algorithm was employed on large-scale sequential data extracted from Occupational Safety and Health Administration accident reports in highway construction. The results show that “operating machines or equipment” and “working at height” are the most common actions leading to severe highway construction accidents. Furthermore, the sequence pattern “operating machines or equipment → crushed by objects → fatality” ranks high among the top accident patterns, with an 80% confidence level. Organizations can use the findings to devise specific safety programs and interventions to mitigate future incidents.

This study focuses on designing and developing a virtual reality (VR) game with decision-making scenarios that require critical thinking and problem-solving skills to enhance users’ understanding of safety best practices in construction safety. The study used professionals’ input, developed a VR game using the Unity platform, and tested the game with construction science students during the initial development phase. The feedback from these students will aid in developing more intricate and advanced scenarios. During the game, the player acts as a construction worker, choosing the right ladder to reach the first floor and get a glass of water from one of three water stations so that choosing the safest water station scores high, not the fastest, risking their safety. The study’s results show that participants without prior safety training found the VR game engaging and recommended it to professionals, and the game positively influenced attitudes toward safety. While participants made mistakes, the study highlights the potential of VR-based safety training to improve safety in the construction industry by introducing decision-making scenarios and recording scores on each decision so that participants can evaluate their performance and improve their self-efficacy by learning or getting safety training. In addition, the employer can equally benefit by evaluating their workers’ performance during the game and focusing on what their workers lack in safety training.

Occupational accidents are an urgent problem in construction. Machine learning (ML) methods for analyzing large amounts of data and the availability of accident report data have generated aspirations for novel learnings. Yet the quality of data in terms of input, inner availability, and output occurs as an issue in many ML development projects. This paper aims at investigating strategies to define, understand, and tackle poor data quality in a contracting company’s accident reports. A selective literature review within software system data quality and ML shows different foci on external or internal data. A set of records of occupational accidents are then analyzed. There are many missing entries on causality, as well as shallow descriptions, which hinder the discovery of new risks—possibly due to the data collection format and procedures. The low number of full entries calls for new repair strategies—both externally and internally.

With more than 900 fatalities and over 200,000 non-fatal injuries annually, construction is undoubtedly one of the most dangerous industries. While efforts have been made to develop autonomous safety surveillance systems, post-accident emergency response remains underexplored. This study addresses this gap by proposing a vision-based autonomous notification system using real-time data from multiple unmanned aerial vehicles (UAVs) at construction sites. The system identifies workers’ hand gestures, alerting a centralized command center with localization data when a distressed worker is detected. The system consists of two modules: (1) hand gesture recognition and interpretation and (2) localization module. A lightweight long short-term memory (LSTM) network was developed for gesture recognition, achieving a 94% accuracy rate in the test set.

Unsafe human behaviors cause over 80% of fatal accidents in the construction industry. Despite the use of technologies for safety control, injury rates remain high due to inadequate safety management systems and lack of analysis of workers' safety performance. This study aims to (1) identify critical construction hazards and the sensor-based technologies that can be used to detect workers' unsafe behaviors; and (2) develop a Sensor-Based Safety Performance Analytic Mobile System (SBSPAMS) to detect, alert, and analyze workers' unsafe behaviors. This study identified the top risky hazards and corresponding unsafe behaviors, along with specific applications of sensor-based technologies for safety management. The developed SBSPAMS is able to monitor and analyze workers’ unsafe behaviors. This study contributes to the body of knowledge of safety management by identifying the critical construction hazards, unsafe behaviors, and the sensor-based technologies for behavior detection. The developed mobile system can be a key instrument for technology adoption in the construction safe management.

Construction workers are vulnerable to physical and mental health challenges, causing illnesses, injuries, and fatalities. This fact stresses the need to closely assess and monitor the health and safety conditions of construction workers. Recently, researchers have used biosensor technology to develop several health monitoring frameworks that can monitor workers’ safety and health status through the acquisition and analysis of workers’ physiological signals. Despite the potential of these frameworks in monitoring subjects’ health status in a controlled lab environment, there is a concern regarding the performance of these frameworks in the field environment. One of the main limiting factors affecting the field performance of these frameworks is motion artifacts in the captured physiological signals. The frequent movements of workers while performing construction tasks can cause motion artifacts during signal acquisition, which will significantly reduce the quality of the captured physiological signals and thus degrade the performance of health monitoring frameworks. To address this gap, this study developed a motion artifacts removal method based on least mean squares adaptive filtering algorithms. To examine the performance, 12 subjects were asked to perform a material delivery construction task while their physiological signals were captured via a wristband-type biosensor, and the proposed method was applied to the signal acquisition process. Results reported that the proposed method removed 61.9% of motion artifacts from the captured EDA, PPG, and ST signals and improved the corresponding signal-to-noise ratio by 51.6%. This study contributes to the establishment of efficient physiological-based health-monitoring frameworks for construction workers.

The construction industry is known for its disproportionately high fatality and injury rates, making it one of the most hazardous industries in the US. Despite the significant risks involved, there is a lack of effective health monitoring in construction jobsites. While wearable physiological sensing and artificial intelligence advancements have introduced unique opportunities to assess workers’ health status, there are still inefficiencies in representing that information to support managers’ decision-making. Recently, the concept of digital twin (DT) has been used in various construction applications. Given the exponential growth of its enabling technologies, DT has great potential to transform worker health monitoring in construction jobsites. Therefore, this research investigates the feasibility of integrating workers’ physiological responses with DT technology to generate health maps that deliver workers’ aggregated health information to managers to reinforce their decision-making. The DT-based health maps are expected to enhance workers’ occupational health and safety at construction jobsites.

Construction safety monitoring increasingly relies on machine learning-based models due to their strong learning capability. However, these models’ performance often degrades when they encounter shifts in data distribution. To address this issue, a tracking algorithm can be used for the detection model to enhance detection performance and ensure high-quality safety monitoring. Most importantly, utilizing such method can lead to robust personalized workplace risk assessment and customized safety training to workers. This paper proposes a novel object tracking approach, called hashing-supported cascaded buffered intersection over union (HC-BIoU) tracking, which addresses low tracking performance due to visual domain shifting. The experimental results demonstrate that the proposed method achieves significant improvements in mean average precision (9.3%) and association accuracy (42.6%) compared to a YoloV8 object detector and a Cascaded Buffered IoU (C-BIoU) tracker, tested on a video sequence of a scaffold dismantlement scene with 1,983 frames and 9,606 ground truth bounding boxes.

Housekeeping is a fundamental component of site safety management. Good housekeeping refers to maintaining a neat and orderly working environment, ensuring the walkway is free from obstruction and properly storing materials and equipment. These good practices improve construction safety and productivity, and they are a reflection of good safety culture. However, poor housekeeping is a common challenge in construction sites. To ensure good housekeeping, supervisors will typically perform on-site manual inspections. However, these inspections are costly and inefficient, and inspectors can be inconsistent in defining good and poor housekeeping. Computer vision approaches can help resolve some of these problems in an automatic manner. Specifically, image classification methods can identify good and poor housekeeping images in video streams supplied by inspection robots or drones. Furthermore, the computer vision system can alert relevant supervisors when it detects poor housekeeping, and the supervisors can tidy up the identified locations. Hence, this study aims to develop a supervised deep learning model to classify good and poor housekeeping images. This study utilized a state-of-the-art vision transformer-based backbone model and tuned for housekeeping images. The study shows that, unlike conventional image classification for general scene images, housekeeping images are highly varied. Nevertheless, despite the challenges, the computer vision models developed in this study achieved satisfactory accuracy for identifying good and poor housekeeping images.

Deep learning-based detection of construction workers wearing or not wearing safety helmets has gained popularity in both research and commercial domains. However, most of the datasets used for training have been created without considering the contextual consistency of construction sites. To address this issue, this study compiled a new SCSI (Safety Construction Site Image) dataset that is suitable for construction environments. The dataset consists of 15,551 images distinguished by two classes, “person with helmet” and “person without helmet,” for similar construction features such as tunnels, commercial buildings, earth works, and others. For validation, YOLOv5s and YOLOv8n were used to confirm the performance of different combinations of datasets with SHEL5K. As a result of training, our dataset produced better results than the SHEL5K dataset in the various construction environments noted above. These results serve to emphasize the importance of creating datasets with consistent image contexts.

This study proposes a method to monitor workers’ personal protective equipment, particularly hardhats, at construction sites with a single query image. Prior studies on this subject required a large dataset for training, which can be time-consuming and labor-intensive since many of them utilize object detection models based on convolutional neural networks (CNNs). The proposed approach involves two steps. First, workers are detected, and their body key points are extracted through human-pose recognition used to crop head images. Second, the cropped head images are compared to a hardhat image using a Siamese network to determine whether a worker complied with the hardhat requirement. The proposed one-shot method showed 69.25% F1-score when validated against 2,000 photos of each worker wearing and not wearing a hardhat. The result indicates that the proposed approach has the potential to reduce the effort required for dataset construction while maintaining performance.

Multiple studies have investigated the use of computer vision to enhance construction workers' safety by detecting personal protective equipment (PPE). However, implementing smart and automated PPE detection in near real time in real practices is still a significant challenge. The performance of PPE detection (e.g., accuracy) in near real-time implementations (i.e., time efficiency) has not been adequately studied to date. Thus, this study proposed an edge computing-based method for detecting PPE gloves in near real time, which can enhance workers' safety and protect data privacy. This study used transfer learning methods to monitor PPE compliance and edge computing to improve time efficiency and protect data privacy. Both edge computing-based and cloud computing-based methods were examined and compared pertaining to time efficiency. The results demonstrated how the developed edge computing-based method can improve safety glove detection in a more time-efficient manner while also maintaining data privacy.

The construction industry is a high-risk environment with numerous potential hazards, emphasizing the need for effective hazard recognition to prevent accidents and injuries. However, the cognitive processes underlying hazard recognition ability are not well understood. Eye-tracking technology can offer valuable insights into the cognitive process, and this study investigates differences in higher and lower hazard recognition ability using eye-tracking technology in virtual reality scenarios. Participants were instructed to report hazards immediately when identified. Results indicate that individuals with lower hazard recognition ability had longer perception or comprehension times in the case of successful identification. These findings underscore the importance of understanding eye movement data to develop effective hazard recognition training programs to improve workers’ ability to identify and respond to potential hazards at jobsites.

Falls from roof (FFR) is a primary concern for most construction operators given the risk associated with such accidents in the US. FFR accounted for about 30% of total fall cases from 2000 to 2020. Therefore, it is paramount to propose and develop systems to help reduce the occurrence of FFR accidents. The main objectives of this study are to (1) identify and determine the causal relationship between hazards and risk factors associated with FFR; and (2) propose an integrated real-time monitoring framework to mitigate these hazards. To achieve these objectives, a literature review and archival analysis were conducted to investigate the root causes of FFR accidents. This investigation led to the identification of the types of falls, key risk factors, and causal relationships. Next, different digital technologies applications were evaluated based on their ability to mitigate the risk factors, and the most effective sensor technologies for real-time monitoring were considered. The primary sensors for the proposed monitoring framework include computer vision, inertial measurement unit (IMU) sensors, and insole sensors. Finally, a protocol for integrating these sensors is proposed. The present study revealed that the lack of PPE use, loss of balance, unsafe behavior, and manual monitoring are the primary root causes of FFR. The proposed framework has the potential to address hazards associated with FFR. Therefore, the present study contributes to fall prevention management research by developing a framework and protocol for enhancing hazard identification and accident prevention.

Forearm- and hand-related musculoskeletal disorders are increasingly becoming a concern in the construction industry. Emerging solutions such as exoskeletons could play a significant role in reducing the risks associated with exerting forearm muscles. Exoskeletons are designed to reduce musculoskeletal disorders (MSDs) in occupational settings by providing support and reducing the strain on the body. However, there are limited studies focused on assessing the impact of exoskeletons designed to target forearm-related disorders. In response to this need, the present study evaluates the effectiveness of a semi-active glove exoskeleton for a simulated drilling task. The evaluation was conducted through a randomized controlled trial with 10 participants, who performed a drilling task with and without the exoskeleton. The participants' muscle activity was measured during both conditions. The results of the study show that the exoskeleton reduced muscle activity in the forearm, including the flexor digitorium superficialis (5.3%), extensor carpi radialis (5.7%), and flexor carpi ulnaris (0.5%). Additionally, the participants perceived the exoskeleton significantly reduced strains in the hand (36%; p-value = 0.016). The findings of this study suggest that forearm exoskeletons have the potential to reduce the risk of MSDs among construction workers during repetitive tasks. Although the reduction in muscle activation was marginal, the results provide impetus for future studies at the intersection of forearm MSD risk reduction and exoskeletons in the construction industry.

Limited skilled labor has been one of the greatest challenges facing the construction industry, especially after COVID. In 2019, around 80% of contractors in the US reported problems in hiring enough skilled workers. Labor retention has been one of the major contributors to skilled labor shortages in construction. The goal of this paper is to study the impact of key construction industry characteristics and economic conditions on skilled labor shortages. This goal is achieved by simulating the flow of skilled workers throughout the construction industry. To this end, a system dynamics (SD) model was developed to represent skilled labor and analyze factors that impact their retention in the industry. After various validity and behavioral tests, the results of the SD simulation were shown to be consistent with the actual collected dataset. This proves the model’s ability to portray the current status of the construction labor market, and thus test various scenarios including economic policies and labor retention regulations on the construction labor market level. This assists industry practitioners and government agencies gain insights into the reasons behind the current shortage of skilled labor in the construction industry and the economic conditions that impact the labor market dynamics in the industry.

While many causes have been identified for worker safety violations, no studies empirically explored the social contagion effect of coworkers’ safety violations on worker safety performance. To fill this gap, this study empirically examined the changes in the workers’ safety compliance and risk-taking behavior in the presence of coworkers who adopted safety violations. The extended reality (XR) environment simulating a roofing task was used to monitor participants’ risk-taking behavior. The data regarding safety compliance and susceptibility to peer pressure were collected using subjective measures. The findings indicated that the extent to which one is susceptible to peer pressure influences the social contagion effect of safety violations and leads to compliance with safety rules at a surface level. This study calls for safety professionals and managers to recognize the differences between deep and surface compliance and the significant effect of peer pressure and social contagion on their implications for safety.

The unsafe behavior of the construction workforce is one of the leading causes of accidents in the construction industry. It has been shown that construction laborers perceive the safety behavior of their co-workers and adjust their personal safety attitudes accordingly. Thus, there has been an increasing attention to studying the impacts of the social influence and norms on the safety behavior of construction workers. While existing studies examined the factors that influence the safety behavior of the construction workforce, there is little-to-no research that specifically focused on modeling the role of social influence on the safety behavior of the workers. This paper addresses this knowledge gap by modeling construction sites as a social interaction environment, where the construction workers often adopt the prevailing behavior of their neighbors. More specifically, this paper used agent-based modeling to simulate the safety behavior of the construction workforce based on a two-dimensional, two-state cellular automaton. Four different social influence rules that could impact the safety behavior of construction workers were considered and compared in the developed simulation model. The results showed that modeling the process of conforming to the safety behavior of the neighboring workers could lead to the formation of clusters of construction laborers having identical safety behavior. The paper also identified the most representative social rule that influences the safety behavior of construction workers. This paper adds to the body of knowledge by enabling contractors better understand the safety behavior of their workers and ultimately improve the safety performance in projects.

Surface miners are exposed to a variety of hazards due to the dynamic working environment. Lack of proficiency in job skills and unfamiliarity with workplaces make new and inexperienced miners (NIMs) more vulnerable to many accidents. The accidents of NIMs are outcomes of miner demographic characteristics, miner activities interacting with the various workplace environmental influences. In order to reduce and prevent injuries among NIMs, this study aims to investigate the characteristics that can influence the accident of NIMs. A logistic regression model was employed to identify the hazard factors from historical accident data. The results show that the miner’s age, occupations, time of day, types of mining, mine machines, miner activities, and nature of injury are the important impact factors for the NIMs’ accidents. The findings of this study can help develop effective and targeted safety precautions and training for NIMs in the surface mining industry.

Workers are required to efficiently distribute their limited attentional resources to the surrounding environment to maintain their situational awareness at hazardous jobsites. While previous studies have argued that the hazard identification performance of workers is significantly affected by their attentional allocation, no study empirically investigated the spatial attentional distribution for a successful search strategy. To address this gap, this study monitored 30 workers’ visual search patterns within immersive 360° construction scenarios using eye-tracking technology. Incorporating a computer vision technique, a new spatial attention metric regarding effective search strategy was developed based on workers’ visual search data. The results indicated that workers tended to distribute their attention to the narrow area limited to their field of view, and often, around 80% of 360° surroundings remained unattended, including behind and overhead areas. These findings highlighted the importance of designing a proper training program to improve workers’ search efficiency and hazard identification performance.

Severe heat waves pose significant health risks to individuals. Under extreme heat conditions, construction workers are susceptible to dual threats from heat stress, which not only reduce their productivity but also have severe implications for their health. Simply raising awareness or monitoring environmental conditions may not suffice to protect individuals from the dangers of heat stress. Given the significance of individual variations in heat tolerance, this research aims to explore heat stress physiological smart monitoring systems, with a particular focus on wearable devices and computer vision (CV) technologies. A systematic literature review was conducted. After evaluating monitoring technologies and their respective strengths and weaknesses, this study proposes the potential for a proactive approach to monitoring thermal stress in construction workers. This approach, which involves using both wearable monitoring devices and CV technology, is suggested as a viable solution to proactively prevent heat stress incidents that could jeopardize human life.

This study aims to investigate the complex interplay between indoor environmental quality (IEQ) factors and their effects on workers’ health, comfort, and productivity. Specifically, the study will identify the optimal range of IEQ conditions that can prevent or minimize sick building syndrome (SBS) and promote healthy buildings while maximizing comfort and productivity. For that, the full study will utilize a longitudinal, naturalistic real-world approach, combining granular data collection of perceived occupants’ health and productivity states using ecological momentary assessments (EMAs) every 1 h, objective continuous heart rate data, and IEQ measurements in office spaces. By integrating physiological data into the study of SBS, the study aims to provide valuable insights into the long-term effects of IEQ on human health and productivity. Ultimately, the findings of this study can inform the development of effective strategies to improve IEQ and prevent SBS in office buildings, leading to healthier and more productive work environments.

Fatigue is a common problem in the construction industry due to the labor-intensive, repetitive, and physically demanding nature of the activities. Most construction projects are carried out in the summer, when the temperature typically ranges from 90^oF to 110^oF, exposing workers to the danger of heightened fatigue and heat exhaustion. Although studies have shown that fatigue and heat have a detrimental effect on the safety of workers, few studies have assessed and predicted workers’ fatigue levels in hot conditions using physiological metrics. Thus, the need for studies focused on developing predictive solutions in this domain. In this study, wearable sensing devices such as electromyography, heart rate, and heart rate variability sensors were used to track the fatigue levels of eight individuals in real time as they performed repetitive tasks in hot conditions (95^oF). In addition, participants observed fatigue was captured during the repetitive task using a step-observational approach. Seventeen features were extracted from the physiological data to train machine learning classifiers such as random forest classifiers, support vector machines, and KNN. The model developed from this study has the capacity to predict workers’ perception of fatigue in hot conditions and could be further integrated into wearable devices that will be used by construction workers and professionals to monitor their fatigue levels in real time.

While researchers have used various off-the-shelf physiological sensors and prevalent machine learning (ML) algorithms to objectively assess construction workers’ health status, there remain specific challenges for consistent and accurate health monitoring on the jobsite. The existing physiological-based data-driven frameworks for predicting workers’ health status in the field are not robust to the distribution shift of physiological signals and face challenges in stability, reliability, and accuracy. To overcome these issues, this paper proposes using an ensemble learning technique implemented on a support vector machine (SVM) with the Adaptive Boosting (AdaBoost) algorithm to develop a resilient predictive performance of the data-driven framework. To examine the performance of the framework, physiological signals were collected from 10 subjects performing material handling tasks with varying levels of physical fatigue. The proposed framework predicted the physical fatigue level with over 88% accuracy, better than single machine learning classifiers. This study has significant implications for improving the accuracy and stability of physiological-sensing-based health monitoring.

Safety in high-rise construction remains an issue that raises concerns in the industry across the globe. Since the industry is labor-intensive, construction workers play a vital role in enhancing the safety of construction sites. Yet, existing training programs are increasingly being criticized for the lack of trainees’ engagement and their emerging sense of exclusion, leading to wide advocacy for the incorporation of VR technology in safety training. Accordingly, this research aims to develop VR-based safety training for the construction of the high-rise buildings using a novel training methodology that is based on the constructivist experiential learning and behaviorist learning theories. The results of this pilot study revealed that the framework has a significant potential in enhancing the learning outcomes of the trainees. However, considerable attention should be given to overcoming controller lags, the provision of personalized feedback, the substitution of AI-based to human-based feedback voice-overs, and the provision of a pre-training theoretical session to further enhance the learning outcomes of the trainees.

Equipment pieces account for major aspects in construction projects, mainly safety, efficiency, time, and cost. Therefore, appropriate planning for their operations in advance of construction execution is of high criticality for project success. Despite the vast existing literature, there is lack of research on construction equipment, especially in regard to material supplies and their planning related to balancing safety and efficiency on complex sites that deal with a high number of interactive equipment pieces. Past research mostly has focused on one aspect only without sufficient consideration of the other. This study aims to develop a generic platform for managing daily equipment operations on construction sites by deploying an agent-based simulation model. During its attempt to improve time and efficiency aspects by proper distribution of material service requests on equipment units and efficient path planning and replanning for each equipment operation, the developed model respects the safety parameter by detecting and preventing any collisions between operating equipment pieces. Validation of the developed platform on an actual construction site with multiple equipment units has proved its significant capabilities in enhancing efficiency and time factors.

The Occupational Safety and Health Administration (OSHA) started requiring businesses to report severe work-related injuries involving an amputation, hospitalization, or the loss of an eye on January 1, 2015. These severe injury reports contain valuable information that can help employers to identify and eliminate serious workplace hazards. Several studies have analyzed this data source to gain insights into severe injuries. However, they have primarily focused on descriptive analyses or reporting of the accidents without investigating the associations between the cause and consequence variables. This study proposes a novel application of association rule mining to the OSHA database to extract associations between a severe accident event, source, injury nature, and affected body parts. The extracted association rules can help employers devise new solutions to eliminate hazards and safeguard employees from the same injuries in the future. Records from the highway, street, and bridge construction are selected to demonstrate the proposed approach.

Construction work exposes workers to high levels of heat, which poses significant risks to their occupational health. In order to mitigate these risks, it is crucial to have a thorough understanding of the heat exposure that workers face. This study proposes a novel framework that leverages the combination of UAV photogrammetry and microclimate simulation to address the previously overlooked risks associated with heat radiation at construction sites. A case study was conducted in College Station, TX, in July 2022, to validate the effectiveness of the proposed framework. The results emphasize the significance of monitoring site-scale radiant heat-induced risks, while conventional indicators like air temperature and heat index do not fully capture the complexities and hazards posed by heat radiation in construction environments.

Object detection has been widely applied for construction safety management, especially personal protective equipment (PPE) detection. Though the existing PPE detection models trained on conventional datasets have achieved excellent results, their performance dramatically declines in extreme construction conditions. A robust detection model NST-YOLOv5 is developed by combining the neural style transfer (NST) and YOLOv5 technologies. Five extreme conditions are considered and simulated via the NST module to endow the detection model with excellent robustness, including low light, intense light, sand dust, fog, and rain. Experiments show that the NST has great potential as a tool for extreme data synthesis since it is better at simulating extreme conditions than other traditional image processing algorithms and helps the NST-YOLOv5 achieve 0.141 and 0.083 mAP_05:95 improvements in synthesized and real-world extreme data. This study provides a new feasible way to obtain a more robust detection model for extreme construction conditions.

Electric vehicles (EVs) have been widely adopted with the expectation of reducing CO₂ emissions. However, the lack of public fast charging stations has hindered the growth of EVs. Despite extensive research on optimal installation of EV charging stations (EVCS), a decision model for a large scale and diverse spatial conditions has been still lacking. This research intends to explore a deep reinforcement learning model using deep Q-network (DQN) algorithms and test the model for optimal planning of fast EVCS at a large scale. The DQN model considers geographic (e.g., building footprints, street network), economic (e.g., capacity of charging station), and environmental (e.g., solar energy) perspectives. The learning model identifies the energy balance between electricity generation and consumption and investigates spatial patterns nearby potential charging stations. This study can aid in decision-making for suitable EVCS sites with advancing the microgrid approach-based infrastructure systems, ultimately enhancing urban sustainability considering vehicle-to-building integration.

The United States maintains about 2 million miles of natural gas pipelines, among which 321,000 miles are oil and gas transmission and gathering pipelines. Most cast-iron pipelines have been in service for more than 100 years. Pipeline failure caused by aging, corrosion, cracks, and damages may result in irreparable societal, economic, and environmental consequences. In this paper, we develop a machine learning-based risk model to predict the failure of pipelines. We train and test the risk model based on the historical data from a report covering 50 years of spillage data in European cross-country oil pipelines. Various factors such as service type, pipeline age, and leak detection are considered in the model. The machine learning-based risk model will provide the pipeline operators with an efficient and effective way to predict the pipeline failure type and guide the pipeline rehabilitation and maintenance procedures. The risk model can also be used to support risk assessment based on the ASME code for pipeline integrity management, help extend the service life of the pipeline, and bring benefits to the environment and the economy.

The potential energy loss of aging buildings traps building owners in a cycle of underfunding operations and overpaying maintenance costs. Energy auditors intending to generate an energy model of a target building for performance assessment may struggle to obtain accurate results as the spatial distribution of temperatures is not considered when calculating the U-value of the building envelope. This paper proposes a pixel-level method based on infrared thermography (IRT) that considers two-dimensional (2D) spatial temperature distributions of the outdoor and indoor surfaces of the target wall to generate a 2D U-value map of the wall. The result supports that the proposed method can better reflect the actual thermal insulation performance of the target wall compared to the current IRT-based methods that use a single-point room temperature as input.

Numerical analyses can aid design exploration, but there are several computational approaches available to consider design options. These range from “brute-force” search to optimization. However, the implementation of optimization can be challenging for the complex, time-intensive analyses required to assess seismic performance. In response to this challenge, this study tests several optimization strategies for the direct displacement-based design of a lateral force-resisting system (LFRS) using mass timber panels with U-shaped flexural plates (UFPs) and post-tensioning high-strength steel rods. The study compares two approaches: (1) a brute-force sampling of designs and data filtering to determine acceptable solutions; and (2) various automated optimization algorithms. The differential evolution algorithm was found to be the most efficient and robust approach, saving 90% of computational cost compared to brute-force sampling while producing comparable solutions. However, every optimization formulation did not return best range of design options, often requiring reformulation or hyperparameter tuning to ensure effectiveness.

Pavement performance models are a handy tool for road management agencies, allowing them to estimate pavement condition and designate the optimal corrective measure for road maintenance. Recent research has presented favorable results when using machine and deep learning techniques in pavement performance modeling. This research proposes a methodology for developing first-phase performance models for urban pavements, managed at the network level, for short-term condition prediction using machine and deep learning techniques in their construction. Specifically, random forest regression (RFR), support vector regression (SVR), gradient boosting regression (GBR), artificial neural networks (ANN), and recurrent neural networks (RNN) are used, and models are built to predict the Chilean urban pavement condition index (UPCI) with these algorithms, using a synthetic database developed in-house for the simulations. As for the obtained results, the iterations offer favorable results for short-term prediction, getting low average prediction errors (0.4% for the GBR algorithm) and RMSE results close to zero (0.063 for the GBR algorithm), which is relevant for each of the alternatives mentioned above, as it places them as recommendable tools for predicting urban pavement performance.

Engineering drawings from various domains, for example, the construction industry, exhibit the geometry of the contained components in certain scales. The drawing’s real-world scale is not necessarily stored as meta information with the drawing file. This information is only implicitly contained by dimension lines. This paper presents a multistage pipeline to infer the true scale using deep learning-based methods and logical reasoning. By using the state-of-the-art object detection method YOLOv7 and the robust optical character recognition model EasyOCR, the proposed approach localizes each dimension line and interprets their respective length. The global scale of the drawing is determined by a voting scheme resulting in the most likely pixel-resolution. The method is tested on bridge construction drawings and shows promising results in all stages of the pipeline. The authors plan to publish the trained model weights together with the source code.

Environmental concerns over embodied carbon—which is generated during the extraction, transportation, manufacturing, construction, and disposal of building materials—have been increasing as the industry shifts to renewable energy and grid decarbonization efforts prevail. Commercial buildings, a rapidly growing sector and a major source of embodied carbon, can contribute immensely to the national climate goals by transitioning into a circular economy (CE). Nevertheless, embodied carbon research is rather dispersed, with sparse data on the actual impact of different CE strategies and how they scale on nationwide commercial building stocks. To address this research need, the goal of this paper is to provide policymakers with a conceptual model that depicts the potential of CE strategy portfolios on embodied carbon reduction of commercial building stocks. Using US commercial buildings data from the Energy Information Administration, the authors (1) developed a systems dynamics model to conceptualize and serve as a baseline for calculating existing embodied emissions; and (2) evaluated the influence of various policy packages in terms of their overall emissions reduction potential over a planning horizon between the years 2022 and 2050. Findings of the study highlight the effectiveness of early design and construction CE interventions as compared to end-of-life strategies such as recycling, as well as traditional and business-as-usual approaches. Ultimately, results of the developed model can aid decision-makers to create multiple “what-if” scenarios for their policies, in addition to capitalizing on the most effective strategies for narrowing material loops and curbing embodied carbon emissions.

The effective control of heating, ventilation, and air conditioning (HVAC) systems can reduce peak energy demand and balance energy usage throughout the day. Co-optimizing energy cost and comfort is one of the main paradigms in HVAC systems’ control. This paper presents a feasibility study of a human-centered co-optimization-based controller. Two methods were utilized to factor in comfort in optimization formulations including deviation from a temperature setpoint and the integration of personalized thermal comfort models. We compared the effectiveness of these two approaches in achieving comfort during peak times by hypothesizing that the integration of personalized comfort measures will result in an improved comfort experience for users during demand response. The controllers’ performance was evaluated using a simulated benchmark small office and various combinations of comfort profiles for single occupancy scenarios. Our results indicate that integrating personalized comfort profiles into an optimization-based controller has the potential to improve users’ comfort experience by considerably decreasing the variance of comfort experience compared to the conventional controller.

Safety planning is currently a manual and labor-intensive task. As a response, there has been active research on automated safety planning approaches. With the emergence of digital twins and the increased temporal resolution of updates on the status of a construction site, automated approaches for safety analysis will be necessary to exploit this increasing quantity of data. Nevertheless, creating a digital representation of the construction site and performing safety planning is computationally complex and intensive. It is not feasible to re-compute the safety assessment of the entire construction situation from scratch for every change that occurs. This work presents a novel approach to capture incremental localized changes happening in the construction situation, that is, when one or more elements are constructed, but the rest of the site remains unchanged. The changes are captured in a graph-based structure where the affected safety elements can efficiently and rapidly be extracted, facilitating a targeted re-computation. This work also demonstrates other benefits of the graph, such as improved decision support for construction planning.

Each year, exposure to radon causes more than 20,000 lung cancer deaths in the United States. People can be exposed to radon as it enters buildings through cracks. The presence of radon in educational buildings can have a negative impact on students. Therefore, it is important to monitor radon levels in educational facilities. To accomplish this goal, a digital twin model of two offices was created to track indoor radon levels and provide suggestions for reducing them. Wireless sensors were used to measure the radon levels. Autodesk Revit was utilized to create a building information model, while Dynamo visual programming was employed to integrate the sensor data. With the assistance of Power BI, a dashboard was created to present the 3D model and visualize radon levels over time. Additionally, the dashboard sends alerts if the radon levels exceed the predetermined thresholds. By implementing this proposed approach, radon levels in the monitored rooms are expected to consistently remain below the established thresholds, thereby reducing students’ risk of lung cancer.

Pipelines are the critical component of energy infrastructure for safely transporting large volumes of oil and hazardous materials over long distances. However, many incidents occurred over the years, leading to shutdowns, disruptions, and economic losses. The Pipeline and Hazardous Materials Safety Administration (PHMSA) maintains those incidents’ reports, constituting a valuable resource for better understanding the underlying associations between those incidents’ causes and shutdown durations. Most previous studies investigating incident databases are focused on bivariate statistical analyses between cause and effect variables, providing little to no insight into their associations. Also, little research has examined the level of disruptions caused by an incident. This study proposes a novel application of association rule mining to the PHMSA database to extract associations and insights between causal factors, background factors (e.g., materials), and their effects, particularly shutdown duration in this study. The results are expected to help improve pipeline infrastructure’s planning, design, and operations.

Pipeline infrastructure is essential to the economy of society and the life quality of the residents. Due to aging as time, corrosion, cracks, or damages induced by earth movements might occur in aging cast-iron pipelines, resulting in a leak or explosion and causing potential losses of lives and properties. Trenchless technology such as cured-in-place pipe (CIPP) structural liner is an effective way to rehabilitate the aging cast-iron pipelines. As an affordable and flexible trenchless alternative, epoxy resin is a widely used CIPP liner material for preventing leakage and increasing the service of cast-iron pipelines. To investigate the effect of epoxy resin liner on underground pipeline maintenance and rehabilitation, we conduct three-dimensional (3D) finite element analysis (FEA) to simulate the mechanical properties of CIPP liner for underground pipeline rehabilitation. A parametric study is performed to identify the effect of factors such as liner thickness, buried depth, and in-pipe pressure on the mechanical performance of the CIPP liner. The numerical analysis results presented in this study can be used to optimize the CIPP liner design and guide the trenchless construction and maintenance procedures for pipeline infrastructure.

Uncertainty of precise and accurate locations of water pipeline damages increases repair and maintenance time and cost and causes a long-term and broad supply interruption. As underground pipeline failures are increasing, traditional maintenance methods relying on two-dimensional drawings are facing technical limitations and conceptual obsolescence. In this respect, this research aims to support effective underground pipeline maintenance by providing an information platform integrating three-dimensional pipeline model in BIM with geographic information data in GIS. The platform provides timely access to visualized and detailed information, including pipeline networks, materials, locations, geometries, and built years of pipelines, guided by a proposed enhanced entity relationship diagram. The platform was applied to a case study in Gansu, China. The usability and user acceptance were then evaluated with surveys based on the technology acceptance model. The results indicate that the proposed platform can enhance state-of-the-practice across perceived usefulness, perceived ease of use, and acceptance.

Hydrogen, recognized as a promising clean energy for the 21st century, has experienced significant growth in demand. Government authorities and private sectors are facing challenges in project planning and execution, particularly with respect to conceptual capital cost. Pipeline cost is contingent on factors such as route, terrain, permits, and regulations, including labor, material, right of way, and other associated facilities, such as compressor stations and mainline valve stations. However, generating a detailed estimate during the planning phase is impractical. This study aims to develop an automatic hydrogen pipeline conceptual estimate system utilizing geographic information systems (GIS). Conceptual costs are determined based on pipeline type, design specifications reflective of community density, area, terrain, right of way acquisitions, environmental regulations, permitting, and owners’ contingency. The proposed system will provide planning authorities with a cost-effective and user-friendly platform for estimating conceptual costs, ultimately contributing to the rapid development of the hydrogen energy industry.

Spatial design greatly influences fire evacuation in buildings. Although the existing research has comprehensively assessed building spatial design in terms of building circulation and overlooked evaluating building evacuation circulation and how building spatial design impacts occupants’ cognitive processes during a fire emergency. Therefore, this paper proposes a framework to investigate the effectiveness of building evacuation impacted by building spatial design by integrating building information model (BIM), fire simulation, and virtual reality (VR) technology. A case study is carried out to evaluate building spatial design in terms of building circulation and emergency signage. Preliminary results demonstrated that most participants preferred to select the path with the shortest safety egress distance, fewer turns, and associated with less accessible spaces. Moreover, insufficient installation of signs will lead a few participants to select the longer evacuation path and waste much effort in searching for helpful information on evacuation-irrelevant objects, which reduces fire evacuation efficiency.

Thermal comfort plays a critical role in ensuring the health and productivity of occupants. However, it is difficult to dynamically assess the thermal comfort at heterogeneous locations because (1) conventional methods merely rely on measurements taken by thermal sensors placed at designated/fixed positions, and (2) many existing facilities still rely on 2D as-built drawings, resulting in low interoperability between the dynamic thermal condition information and the 2D plans. To address these challenges, this paper proposes a framework that integrates digital twins, IoT (Internet of Things), robotics, and indoor positioning to better support real-time thermal comfort assessment and visualization in as-built facilities. The following steps were conducted: (1) constructing a 3D virtual digital twin representation of a facility, (2) creating a smart thermal sensing unit to measure the real-time air temperature and relative humidity using IoT-enabled devices, (3) configuring an unmanned ground vehicle (UGV) (i.e., a robot) to enable the mobility of the sensing unit by autonomously navigating to different locations, and (4) developing an indoor positioning system for assisting in dynamically visualizing the thermal condition in the created 3D digital twin. The developed framework was tested in a real-world application at the authors’ institution to validate its applicability and performance. The results showed that the proposed framework has satisfying performance on real-time thermal condition assessment and visualization. Ultimately, this paper contributes to the body of knowledge by offering a novel thermal comfort assessment and visualization system through integrating digital twins, IoT, robotics, and indoor localization to facilitate smarter facility management.

Controlling and monitoring temperature in environmental chambers is very crucial for testing the thermal performance of building envelope systems. However, off-the-shelf options are costly, non-customized, and hard to maintain. Therefore, the goal of this study is to develop a cost-effective, scalable, and user-friendly temperature control module that is capable of monitoring, analyzing, and controlling the temperature in environmental chambers. The PID temperature control module was developed in Python by using an open-source library coupled with an Arduino Uno Rev 3 that interacts with a digital dimmer meant for alternating current. The system was configured to dim incandescent lights which were programmed to increase or decrease the temperature in a confined chamber. Moreover, this module was designed and validated by applying it on an environmental chamber to control and monitor its temperature, and thus the module’s design accuracy and efficiency. The developed module can help future users perform built envelope thermal performance analysis with minimum cost and least design effort.

Building occupants have different thermal comforts at the static indoor set-point temperatures due to inter- and intra-individual differences in preferences, needs, and activities. Despite advances in intelligent building management/control systems, occupants’ dissatisfaction with the indoor thermal environment is still widespread. This study developed a personalized thermal comfort model to predict individual thermal preferences in multiple occupancy. The proposed model integrates both physiological signals (e.g., heart rate, skin temperature, and skin conductance) and environmental parameters (e.g., air temperature, relative humidity, and CO₂) in the process of predicting individual thermal preferences. The collected data was trained through the proposed model using various machine learning-driven classification algorithms and sought the most reliable personalized predictive model. To evaluate the performance of the proposed model, a case study is conducted in multiple occupancies in an office building. The results demonstrate that each person has a different powerful classification model to accurately predict their thermal preferences.

Despite the growing interest in moisture behavior in buildings, building envelope designs rarely undergo assessment for hygrothermal control, especially when it comes to energy efficiency retrofits. Timber frame structures often incorporate a continuous air cavity that separates the exterior cladding from the wall sheathing. Indeed, using an air cavity increases thermal performance. To date, however, few studies have examined the impacts of air cavities on the hygrothermal performance in retrofitting projects. This study used WUFI Pro software to assess the influence of air cavities on the hygrothermal performance of existing building envelope components. We tested various ventilation rates for each wall assembly across six climates of the US. Findings show that adding an air cavity on the exterior side of the wall can reduce water accumulation during the simulation period. This element can prove especially beneficial in hot and humid climates since it increases dryness rates in the assembly.

Drainage infrastructure maintenance projects are resource-intensive undertakings at the clients’ premises. Their projects’ domain consists of multiple one-of-a-kind small projects scattered over various locations and subject to constant changes over time (emergent projects, deadline changes). Conventional project planning models relying on a sequential phase-to-phase (waterfall) methodology do not lend themselves well to plan crew allocation and job schedules in response to dynamic customers’ demands. This research introduces agile project management methods originated in the software development industry; further a practical agile project management approach is conceptualized to allocate a finite number of crews to perform routine jobs by their respective deadlines while dealing with high-priority jobs or emergency jobs, all subject to the updated constraints on crews’ availability. A drainage service case study is provided to demonstrate method application.

Wildfire damage related to residential areas causes enormous economic and social losses. Therefore, appropriate measures must be taken at the appropriate time to reduce the damage to residential infrastructure caused by the disaster. Post-wildfire assessment of a home can play an important role in understanding the overall extent of damage and developing a disaster mitigation plan for the future. Unfortunately, these assessments require extensive on-site investigation, which takes a lot of human resources and time. As a way to compensate for this, this study provides an efficient and effective methodology to perform damage assessments of housing after a wildfire using deep learning techniques to efficiently analyze unmanned aircraft systems’ (UAS) imagery of residential areas. Application of this efficient methodology reduces the need for on-site investigation and associated safety risks as well as enables decision-makers to rapidly assess a large area to determine the relative degree of damage to structures and develop cost estimates for a community for damages after a wildfire.

Traditional bridge inspection is a manually performed visual process that is time-consuming, costly, and requires significant support from equipment and resources. Recent advances in artificial intelligence (AI), represented by image-based deep learning (DL) methods, have accelerated the advances toward inspection automation enabled by robotic imaging and machine vision. Many researchers have reported that damage types and their locations can be identified in two-dimensional (2D) images with complex scenes through a DL-based semantic object detection or segmentation process. However, recent efforts in image-based structural damage assessment have not achieved a level of applicability as meaningful as those from traditional bridge inspection. We state that structural elements and damage patterns must be identified in a 3D space to approach the intelligence level of engineer-produced inspection. In this work, we adopt the data-centric principle of AI development by exploiting off-the-shelf AI models and then develop a novel 3D structural identification framework. Combining a 3D geometric learning network and a transformer-based object-detection network, the proposed DL models can output 3D maps of structural elements and structural damage, enabling truly engineering-meaningful damage assessment and mapping. By conducting numerical experiments and validation, it is expected that the proposed methodology can contribute to the vision of realizing autonomous and low-cost structural inspection in the near future.

Despite the growing body of research on pavement quality assessment in civil and infrastructure engineering, the issue of achieving rapid analysis of nondestructive structural damage remains unresolved. The application of nondestructive testing, such as ground penetrating radar (GPR) in highway engineering, provides opportunities to analyze distress hidden under pavement surfaces. However, the traditional analysis process of GPR image databases mainly relies on technicians’ experience. Furthermore, existing methods for the recognition and identification of internal damage, such as looseness and cracks, require a high technical threshold and a tremendous amount of time, leading to low efficiency and unstable accuracy of the results. To address this issue, this paper developed an automatic identification method by combining nondestructive testing and computer vision for pavement structural damage. First, the GPR images of the internal damage were calibrated by comparing them to the cores drilled onsite. Then, an object detection (OB) algorithm, called YOLOv5, was adapted to construct machine learning models. A total of 1,000 labeled images were fed into the model, which was trained and optimized to learn and identify the images’ features. The result indicated that the developed method achieved high-speed and accurate identification of the location and type of pavement structural damage. The prediction accuracy of the OB model reached 88.2%, and the GPR image dataset of 10 km pavement with 1,000 samples can be processed in 1 min. The contribution of this research facilitates the development of pavement maintenance decision systems to provide a real-time assessment method for onsite pavement structures.

Research has demonstrated the critical role of reinforced concrete (RC) columns in major built infrastructure systems such as bridges and high-rise buildings. To build upon this, this study aims to demonstrate the effectiveness of the support vector machine (SVM) learning technique in predicting the potential failure mode of laterally loaded spiral RC columns. The Pacific Earthquake Engineering Research Center’s laboratory test data of 159 spiral RC columns was used for this data-driven study, and the SVM model adopted the radial basis function kernel and optimized values of the hyperparameters cost and gamma to develop a robust SVM classification model. The model has an accuracy of 94.59% and a Cohen’s kappa coefficient of 0.92. Overall, this study has successfully demonstrated machine learning tools’ effectiveness and usability in data-driven engineering problem-solving and would contribute immensely to the domain knowledge on using technological tools to solve society’s most pressing challenges.

Although prior research has shown that diverse high-level management teams can improve organizational performance, there is limited knowledge regarding the racial diversity among executive roles in US construction companies. This study analyzed the LinkedIn profiles of over 2,800 US executives from the Engineering News Record’s Top 400 list to assess racial diversity in executive roles within large construction companies. Researchers used Python deep face recognition to identify the profile race with over 97% accuracy. Results showed that less than 21% of construction executives came from non-white communities, with women occupying less than 10% of executive positions. Statistical analysis revealed no significant differences in how racial groups achieved leadership positions. However, male leaders from underrepresented racial groups had a greater gap compared to white male leaders than female leaders from different racial groups. The study provides insights that can help early and midlevel professionals model their career journeys, and diversification in leadership can benefit companies and their customers from an industry perspective.

Automated visual inspection of infrastructure and the built environment using unmanned aerial systems (UAS) has found considerable traction among infrastructure owners and managers. These inspections rely on the limited battery and payload capacity of the UAS, and therefore optimizing the inspection path is essential for the scalable deployment of these systems. This preliminary study proposes a framework for planning UAS-based inspections by considering visual coverage as a constraint and an optimization objective function that could include path length as well as other performance metrics. The path parameters were optimized accordingly, and the influence of the hyperparameters on the path optimization results and their tradeoffs were also studied. The proposed framework lays the foundation for integrating structural engineering domain expertise into UAS-based inspection path planning.

Infrastructure maintenance projects involve activities scattered over a large geographical area (e.g., scattered road segment maintenance). Existing scheduling techniques fall short of addressing the unique challenges of the scattered nature of these projects while accounting for resources availability constraints. This paper presents a resources-based planning framework for scattered repetitive projects. The framework includes (1) adaptation of queuing theory in prioritizing and assigning resources to activities, (2) integration of multimode resources in combination with different construction methods for activities, and (3) GIS-based analysis of resources mobility. The proposed framework is applied in an agent-based simulation model using Anylogic software. The model was tested on several examples from literature where it successfully developed resources allocation plans and project schedules. The proposed framework enables efficient resources-based planning of scattered projects on a large scale as an added value for the planning of infrastructure maintenance and rehabilitation projects.

Model trees combine the strengths of two interpretable models—namely, decision trees and multiple linear regression—to handle complex nonlinear problems, making them a popular machine learning strategy in civil engineering. Analyzing the error terms between predicted and target outputs is the standard method for evaluating the model’s accuracy. However, the model does not explain the variance associated with the predicted output for a given set of inputs, which may be deemed insufficient in many civil engineering applications. This research proposes a framework that combines error propagation theory with the model tree algorithm to quantify the variance of the model predicted output and provide explainable artificial intelligence (XAI) for engineering applications. A metric based on the variance estimate is proposed to decide whether the model predicted output is acceptable. A concrete strength prediction model is presented to demonstrate the application of the proposed methodology.

Data-driven autonomous anomaly detection, empowered by advanced sensing and machine learning techniques, has the potential to significantly reduce costs associated with infrastructure maintenance. In this paper, we studied the impact of probabilistic anomaly detection in the context of a proposed deep learning-empowered framework for multivariate time series anomaly detection. In this framework, two-dimensional convolutional autoencoders are utilized to learn the patterns of normal data and detect unseen anomalous data points that fail to be reconstructed by the well-trained autoencoders. Kernel density estimation was utilized to generate a probability distribution function based on raw window-based anomaly predictions. Our assessments showed promising performance in a real-world case study of Class I railroad track anomaly detection. We conducted a comparative study that proved our hypothesis that probabilistic anomaly detection outperforms the deterministic approach of anomaly detection using direct autoencoder output. Furthermore, the impact of multivariate time series data processing methods, autoencoder architecture, and anomaly score thresholds was also investigated.

A deep learning method is proposed to act on point clouds for segmentation, which can feed the data into a built network based on an encoder-decoder architecture coupled with an improved 3D dual attention module to extract and learn features. To verify the effectiveness and feasibility of the proposed model, a tunnel point cloud dataset collected in a metro tunnel project is used. The experimental results show that the proposed model has a plausible performance with an Intersection over Union (MIoU) of 0.8597, and it outperforms other state-of-the-art methods such as PointNet and DGCNN. Overall, the proposed model shows excellent performance and provides effective and accurate results for multi-class segmentation on 3D tunnel point clouds.

Green building rating systems have grown exponentially in popularity since the 1990s due to the increasing need to minimize the negative impact of buildings on the environment. These systems provide a framework for healthy, highly efficient buildings, lowering operational costs, and greenhouse gas emissions while improving the quality of buildings for occupants. This study describes the importance of sustainable building rating systems, analyzes the trends in the adoption of the Leadership in Energy and Environmental Design (LEED) rating system, and critiques key factors that influence the adoption of sustainable building rating systems. Cross tabulations and trend analysis were completed to analyze the patterns in the adoption of the LEED rating system. One of the trends was a surge in registration of projects due to anticipated changes to the LEED rating system. The insights generated by this research are valuable resources for fellow investigators seeking updated information on LEED certification practices.

Natural disasters (e.g., winter storms, heat waves, and hurricanes) threaten residents’ energy services due to associated limited power supply or even power outages. Maintaining the desired level of residential energy services is critical. However, it is still unclear about the importance of each energy service to residents during disasters and how these services differ from normal days. In this study, we developed and distributed a large-scale survey to Texas residents to investigate the perceived importance of different energy end-uses during normal and disaster situations. The results illustrate that during disasters, residents prefer reducing energy use but keeping the continuous power supply, and the importance ratings on all energy services have significantly decreased. Based on the degree of compromise, measured as the difference in perceived importance between normal days and disasters, residential energy services are classified into two categories: critical services (i.e., HVAC, cooking, refrigerator, and freezer) and noncritical services (e.g., lighting, water heater, TV, dishwasher, and laundry). Critical services are less likely to be compromised during disasters than noncritical services. Additionally, one-way ANOVA results illustrate that race, income, education level, family size, and climate and energy beliefs influence end-use importance and compromise. Generally, these factors have larger impacts on noncritical services and vary between normal or disaster situations. In particular, higher education levels or stronger climate and energy beliefs contribute to greater compromise on energy services. The findings fill in the gaps to understand the varied importance of different energy services during disasters and inform the improved design of energy supply schemes under climate change, increasing energy service resilience.

Buildings account for 30% of global energy use and are expected to continue to increase as urbanization progresses. Among the various drivers, urban form is known to have a non-negligible effect on building energy consumption. Past studies have focused on modeling the physical obstructions and microclimate around the building. However, large-scale modeling is usually computationally prohibitive. In this paper, we aim to empirically evaluate how urban form affects energy use of residential and commercial buildings in Santa Clara, CA. We combined a large-scale building energy dataset with information on building properties and urban environment. We build various regression models and analyze the non-linearity and interaction effects between variables. Results show that the relationship between density and building energy use is non-linear and the monthly energy use intensity could drop by up to 50%, 18%, and 17% for single-family residential, multi-family residential, and commercial buildings, respectively, if we increase the density in surrounding areas. These results can assist urban infrastructure planning and policy-making professionals to make informed decisions regarding land-use and building decarbonization goals.

Of US households, 44%—about 50 million—are defined as low income. Low-income households (LIH) typically inhabit affordable homes, often the poorest-quality housing units. Compared with other properties, affordable housing is less energy efficient. Because of high affordable housing energy bills, low-income households face a disproportionately high energy burden, that is, the percentage of household income spent on utility costs. Therefore, much more attention is required to develop efficient long-term solutions to reduce the energy burden in low-income households. This study aims to investigate this issue by providing a statistical analysis of how the energy-related characteristics of affordable homes impact the energy burden in low-income households. The 2019 American Housing Survey data has been used in this study to represent the US housing stock. The results show that the average energy burden in low-income households is 11.0%, three times higher than in non-low-income households (3.5%). The results of this study can enhance the awareness of policy-makers, urban planners, and building designers toward designing and creating more energy-equitable decisions.

State departments of transportation (DOTs) face several challenges in managing their transportation asset data and maintaining the asset condition to a certain level of required serviceability. In the current practices, asset information is siloed and isolated throughout the asset life cycle, resulting in data stagnation and loss. Conversely, understanding data architecture is critical to leverage the value of asset data. Thus, this paper aims to develop a Transportation Asset Data Lifecycle Ecosystem (TADLE) that compiles the information associated with and required for each data life-cycle stage. Furthermore, this study presents a case study to validate and implement TADLE to determine the essential information and data flow associated with managing an ancillary asset throughout its life cycle by a state DOT. The developed framework will support state DOTs in advancing the data and information systems necessary for efficient transportation asset management practices to allow for data computing, analysis, and informed decision-making.

To analyze the human-building interaction, it is crucial to accurately detect the presence of occupants inside buildings. Current occupancy detection systems usually require expensive and complex sensors, while individual occupancy detection algorithms might not provide enough accuracy. This study aims to develop a cost-effective occupancy detection model using machine learning algorithms based on individual plug load data. The models are trained to detect real-time occupancy, validated by a battery-operated seat pressure sensor in an office space test bed. The test bed was a student office station at Louisiana State University, equipped with energy-monitoring wallplugs and seat pressure sensors. Two weeks of data were collected and used for training support vector machine and artificial neural network models. The developed models are also validated using the ASHRAE occupant behavior plug-in device load database in a public building located in the US. The results showed an accuracy of 98.72% in detecting occupant presence in real-time using plug load profiles.

Commercial buildings often face challenges in coordinating heating, ventilation, and air conditioning (HVAC) systems due to varying occupant preferences, resulting in thermal discomfort and energy waste. Balancing comfort and efficiency requires understanding comfort profiles and energy consumption at different temperatures while accounting for uncertain disturbances like outdoor temperatures and extra heat. Furthermore, control algorithms (e.g., model predictive control) are typically computationally expensive, limiting large-scale building applications. To address these challenges, this paper presents a robust HVAC control framework ensuring occupant comfort and energy efficiency despite external disturbances. By solving an optimization problem, the approach determines temperature setpoints that minimize energy usage while maintaining desired comfort probability. Specifically, a probabilistic certificate guarantees long-term comfort under disturbances, and a myopic method enhances computational efficiency. Tested in a 98-room real-world building, the proposed method effectively ensures comfort and energy efficiency, surpassing the baseline model.

Data-driven HVAC energy consumption modeling could play an important role in operating energy-efficient buildings. However, (1) existing data-driven approaches rely on rigid sensor networks to collect indoor physical parameter data, often deployed in an ad-hoc manner or based on engineering judgment; and (2) the impact of this practice on model performance is often poorly understood. To address this gap, this paper aims to assess the impact of sensor deployment configurations on HVAC energy consumption modeling, where a configuration is defined in terms of the number and locations of sensors and their flexibility (i.e., fixed or can change over time). Indoor temperature and humidity data were collected from an office building. Several configurations were defined and evaluated in predicting the HVAC consumption, using an XGBoost-based model. The results showed that sensor configurations could significantly impact the performance of HVAC consumption modeling, and that periodic changes in sensor locations could improve the performance of traditional rigid methods. The findings from this work could help define improved sensor configurations for enhanced HVAC management and operation.

The conventional design process of pharmaceutical clean rooms is typically based on 2D CAD drawings, which can rarely generate simulation models of temperature distribution to verify the effectiveness of HVAC systems. Such design process is likely to cause errors in the operation stage, eventually failing the design requirements. To address the design challenges in pharmaceutical projects, this paper presents a BIM-based approach to optimize HVAC design with computational fluid dynamics (CFD). By utilizing CFD to simulate the dynamic conditions of airflow in clean rooms, the effectiveness of HVAC systems can be verified. A case study of a typical clean room design is presented to demonstrate the workflow of the approach and validate its functionalities. The results of the case study have shown that CFD can successfully simulate the design intentions of indoor air quality in BIM models and suggest optimized HVAC systems for clean room design. The findings of this paper contribute to the body of knowledge on overcoming the limitations of the traditional CAD-based HVAC design process and provide valuable insights on optimizing HVAC design with BIM and CFD technologies.

Challenged by the prevalent workforce shortage, the construction industry is picking up interest in using robotic arms in construction operations, especially in the context of modular construction and prefabrication. However, the lack of systematic investigations into integrating robotic arms with mobile systems to enhance mobility and operational range has been identified as one main research gap. Stationary robotic arms have inherent limitations in their range, making mobility a critical need. To address that issue, in this paper, the authors proposed a mobile construction robotic system to facilitate their use in the automation of timber frame assembly operation. The authors simulated the system to assess the interactions and coordination among its various components, and to identify potential areas for improvement. This study showcased the effectiveness of the new system design in improving the timber construction automation process and reveals its potential for further exploration in the realm of mobile construction robotics.

This study presents an autoregressive method for forecasting construction labor productivity metrics. Productivity is an essential parameter to monitor progress in construction projects as it can determine whether the project succeeds or fails in terms of cost and time. However, collecting productivity data, or data correlating with productivity, takes time and effort. Furthermore, the collected productivity data offers limited insights into the current and past performance of the construction activities, which can be valuable to project managers but are often too late to act on due to the transitory nature of construction projects. To increase the amount of information available for decision-makers and analyses, this paper investigates the possibility of using forecasting methods to estimate future values of direct work, indirect work, and waste. Four models are developed and evaluated on a dataset collected on Danish construction sites. Being able to forecast these metrics will add value for decision makers.

Distributed solar generation (DSG), such as residential photo-voltaic (PV) solar panels, offers many benefits to consumers and can improve the sustainability and resilience of the electric power infrastructure. As such, many local and national authorities implemented policy incentives, such as rebates and loans with reduced interests, to promote the adoption of DSG. However, the increasing penetration of DSG is creating new obstacles for system operators due to the uncertainty in forecasting electrical power demand. There is a need to investigate how incentives affect DSG adoption. To achieve that goal, a complex system-of-systems (SoS) model is developed using agent-based modeling (ABM). The model can simulate the complex relationships between utilities and generating companies in wholesale power markets where consumers can make an economic decision to install DSG systems. The model is tested using a case study based on a modified IEEE 6-bus system. The findings highlight the need to carefully consider the rippling effect of incentives on the grid to ensure sustainable DSG diffusion as demonstrated using ABM.

As structures age, the continuous surveying of defects to structural elements becomes a central part of the life cycle. Automated and frequent inspections could allow for an enriched defect history for faster decisions and reduce the time until problems are first reported. This requires recording the defect on an autonomous platform and automatically classifying areas of interest. We propose and test a prototype that uses a defect segmentation network for classifying defect instances deployed on a Spot mobile robot for autonomous maintenance missions. The robot maps its surroundings using a continuous laser scan and annotates the resulting environment mesh using the defect segmentation network. We test the feasibility of entirely autonomous, repeat missions and the possible use cases of such a system.

The service environments (e.g., climatic environments, water environments) of steel bridges have severe impacts on the performance of their protective coatings, thus affecting the safety, integrity, and longevity of the steel bridge elements. By leveraging existing data on both the bridge-related features (e.g., bridge age, type of service under bridge) and environmental features (e.g., chloride, moisture, atmospheric pollutants, temperature), we can develop a data-driven understanding of bridge coating deterioration patterns as well as how such patterns are impacted by the service environments of bridges. This paper focuses on presenting a multilayer perceptron (MLP)-based artificial neural network (ANN) model that predicts coating conditions of the girder elements of steel bridges in Florida. The model integrates both the bridge-related features and environmental features. It uses data from multiple sources, including Florida Department of Transportation and national environmental agencies [e.g., National Atmospheric Deposition Program (NADP), Florida Automated Weather Network (FAWN), United States Environmental Protection Agency (USEPA)]. The results show that the proposed model can effectively predict the coating conditions of steel bridges with a minimum mean absolute error (MAE) of 0.0807. As compared to the MAE of the model that only accounts for bridge-related features, the MAE of the proposed model was significantly reduced by approximately 11.36%. Thus, by offering effective predictions on steel bridge coating performance, the proposed model has the potential to allow bridge owners to better maintain, repair, and/or replace steel bridge coating systems with less effort, time, cost, and risk.

Urban clusters are areas with high building and infrastructure concentration and limited green and natural landscapes. Since buildings, bridges, roads, and other infrastructure absorb heat from the sun and re-emit it to the atmosphere, urban areas are becoming “islands” that have higher temperature compared to rural surrounding areas. This phenomenon is known as the urban heat island (UHI) effect, which is associated with global warming and the heat generated from increased human urban activities. While previous studies examined potential factors that could affect heat-related challenges, little-to-no research prioritized these factors using a data-driven approach. To address this knowledge gap, this paper determined the most critical factors that contribute to UHIs using feature selection- and machine learning-based methods. First, a dataset was collected and developed for all the census tracts in the state of New Jersey in relation to 14 factors that could impact the UHI effect. These factors are considered as independent variables, while the average day UHI intensity for each census tract is considered as the dependent variable. Second, feature engineering algorithms were implemented to extract or identify the prominent factors. Third, the most critical factors were ranked and prioritized. The results compared the prioritization and ranking of the different factors based on multiple feature engineering algorithms so that robust conclusions could be obtained. This study serves as a useful reference for decision-makers involved in addressing the challenges resulting from UHIs by offering a better understanding of the most important factors that need to be better controlled or managed.

Multi-layered and cross-referred disaster resilience regulations and building codes have been a critical obstacle to holistically evaluating long-term resilience of a residential building according to various disaster scenarios. This challenge is caused and worsen by the primary knowledge gap, which is the lack of integrated and executable regulatory rules and specifications related to disaster resilience. To tackle this issue, the authors have identified fundamental resilience rules and built a disaster resilience validation framework that incorporates formalized, but implementable resilience requirements. To accomplish this goal, we thoroughly investigated multi-layered flood-related regulations and revealed their inherent rules for validating the disaster resilience of a single-family house. We identified 42 resilience-related rules and developed them in an implementable format using the eXtensible Markup Language (XML) schema for broadly disseminating them to related domains. The rules include the flood resilience features dealing with lowest floor elevation and freeboard requirements, openings of building walls, breakaway walls for elevated buildings, and obstruction-free buildings in coastal areas. The revealed rules play a pivotal role as a foundational baseline to establish a web-based automated validation system that can assist the public to evaluate building information modelling data. The proposed approach is expected to help logically assess the resilience of buildings and consistently share their updated regulations.

Academic buildings serve important roles for college students. Maintaining a comfortable built environment can benefit the well-being of college students and promote their performance and productivity. The most common way of maintaining a comfortable built environment is still highly manual and reactive; for example, students adjust the light levels when they have already had visibility issues. Existing Internet of Things (IoT)-based efforts facilitate real-time collection and analyzing indoor environmental quality (IEQ) data. However, they failed to incorporate feedback from students. To address this problem, this paper proposes the use of artificial intelligence of things (AIoT) technologies to support academic building operations that can automatically and proactively adapt to the comfort needs of college students. The proposed framework consists of three steps: (1) collecting indoor environment data using IoT, (2) gathering direct and indirect feedback data from students, and (3) developing a transformer-based model for analyzing and predicting IEQ conditions.

Monitoring the condition of road furniture, such as traffic lights, signposts, and guardrails, is critical for road safety and driving comfort. However, due to limited resources and the vast size of road infrastructure, comprehensive monitoring can be challenging. To address this issue, this study proposes a framework that automatically identifies damaged road furniture and displays their locations on the user interface, allowing inspectors to remotely monitor road furniture conditions. The framework leverages a road furniture digital twin, containing real-time information about road infrastructure, landmarks, and GPS coordinates. The framework was tested on the NDSU campus road, successfully detecting traffic poles with real-time information displayed on the user interface, reducing the need for manual inspections, and enabling proactive maintenance to improve road safety and driving comfort. The proposed framework is expected to facilitate the real-time remote monitoring of road furniture, minimizing the time, cost, and effort required for inspections.

The increasing availability of low-cost sensors, such as smartphones and the cameras embedded in cars for driving assistance, provides new opportunities to enhance pavement health monitoring. While computer vision techniques have been successfully used to detect pavement distresses using 2D images, distresses involving depth measurements, such as rutting, remain a challenge. The objective of this study is to develop an object detection model to automatically detect instances of rutting in pavements. This study leverages pavement images and distress data collected from the United States Federal Highway Administration Long-Term Pavement Performance (LTPP) database to train a rutting detection model. Transfer learning is used to analyze the generalizability of the model. Results indicate that street-level images are highly suitable for detecting pavement rutting using computer vision algorithms. Future research is needed to quantify the severity of rutting using image processing techniques.

Municipalities and governmental agencies continue to mainly rely on responsive treatments such as road resurfacing or reconstruction which are costly. This study found that there is a knowledge gap about the long-term economic benefits of preventive treatments at the road network level. To address this knowledge gap, this study presents a long-term and life cycle economic feasibility analysis of implementing preventive treatments for a municipal highway. This paper examines three main aspects: (1) the effectiveness of preventive maintenance treatments, (2) the degree of cost reduction associated with different types of preventive treatments, and (3) the extent of preventive treatment required to achieve the desired effect. According to the analysis, the combination of preventive maintenance and rehabilitation treatments can result in a substantial decrease of 34% in overall expenses over a 30-year timeframe, regardless of the preventive types.

The construction sector is among the most hazardous workplace where workers are more likely to be at risk than in other jobsites. Generally, construction accident reports possess a wealth of empirical knowledge in the form of text summarizing related events. However, analyzing this information source typically requires the high cost of manual content analysis as accident reports are often voluminous and presented in an unstructured format. Text mining and machine learning have recently been applied to extract and leverage valuable information from accident reports. However, few studies have focused on automatically identifying accident causes from highway construction incident reports. This study aims to extract highway construction critical accident causes from a large narrative dataset obtained from the Occupational Safety and Health Administration by adopting the Latent Dirichlet Allocation algorithm. As a result of this implementation, 12 critical accident causes were identified, which were subsequently classified into five groups of accident causes: management factors, human factors, unsafe behavior, environmental factors, and material factors. This study is expected to empower organizations to rapidly analyze and obtain reliable critical highway accident causes from their accident report databases with minimum expert involvement. Managers can use such outcomes to formulate appropriate safety strategies to reduce catastrophes.

The number of fatalities involving vulnerable road users (VRUs) has reached record highs in recent years. Due to the complex and costly process of naturalistic studies, there is a lack of understanding on the behavior of road users when they interact with each other in different environments. However, advancements in virtual reality have made it possible to conduct distributed simulations, enabling multiple users to participate in a virtual re-creation of an entire system. This development has promising applications in human-factor research, especially in evaluating behavioral responses in complex environments. This paper presents a roadmap outlining major technical challenges and requirements for real-time distributed and multi-user simulation in the context of road users’ safety along with a case study of distributed multi-agent VR. This paper contributes to the field by providing feasibility benchmarks and research details on low-cost distributed technology for simulating VRU interactions.