International Conference on Transportation and Development 2024: Transportation Safety and Emerging Technologies

Work zone safety, especially amid increasing highway construction and repair, is paramount in ensuring the well-being of construction workers. Between 2003 and 2019, road construction sites saw a total of 2,103 worker fatalities, averaging 131 deaths annually. During this period, Texas had the highest number of road construction-related worker deaths. With traffic congestion intensifying and a rise in on-site hazards, there is a pressing need for workers to be adequately educated on the multitude of risks they face, from physical injuries to exposure to toxic substances. However, the efficacy of current training programs in inculcating proper safety practices remains underexplored. This study seeks to develop comprehensive training materials highlighting the myriad of hazards in work zones and the preventive measures to be adopted. An experimental highway construction worker group in Texas underwent this tailored training. Their knowledge was gauged before and after the session through a standardized questionnaire comprising 10 questions aligned with the training content. The data was rigorously analyzed, unveiling a marked elevation in the participants’ safety awareness post-training. These findings can guide construction firms and regulatory bodies in instituting training measures that heighten the safety quotient in work zones.

The mental health of construction workers is often neglected despite their overrepresentation in deaths associated with suicide and other mental health conditions. The nature of the construction sector is inherently stressful, and failure to manage individual stress levels can manifest in depression, which can eventually lead to suicide ideation and suicide. This study identifies prevalent mental health conditions within the industry, defines mental health stressors on construction worksites and determines their frequency of citation, and discusses best practices for the promotion of a healthy work environment. Through the identification of 13 factors negatively affecting their mental health, the findings highlight the significance of protecting not only the physical but also the mental health of construction personnel. The most frequently cited mental health stressors were found to be excessive work demand, work life imbalance, and poor working environments. Additionally, the study showed that the effective management of workplace mental health involves the implementation of a strategy considering the integration of primary, secondary, and tertiary interventions. The promotion of a help-seeking culture may also help reduce the stigma around mental health and diminish internal shame, specifically for male construction workers.

This study used a driving simulator to investigate driver behavior near bike lanes. The interaction between cars and bicyclists during passing maneuvers is a key factor in bike-related accidents. The study tested two scenarios: one with a road containing a bike lane and one without. Twenty-nine participants had two driving sessions, each matching one of the scenarios. The study used car and bike simulators allowing two participants to drive simultaneously on the same road network and interact live. Driver performance was evaluated by analyzing their interactions during 30-s overtaking maneuvers. Statistical analysis assessed the bike lane’s impact on driver behavior. The presence of a bicyclist led to significant changes in driver lateral movement and speed. When more space was available from the bike lane, drivers felt safer and increased speed to overtake bicycles. In conclusion, this study revealed that most drivers gradually approached bicyclists and accelerated to pass them quickly.

This research evaluated the impacts of commercial driveway and connecting street features on pedestrian and bicycle safety through crash frequency and severity analyses. Data for driveway-related crashes involving pedestrians or bicycles in Florida over a five-year period were analyzed. The study also reviewed the characteristics of commercial driveways along selected major corridors with high pedestrian and bicycle crash frequencies, which were included in the analysis. The results revealed several significant factors affecting pedestrian and bicycle crash occurrence at commercial driveways, including the number of lanes on connecting streets, driveway design features, median opening type, traffic control device type, and painted bike lanes. Significant factors affecting crash severity were also identified, including shoulder type, alcohol or drug involvement, driveway number of lanes, bike lane type, driveway throat length, and connecting street annual average daily traffic (AADT) in the crash year. These findings provide important insights on addressing pedestrian and bicycle safety issues at commercial driveways, which could be used as a reference for future access management guideline improvement.

Non-motorists using personal conveyances, like skateboards, face increased safety challenges due to reduced visibility, particularly at intersections and in low light conditions. Safety studies on non-motorists using personal conveyances rely on limited hospital datasets or controlled/naturalistic riding conditions, lacking comprehensive identification of e-scooter riding risk factors. To address this research gap, this study collected real-world traffic crash data and quantified safety risks related to key contributing factors. Utilizing the 2020−2021 Fatality Analysis Reporting System (FARS) fatal crash data, this study examined the patterns of crashes associated with non-motorists using personal conveyances at segments and intersections. The findings provide valuable insights into key risk factors that can guide stakeholders, municipalities, and campus administrators in developing effective mitigation strategies to reduce safety risks associated with non-motorists using personal conveyances. By addressing these safety concerns, personal conveyances devices can be integrated as a safe and sustainable shared mobility option in urban and campus environments.

Commercial development typically includes offices, shopping centers, gas stations, restaurants, and other uses that generate turning movements and traffic. As a result, commercial driveways and access streets connecting to major roadways are a key source of vehicular, bicycle, and pedestrian conflicts. Case studies are one of many methods for understanding the safety effects of commercial driveways. This paper includes six exploratory case studies conducted as a part of a statewide study of commercial driveway crashes in Florida. Case study applications for this safety analysis included: (1) illustrating safety issues with high levels of detail, (2) verifying research findings from high-level analysis, (3) exploring the root causes of identified safety issues, and (4) identifying potential countermeasures for specific safety issues. The case study sites represent a subsection of sites identified as high-crash commercial corridors. The results reveal several safety issues resulting from commercial driveway type and location in relation to intersections and interchange ramps, driveway alignment with median openings, driveway density, and traveler behavior (including motorists, bicyclists, and pedestrians). Although conducted in Florida, the approaches, lessons learned, and findings of this study are relevant to understanding commercial driveway safety across the US and internationally.

Road accidents pose public health concerns, incurring substantial costs. While crash barriers are recognized for preventing accidents, their impact on driver behavior in mild road curves remains understudied. This study addresses the gap by investigating crash barrier influence on vehicle speed and lateral displacement. A Class “AB” road with consecutive mild curves, one with a crash barrier and one without, underwent controlled analysis. Drones captured speeds and lateral positions. SPSS t-tests revealed significant differences (P < 0.05) in speeds and displacements at curves with and without barriers. Regression analysis showed road conditions did not significantly affect speeds; middle and exit speeds depended on entering speed and crash barrier presence. Vehicle speed at curves with barriers decreased, indicating heightened safety perception. Vehicles approached the centerline with barriers, posing head-to-head collision risks. Recommendations include widening lanes near crash barriers. In conclusion, crash barriers can have both positive and negative effects on driver behavior, requiring careful consideration.

In the study, a methodology was developed for identifying stopping sight distance (SSD) deficient vertical alignments on highways in Alabama and for assessing the safety impact of such deficiencies. The roadway geometric data were obtained from a database developed by Reiker Incorporated using their Curve Advisory Reporting Service (CARS) system. In addition to using CARS, this study used Google Earth and Google Maps to collect roadway geometric data. The geometric data from CARS were processed and exported to a computer-aided design (CAD) program to graphically measure available SSDs on vertical roadway alignments. The available SSDs were then compared with the recommended SSDs based on the posted speeds. A total of 29 roadway segments were identified that did not meet the minimum SSD requirements. The method developed in this study is expected to assist the responsible agencies in prioritizing deficient vertical alignments for upgrading to current design standard.

This study investigates the impact of COVID-19 pandemic on distraction-related motorcycle crashes in Kentucky while comparing pre-pandemic (2015−2022) and post-pandemic (2020−2022) periods. Recent eight years (2015 through 2022) of distraction-related motorcycle crash records and associated roadway information were retrieved from the Kentucky Transportation Cabinet (KYTC). Statistical tests including the chi-square test of independence, Z-test of proportions, and odds ratio (OR) were used to identify the factors affecting the severity of distraction-related motorcycle crashes pre- and post-pandemic. The chi-square test indicated that crash characteristics (e.g., manner of collision and speeding-related), at-fault driver characteristics (driver age and DUI), roadway characteristics (area type, crash location, road functional class, presence of horizontal curve, and vertical gradient), and environmental characteristics (weather and lighting conditions) were significantly associated with the severity of distraction-related motorcycle crashes pre- and post-pandemic. In addition, the presence of horizontal curve significantly affected the severity of distraction-related motorcycle crashes post-pandemic. Results from the Z-test comparing the proportions of severe crashes pre- and post-pandemic revealed that the percentage of severe distraction-related motorcycle crashes significantly increased post-pandemic. Furthermore, the proportion of severe distraction-related motorcycle crashes for very young at-fault drivers (16−20 years old), presence of horizontal curves, speeding, clear/cloudy weather, and at intersections significantly increased post-pandemic. The OR results revealed that the odds of severe distraction-related motorcycle crashes involving speeding, with presence of horizontal curves, and for very young drivers increased by 67%, 64%, and 48%, respectively, post-pandemic. Based on the study findings, removal of distraction-related road elements (e.g., billboards) at high motorcycle crash risk locations, and ensuring installations of rumble strips on edge line, could help reduce the severity of distraction-related motorcycle crashes on Kentucky roads.

Ensuring bridge safety is crucial for transportation in many countries, including the US. In fact, the ASCE gave an overall cumulative grade of C− for US infrastructure, in large part due to deteriorating bridges. Unsafe bridges within transportation routes pose significant challenges to stakeholders including traffic safety concerns, environmental hazards, and economic implications. The collapse of Minneapolis I-35W bridge highlights these issues. Our research encourages stakeholders to make informed decisions by taking into account bridge risk factors and the sensitivity of transported materials (e.g., hazardous materials, heavy equipment). This approach enables strategies like load balancing, rerouting enforcement, and prioritizing maintenance tasks to be applied effectively. We generated a bridge-health enriched geographic information system (GIS) dataset that integrates bridge information from the National Bridge Inventory (NBI) with GIS data from OpenStreetMap. We analyzed example origin-destination (O-D) pairs in Nebraska to demonstrate how different routes in Nebraska exhibit, besides differences in travel distance, varying statistics on bridge health scores. We also introduce a simulation platform for route planning and visualization currently in active development. We finally discuss the challenges we faced and opportunities ahead in combining these distinct datasets for sustainable route planning and infrastructure maintenance.

Highway work zones present significant safety concerns. This study aims to identify spatial and temporal factors that affect the number of highway work zone crashes in North Carolina (NC). The data from 2000 to 2020 were provided by the NC Department of Transportation (NCDOT). Descriptive statistics, exploratory analysis, and time series analysis were employed. It was found that crashes occurred most frequently in construction-related work zones, work zones with ongoing activities, during clear weather conditions, dry surface conditions, daytime, and right next to the work area. The time series analysis identified multiple trends and a seasonal pattern for monthly crash frequencies, which were used to develop a Holt-Winters time series crash prediction model. The model findings indicate that the crash trend and seasonality are important factors in highway work zone crash modeling along with geometric (e.g., lane width) and traffic variables (e.g., volume).

Low-cost safety improvements, road diets, and project upgrades are three typical solutions to integrate the concept of Complete Streets into pavement preservation programs of state departments of transportation (DOTs). Speed management, as a low-cost safety improvement method, could be implemented during pavement preservation to reduce crash risk due to speeding issues. The current study collected a list of preservation projects funded by Louisiana State DOT from 2018 to 2020 and conducted a project-based crash analysis to examine the change of crash rate before and after pavement preservation to evaluate the needs and feasibility of applying speed management. The before-and-after comparison and the regression analysis results revealed a significant increase of crashes after pavement preservation works started, particularly on rural collectors. Radar speed feedback signs (RSFS) were identified as a potential countermeasure to reduce crash risk during pavement preservation on rural collectors. The subsequent benefit-cost analysis results show that implementing speed management within two years after pavement preservation work starts would yield large positive returns on investment as evidenced by a benefit-cost ratio (BCR) over three. The study findings, along with the speed management technology reviewed, are expected to be useful for state DOTs in reducing crash risk related to speeding, improving work zone safety, and better integrating the concept of Complete Streets into their pavement preservation programs.

Highway safety is impacted by pavement surface characteristics, especially microtexture and macrotexture. This research proposed a new approach to consider the combined effects of pavement macrotexture and low slip friction (mainly impacted by microtexture) to analyze pavement safety performance functions (SPFs). The study investigated the impact of microtexture and macrotexture into a single indicator, allowing a more comprehensive assessment of pavement safety. The implementation of estimated available friction at different speeds was analyzed using the speed correction equation in ASTM E1960-07 to model the expected number of crashes on the interstate freeway facility in the United States. The analysis showed that the estimated available friction at 120 km/h (FRS120) significantly combines the effect of friction and macrotexture on freeway crashes. This index allowed developing crash modification factors (CMFs) by effectively capturing potential reduction in expected crashes due to changes in both SFN40 (friction at a low slip speed) and macrotexture.

A safety performance function (SPF) is a tool to identify explanatory variables prone to traffic crashes in safety systemic analysis. This study explores advanced statistical methods for developing SPFs. Traditional negative binomial regression, while linear and interpretable, could have inaccurate performance with an increasing number of variables included in SPFs. We explore alternative methods and compare them to the traditional method, including best subset selection, shrinkage, generalized additive model, and tree-based method. Using Florida curve crash data, the study examines various risk factors such as traffic volume, demographics, curve characteristics, roadway characteristics, and spatial relationships, and develops SPFs. The study assesses the contributing risk factors, the overall SPFs performance, the methods’ capabilities in handling statistical challenges, and variable interpretations. The findings show that complex methods, such as generalized additive model, outperform the traditional approach, suggesting their potential application in SPF development for more accurate crash predictions and variable interpretations.

This study delves into the causes of motorcycle crashes in California by analyzing data from the California Statewide Integrated Traffic Records System (SWITRS), comprising 111,282 motorcycle crashes between 2012 and 2022. Analysis of main violations reported for motorcycle crashes indicated the need to focus on deterring riding under the influence, managing speed, promoting safe turning and lane-changing practices, and addressing traffic signal compliance and improper passing. The study also explored the relationship between collision types and crash severity, emphasizing the significance of addressing issues related to broadside and head-on collisions in fatal crashes, overturning incidents in injury-related crashes, and sideswipe and rear-end collisions in property damage only. Utilizing an ordered logit model with random coefficient decomposition, the research identified statistically significant variables affecting crash severity. High-severity outcomes were associated with crashes occurring on highways, non-compliance with laws, collisions with other motor vehicles, and wet road surfaces. Age and gender further influenced the severity of crashes. Findings can collectively contribute to reducing the frequency and severity of motorcycle crashes, making the roads safer for motorcyclists and other road users alike.

This study combines econometric and geospatial methods to analyze the impact of cellular coverage on distracted driving incidents, crashes, and injuries in Oregon from 2017 to 2020, focusing on cell phone-related crashes. Despite reduced travel in 2020, these crashes remained high. Geospatial tools identified urban hotspots like Portland and Salem. We used the mixed logit model to evaluate factors like driver demographics, vehicle characteristics, and environmental conditions, shedding light on the economic aspects of injury severity. Results highlight the crucial role of seatbelt use in reducing injury severity. The study underlines the need for comprehensive strategies to combat distracted driving in Oregon for better road safety and to lower economic costs associated with such incidents.

This research investigates driver-yielding behaviour at unsignalized crosswalks, focusing on pedestrian characteristics, and oncoming vehicle factors. The study uses data from two-lane, two-way divided roads, and drone video footage to capture vehicle trajectories and pedestrian behaviour at the median or further lanes. Driver-yielding rates are accurately captured using TRACKER 5.0 software, and the behavior is modelled using binary logistic regression. The study found that yielding patterns are similar across different vehicle categories, but factors influencing drivers’ willingness include the number of female pedestrians, the number of pedestrians in the median, the number of injured pedestrians, and the vehicle position/vehicle approaching the lane. The research contributes to understanding and addressing the complexities of unsignalized pedestrian crossings, laying the groundwork for future traffic management strategies to create safer, more efficient urban environments for pedestrians and drivers.

This study provides a comprehensive review regarding the issue of red-light running in the US, highlighting its significant impact on road safety. Through a thorough examination of pertinent literature, the research sheds light on the challenges posed by red-light running and explores diverse strategies adopted by major US cities to curb this behavior, including through the deployment of red-light cameras. This study discusses the effectiveness of measures like red-light cameras in reducing intersection crashes, notably right-angle collisions, despite potential concerns like a slight increase in rear-end collisions. Emphasizing a comprehensive approach, it reviews punitive measures for red light running across various US cities. Additionally, the research investigates Vision Zero action plans in different cities, examining their strategies for reducing red light running crashes. By providing these valuable insights, this study aims to guide researchers, practitioners, and policymakers in developing holistic solutions to address and mitigate the challenges associated with red-light running.

This study explores how “No Turn on Red” (NTOR) signs impact driver compliance and behavior at signalized intersections using a driving simulator. Data from 67 participants, navigating three scenarios (no sign, static sign, and dynamic sign), were analyzed to assess driver performance. The primary aim was to understand driver behavior concerning various NTOR signs, evaluating their speed profiles when approaching and turning at signalized intersections. Statistical and compliance analyses were conducted to assess driver performance. Results indicated that male drivers exhibited lower speeds in the presence of dynamic signs, suggesting heightened attention and compliance. Compliance analysis revealed a higher adherence rate in the dynamic sign scenario, emphasizing the effectiveness of its clear and easily understandable message in influencing driver behavior. These findings have implications for road safety, NTOR sign design, and educational strategies, offering valuable insights for future research and decision-making in these domains.

Emerging urban parcel delivery (UPD) modes are anticipated to decrease surface UPD truck trips and stops, thus leading to less exposure of UPD trucks on surface roads and reduced UPD crashes. This paper evaluated the safety impacts of innovative last-mile delivery strategies in urban areas. The geographically weighted negative binominal regression (GWNBR) model was developed at zone levels based on the roadway, traffic, and demographic data collected in Hillsborough County, Florida. Future UPD scenarios were projected for coming years (2030, 2040, and 2050) with different replacement rates (10%, 30%, and 50%) of UPD truck stops by emerging UPD modes. The developed GWNBR model was used to predict UPD crashes for future scenarios. The results indicate that emerging UPD technologies cause a decrease in delivery truck stops and reduce UPD crashes by 3%, 11%, and 20% for 2030, 2040, and 2050, respectively.

Urban intersections often suffer from high traffic congestion and emissions due to frequent braking and idling. This study proposes a Speed Control Guidance (SCG) system to optimize vehicle trajectories in mixed traffic, including both internal combustion engine vehicles (ICEVs) and battery electric vehicles (BEVs). A 3D driving simulator assesses driver responses to real-time SCG guidance at signalized intersections. Participants receive color-coded speed recommendations along their route in various scenarios, and their driving behavior is compared to scenarios without SCG. Sociodemographic factors like gender and age influence SCG effectiveness. Female drivers show lower compliance with speed guidance, and older drivers struggle to follow recommendations. To enhance SCG’s effectiveness, it is crucial for researchers and vehicle manufacturers to develop strategies that address the specific needs and preferences of diverse driver groups.

Crashes occur due to the influence of the interaction between three factors: humans, vehicles, and the road environment. However, the human factor is often found to be the most dominant factor responsible for crash occurrences. The objective of this study is to assess the contribution of human behavior to the occurrence of road crashes by analyzing crash occurrence in highly similar road environments. Homogeneity in the road environment was achieved by clustering road segments with highly similar characteristics and by controlling the effect of vehicle-related factors, and the variation in crash occurrence between and within clusters was assessed to explore the role of human behavior. The approach developed in this study can be used to understand the variability of the human factor and its contribution to crash occurrence without knowing drivers’ actual behavior. Furthermore, the results can be applied by road management authorities to mitigate road crashes and to prioritize and select appropriate interventions.

Accurate vehicle weight monitoring is essential for efficient and effective traffic management and road maintenance, particularly in preventing potential damage from overweight vehicles to road infrastructure. Weigh-in-motion technology plays a pivotal role in traffic engineering due to its ability to rapidly collect data with minimal traffic disruption. However, the accuracy of weigh-in-motion stations need enhancement and may be affected by weather conditions, especially fluctuations in temperature. To address this concern, a field test was conducted, incorporating controlled vehicle weight, speed, and driving behavior. Additionally, electrical resistance strain transducers were configured for data collection, strategically positioned at the location of maximum bending, the bottom of the asphalt layer, 5 in. beneath the asphalt road surface. This field test, spanning 10 experiments over two years, was conducted to analyze the impact of weather factors on measured vehicle weights by strain sensor. The findings of this investigation will provide insights into the benefits of enhancing the accuracy and reliability of vehicle weight measurements, contributing to more effective traffic management and improved road maintenance practices.

Motor vehicle crashes, claiming over 100 lives daily in the US, pose a critical public health risk. This study aims to examine relationships between a five-year period of crash data and disadvantaged communities (DACs) in the city of Atlanta to assess the equity aspects of crash safety. The study used regression analysis to evaluate the association between transportation, health, economic, equity, and resilience indicators and crashes. Results indicate a positive relationship between DACs and crashes, with transportation disadvantage emerging as the most influential predictor. The findings emphasize the need for targeted policies addressing underlying factors contributing to elevated crash rates in DACs, such as poverty and limited access to transportation and healthcare. Furthermore, the study underscores the importance of considering community insights in traffic safety planning and informs the development of interventions to enhance road safety and equity.

Pedestrian safety is a significant concern due to the severity of pedestrian-involved crashes. Pedestrian fatalities have increased by 77% over the past decade, constituting a 5% increase in pedestrian fatalities per the overall number of traffic-related fatalities. Prior studies have explored various regulatory traffic signs; however, there have been limited studies on in-crosswalk signs. This study compared the effectiveness of in-crosswalk “YIELD TO PEDESTRIANS” and “STOP FOR PEDESTRIANS” traffic signs. Yielding behavior and vehicle speed were used as safety indicators, and impacts were measured during different school sessions and times of day. The traffic signs’ impact was evaluated through independent t-tests and two-proportion z-tests at a 95% confidence level. The results demonstrate that both traffic signs led to significant vehicle speed reductions and yield to pedestrian enhancements across all situations; however, the difference between the two traffic signs’ impact was insignificant, suggesting they can be used interchangeably.

Ensuring the safety of roadways for travelers of all modes is a collaborative process that involves stakeholders across state and local government agencies, advocacy groups, research communities, and between the private and the public sectors. To better understand the policies and practices regarding data-driven safety coordination between counties and first responder agencies—especially coordination activities that leverage available data assets of local jurisdictions to benefit and maximize the resources of the first responder agencies—a survey was conducted of 24 of the most populous counties in Florida, representing the largest third of counties in the state. An in-depth literature review was completed to supplement information gathered from the survey and to assist in analyzing the results. From the literature review and survey findings, a set of suggestions to enhance coordination between counties and local first responder agencies on transportation safety were developed. Although this study was conducted in Florida, the findings and suggestions of this study are relevant to understanding agency coordination on transportation in both the US and internationally.

This study investigates factors influencing the severity of bicycle-involved crashes in the US, particularly the Southern California Association of Governments (SCAG) region from 2013 to 2017, using Highway Safety Information System (HSIS) data. Employing a Bayesian network model with rigorous validation, the study yields a low error rate, emphasizing its effectiveness in analyzing crash data and enhancing rider safety insights. Two scenarios explore variables affecting the probability of fatal crashes, revealing the positive impact of proper lighting and surface on visibility. Additionally, the need for infrastructure capable of handling wet surfaces and providing adequate drainage is underscored. The results demonstrate the importance of effective infrastructure design, emphasizing proper lighting and visibility for cyclists to mitigate fatal crash risks. The study’s implications extend to informing policymakers and transportation engineers on prioritizing safety measures, showcasing the Bayesian network model’s efficacy in identifying critical factors in bicycle-involved crashes.

Interstates in New Jersey experience a high volume of traffic, and unfortunately, around 15,000 crashes occur on these roads every year. Statistics show that 21% of these crashes result in fatalities or severe injuries. Therefore, it is critical to understand the factors contributing to this high number of crashes. This study aims to identify these factors through the application of statistical techniques. Data was collected from the heavily traveled interstate corridors. Utilizing a negative binomial model, we analyzed the total number of crashes in relation to traffic, roadway geometry, and operational conditions. As expected, factors associated with exposure, such as traffic volume and segment length, were significant. As these factors increase, crashes tend to increase. Additionally, interstates with concrete barriers exhibited fewer crashes compared to those with curbed barriers, indicating the potential benefits of concrete barriers in reducing crashes. Interestingly, our study uncovered an aspect that is overlooked in the widely utilized safety assessment guidebook, the Highway Safety Manual—the impact of speed limits on crash frequency. Through a variable prioritization process, we identified speed limit as the second most important factor. Further exploration of the dataset revealed that crashes were more prevalent when the speed limit decreased from the previous segment on a specific interstate route. Such findings have practical implications for transportation agencies and policymakers. The insights gained can be instrumental in devising effective countermeasures, such as leveraging intelligent transportation systems (ITS) to provide advance signs for speed limit changes. Implementing such measures can help mitigate crashes, particularly those associated with speed limit alterations. Overall, this study identified the factors contributing to the high crashes in New Jersey’s interstates. Understanding these factors can aid in formulating targeted strategies to improve the safety of these roads.

Lane departure crashes are among the most common crashes in Florida and in the United States. Many departments of transportation (DOTs) prioritize reducing both the frequency and severity of these accidents. Auditory vibratory treatments (AVTs), or rumble strips, are frequently used to prevent or reduce lane departure incidents. Traditional rumble strips produce exterior noises that become a burden to nearby residents. To address the noise issues, several DOTs considered a new type of AVT called sinusoidal rumble strips. The research team conducted an in-depth literature review, agency interviews, a focus group, and an assessment of studies conducted by the Florida Department of Transportation (FDOT) to evaluate the noise reduction and bicycle accommodation of sinusoidal rumble strips installed in Florida and across the nation. The research concludes that sinusoidal rumble strips can effectively mitigate external noise problems while effectively notifying drivers and occupants of lane departures.

Speeding is the major cause of severe crashes in the United States. Speed management is essential to reduce fatalities and serious injuries for all road users in any crash, especially for vulnerable road users (VRUs), such as pedestrians and bicyclists. This study evaluated the effects of law enforcement presence on vehicle speeds for roadway segments with speeding issues. The project team used the citations and warning data collected through the Florida Department of Transportation (FDOT) District 7 Enhanced Law Enforcement Engagement (ELEE) project to assess the effects of law enforcement presence, frequency, and duration on vehicle speeds on study corridors. The research findings demonstrate that the effects of police presence and number of citations and warnings on vehicle speeds were instantaneous, but no lasting impact of police presence was observed. The effects of police presence on vehicle speed may be limited if the law enforcement officers are not visible to the public, or limited citations and warnings are issued by the police officers.

Excessive speeding poses significant risks to road safety, impacting a driver’s ability to maneuver safely around obstacles and leading to longer stopping distances and delayed reactions to hazardous situations. It is a major contributor to fatal and serious road trauma, accounting for over 20% of such incidents in the US. Eliminating speeding entirely could potentially substantially reduce fatal injuries by 20% or more. This study leveraged the comprehensive National Highway Traffic Safety Administration’s (NHTSA) Fatality Analysis Reporting System (FARS) data, focusing on 50,081 speed-related motor vehicle traffic fatal crashes that occurred between 2016 and 2021, and a probabilistic graphical model to investigate the causal associations between key contributing factors involved in these speeding incidents. Ultimately, this high-impact research advanced our understanding of the risks posed by speeding and impaired driving, guiding the way for evidence-based interventions and transformative policies to build safer roads and protect all road users from preventable tragedies.

Lessons learned after hurricanes revealed that although most traffic signal infrastructure withstood the impacts of the storms, signalized intersections were not operational due to lack of utility power. Current traffic cabinets employ a standard uninterrupted power supply (UPS) system, which is powered by lead acid batteries, an outdated technology that provides power for only up to 8 h of operation. For this study, a liquid propane generator was tested successfully to provide constant power without issues. A system that included a battery + solar backup was tested successfully to provide power for several days without issues. This system is not usually an issue after hurricanes, especially if adequate solar energy generation is provided to charge the battery during available daytime hours. The result of the testing provides options for transportation agencies for extended runtime at signalized intersections using backup power after the loss of utility power.

This study critically assesses the impact of converting the Dhaka-Chattogram National Highway (NH-1) in Bangladesh from an undivided two-lane to a four-lane highway with a median on crash rates and patterns. Crash data pre- and post-conversion was sourced from the Accident Research Institute (ARI) and the Bangladesh police headquarters for detailed analysis. The transition resulted in a notable decline in head-on collision-related fatal crashes. However, concerning trends emerged post-conversion, revealing an increase in various other significant crash types and associated injuries and fatalities, particularly those involving pedestrians. Effective countermeasures are recommended based on the study findings to mitigate highway crashes and their severity. This investigation provides critical insights into the dynamics of highway safety post-expansion, highlighting the necessity for targeted interventions to address emerging risks and enhance overall safety on NH-1.

Multi-modal transportation relies on efficient network connectivity among active transportation services, such as biking and transit stops. In many cases, road users must walk or bike varying distances to reach a transit stop and then continue their journey to their final destinations. As shared mobility services like installations of shared bike docking stations increase in urban settings, planning additional docking stations that efficiently connect with existing bike network and transit stops becomes a significant challenge. This paper proposes a machine learning (ML)-based multi-modal transportation network planner to address the optimization challenge of expanding shared bike docking stations (Cincinnati Red Bike) that connect with the existing transit network (Cincinnati Go Metro) and amenities within neighborhoods of the city of Cincinnati. We identify relevant data sources for the planning problem, such as existing bike share ridership data, transit data from General Transit Feed Specification (GTFS), demographic data from the Census, and built environment factors such as amenities and land use data. The K-means clustering algorithm is used to model the region’s features across the selected variables and identify potential areas to expand into. Network analysis is performed to observe the network effects of expansion and determine candidate locations where future bike share stations may be built.

The control of autonomous vehicles (AV) within complex signalized intersection scenarios poses a substantial challenge, requiring a significant amount of data for effective training of autonomous driving models. This challenge is compounded by limitations in relying solely on a single AV for data, which may fall short of capturing the diverse and dynamic nature of complex intersection settings. The advent of connected and autonomous vehicles (CAVs) introduces a transformative concept, enabling real-time traffic information exchange among interconnected vehicle networks. This shift empowers CAVs to exhibit vastly improved driving behaviors, leveraging shared information for enhanced accuracy, reliability, and efficiency. However, the imperative of preserving user privacy and ensuring data security in the context of AI-enabled CAVs remains a significant challenge. This paper proposes a federated reinforcement learning framework designed to establish a privacy-preserving knowledge-sharing strategy. The proposed framework efficiently aggregates and reuses knowledge learned by diverse CAVs, operating in varying intersection environments. This collective knowledge enhances decision-making precision and adaptability, contributing to safer and more efficient autonomous driving at signalized intersections.

Numerous limitations of proper lands, especially in offshores, bring about outcomes to consider soil improvement methods. The deep soil mixing (DSM) method is an efficient ground improvement procedure using in situ soils for treatment plans. In order to ensure the quality of the treated soil, the DSM elements should be evaluated using unconfined compressive strength (UCS) tests on only 2%−4% of executed columns, based on the recommendations of the FHWA. Therefore, there are significant uncertainties regarding the UCS of the other 96% of the untested DSM elements. To reduce these deficiencies and bring an insight into the uncored DSM columns indirectly, artificial intelligence (AI) and machine learning (ML) methods (XGBoost, random forest, artificial neural network, and lasso regression) were implemented to predict a correlation algorithm between UCS results and several drilling rig machine parameters, including drilling, lifting, rotary speeds, rotary pressures, and the added water during pre-drilling phase. The extensive results of over 3,780 UCS tests and corresponding machine data of a DSM soil improvement project in Hormozgan, Iran, were considered to train and test the ML algorithms. In this case, instead of conducting a probability analysis to find out a DSM column UCS for using in design procedure, the predicted strength of each column by the well-trained ML method can be considered.

Microscopic traffic simulators, such as VISSIM, use time-based step-by-step vehicle movements to predict outcome of proposed network configuration. This process is inherently slow and impractical for use in stochastic large network optimizations. Our solution is to use a mesoscopic traffic simulator with one-step movement approach between nodes in traffic network. Our model uses a predictor function for each road segment which predicts travel time distribution for traffic conditions using machine learning (AI). Our simulator chooses a random sample from predicted distribution as travel time. Travel time distributions may be sensitively dependent on parameters such as traffic and road conditions, and traffic behavior patterns which may be dependent on specific road. Experimental results show that our simulator’s fidelity is like that of VISSIM for various traffic conditions. We demonstrate that our simulator is more than 100 times faster than VISSIM and provides network performance results that are comparable to these models.

Frost effects, such as thaw weakening and frost heave, can significantly degrade pavements. Effective mitigation relies on accurate prediction of soil thermal properties. Current prediction methods use empirical equations or finite element analysis, with recent progress using machine learning. Recurrent neural networks, a type of machine learning model designed for prediction of sequential data, may be appropriate. Further, the model’s output can be fed back into itself recursively to make long-term predictions. Using this method and training data from Hill Air Force Base, Utah, and Air Force Academy, Colorado, we were able to forecast soil parameters including temperature, thermal conductivity, and moisture content for frost susceptible soils. The model had a deviation from experimental values of no more than 10%, with the most significant contributions to accuracy being the use of gated recurrent unit neurons and the incorporation of multiple soil parameters.

This study investigates pedestrian crash severity through automated machine learning (AutoML), offering a streamlined and accessible method for analyzing critical factors. Utilizing a detailed dataset from Utah spanning 2010−2021, the research employs AutoML to assess the effects of various explanatory variables on crash outcomes. The study incorporates SHAP (SHapley Additive exPlanations) to interpret the contributions of individual features in the predictive model, enhancing the understanding of influential factors such as lighting conditions, road type, and weather on pedestrian crash severity. Emphasizing the efficiency and democratization of data-driven methodologies, the paper discusses the benefits of using AutoML in traffic safety analysis. This integration of AutoML with SHAP analysis not only bolsters predictive accuracy but also improves interpretability, offering critical insights into effective pedestrian safety measures. The findings highlight the potential of this approach in advancing the analysis of pedestrian crash severity.

Traffic conflict analysis has gained significant attention over the last few years. Recent research has made significant progress in the development of detection and tracking systems by utilizing various technologies such as closed-circuit televisions (CCTVs), unmanned aerial vehicles (UAV), and sensor fusion exploiting light detection and ranging (LiDAR). Challenges impeding widespread adoption of artificial intelligence-based traffic conflict analysis include issues with monitoring devices, detection algorithms, and output processing. To address these, this study employs a convolutional neural network (CNN)-based algorithm designed for autonomous vehicles, enhancing detection and tracking of road users at signalized intersections. The algorithm’s strength lies in depicting road users with three-dimensional bounding cuboids, streamlining data processing. Post-processing algorithms refine tracking precision, enabling seamless trajectory reconstruction and accurate vehicle state estimation. Tested on video footage, the algorithm excels in high-traffic scenarios, achieving 95.08% precision and 92.81% recall. It identifies diverse conflict instances using four unique traffic conflict indicators.

Weather conditions significantly influence traffic safety, with adverse conditions often leading to hazardous driving environments. This study presents the development of a weather detection model based on convolutional neural networks (CNNs), specifically trained on in-vehicle dash camera videos. Using thousands of images sourced from the SHRP2 trajectories level dataset, seven distinct weather categories were identified: clear, light snow, heavy snow, light rain, heavy rain, distant fog, and near fog. For enhanced training efficiency, images were strategically segmented. An innovative approach was employed where, on average, nine images were selected from every minute of the video; these images underwent processing using various pre-trained CNN architecture, such as “AlexNet,” “GoogleNet,” “ResNet,” among others. The predominant detection results then informed the overall weather condition for the given video segment. The model exhibited promising performance, underscoring its viability for real-time in-vehicle weather detection and consequent enhancement of road safety.

Over the past decade, researchers have shown significant interest in traffic monitoring using various devices such as surveillance cameras, unmanned aerial vehicles (UAV), loop detectors, and sensor fusion exploiting light detection and ranging (LiDAR). These techniques, widely applied for vehicle detection and tracking, have established robust safety assessment methodologies. However, nationwide deployment is hindered by constraints in monitoring devices, detection algorithms, and output processing. This study addresses these limitations by introducing an innovative approach that integrates three surveillance cameras with meticulously designed framework. Leveraging two distinct convolutional neural network (CNN)-based detection algorithms, this method accurately detects and tracks vehicles, providing reliable outputs that are further analyzed, smoothed, and integrated for a comprehensive reconstruction of traffic scenes at signalized intersections. The proposed framework enables the extraction of crucial traffic flow features, facilitating the identification of traffic conflicts instances. Through the synergistic integration of surveillance cameras and CNN-based algorithms, this research contributes significantly to advancing traffic safety practices, offering an efficient and accurate tool for safety assessment.

Pedestrian detection poses significant challenges for traffic safety researchers, given diverse motion patterns, clothing colors, partial occlusions, and varying positions concerning detection devices. Surveillance cameras, light detection and ranging (LiDAR), and microwaves beams have been utilized in pedestrians’ detection, leading to robust safety assessment methodologies. Despite their efficacy, widespread deployment encounters constraints. This study introduces an optimized approach by combining surveillance cameras with a distinctive convolutional neural network (CNN)-based pose estimation algorithm for precise pedestrian detection at signalized intersections. A geometrical spatial proximity method, grounded in linear and curvilinear perspectives, restores pedestrian joint coordinates from the image plane to a top-down view. These coordinates are clustered and integrated to construct pedestrian trajectories. The proposed framework serves as an efficient and accurate tool for assessing pedestrian motions, addressing challenges in widespread deployment and contributing significantly to advancing traffic safety practices at signalized intersections.

Numerous studies have highlighted the significant influence of adjacent geographic units on rear-end crashes, termed as the spatial spillover effect. This paper introduces an innovative systematic model based on Bayes-GLM to illustrate the propensity for rear-end crashes for improving safety countermeasures on a broader scale. The approach integrates hotspot analysis with K-means clustering into the crash causation model, creating a multi-layered strategy, seeking to improve functionality and scalability by systematically screening potential factors contributing to rear-end crashes. It further identifies relationships between the spillover effect and microscopic impacting factors. The primary advantage of employing such an integrated macro- and microscopic approach lies in the swift identification of critical areas. Moreover, a heuristic analysis of vehicle-to-vehicle interactions, reflecting pre-crash behaviors, is systematically connected to the spillover effect. This linkage enables the integration of a larger scale of crash-influencing factors, thereby enhancing the understanding of the rear-end crash occurrence mechanism.

Highway abandoned objects have seriously threatened traffic safety, which may lead to vehicle crashes and block roadways. It is significant for transportation agencies to recognize and clear them as soon as possible. However, it is challenging to detect these abandoned objects because they have various sizes/categories and appear at different locations. Although computer-vision-based object detection algorithms have succeeded in traffic monitoring and autonomous driving, most SOTA algorithms cannot satisfy the demand for highway abandoned object detection due to undetermined target categories. Inspired by impressive progress in contrastive learning and large language models, this paper proposes a two-stage open-vocabulary object detection algorithm for highway abandoned object detection. The first stage is the region proposal, with the pre-trained image encoder presenting all region candidates on multiple scales. Then, the second stage adopts CLIP and computes similarities of embeddings from proposed regions. In contrast to conventional object detection algorithms, this algorithm is not constrained by limited training data categories. The proposed algorithm is validated using real-world highway surveillance videos collected in Nanjing, China, containing varying light conditions and multiple video angles. Compared with SOTA object detection models, the proposed one has shown extraordinary performance and generalization abilities for abandoned object detection, with higher precision and running efficiency, that is, in low-light conditions. This approach is helpful for transportation agencies to provide early warning in time and find accident perpetrators.

Transportation equity is an interdisciplinary agenda that requires both transportation and social inputs. Traditionally, transportation equity information is sourced from public libraries, conferences, television, and social media, among others. Artificial intelligence (AI) tools, including advanced language models such as ChatGPT, are becoming favorite information sources. However, their credibility has not been well explored. This study explored the content and usefulness of ChatGPT-generated information related to transportation equity. It utilized 152 papers retrieved through the Web of Science (WoS) repository. The prompt was crafted for ChatGPT to provide an abstract given the paper’s title. The ChatGPT’s abstracts were then compared to human-written abstracts using statistical tools and unsupervised text mining. The results indicate a weak similarity between ChatGPT and human-written abstracts. On average, the human-written and ChatGPT-generated abstracts were about 58% similar, with a maximum and minimum of 97% and 1.4%, respectively. The keywords from the abstracts of papers with over the mean similarity score were more likely to be similar, whereas those below the average score were less likely to be similar. Themes with high similarity scores include access, public transit, and policy, among others. Contrarily, the findings from collocated keywords were inconclusive. The study findings suggest that ChatGPT has the potential to be a source of transportation equity information. However, currently, a great amount of attention is needed before a user can utilize materials from ChatGPT.

Safety is the predominant concern in the advancement of automated vehicles (AVs); therefore, extensive real-world testing is underway to ensure their secure operation. Despite the widespread belief that they will decrease the frequency of accidents, there remains uncertainty about their impact on the severity of crashes in which they are involved. The primary objective of this study is to use the bagging classifier technique to predict the likelihood of injuries in accidents involving AVs. This was accomplished by conducting an in-depth examination of a wide range of independent variables and an analysis of injuries sustained in crash incidents involving AVs from 2014 to July 2023. The bagging classifier model showed notable effectiveness, achieving a balanced accuracy of 0.59, along with high precision and recall values of 0.94 and 0.97, respectively. These metrics indicate the model’s strong capability for accurately identifying severe crash outcomes and minimizing false positives. The precision-recall curve and a modified F1 score of 2.39 further endorse the model’s performance, particularly highlighting its efficiency in handling the class imbalance present in the dataset. The validation and learning curves underscore the model’s optimal complexity, displaying its proficiency in identifying essential patterns without succumbing to overfitting. Collectively, these metrics underscore the model’s success in predicting injury outcomes in AV crashes with a high level of accuracy. This study contributes to the literature on AV safety by providing valuable insights for manufacturers and policymakers that will enable them to develop effective safety features and strategies, thereby enhancing traffic safety.

Public opinion regarding autonomous vehicles (AV) heavily influences how quickly the technology will be implemented and adopted in the future. This research utilized data collected from riders of Rideshare Automation Payment Integration Demonstration (RAPID), a shared autonomous vehicle (SAV) service in Arlington, Texas, and performed exploratory data analysis to identify the monthly ridership trends of the riders. Within the scope of this project, AVs provided service to the UTA campus and downtown Arlington. The findings revealed that students are the most frequent riders of the RAPID service when compared to non-students. It was also revealed that the usage frequency of RAPID service was more on Tuesdays and Wednesdays when compared to other working days of the week. Additionally, the findings indicate that the service was less utilized during early mornings before 9:00 am and evenings after 6:00 pm when compared to other times of the day. This study offers critical insights toward ridership trends of shared AVs that will assist local, state, and federal transit authorities and planners in formulating policies and transportation strategies to target SAV ridership when the service is more widely offered.

As automated vehicles (AVs) gradually gain prevalence on public roads, understanding their distinctive driving behavior is crucial for traffic management and planning. This study conducted field experiments using an SAE Level-3/4 AV and collected driving data of AVs and human drivers on public roads using sensors including GPS, radar, camera, and LiDAR. The Wilcoxon rank-sum test is used to identify the difference in the behavior between AVs and human drivers. In addition, logistic regression and Extreme Gradient Boosting (XGBoost) are used to classify AVs and human drivers. Results suggest that there exists a significant difference in driving behavior between AVs and human drivers. Moreover, features including the mean speed and the distance from the vehicle to the detected objects are positively related to the probability of the vehicles being AVs, while the standard deviation of speed and the mean acceleration are negatively associated with it. Furthermore, XGBoost accurately identifies AVs and human drivers using the extracted features with an average area under the curve of 0.92. Results from interpreting results from XGBoost indicate that it performs better when the mean speed is either in the low or high ranges. Moreover, AVs and human drivers are hard to differentiate using the model when the vehicle is too far from other objects. This study underscores the substantial divergence in driving behavior between AVs and human drivers, offering valuable insights for the evaluation of the impact of AVs on traffic conditions.

Natural disasters like hurricanes, wildfires, and floods pose immediate hazards. Such events necessitate prompt and well-planned emergency evacuations to save lives and reduce fatalities, injuries, and property damage. This comparative study explores the potential impact of autonomous vehicles on evacuation travel times, citing the significance of swift evacuations amid the increasing frequency, severity, and global reach of natural disasters. The study rigorously models evacuation scenarios using GIS and VISSIM simulations to assess the feasibility of integrating autonomous cars into evacuation strategies for various natural disasters. The methodology focuses on evaluating a historical wildfire evacuation event, by employing a connected vehicle dataset, to showcase how autonomous vehicles navigate various emergency conditions. The results indicate a promising reduction in evacuation travel times and highlight the potential of autonomous vehicles in optimizing emergency evacuations and reducing clearance time. The findings also highlight the importance of integrating emerging technologies like connected and autonomous vehicles (CAVs) to enhance evacuation strategies and preparedness for emergencies.

In the era of connected and autonomous vehicles (CAVs), optimizing transportation infrastructure is crucial to unlock the full potential of connected and autonomous technology (CAT), particularly in the context of advanced operations like platooning. Considering the case, this study focused on infrastructure modifications, with a specific focus on low-automation CAVs. The study employed PTV VISSIM to model a variety of scenarios on a freeway segment, considering six distinct lane configuration scenarios, varying market penetration rates (MPRs), and factors such as platoon allowance/restriction within a 1-mi buffer upstream of lane closures, as well as open or closed lane conditions (dedicated lanes). The findings revealed that, for a two-lane segment, restricting platoons to open lanes and allowing them on upstream 1-mi resulted in a 28% reduction in the total number of conflicts, with time-to-collision (TTC) serving as the conflict indicator. The worst-case scenario involves allowing platoons without any lane restrictions, leading to a substantial increase in conflicts at higher MPRs.

The ongoing progress in automation technologies holds the potential to mitigate human errors and create a safer transportation environment. Soon, the roads will see a mix of human driven vehicles alongside connected and automated vehicles (CAVs) with varying market penetration and level of autonomy (MPLA). As technology advances, it becomes clear that as market penetration of CAV increases, so does their level of autonomy. This study employs a VISSIM-based microsimulation model to mimic these mixed traffic scenarios and investigate its effect on safety and operational efficiency. Altogether 27 scenarios with varying traffic volumes and MPLA were explored, assessing safety using the time to collision indicator with three different thresholds. The findings showed that CAVs with mid to high autonomy and a market penetration of about 50% reduced conflicts by about 50% in medium to high traffic volumes while improving network efficiency. At higher MPLA scenarios, the system performed consistently well, with minimal variations in safety and operational impacts across different traffic volumes.

Freeway work zones, particularly in merging areas, experience serious traffic safety and operational difficulties. The advent of connected and autonomous vehicles (CAVs), capable of complex maneuvers like platooning, adds new layers of complexity. Despite numerous studies exploring operational aspects of heavy vehicle platooning in work zones, safety evaluations remain scant. This research, using VISSIM microsimulation, addresses this gap by examining the safety implications of cooperative adaptive cruise control (CACC) truck platooning in such areas. The study simulated varying scenarios by positioning lane change warning signs at distinct distances upstream of work zones. Findings reveal that early merge strategies—signs at least 1 mi upstream—decreased traffic conflicts by 300% compared to late merge conditions. This study also emphasizes that while platooning provides significant benefits in controlled traffic flow situations, infrastructure owners and operators (IOOs) should work closely with other stakeholders such as trucking companies, and law enforcement agencies to make necessary changes to their infrastructure to accommodate truck platooning.

The long-tail distribution of real driving data poses challenges for training and testing autonomous vehicles (AV), where rare yet crucial safety-critical scenarios are infrequent. And virtual simulation offers a low-cost and efficient solution. This paper proposes a near-miss focused training framework for AV. Utilizing the driving scenario information provided by sensors in the simulator, we design novel reward functions, which enable background vehicles (BV) to generate near-miss scenarios and ensure gradients exist not only in collision-free scenes but also in collision scenarios. And then leveraging the robust adversarial reinforcement learning (RARL) framework for simultaneous training of AV and BV to gradually enhance AV and BV capabilities, as well as generating near-miss scenarios tailored to different levels of AV capabilities. Results from three testing strategies indicate that the proposed method generates scenarios closer to near-miss, thus enhancing the capabilities of both AVs and BVs throughout training.

Amid the COVID-19 pandemic, automated delivery vehicles (ADVs) have gained even more significance due to their touchless capabilities, aligning perfectly with the need for contactless delivery services to safeguard public health. This study collected and conducted quantitative analysis on the most relevant 86 news articles focused on ADVs by employing sophisticated text network and collocation analyses to extract valuable insights. The findings indicated a noteworthy trend: express delivery services have forged partnerships with various ADV companies, leveraging their capabilities to tackle the challenging “last mile” delivery stage. Furthermore, the study revealed a growing integration of ADVs into diverse delivery sectors, particularly in food, grocery, and other on-demand services. Overall, this study provides invaluable insights into the current state of ADVs and their widespread adoption across various delivery sectors.

With the rapid development of autonomous driving technology, it is imperative to study the impact of this technology on the resilience of urban road transportation networks. This study focuses on virtual traffic signal control (VTS) as a representative of autonomous driving technology and proposes a multi-dimensional resilience evaluation system to compare the impacts of three different signal-timing strategies on the resilience of urban traffic networks under various scenarios. Simulation results demonstrate that the average vehicle speed under VTS control is 38.1% and 40.1% higher than the other two control modes, respectively. In addition, the VTS approach significantly reduces delays and minimizes fuel consumption compared to the other two approaches. Moreover, this paper delves into the intrinsic mechanisms by which VTS enhances traffic resilience. It highlights the importance of integrating autonomous driving technology into traffic strategies to achieve carbon emissions reduction and promote sustainable transportation. The network resilience enhanced by VTS enables cities to effectively respond to unexpected events, thereby ensuring smooth traffic flow and safety. This research provides valuable insights into existing knowledge and promotes further research in the field of traffic resilience.

Given the promising future of autonomous vehicles, it is foreseeable that self-driving cars will soon emerge as the predominant mode of transportation. While autonomous vehicles offer enhanced efficiency, they remain vulnerable to external attacks. In this research, we sought to investigate the potential impact of cyberattacks on traffic patterns. To achieve this, we conducted simulations where cyberattacks were simulated on connected vehicles by disseminating false information to either a single vehicle or vehicle platoons. The primary objective of this research is to assess the cybersecurity challenges confronting connected and automated vehicles and propose practical solutions to minimize the adverse effects of malicious external information. In the simulation, we have implemented an innovative car-following model for the simulation of connected self-driving vehicles. This model continually monitors data received from preceding vehicles and optimizes various actions, such as acceleration, and deceleration, with the aim of maximizing overall traffic efficiency and safety.

This study presents a comprehensive analysis of automated vehicle (AV) disengagements in California, leveraging traffic conditions and employing the XGBoost model and explainable AI techniques. The study investigates six distinct categories of disengagement initiation: driver-initiated, system-initiated, freeway-based, arterial-based, driverless-capable, and driverless-incapable. The findings reveal that predictors of each disengagement type differ significantly. Generally, the day of the week and vehicle miles of travel showed a great contribution to the prediction of overall AV disengagements. On the other hand, human-driven VMT was the main predictor for driver-initiated disengagement, while human-driven vehicle crashes were the highest-ranked factor for system-initiated, arterial roadways, and driverless-capable AV disengagement. For freeway-based disengagements, the year of operation was the major factor. The study contributes to the existing body of knowledge by providing a nuanced understanding of AV disengagements, emphasizing the need for targeted interventions and strategies. The research serves as a pioneering step toward unveiling hidden patterns in AV disengagements, showcasing the potential of explainable AI in unraveling complex AV disengagement scenarios, and providing a solid foundation for future advancements in AV technology and policy formulation. The findings point to further investigation of AV disengagements to understand the safe operation of AVs and better predict their maturity, thereby aiding policymakers in making informed decisions about AV testing and deployment.

Parking management presents a complex challenge in urban cities, as a scarcity of parking spaces and the ever-increasing vehicular traffic have led to congestion, environmental pollution, and overall reduced urban productivity. Addressing the problem requires predicting the exact number of available parking spaces and categorizing parking occupancy levels. This study aims to achieve these tasks by employing machine learning models to accurately predict occupancy, thus optimizing parking resource allocation and enhancing the urban parking experience. A dataset derived from a college campus garage for a period spanning from January 2022 to June 2023 was used to analyze the performance of various predictive models, including random forest, decision tree, linear regression, and support vector machine. The models were compared using multiple evaluation metrics, and the results revealed that the random forest model was the most reliable. Its strong performance in regression analysis translated into precise estimations of available parking spaces. Similarly, its capability in classification analysis proved essential for categorizing parking occupancy into distinct levels, enhancing communication and streamlining decision-making processes. These findings are significant for improving parking management systems and contributing to the development of efficient and sustainable parking solutions in urban environments.

Accurate pedestrian trajectory prediction is crucial for enhancing pedestrian and autonomous vehicle safety. In comparison to the bird-eye view, ego-centric trajectory prediction is more challenging because the ego-motion of the camera can dynamically alter the scale of pedestrian bounding boxes. As the distance between the vehicle and pedestrians decreases, pedestrian bounding boxes are magnified, subsequently influencing the predicted pedestrian trajectories within the image plane. To address this challenge, the paper proposes a deep learning network based on LSTM encoder-decoder architecture that leverages trajectory, ego-vehicle motion information, and local-visual context as the additional input information. Specifically, the motion parameters of the in-vehicle camera are introduced to represent self-vehicle motion information at a coarse granularity level, reducing the impact of noise when representing self-vehicle speed by optical flow. Meanwhile, due to the disproportionate impact of pedestrians at different distances on the optimization process, the loss function is normalized using the actual width and height of the pedestrian’s bounding box, effectively reducing the weight assigned to pedestrians in proximity during optimization. The attention mechanism is utilized to better capture the long-term temporal variations. Several typical pedestrian trajectory prediction benchmark datasets are used for the algorithm validation, such as JAAD. According to the results, the proposed method outperformed the selected state-of-the-art algorithms with about 19% improvements in trajectory prediction accuracy.

This study had a primary objective of creating a versatile and scalable data dashboard prototype that highlights transportation and community health data, using El Paso as a case study. The dashboard serves as a centralized hub for various data sources within the city, facilitating visual analysis to uncover the intricate connections between these two crucial domains. It’s important to note that the primary aim was not to develop a final product for the city, but rather to establish a proof-of-concept, engage potential users, and conduct scenario analysis. The true impact of this study is expected to extend beyond the creation of the dashboard itself, as it has the potential to influence significant city-wide or regional policy changes based on the valuable insights derived from the comprehensive data analysis. This initiative aims to provide a robust tool for decision-makers and stakeholders to make informed choices that could ultimately benefit the community’s transportation and health outcomes.

The paper discusses a vision-based sensor methodology for tracking displacements in structures using videos and computer vision algorithms to detect and track displacements at various scales. This method is tested on a full-scale bridge column footage from earthquake simulations, demonstrating its ability to capture both global and large-scale displacements and detailed local deformations. The paper compares these video-based measurements with traditional sensor data, proving the method’s effectiveness. The study highlights the potential of video analysis in expanding data collection to areas not typically covered by sensors. It suggests that this method can enrich datasets and improve the accuracy of structural health monitoring.

This study pioneers the use of synthetic image datasets, generated via the Unity game engine, to train deep learning models for construction work zone scene understanding. This innovative approach simplifies data acquisition and ensures a rich, diverse dataset that includes various construction scenarios, both hazardous and typical. We created 12,360 images with accurate bounding box annotations, ensuring high-quality, consistent data crucial for model training and validation, and effectively addressing the ambiguities often found in real-world datasets. An object detection benchmark was established using this dataset alongside eight state-of-the-art object detectors. This benchmark thoroughly evaluates the performance of these detectors on a wide range of construction site images, enabling comparisons and analyses of different models. It highlights their respective strengths and weaknesses in construction site applications. Notably, YOLOv8-L demonstrated exceptional performance, achieving a mean average precision (mAP) of 70.7% on the validation set and 69.8% on the testing set. These results underscore the efficacy of synthetic datasets in training models for complex scene understanding. This integration of synthetic and real-world imagery has the potential to revolutionize scene comprehension in construction zones, significantly enhancing safety and efficiency in the construction industry.

The study utilizes deep learning technology to improve crack detection in tunnels, surpassing the limitations of manual inspection methods. The deep learning model identifies cracks in high-resolution tunnel images, detecting even minor defects often overlooked in manual inspections. The model ensures unparalleled accuracy and consistency and is customized to tackle the unique challenges of tunnel environments, such as complex surface textures and varying lighting. The approach’s effectiveness is evident from its remarkable performance metrics, including an F1 score of 0.873, a mean Intersection over Union (mIoU) of 0.794, a precision of 0.760, and a recall of 0.746. These exceptional results represent a significant advancement in tunnel inspection, leading to more reliable and comprehensive evaluations of structural integrity. The integration of deep learning technology for tunnel safety marks a crucial milestone in enabling more effective maintenance strategies and reducing risks associated with structural defects.

Bridge approach slabs often experience differential settlement due to a higher settlement of embankment than bridge abutment under traffic. Severe differential settlement produces bumps and distress at approach slabs and joints, leading to riding discomfort and safety risks to public travelers. Surveying the condition of approach slabs and joints is crucial to evaluate the field performance of current designs and countermeasures to reduce approach bumps and distress throughout the lifespan of approach slabs. In this study, a total of 377 bridges located across Oklahoma with different designs and conditions of approach slabs were surveyed. The recent Pave3D 8K system featuring sub-mm 3D laser technology and inertial profiler was used to collect the 2D/3D images and inertial data from these bridges for evaluating approach bump and distress. Based on the bump levels of all approach slabs, the performance of the slab design, including slab length and skew angle, and bump mitigation method via asphalt overlay were evaluated. The results show that (1) approach slabs with lengths of 60−80 ft yield the least level of bumps; (2) approach slabs with a skew angle of 60° or departure slabs with 120° show less severe bumps than those with other angles; and (3) asphalt overlay is an effective method to minimize the bumps at bridge ends connecting with flexible pavement, while the road, approach slab, and bridge deck constructed entirely of concrete exhibit the smoothest transition.

Most of the researchers who have explored the factors that affect the severity of crashes involving autonomous vehicles (AVs) by analyzing the data made available by the California Department of Motor Vehicles (CA DMV) have done so in general terms. A comprehensive investigation is lacking in the literature, and the purpose of this study is to remedy that by examining accidents and resulting injuries, if any, that involved at least one AV and identifying and categorizing the factors that affected the severity of the crashes. Strategies for addressing challenges such as public acceptance and unavoidable accidents are also suggested. The findings support the frequently touted belief that AVs improve traffic safety and show that they are not responsible for most of the collisions in which they are involved. This study will benefit AV manufacturers and policymakers by providing them with a better understanding of AV crashes and recommending strategies that can help manage the challenges of AVs.

Traffic-related fatalities are a significant global challenge that is primarily attributed to human error. Autonomous vehicles (AVs) hold the promise of mitigating such fatalities by removing the potential for human error and thereby enhancing road safety; however, they too are involved in crashes that result in injuries. Prior research, limited by the scope of the dataset and depth of analysis, has not fully explored these factors, resulting in a lingering lack of clarity about the cause of the crashes. This study analyzed 358 AV crash records from California’s Department of Motor Vehicles’ database for the period spanning from 2014 to July 2023. A bagging classifier model was utilized to predict the likelihood of injuries occurring in AV crashes, and various features of the vehicles were evaluated to determine their impact on crash outcomes. The findings revealed vehicle damage as the key predictor of injury risk, with a non-linear damage-injury relationship. Characteristics specific to manufacturers were also found to play an important role, demonstrating their differences in technology and safety features. The study underscored the influence of vehicle and collision types on injury occurrences, particularly highlighting the vulnerability of two-wheeler drivers and the prevalence of sideswipe and rear-end collisions. Additionally, it revealed pre-collision movement patterns, especially those involving stationary vehicles, as crucial risk factors. The impact of different road types, like streets and avenues, on crash severity was also noted, further enriching the analysis. The findings offer valuable implications for AV manufacturers, policymakers, and urban planners in enhancing vehicle safety features.

About 20% of total fatalities, reported by National Highway Traffic Safety Administration (NHTSA), in 2022 are pedestrians and pedal cyclists. In a lot of instances, pedestrian crashes and injuries are not accurately reported which resulted in lack of accurate vulnerable road user (VRU) data. As computer vision models evolve, there is an opportunity to collect more accurate road user data and develop real-time safety response systems. The Federal Highway Administration (FHWA) and Intelligent Transportation Systems Joint Program Office (ITSJPO) initiative, connected and automated vehicle education (CAVe-in-a-Box), allows for the seamless adaptation of existing computer vision and machine learning models used in transportation data analytics. At its core, it utilizes vehicle-to-everything (V2X) Hub and a cellular-vehicle-to-everything (C-V2X) road-side unit (RSU). The C-V2X framework supports the use of long-term evolution (LTE) and other wireless communication protocols using sidelink to establish robust communication between vehicles and infrastructure. Although in an early stage of implementation, C-V2X has been gaining popularity over the dedicated short-range communication (DSRC) protocol due to different channelization schemes, timing requirements, and a wireless architecture similar to that used by cell phones. Contrary to DSRC, C-V2X uses multiple carriers in blocks of smaller subframes. In this research study, CAVe-in-a-Box infrastructure and mobile kits are assembled and integrated with a real-time computer vision-based pedestrian detector to detect, localize, and classify VRUs, such as pedestrians and bike-pedestrians in addition to vehicles. Furthermore, real-time alerts are generated in two ways: (1) by sending pedestrian detection calls to a traffic signal controller via synchronous data link control (SDLC) and (2) by transmitting V2X personal safety messages (PSMs) to RSUs.

The escalating apprehension regarding environmental issues is driving nations to adopt renewable energy technologies to mitigate the release of greenhouse emissions from vehicles. Electric vehicles (EVs) provide a practical and eco-friendly alternative that can help transition to an environmentally friendly, low-emissions transportation system while protecting the environment. This study employed the theory of planned behavior (TPB) and incorporated additional factors such as price value and moral norms to examine consumers’ intention to adopt EVs. A survey was administered to prospective consumers in March 2023, and the 733 responses were analyzed, using a structural equation model (SEM), to determine which factors have the greatest impact on EV adoption. The results revealed that attitudes, perceived behavior control, subjective norms, moral norms, and price value positively and significantly influenced consumers’ intentions to adopt EVs.

This study investigates the impact of traffic signal failures on traffic flow, safety, and congestion at a signalized intersection with high rate of crashes in Baltimore City. Analyzing data from a malfunctioning signal, collected through a LiDAR sensor during a temporary outage, the research evaluates vehicle and pedestrian counts, V2V and V2P conflicts, and jaywalking events. Findings revealed that malfunctioning signals disrupt traffic flow and safety, emphasizing the necessity of adaptive signal systems. The analysis underscores the role of signal optimization in reducing congestion and improving traffic flow, promoting sustainability in urban transportation. Ultimately, this study emphasizes the crucial role of traffic signals in managing congestion and highlights the need for innovative, evidence-based strategies in urban planning to create more efficient and sustainable transportation systems. Policymakers and transportation authorities can use the insights gained to develop improved traffic management solutions to improve traffic safety at signalized intersections.

With the rise in highway maintenance demands, work zones have proliferated across the United States, posing traffic and safety challenges. Dissemination of work zone information to the public is essential to improve traffic flow and efficiency in work zones. Central to the success of smart work zones (SWZ) public information is real-time data availability. This study compares two alternative probe data types to ground truth data from field sensors. The two emerging probe vehicle datasets tested in this study are connected vehicle (CV) data supplied by a third-party data vendor and segment average speeds (for each minute) obtained from a different vendor. The mean and minimum speeds of the probe data types were analyzed using frequency distributions and descriptive statistics. The CV data displayed a high amount of variability in minimum and mean speeds in both travel directions compared to segment average speeds, making them more similar to the ground truth data. This finding is useful in providing precise real-time information to the driving public and will improve the accuracy with which work zone information is relayed.

Machine learning models are increasingly being utilized to develop predictive models for Transportation Systems Management and Operations (TSMO) applications. These models are often assessed based on a global performance metric that evaluates the model’s performance when the entire testing dataset is presented to the model. A TSMO application is expected to perform reliably and consistently in various situations and roadway conditions. The reliability and consistency of the model predictions for various scenarios are critical to the success of transportation agencies’ efforts to address mobility and safety issues. Performance bias might be influencing the model when a model’s performance is inconsistent for different scenarios. This paper investigates the performance bias that the traffic management center may face when applying machine learning methods to predict incident clearance time. Additionally, dimension-reduction techniques are employed as mitigation techniques in the model development process. This paper investigates the impact of two common dimension-reduction methods, important feature selection and principal component analysis, on performance bias. In a case study, this paper investigates the performance bias of RF, BRNN, SVR, KNN, XGB, GP, and NNET in incident clearance time prediction. Incident data from three interstate corridors in Missouri, USA, were utilized to develop and evaluate the models. Repeated k-fold cross-validation was used to prepare 20 training and testing sets to demonstrate and assess the learners’ performance variations due to data splits. The results indicated that the seven learners suffered from performance bias. The analysis of the impact of dimension-reduction models revealed that the important feature selection method did not significantly mitigate the performance bias. On the other hand, the principal component analysis method significantly mitigated this bias for all learners, with poor-performing learners gaining the most improvements. In addition to contributing to reducing the performance bias, the principal component analysis significantly reduced the learners’ global (i.e., overall) error metrics.

Incident clearance time prediction is a key task in Advanced Transportation Management and Advanced Traveler Information Systems, as it informs mitigation and response strategies adopted by transportation agencies. Incident clearance time prediction is not a trivial task given that the duration of an incident is influenced by various factors, some of which are difficult to measure or not measured at all. Such constraints in the data might limit the accuracy of the models developed for predicting clearance time. Over the past decades, researchers have employed various statistical methods and machine learning approaches to overcome these limitations and improve the accuracy of the incident clearance time prediction models. While these efforts have successfully improved the overall accuracy of models, very few studies have focused on assessing and improving the generalization ability of incident clearance time prediction models. This paper proposes the hierarchical classification-regression (HiClassR) framework to improve the predictive performance of common machine-learning algorithms in predicting incident clearance time. HiClassR is a framework for developing a bi-level model in which the incidents are first assigned to an incident class; then, a class-specific model estimates the incident clearance time. While the HiClassR method is expected to improve error measures such as the RMSE over the entire dataset, it is also expected to improve the model’s generalization ability across various incident scenarios. In a case study, the HiClassR method was applied to six conventional predictive methods to develop incident clearance time prediction models for St. Louis, MO, USA. The case study results indicate that the HiClassR model consistently outperforms the conventional base models and effectively generalizes the overall prediction power of the model to various incident categories of different conventional predictive methods. On average, the preferred HiClassR model reduced the RMSE from 22.7 to 17.6 min for the random forest model, from 22.5 to 19 min for Bayesian regularization neural networks, from 23.7 to 16.9 min for support vector machines, from 25.2 to 21.1 min for K-nearest neighbors, from 24.5 to 19.7 min for XGB, and from 23.9 to 18.3 min for neural networks.

The development of electric vehicles (EVs) has accelerated in recent years, in part because of significant governmental support. Despite this support, however, the market share of EVs in the United States remains relatively small, and a large majority of consumers exhibit reluctance toward adopting them. This study aims to facilitate the effective development of EVs by identifying the barriers that affect their adoption in the United States and analyzing responses to an online questionnaire designed to provide insight into various perspectives of the barriers. A literature search identified 17 barriers that served as the basis of the survey questionnaire. A total of 733 individuals provided complete responses, and cluster analysis was performed. The results revealed three cluster segments and indicated that middle-aged males who possess higher levels of education and income exhibit greater levels of interest in the adoption of electric vehicles. The outcomes of this study may offer direction to policymakers in formulating efficacious energy and transportation policies and guide designers of electric vehicles to align the requirements and preferences of prospective consumers with their designs.

This study explores the growing popularity of smartwatches since 2015, particularly their application in various transportation sectors. Utilizing the PRISMA guideline, the research systematically reviews 33 articles focusing on smart wearable devices, with an emphasis on smartwatches. The evaluation encompasses goals, methods, findings, advantages, and disadvantages, drawing comparisons with smartphones. Studies are categorized into safety and driver behavior, health and physical activities, equity, comparison of smartwatches and smartphones, and workers and construction. Results indicate a surge in research interest in safety, driver behavior, and health-related aspects of smartwatches, with a noticeable rise in popularity in recent years. The study highlights the potential for smartwatches, given advancements in smartphone health sensors, especially in examining the physical activity and health issues of cyclists, scooter riders, and pedestrians in transportation engineering. This comprehensive review aids researchers in selecting quantification methods aligned with their objectives and study constraints.

The long-term monitoring of transportation infrastructure assets, such as pavements, highway embankments, and earth retaining systems, at a lower cost, and with short mobilization time, is of interest to owners and stakeholders. Due to the improvement in spatial and temporal resolution of synthetic aperture radar (SAR) remote sensing systems as well as the drastic reduction in the price of data acquisition, SAR has now become a viable method to provide an economical and rapid condition assessment of transportation assets. This paper presents an example of this through a case study on the inspection and characterization of a pavement surface based on the amplitude of backscattering from an X-band radar. In situ characterization of the test site was first performed using aerial photogrammetry techniques with uncrewed aerial vehicle (UAV). The pavement features, extracted from the digital twin model created using photogrammetry, were compared with the radar cross section (RCS) of the pavement. The results show that problematic areas evident during in situ characterization can be delineated and quantified based on the normalized radar cross-section of the pavement surface. Overall, the outcomes of this research demonstrate the potential of SAR for future transportation asset management undertakings and can be of significant interest to state engineers and practitioners.

The authors have conducted detailed pavement inspections at over 100 airports using uncrewed aerial systems with high resolution photogrammetry to identify pavement distresses and calculate pavement condition index using the formula of ASTM D-5340. New aerial vehicles with new 100-megapixel cameras allow the team to collect the same 1.5 mm pixel GSD imagery three times faster than the previously reported data collection system. The team uses artificial intelligence coupled with deep machine learning models to automatically identify both distress type and distress severity levels. The large library of distresses from 140 airports that have previously been flown and verified has greatly improved the accuracy of the distress identity and distress severity determination. Although the AI automatically generated distress determinations are reviewed by technicians, the quality of the AI generated distresses and severities have greatly improved from what we reported two years ago. The result is that distresses that may have gone unnoticed with a visual scan are being discovered and correctly identified by the AI. Although an FAA report has claimed that the UAS technology is not progressed enough to identify several low severity distresses (alkali-silica reaction, corner spalling, joint spalling, joint seal damage, depression, raveling, swell and weathering), the authors’ analyses using AI and deep machine learning have identified each of these low severity distresses without additional imagery. Now that the team is scanning many airports for a second time, we are using change detection at the pixel level to see changes that are happening to distresses and distress severities. Using AI and change detection software together has yielded important results. It provides the ability to visually see in what geospatial area specific distresses are increasing. A change of 1−3 PCI points is normally expected in year-to-year inspections. However, determining what types of distress changes are taking place is much more meaningful to the pavement management program (PMP).

Airfield pavement inspection and maintenance are critical aspects of aviation infrastructure, representing a substantial portion of life-cycle costs. Longitudinal, transverse, and diagonal (LTD) cracks; corner breaks; shattered slabs in Portland cement concrete (PCC) pavement; and longitudinal and transverse (L&T) cracks of asphalt concrete (AC) pavement consist of most of the airfield pavement distresses. Traditional airfield pavement inspection methods are manual, time-consuming, laborious, and reliant on the inspector’s experience, leading to increased expenses and safety risks. This research explores the potential to automatically identify those distresses in red-green-blue images using four variants of deep learning (DL) model YOLOv8, ranging from nano to large. YOLOv8 is a widely used off-the-shelf DL object detection model that allows rapid training and easy execution. A DL training dataset of 5,273 small uncrewed aircraft systems (sUAS) collected images was developed. The transfer learning technique was used, and the dataset passed through each model 100 times for adequate training. The model exhibits mean average precision values exceeding 0.65, with varying processing times. Such accuracy showed that crack-related distress detection using DL models could enhance airfield pavement inspection efficiency.