Safety through Disruption (SAFE-D) University Transportation Center (UTC)

Permanent URI for this collection

https://hdl.handle.net/10919/86341

Browse

Now showing 1 - 20 of 102

Development of an Infrastructure Based Data Acquisition System to Naturalistically Collect the Roadway Environment
Sarkar, Abhijit; Papakis, Ioannis; Herbers, Eileen; Viray, Reginald (2023-12)
Automatic traffic monitoring is becoming an important investment for transportation specialists, especially as the overall volume of traffic continues to increase, as do crashes at intersections. Infrastructure cameras can be a good source of information for automatic monitoring of traffic situations at intersections. However, efficient computer vision methods that can process the video data effectively are required for this endeavor. Intersection cameras often record video data that are of low quality and low frame rate, making them challenging to use. In this project, we have demonstrated how traffic cameras can be used to automatically track roadway agents, find their kinematic behavior, and devise a safety measurement strategy, leveraging recent advancements in computer vision and deep learning. In this process, we have specifically focused on the Virginia Beach area and used publicly available traffic data to demonstrate our results. We developed a full computer vision pipeline that trains a custom object detector specifically using traffic data. We also used an optical flow method and a graph neural network to improve the accuracy of object tracking. The tracked objects from the image frames were further used as a point source and mapped to their GPS locations. Finally, the speed of each object was calculated to understand the traffic dynamics and determine possible crash predictors. This information can be used to quickly alert traffic control operators to a specific intersection that likely needs their attention so that crashes can be mitigated.
Investigating and Developing Methods for Traditional Participant-based Data Collection with Remote Experimenters
Miller, Marty (2023-11-09)
This project investigated and developed methods and technologies to allow experimenters to conduct and monitor data collection from a remote location. The technologies developed consist of a desktop application that allows researchers to view the vehicle cabin and various vehicle parameters (like speed, acceleration, etc.) in real time. Physically removing researchers from the vehicle during experiments can increase realism and offer naturalistic observations in traditional, experimenter-conducted studies. The remote experimentation methods and technologies developed can be particularly helpful for studying automated driving by creating a more natural environment while still maintaining oversight and control of the experiment.
Autonomous Vehicles for Small Towns: Exploring Perception, Accessibility, and Safety
Li, Wei; Ye, Xinyue; Li, Xiao; Dadashova, Bahar; Ory, Marcia G.; Lee, Chanam; Rathinam, Sivakumar; Usman, Muhammad; Chen, Andong; Bian, Jiahe; Li, Shuojia; Du, Jiaxin (Safe-D University Transportation Center, 2023-09)
As of 2021, there were 18,696 small towns in the US with a population of less than 50,000. These communities typically have a low population density, few public transport services, and limited accessibility to daily services. This can pose significant challenges for residents trying to fulfill essential travel needs and access healthcare. Autonomous vehicles (AVs) have the potential to provide a convenient and safe way to get around without requiring human drivers, making them a promising transportation solution for these small towns. AV technology can become a first-line mobility option for people who are unable to drive, such as older adults and those with disabilities, while also reducing the cost of transportation for both individuals with special needs and municipalities. The report includes our research findings on 1) how residents in small towns perceive AV, including both positive and negative aspects; 2) the impacts of ENDEAVRide—a novel “Transport + Telemedicine 2-in-1” microtransit service delivered on a self-driving van in central Texas—on older adults’ travel and quality of life; and 3) the potential safety implications of AVs in small towns. This report will help municipal leaders, transportation professionals, and researchers gain a better understanding of how AV deployment can serve small towns.
Simulation-based approach to investigate the electric scooter rider protection during traffic accidents
Untaroiu, Costin D.; Untaroiu, Alexandrina; Chontos, Rafael; Grindle, Daniel (Safe-D University Transportation Center, 2023-12)
The recent emergence of electric scooter (e-scooter) rideshare companies has greatly increased the use of escooters around the world, which has increased the number of injuries associated with their use. A primary cause of e-scooter crashes is front-wheel collisions with a vertical surface. This research numerically simulated various e-scooter-stopper crashes across different impact speeds, approach angles, and stopper heights to characterize their influence on rider injury risk during falls. A finite element (FE) model of a standing Hybrid III anthropomorphic test device was used as the rider. The angle of approach was found to have the greatest effect on injury risk to the rider. Additionally, arm bracing was shown to reduce the risk of serious injury in two thirds of the impact scenarios. Most e-scooter rider fatalities are recorded in collisions between a car and an e-scooter. Therefore, crashes between an e-scooter and a sedan and between an e-scooter and a sports utility vehicle were simulated using FE models. The vehicles impacted the e-scooter at a speed of 30 km/hr in a perpendicular collision and at 15° towards the vehicle. The risks of serious injury to the rider were low for the head, brain, and neck, but femur/tibia fractures were observed in all simulations.
Enhancing Automated Vehicle Safety Through Testing with Realistic Driver Models
Garcia, Alfredo; McDonald, Anthony D. (Safe-D University Transportation Center, 2023-12)
Driver process models play a central role in the testing, verification, and development of automated and autonomous vehicle technologies. Prior models developed from control theory and physics-based rules are limited in automated vehicle applications due to their restricted behavioral repertoire. Data-driven machine learning models are more capable than rule-based models but are limited by the need for large training datasets and their lack of interpretability. In this project we developed a novel car following modeling approach using active inference, which has comparable behavioral flexibility to data-driven models while maintaining interpretability. We assessed the proposed model, the Active Inference Driving Agent (AIDA), through a benchmark analysis against several benchmarks. The models were trained and tested on a real-world driving dataset using a consistent process. The testing results showed that the AIDA predicted driving controls significantly better than the rule-based Intelligent Driver Model and had similar accuracy to the data-driven neural network models in three out of four evaluations. Subsequent interpretability analyses illustrated that the AIDA's learned distributions were consistent with driver behavior theory and that visualizations of the distributions could be used to directly comprehend the model's decision-making process and correct model errors attributable to limited training data.
Developing a Framework for Prioritizing Bicycle Safety Improvement Projects using Crowdsourced and Image-Based Data
Sadeghi, Amir Reza; Jahangiri, Arash; Machiani, Sahar Ghanipoor; Hankey, Steve; Abdollahpour, Seyed Sajjad (Safe-D University Transportation Center, 2023-08)
Active transportation, including walking and cycling, has gained popularity due to the economic, environmental, and energy-efficient benefits. However, the rise of active transportation has also led to an increase in fatalities, particularly for bicyclists. A crash-risk scoring method was proposed to prioritize bicycle safety improvement projects for 50 bridges located in San Diego County. This study employs surrogate safety measures to estimate crash risk, addressing the limitations of traditional data collection methods, and incorporates transportation equity factors into the safety measure scoring method. To identify significant factors contributing to the likelihood of bicyclists exceeding 10 mph on bridges, binomial logistic regression models were employed, with three models focusing on different predictor variables. The results showed that factors such as race, home-to-work travel patterns, education levels, and crime rates influenced bicyclists' speeds on bridges. This study provides a foundation for understanding the factors associated with bicyclists' speeds on bridges and can inform future safety improvement projects in San Diego County and beyond. The findings highlight the importance of considering a range of factors to improve bicyclist safety and can ultimately lead to safer and more equitable transportation for all.
Developing an Intelligent Transportation Management Center (ITMC) with a Safety Evaluation Focus for Smart Cities
Salehipour, Sina; Jahangiri, Arash; Paolini, Christopher P.; Machiani, Sahar Ghanipoor; Bergcollins, Django (Safe-D University Transportation Center, 2024-01)
In the context of smart cities, ensuring transportation safety is a complex task that involves understanding the impact of new technologies, measuring the effectiveness of safety measures, and identifying high-risk locations. However, recent advances in communication and big data analytics have made it possible to address these challenges in a more efficient manner. Traditional transportation management centers (TMCs) are limited in their ability to analyze large amounts of data for safety evaluation. To overcome this limitation, this project aims to develop an intelligent transportation management center (ITMC) that utilizes automated video analysis to assess safety. By leveraging Intelligent Transportation Systems (ITS) technologies and big data analytics, the proposed ITMC can proactively evaluate safety at signalized intersections. Unlike conventional methods that rely on crash data, the ITMC uses safety surrogate measures (SSMs) to identify near-crash situations and calculate proactive risk. In this study, the results obtained from a machine vision model were used along with the Post Encroachment Time (PET) safety surrogate measure (SSM) to assess safety proactively at a selected signalized intersection. The study utilized the latest YOLO series model, YOLOX, for deep learning to detect and classify road users in video frames from four intersection traffic cameras.
Critical Areas in Advanced Driver Assistance Systems Safety: Point of Sale and Crash Reporting
Goddard, Tara B. (Safe-D University Transportation Center, 2023-09)
Automated assistive vehicle technologies vary from simple alerts to partially automated driving tasks that are increasingly available in today’s vehicles. Advanced driver assistance systems (ADAS) seek to alert a driver to critical events (e.g., forward collision warning) or even intervene (e.g., emergency braking, lane-keeping steering) to prevent crashes. These technologies, however, are not available equally across the passenger vehicle fleet, nor is there standardization in how they are marketed to potential buyers or demonstrated at point of sale, including by increasingly popular online “dealerships” like Vroom and Carvana. The proliferation of ADAS has also outpaced current crash investigation forms. The current Model Minimum Uniform Crash Criteria includes limited guidance on crash avoidance technologies and most state crash reports do not include ADAS variables. Realizing the full benefit of ADAS relies on salespeople, consumers, and law enforcement understanding their benefits and limitations in improving traffic safety. This project investigated the state of knowledge and current practices on how ADAS technologies are marketed and sold, how ADAS are notated in crash reports, and what existing crash data reveal about ADAS in crash involvement to help illuminate and address gaps in current pre-sale and postcrash ADAS research.
Development of a Roadside LiDAR-Based Situational Awareness System for Work Zone Safety: Proof-of-Concept Study
Wu, Jayson (Dayong); Le, Minh; Ullman, Jerry; Huang, Tianchen; Darwesh, Amir; Saripalli, Srikanth (Safe-D University Transportation Center, 2023-09)
Roadway construction and maintenance have become increasingly common as the U.S. transportation system ages and the population and traffic volume increase. This places more and more work zone workers near high-speed vehicles and increases the probability of being struck by them. This project innovatively deployed 360-degree LiDAR sensors at the roadside and tested their potential to provide work zone safety in terms of detection accuracy, efficiency, and ease of use. Researchers developed a set of algorithms to collect and interpret real-time information for each approaching vehicle and worker (e.g., location, speed, and direction) in and outside work zones using roadside LiDAR. Ultimately, the outcome of this pilot study could lead to developing a full-scale warning system deployable in a real work zone environment. Such a system could detect and analyze live traffic and work zone activity, activate the appropriate warning scheme, and deliver information to roadway workers in work zones in a timely manner so they can take evasive actions instead of relying on traditional “passive” safety countermeasures. This kind of panoramic, trajectory-level data for work zone actors can be used to develop a next-generation work zone situational awareness system.
Characterizing Level 2 Automation in a Naturalistic Driving Fleet
Perez, Miguel A.; Terranova, Paolo; Metrey, Mariette; Bragg, Haden; Britten, Nicholas (Safe-D University Transportation Center, 2024-01)
Introducing automation into the vehicle fleet disrupts how vehicles operate and potentially affects what drivers do with these features and expect from vehicle performance. Therefore, it is imperative to study driver adaptations in response to these innovations. This investigation leveraged 47 vehicles from the Virginia Tech Transportation Institute Level 2 (L2) Naturalistic Driving Study to analyze driver behavior with L2 automation features. Results showed no sizeable differences between periods of L2 feature usage and general driving periods with respect to time-of-day and calendar-related metrics. Most L2 feature usage occurred on motorways, following design expectations. L2 features were activated for 7.2 minutes in trips lasting an average of 22.8 minutes, or about 32% of the L2 trip duration. Driver-initiated overrides were predominantly done by braking or accelerating the vehicle, with steering-based overrides being minimal and likely involving lane changes without using a turn signal. Intervention requests were the most common takeover request, followed by requests due to insufficient driver hand contact with the steering wheel. Findings suggest that as L2 features penetrate the U.S. fleet in non-luxury consumer vehicles, system usage will be common and comparable with previous findings for luxury offerings. While evidence of potential system misuse was observed, future work may further operationalize system misuse and assess the prevalence of such behaviors.
Evaluation of Eyes Off Road During L2 Activation on Uncontrolled Access Roadways
Anderson, Gabrial T.; Glaser, Yi; Klauer, Charlie (2024-01)
The current study investigated eyes-off-road (EOR) behavior of drivers when traveling on uncontrolled access roadways in vehicles equipped with SAE Level 2 (L2) automated features. Previously collected naturalistic driving data were analyzed. Events were split between L2 features being active or available but inactive and matched across a spectrum of criteria (e.g., time of day). Primary analyses focused on L2 activation status and intersection type (no intersection, straight through intersection, and turning) and any interaction between those variables. EOR glances were operationalized in two ways: EOR 1, only forward was considered on road; and EOR 2, all driving-related glances were considered on road. EOR metrics involved total EOR, mean EOR, single longest glance, and number of glances per event. Overall, results for the primary research questions indicated that EOR behavior was higher when L2 was active across all EOR metrics, that intersection type affected EOR behavior on some metrics, and that there was an interaction between these variables for select metrics. Ancillary analyses represented differences for single longest glance when excluding slower speed segments, higher EOR behavior when speeds were below 37 mph, and increased hands-off-wheel behavior when L2 systems were active.
A Diagnostic Machine Learning Model for Air Brake Systems in Commercial Vehicles
Darbha, Swaroop; Rajagopal, K. R. (Safe-D University Transportation Center, 2023-11)
Safely introducing autonomy to trucks requires monitoring their brake systems continuously. Out-of-adjustment push rods and leakages in the air brake system are two major reasons for increased braking distances in trucks, resulting in safety violations. Air leakages can occur due to small cracks or loose/improperly fit couplings, which do not affect overall braking capacity but contribute greatly to increased braking lag and reduced maximum braking torque at the wheels. Similarly, an increased stroke of push rod leads to a larger delay in brake response and a smaller brake torque value at the wheels. Currently, an air brake system’s condition is monitored manually by measuring the push rod offset and inspecting the system’s couplings and hoses for air leakages. These inspections are highly labor intensive, subjective, time consuming, and inaccurate in quantifying adversely affected braking systems. An onboard diagnostic device that can monitor air brake health would be crucial in preventing road accidents. The focus of this report is to help develop a diagnostic system that facilitates enforcement and pre-trip inspections and continuous onboard monitoring of trucks by developing a model for its multi-chamber braking system using machine learning; this model can be used to estimate the severity of leakage and the push rod stroke using real-time brake pressure transients. The novel approach of a gradient descent model that predicts the air brake system air leakage rate using pressure transients at the brake chamber was developed and experimentally corroborated.
Using Health Behavior Theory and Relative Risk Information to Increase and Inform Use of Alternative Transportation
Glenn, Laurel; Sinha, Nishita; Dopp, Lia; Shipp, Eva; Jiles, Kristina; Edwards, Samantha; Hosig, Kathy; Wu, Lingtao; Villani, Domenique; Quint, Nicholas; Ogieriakhi, Macson; Perez, Marcelina; Woodson, Caitlin; Martin, Michael; Ramezani, Mahin (2024-01)
One of the main goals of the Virginia Tech (VT) Alternative Transportation Department is encouraging the VT community (including students, faculty, and staff) to walk, use the bus, carpool, or ride bicycles for alternative transportation to decrease dependency on vehicle use and traffic around campus and increase overall safety. This project develops an intervention and education program to encourage alternative transportation to, from, and around campus to reduce campus traffic. In addition, since there is currently no standardized approach for computing the injury rates for non-vehicle roadway users, this project also refines and assesses a methodology for estimating injury rates for pedestrians and pedalcyclists, which was used to inform the developed educational alternative transportation safety course.
Micromobility Regulation: Best Practices Review
Stoeltje, Gretchen; Hansen, Todd; Hwang, William; Geislebrecht, Tina (Safe-D University Transportation Center, 2023-09)
As rented and shared micromobility options, e-scooters are new and potentially transformative app-based modes that promise to alleviate first mile/last mile mobility issues, congestion, and more. Yet their safe deployment has not yet been systematically understood or standardized by users, cities, or operators. From 2017 to 2021, 267,700 people were treated for injuries in emergency departments and 129 were killed in micromobility product-related crashes. These devices are not yet regulated by a federal agency like the National Highway Transportation Safety Administration or the Consumer Product Safety Commission, and their use is not uniformly regulated at the municipal level. Some jurisdictions are imposing strict regulations across a region, regardless of density levels or urban design, while others have not imposed any rules at all. Without further understanding of what constitutes effective local regulation, the safe operation of these devices may not improve. This project explores what types of regulations municipalities and regions are imposing in an effort to address the safe deployment of these micromobility options.
Evaluating the Safe Routes to School (SRTS) Transportation Program in Socially Vulnerable Communities in San Diego County, California
Fernandez, Gabriela; Etaati, Bita; Mercado, Andrick; Jahangiri, Arash; Machiani, Sahar Ghanipoor; Tsou, Ming-Hsiang; Mejia, Christian (Safe-D University Transportation Center, 2023-05)
Child safety concerns are among the strongest impediments to children walking or biking to school, but some students must walk or bike due to financial or other circumstances. These travel modes are more than twice as common among students from low-income households than students from higher income households. The Safe Routes to School (SRTS) program fosters opportunities for students to walk and bike to school safely and routinely. This study provides insights into the SRTS program’s effectiveness and potential to improve walking and biking safety in socially vulnerable communities by evaluating the program’s impact on schools in the Chula Vista Elementary School District, a vulnerable area in San Diego County. (i) A linear regression model was used to assess the program’s impact on each school, and a logistic regression model was employed to identify factors influencing students’ walking behavior. (ii) An SRTS web-based interactive tool (ArcGIS Experience) was developed to identify traffic incident hot spots and facilitate future routing improvements. (iii) A virtual reality (VR) road safety training tool for children was developed, and a case study at Feaster Charter Elementary School was conducted to assess its effectiveness. Twenty-six students played the VR game before and after watching traffic safety educational videos, and observations from the VR session were recorded. (iv) The outreach and deliverables from this study strengthened community collaboration across San Diego County.
Developing AI-Driven Safe Navigation Tool
Das, Subasish; Sohrabi, Soheil; Tsapakis, Ioannis; Ye, Xinyue; Weng, Yanmo; Li, Shoujia; Torbic, Darren (Safe-D University Transportation Center, 2023-09)
Popular navigation applications such as Google Maps and Apple Maps provide distance-based or travel timebased alternative routes with no real-time risk scoring. There is a need for a real-time navigation system that can provide the data-driven decision on the safest path or route. By leveraging data from a diverse range of historical and real-time sources, this study successfully developed a user interface for a navigation tool or application that offers informed and data-driven decisions regarding the safest navigation options. The interface considers multiple scoring factors, including safety, distance, travel time, and an overall scoring metric. This study made a distinctive and valuable contribution by designing and implementing a robust safe navigation tool driven by artificial intelligence. Unlike existing navigation tools that offer multiple uninformed route options, this tool provides users with an informed decision on the safest route. By leveraging advanced AI algorithms and integrating various data sources, this navigation tool enhances the accuracy and reliability of route selection, thereby improving overall road safety and ensuring users can make informed decisions for their journeys.
Building Equitable Safe Streets for All: Data-Driven Approach and Computational Tools
Dadashova, Bahar; Zhu, Chunwu; Ye, Xinyue; Sohrabi, Soheil; Brown, Charles; Potts, Ingrid (Safe-D University Transportation Center, 2023-09)
Roadway safety in low-income and ethnically diverse U.S. communities has long been a major concern. This research was designed to address this issue by developing a data-driven approach and computational tools to quantify equity issues in roadway safety. This report employed data from Houston, Texas, to explore (1) the relationship between road infrastructure and communities’ socioeconomic and demographic characteristics and its association with traffic safety in low-income, ethnically diverse communities and (2) the type of driver behaviors and characteristics that affect crash risks in underserved communities. The team first built an inclusive road infrastructure inventory database by employing remote sensing and image processing techniques. Then, the relationship between communities’ socioeconomic and demographic characteristics and traffic safety was investigated through the lens of road infrastructure characteristics using data mining, deep learning tools, and statistical and econometric models. Clustering analysis was used to uncover the role in underserved communities of socioeconomic and demographic characteristics of drivers and victims involved in crashes. Structural equation models were then used to explore the association between neighborhood disadvantage, transportation infrastructure, and roadway crashes. Findings shed light on road safety inequity and sources of these disparities among communities using data-driven methods.
Countermeasures to Detect and Combat Inattention While Driving Partially Automated Systems
Paras, Carolina Rodriguez; Ferris, Thomas (Safe-D University Transportation Center, 2023-09)
Vehicle manufacturers are introducing increasingly sophisticated vehicle automation systems to improve driving efficiency, comfort, and safety. Despite these improvements, partially and fully automated vehicles introduce new safety risks to the driving environment. Driver inattention can contribute to increased risk, especially when control transfers from automation to the human driver. To combat inattention and ensure safe and timely transitions of control, this study investigated the effectiveness of a vehicle cuing system that engages different sensory modalities (e.g., visual, auditory, and tactile) and both simple and complex cue messages to announce the need for manual takeover. Twenty-four participants completed a driving simulator study involving scripted driving sections with and without partial automation. Participants navigated six scripted automation failure events, some preceded by takeover cues. Measures of driving performance, safety, secondary task performance, and physiological indices of workload did not differ significantly based on display type or complexity. However, a clear trend showed that, compared to events not associated with takeover cues, driver reaction time to automation failure is substantially faster when preceded by cues of any type or complexity. This study provides evidence of the benefit of supporting driver situational awareness, safety, and performance by issuing cues and guiding drivers in taking control when the vehicle system predicts a likely automation failure.
Connected Vehicle Data Safety Applications
Martin, Michael; Wu, Lingtao; Ramezani, Mahin; Li, Xiao; Turner, Shawn; Stutes, Sophia; Hasan, Faiza; Potter, Michael (2023-09)
The large-scale assessment of how driving behavior affects traffic safety and ongoing surveillance is hindered by data collection difficulties, small sample sizes, and high costs. Connected vehicles (CV) now offer massive volumes of observed driving behavior data from newer vehicles with myriad electronics and sensors that monitor the state of the vehicle, environmental conditions, and the driver’s actions. This project evaluated the viability of CV data in roadway safety applications with the objective of improving existing predictive crash methods, measuring traffic speed and its relationship to crashes, and determining whether CV data could be used to evaluate pavement marking products. The research team developed safety performance functions (SPFs) for rural two-lane segments and urban intersections in Texas. The results showed that the SPFs improved with the addition of hard braking and hard acceleration counts in a majority of areas. Further, a variety of CV speed measures were generated from the CV data and were shown to have conflicting correlations with crash risk and counts. Lastly, the research team developed the data processing methods for evaluating pavement marking products but was unable to perform an evaluation due to the lack of detailed construction project records.
E-Scooter Safety Assessment and Campus Deployment Planning
White, Elizabeth; Mollenhauer, Michael A.; Robinson, Sarah; Novotny, Adam (Safe-D University Transportation Center, 2023-12)
E-Scooters are a popular new service that provide last mile transportation, but there are reports of safety concerns for riders and impingement on other users of rights of way. Little formal research has been conducted on E-Scooter safety or the optimal approach to deployment to decrease nuisance issues. To address this, VTTI and Spin deployed a fleet of E-Scooters on the Virginia Tech campus through an exclusive, controlled research program. Through on-scooter data acquisition systems, fixed infrastructure cameras, anecdotal injury reports, and surveys, data was collected to assess safety impact as well as to understand beneficial and problematic user behaviors and patterns for subsequent countermeasure development and deployment recommendations. The resulting naturalistic dataset includes over 9,000 miles of riding data. Overall, the E-Scooter deployment on the Virginia Tech campus was safer than other reported deployments. The operational constraints that were put in place were largely effective, and with the additional results from this study, some additional constraints and expanded outreach programs may make future deployments even safer. The campus community largely considered the deployment of E-Scooters a clean alternative transportation option and viewed the service favorably.

Browse

Recent Submissions