Abstract
Most light-duty vehicle (LDV) crashes occur due to human error. The National Highway Safety Administration (NHTSA) reports that eight percent of fatal crashes in 2018 were distraction-affected crashes, while close to ninety-four percent of all crashes occur in part due to human error. Crash avoidance features could reduce both the frequency and severity of light and heavy-duty vehicle crashes, primarily caused by distracted driving behaviors and/or human error by assisting in maintaining control or issuing alerts if a potentially dangerous situation is detected. As the automobile industry transitions to partial vehicle automation, newer crash avoidance technologies are beginning to appear more frequently in non-luxury vehicles such as the Honda Accord and Mazda CX-9. Additionally, the market penetration of electric vehicles (EVs) is increasing, in turn increasing the weight and size of vehicles on the road. However, the patterns and characteristics of crashes involving EVs or vehicles automated features have not been explored in much detail. This project develops a replicable, open, deployable model that can: 1) assess the distribution of crashes with automated features across factors such as weather conditions, vehicle speed, crash severity, pre-crash movement, facility type, and time of day, 2) estimate the relationship between contributing factors and the severity of crashes involving vehicles with automated features using regression analysis, and 3) assess the patterns and characteristics of crashes involving EVs.
Our hypothesis is that current and past automated vehicles (AV) perform well in certain driving scenarios, facility types, and weather conditions but not others. However, there are not many frameworks and tools to help federal, state, and local agencies better understand the causes of crashes involving AVs, the locations of crash hotspots, and the infrastructure improvements and policies that could enhance road safety with AVs. Additionally, this study should help technology providers better understand under which scenarios and conditions further testing and technology development is needed to improve safety and performance. We also hypothesize that the characteristics and patterns of crashes involving EVs differ from crashes not involving this technology.
Most studies that have assessed the contributing factors in crashes involving vehicles with automated features have mostly focused on California (e.g., Kutela et al., 2022; Liu et al., 2021, 2024; Xu et al., 2019). Our study will add to the previous literature due to the overall comprehensiveness of the crash data, which is representative of all 50 states and the District of Columbia.
In the first part of this project, we will compile crash data on advanced driver assistance systems (i.e., Level 2 automation) and automated driving systems (i.e., Level 3 and above) to better understand the frequency of crashes with automated features across contributing factors (e.g., weather or vehicle speed). Second, we will incorporate this data into a regression model to better understand the relationship between the contributing factors and the frequency and severity of crashes. Finally, we will conduct an exploratory analysis on the contributing factors for crashes involving EVs. To do this analysis, we will utilize several publicly available datasets such as NHTSA’s AV crash reports, which provides information on crashes involving advanced driver assistance systems and automated driving systems (NHTSA, 2023) and the 2022 Fatality Analysis Reporting System.
In this phase of this project, we will mainly focus on contributing factors for AV crashes and will use similar datasets to conduct a more in-depth analysis on the contributing factors for EV crashes in future iterations. Because of Tesla’s dominance in both the EV and automated vehicle (AV) market space it’s important to consider both the role that EVs and AVs play in crash safety. By focusing on both technologies independently, we can better understand the different ways they are impacting crash safety and how to mitigate any negative effects through policy.
References
Kutela, B., Das, S., & Dadashova, B. (2022). Mining patterns of autonomous vehicle crashes involving vulnerable road users to understand the associated factors. Accident Analysis & Prevention, 165, 106473. https://doi.org/10.1016/j.aap.2021.106473
Liu, Q., Wang, X., Liu, S., Yu, C., & Glaser, Y. (2024). Analysis of pre-crash scenarios and contributing factors for autonomous vehicle crashes at intersections. Accident Analysis & Prevention, 195, 107383. https://doi.org/10.1016/j.aap.2023.107383
Liu, Q., Wang, X., Wu, X., Glaser, Y., & He, L. (2021). Crash comparison of autonomous and conventional vehicles using pre-crash scenario typology. Accident Analysis & Prevention, 159, 106281. https://doi.org/10.1016/j.aap.2021.106281
NHTSA. (2023). Standing General Order on Crash Reporting [Text]. https://www.nhtsa.gov/laws-regulations/standing-general-order-crash-reporting
Xu, C., Ding, Z., Wang, C., & Li, Z. (2019). Statistical analysis of the patterns and characteristics of connected and autonomous vehicle involved crashes. Journal of Safety Research, 71, 41–47. https://doi.org/10.1016/j.jsr.2019.09.001
Description
Timeline
Strategic Description / RD&T
This proposal addresses the USDOT’s research priority of “Data-driven System Safety” with the research objective of advancing transportation safety through “safe technology” (pg. 19). Specifically, this project will provide policymakers with a better understanding of the effect that AVs and EVs have on overall road safety, including those in vulnerable populations (e.g., pedestrians and cyclists). Critical research topics (found on pg. 16) such as vehicle and aircraft safety, automation, and connectivity and safety risk analysis methods are also addressed by this project. Finally, the project focuses upon the safety and intelligent vehicle systems goals of the Safety21 University Transportation Center.
Deployment Plan
Quarter 1: Begin compiling safety data and meet with PennDOT on state AV and EV road safety issues and needs.
Quarter 2: Exploratory and regression analysis framework outlined and meet with City of Pittsburgh on local AV and EV road safety issues and needs.
Quarter 3: Compile initial results of analysis. Present initial results to deployment partners (i.e., PennDOT and City of Pittsburgh). Presentation at the Transportation Research Board with results of our initial modeling efforts
Quarter 4: Policy brief providing recommendations to policymakers on ways promote a smooth transition to a safe and sustainable transportation system with EVs and AVs. Journal paper on AV safety submitted for review. Exploratory analysis on the contributing factors of crashes involving EVs.
Expected Outcomes/Impacts
This project will have a variety of outcomes and impacts:
1. Designing the framework to assess the contributing factors for crashes involving AVs and EVs lays the foundation for federal, state, and local organizations to better understand the infrastructure improvements and policies that are needed to enhance AV and EV safety.
2. Working closely with the City of Pittsburgh and PennDOT could help inform local and state crash safety policy as the number of EVs and AVs on the road increase. Additionally, this will provide an opportunity for students to network and learn the missions and goals of City of Pittsburgh and PennDOT, providing them with pathways to obtain internships and higher paying jobs after graduation (e.g., director of planning, policy analyst).
3. Journal papers and policy briefs can inform decision makers and researchers about pathways to help achieve safety goals such as zero roadway fatalities.
Expected Outputs
Anticipated outputs from this proposal include:
1. Novel methods to evaluate the safety implications of EVs and AVs
2. Data sets, modeling results, and online/offline tools to deploy and validate the proposed methods.
3. Memorandum to the City of Pittsburgh and PennDOT and the development of a policy brief to inform future policy-making.
4. Journal paper to disseminate results to research community.
TRID
A TRID search was conducted for the different aspects of the project.
First, a search with the keywords ‘automated vehicle crashes contributing factors’ was done using the TRID database which yielded 51 results. Out of the 51 results there were six projects sponsored by USDOT. The most relevant research project was by Kusano et al. (2023), which developed a conflict topology that included contributing factors for use in automated driving system safety evaluation (Kusano et al., 2023). While Kusano et al. (2023)’s conflict typology allows researchers to categorize crashes based on a unique set of scenarios, their study is based on naturalistic driving data from non-AVs. We contribute to the literature by using national AV crash datasets to better understand the factors contributing to AV crashes.
Finally, we conducted a search with the keywords ‘electric vehicle crash characteristics’ using the TRID database and yielded 119 results. Out of the 119 results there were 17 projects sponsored by USDOT. The most relevant study was done by Chen et al. (2015), which based their analysis on crash data from 1999 to 2013, a point in time with very low EV penetration. We contribute to the literature by using more recent crash data.
References
Chen, R., Choi, K. S., Daniello, A., & Gabler, H. (2015). An analysis of hybrid and electric vehicle crashes in the US. In 24th International Technical Conference on the Enhanced Safety of Vehicles (ESV) National Highway Traffic Safety Administration (No. 15-0210).
Kusano, K. D., Scanlon, J. M., Brannstrom, M., Engstrom, J., & Victor, T. (2023). Framework For a Conflict Typology Including Contributing Factors For Use In ADS Safety Evaluation. 27th International Technical Conference on the Enhanced Safety of Vehicles, Yokohama, Japan.
Individuals Involved
| Email |
Name |
Affiliation |
Role |
Position |
| hcain@andrew.cmu.edu |
Cain, Heather |
CMU CEE |
Other |
Staff - Business Manager |
| cdharper@andrew.cmu.edu |
Harper, Corey |
CMU CEE |
PI |
Faculty - Untenured, Tenure Track |
| cth@andrew.cmu.edu |
Hendrickson, Chris |
Carnegie Mellon Heinz College |
Co-PI |
Faculty - Tenured |
Budget
Amount of UTC Funds Awarded
$95000.00
Total Project Budget (from all funding sources)
$200000.00
Documents
Match Sources
No match sources!
Partners
| Name |
Type |
| City of Pittsburgh |
Deployment & Equity Partner Deployment & Equity Partner |
| PennDOT |
Deployment & Equity Partner Deployment & Equity Partner |
