Investigation of Key Safety Measures for Pre and Post-Deployment of Connected and Automated Vehicles

Details

• Creators:
  Kondyli, Alexandra ; Schrock, Steven ; Bakhti, Bahareh
• Corporate Creators:
  Mid-America Transportation Center for Transportation Safety and Equity (MATC-TSE) Region 7 University Transportation Center (UTC) ; University of Kansas
• Corporate Contributors:
  United States. Department of Transportation. University Transportation Centers (UTC) Program ; United States. Department of Transportation. Office of the Assistant Secretary for Research and Technology ; University of Nebraska-Lincoln
• Subject/TRT Terms:
  Autonomous Vehicles Connected Vehicles Forecasting Highway Safety Metrics (Quantitative Assessment) Performance Measurement Predictive Models
• Publication/ Report Number:
  25-1121-3002-101
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States
• Edition:
  June 1, 2023 – May 31, 2025
• Corporate Publisher:
  Mid-America Transportation Center for Transportation Safety and Equity (MATC-TSE) Region 7 University Transportation Center (UTC)
• Abstract:
  In recent years, automated vehicles (AVs) are increasingly penetrating road networks with the main purpose of reducing driver error. Since around 94% of traffic crashes are due to driver errors, automated vehicles have the potential to enhance road safety by eliminating human drivers' tasks. Despite claims that these vehicles will increase road safety, their ability remains unclear. However, it is essential to address the safety effectiveness of these vehicles. To date, researchers have used surrogate safety measures, such as time-to-collision (TTC) to quantify the near-crash occurrence of vehicles, where 1.5 seconds is typically considered the critical TTC thresholds for both AVs and human-driven vehicles (HDVs). Although AVs can travel with shorter headway due to their enhanced safety features, studies continue to use the same threshold value in their analysis. This study aims to evaluate the safety effectiveness of AVs through accomplishing several phases. First, the reaction time of AVs, while they are traveling in car-following mode along freeway and arterial facilities, was measured through two different approaches, cross-correlation and visibility graph algorithm. In the next phase, multiple machine learning models were tested to model car-following behavior of AVs. Finally, a mixed environment simulation framework was developed to investigate the effects of different penetration rates of AVs on safety, using the derived reaction times for AVs. The results indicate that the mean reaction time of AVs is 0.95 seconds and 1.05 seconds on freeways and arterials, respectively. Considering these values as the critical TTC thresholds, this study found no incidents of potential conflicts for AVs in the dataset provided, which aligns with previous studies, proving the safety benefits of AVs. Furthermore, it is found that the LSTM algorithm outperforms other models in replicating AV behavior. Further research is needed to implement the proposed simulation framework within VISSIM to estimate the safety effects of AVs, considering the critical TTC thresholds based on our analysis.
• Format:
  PDF
• Funding:
  69A3552348307
• Collection(s):
  University Transportation Centers Autonomous Vehicles
• Main Document Checksum:
  urn:sha-512:d8139948dc761960ede0f53f2c09da6cd799fa6328a8b475f39855bfa9feb9b1b3ebc186d5a05e1c9e294fd16713a15d82c2829ebd3f17a7292c51e51b62d7b2
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/85612/dot_85612_DS1.pdf
• File Type:
  [PDF - 1.38 MB ]