Evaluation of Driver Behavior Influenced by Traffic Safety Messages From Roadside Units in Connected and Autonomous Systems

Details

• Creators:
  Jeihani, Mansoureh ; Ansariyar, Alireza ; Yang, Di
• Corporate Creators:
  Morgan State University. Safety and Mobility Advancements Regional Transportation and Economics Research Center
• 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
• Subject/TRT Terms:
  Analysis Autonomous Vehicles Connected Vehicles Cyclists Driving Behavior Driving Behavior Analysis Machine Learning Pedestrians Safety Messages Signalized Intersections Test Beds Transportation Safety Urban Areas Vulnerable Road Users

  More +
• Publication/ Report Number:
  SM02
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States
• Edition:
  Final Report: September 2023 - September 2024
• Corporate Publisher:
  Morgan State University. Safety and Mobility Advancements Regional Transportation and Economics Research Center
• Abstract:
  The safety of vulnerable road users (VRUs) at signalized intersections is a significant concern in urban traffic management. This study aims to enhance VRUs safety through the deployment of advanced communication and machine learning technologies. Using a Connected and Autonomous Vehicle (CAV) testbed, this research explores real-time communication between LiDAR sensors, Roadside Units (RSUs), On-Board Units (OBUs), and traffic controllers. This integration allows for the timely dissemination of safety messages and enables drivers to adjust their behavior to prevent potential crashes. The CAV testbed offers a controlled environment for studying these interactions, facilitating precise data collection and analysis. By collecting driving data from 32 participants under three scenarios with two different sets of safety messages broadcasted by RSUs on the testbed, the study investigated the participants' driving behavior in interactions with pedestrians and bicyclists. The study utilized machine learning models, including Logistic Regression, Random Forest, and Support Vector Machine (SVM), to predict driver behavior under various scenarios. The SVM model demonstrated the highest accuracy, particularly in predicting lateral distance, with an accuracy rate of 0.88 in response to the "Keep to the Right Lane" message. Key findings emphasize the importance of factors such as driver demographics, experience, and familiarity with CAV technology in shaping responses to safety messages. Additionally, the male drivers showed significant adjustments in speed and acceleration, while female drivers exhibited a more consistent, cautious approach, with minimal changes in speed, acceleration, and lateral distance.
• Format:
  PDF
• Funding:
  69A3552348303
• Collection(s):
  Autonomous Vehicles University Transportation Centers
• Main Document Checksum:
  urn:sha-512:ef6aaac30ab958a34376e5265366f4d2869f75983bbd81adba79ba348577e304831ba3eaf4adfb8f837d976533b3181d68521cae88de65ff02dfcd803bb99329
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/92681/dot_92681_DS1.pdf
• File Type:
  [PDF - 729.99 KB ]