基于交通流轨迹数据的网联自动驾驶车辆运行安全分析与评估研究

本项目面向未来智能网联自动驾驶车辆（Connected and Automated Vehicle, CAV）推广普及的核心安全需求，提出基于交通流轨迹数据的CAV运行安全分析与评估方法，通过交通视频采集与处理获取人工驾驶车辆微观轨迹数据，基于交通流轨迹数据进行不同交通场景的自动判别划分，依托各类复杂交通场景为背景环境生成CAV轨迹数据，通过建立的基于轨迹数据的事故风险评估模型对比分析CAV的运行安全并挖掘原始高风险状态，基于重平衡理论强化拓展高风险状态数据并进行数据融合，采用融合高风险状态数据进行CAV运行安全的最终评估。项目的研究成果为量化评估CAV自身运行状态与安全状态提供理论依据，可促进CAV的市场化发展，具有显著的社会经济效益和工程应用价值。

Based on the core popularization requirement of connected and automated vehicle (CAV) in the near future, this project proposes methods for CAV’s safety analysis and evaluation via trajectory data of traffic flow. Trajectory data of manually driven vehicles are first collected through traffic video acquisition and processing techniques. With a large number of trajectory data, automatic classification algorithms are developed to categorize various traffic scenarios. Focusing on each scenario, trajectory data of CAV are generated with different scenarios as the background environment. Then, the crash risk evaluation models are established based on trajectories to compare and analyze the original high risk states. The high risk states are expanded further and data fusion is conducted based on rebalance theory, which provides data for the final safety evaluation of CAV. Results of this project provide theoretical basis for quantitative evaluation of CAV’s safety, promote the market-oriented development of CAV, and generate significant social and economic benefits as well as practical application value.