Takeover Behavior Patterns for Autonomous Driving in Crash Scenarios
Journal of transportation safety & security(2022)
摘要
It is necessary to identify takeover behavior patterns of conditional autonomous driving. In this paper, using driving simulations, takeover request lead time (5 s and 10 s) and nondriving-related tasks (working task and entertainment task) are designed to study the takeover behavior pattern in crash scenarios. Through driving simulation experiment, the number of takeover behavior patterns is eleven and the number of first takeover behaviors is three. Results showed that the first takeover behavior has a significant impact on the first takeover reaction time, speed, lateral offset, and minimum TTC, but the first takeover behavior has no significant effect on the takeover correct time. The takeover request lead time (TORlt) has a significant impact on the pattern and the first takeover behavior, while the non-driving-related task (NDRT) has no significant effect on the pattern and the first takeover behavior. In addition, this paper constructs a maps of takeover operation behavior, which more intuitively shows the behavior changes during a takeover.
更多查看译文
关键词
Autonomous driving,takeover behavior pattern,first takeover behavior,crash scenarios
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要