Analysis of the Impact of Traffic Hazard Scenarios on the Driver's Take-over Behavior of Autonomous Vehicle

The take-over human-machine co-driving technol-ogy will be still the main before the fully automated autonomous driving is achieved. Due to the presence of humans, the driver's psychological and physiological state is easily affected by the degree of danger of the surrounding traffic environment. The varying take-over reaction time and manipulation behavior will greatly affect vehicle safety. Currently, there are few studies analyzing the impact of the number and type of hazards on driver tak-eover behavior. To this end, this paper analyzes the impact of hazards on reaction time and manipulation behavior during take-over process. This paper collects driving behavior data and eye movement characteristics data to study the impact of differences in traffic hazard scenarios on the takeover behavior and hazard perception of autonomous drivers. The results show that drivers have efficient perception and response capabilities for explicit and double hazard emergency Take-over scenarios. The driver's reaction time is significantly reduced, the vehicle acceleration is significantly reduced. The number and visual salience of hazards significantly impact drivers' visual search activity.