Human-Like Autonomous Lane Change Trajectory Planning Based on Driver Risk Perception Model

Unmanned driving technologies have great potential in improving traffic safety and reducing driver workload, but the human driving mechanism is rarely considered. Human-like unmanned driving could meet the expectations of passengers and pedestrians, so that unmanned vehicles can be more widely accepted. However, current human-like unmanned driving methods rely heavily on historical data, only imitating the driver's behavior without fundamentally explaining the motivation behind his behavior. This study develops a driver risk perception model that describes drivers' perceptions of risk for providing an in-depth explanation of human driving mechanisms to lay a theoretical foundation for human-like driving methods. It also reveals the characteristics of the driver's perceived risk index during lane changing and its relationship with the lane change trajectory. Finally, it proposes a human-like autonomous lane change trajectory planning method, which achieves autonomous planning of human-like lane change trajectory with high fidelity by using the position and speed information of the ego vehicle and obstacles.