Game-Theoretic Decision-Making Method and Motion Planning for Autonomous Vehicles in Overtaking

Overtaking is a common driving behaviour used by human drivers while driving. Therefore, the decision on overtaking is very important in the automatic driving decision. To be able to improve the passing efficiency of intelligent vehicles, it is crucial to be able to interact with oncoming vehicles with different driving styles. An overtaking decision needs to be adapted to the situation where the vehicle being overtaken is potentially stationary or moving. Therefore, this paper proposes an overtaking decision method based on potential conflict area based on the requirements. Firstly, the planning method for each stage is given, and the generation method of the potential conflict area is proposed. Second, the interaction process between the host vehicle and the opposite oncoming vehicle is modelled by a dynamic game based on the potential conflict area. A driving style assessment method for oncoming vehicles based on potential conflict area is proposed. Thirdly, the priority of passing the potential conflict area of multiple oncoming vehicles is divided to correct the speed planning in the waiting and the speed payoff function in the game. Finally, the overtaking decision is simulated and validated by Virtual Test Drive.