Research On Trajectory Planning Algorithm Of Autonomous Vehicle Based On Polynomial And Game Theory

Automobile market of China is broad,urban road traffic conflicts are increasingly prominent,and low-carbon,electric,intelligent autonomous vehicle have received widespread attention.Autonomous lane change decision and trajectory planning,as an important functional block to realize the automatic driving function of vehicles,have become the focus of research in this field.This paper takes driverless vehicles as the research object,establishes a lane change decision system for changing driving strategies,proposes a trajectory planning method based on quintic polynomial,and evaluates the trajectory based on game theory to improve lane change safety,comfort and acceleration potential.The specific research contents are as follows:(1)The decision threshold is obtained by analyzing the state of the surrounding vehicles and the vehicle under each driving state,and the decision tree is established accordingly,so that the autonomous vehicle has the ability to change the driving strategy in the changing vehicle and road environment,and grasp the lane change opportunity.(2)The vehicle dynamics model in the Frenet system is established to determine the trajectory constraints.Under the constraint conditions such as initial state and target state,the horizontal and vertical tracks are planned according to the quintic polynomial,the set of alternative tracks is created,and each alternative track in the track set is evaluated according to the track cost function.(3)In order to improve the ability of lane change and obstacle avoidance,and improve the safety of lane change,the game income matrix is used to conduct multi-dimensional evaluation on the alternative tracks in the track set.According to the self-defined weight and two evaluation methods,the optimal trajectory with comprehensive optimal lane change safety,comfort and acceleration potential is selected to complete the trajectory planning.(4)In the software joint simulation environment,the effectiveness of the lane change decision and trajectory planning method proposed in this paper is verified,and the results show that the planning trajectory obtained in this paper has good comfort and effectiveness.(5)The hardware-in-the-loop automatic driving simulator is established to verify the effectiveness of lane change decision and trajectory planning methods under comprehensive road conditions.The results show that the planning trajectory obtained in this paper can basically realize the automatic driving of vehicles under comprehensive road condition.