Parallel Parking Path Planning Based on Improved Arctangent Function Optimization

Aiming at the problems of small parking space and poor parking effect for existing parallel parking, this paper proposes a parallel parking path planning based on improved arctangent function optimization. Firstly, the vehicle parking kinematics model is established, and the vehicle parameters are determined according to the three classical models. Secondly, the parking space model is established, and the most reasonable parking space parameters are selected according to the minimum parking space. Then, aiming at the problems of abrupt curvature of the designed arc-line-arc initial path, unsmooth path, large yaw angle at the end and large parking space, an improved arctangent function model is proposed, the parking constraints are established, the absolute value of vehicle yaw angle is taken as the objective function, and the parameters are optimized by genetic algorithm. Finally, it is verified by simulation experiments. The results show that the method can achieve smoother path, smaller parking space and more ideal parking posture, meet the requirements of stability, safety and comfort in the process of parallel parking, and improve the ability of parallel parking. Therefore, this method can provide a theoretical reference for path planning in automatic parking technology.