Autonomous parking method based on improved A* algorithm and model predictive control

To improve the search efficiency, smoothness, and parking path tracking accuracy of the existing parking methods, we propose a novel autonomous parking method based on A* algorithm, Bezier curve, and model predictive control (MPC) method. Firstly, a vehicle's low-speed kinematics model is constructed which describes the relationship between the front wheel and rear wheel to obtain the velocity of the rear wheel and the heading angle which is necessary for parking. Then, the vehicle parking environment is introduced for constructing three parking scenarios. An improved A* autonomous parking path planning algorithm is presented. This algorithm adopts a novel cost evaluation function to address the issue of excessive search nodes to enhance search efficiency. Meanwhile, to ensure the continuity and comfort of the planned path curvature, a third-order Bezier curve is used to smooth the planned path. Based on the established vehicle kinematics model, a first-order holder-based MPC method for autonomous parking path tracking control is proposed to realize autonomous parking with higher tracking accuracy. Finally, simulations and tests are conducted to certify the proposed autonomous parking method. The results verify that our method has significant improvement in theory and practice.