作 者：
ZHAO Jiawei;WANG Zhaoqiang;MA Siqun;ZHANG Liming;School of Mechanical and Automotive Engineering, Shanghai University of Engineering Science;State Grid Zhejiang Tongxiang Power Supply Company;
关键词：
parallel parking;;path planning;;track control;;model predictive control;;Bézier curve
摘 要：
With the prosperous development of cities, parking is becoming more and more difficult. Automatic parking technology can effectively alleviate the problem of parking anxiety for novice drivers and ensure parking safety while improving efficiency. The implementation of automatic parking technology mainly needs to solve the two key problems of parking path planning and path tracking control. For the problem that the vehicle pose adjustment is not smooth and unreliable path tracking due to discontinuous path curvature,taking into account the environmental constraints in the parallel parking process and the structural constraints of the vehicle itself, a backpropagation method is proposed to form a collision-free parking path with continuous curvature by using a two-segment third-order Bessel curve. For the path tracking control, the problems of tracking error accuracy is mainly considered, a model predictive control algorithm is used to design the path tracker, and the front wheel turning angle and the speed change rate of the vehicle are used as constraints to find the optimal control solution. Finally, a simulation platform is built using CarSim and Simulink. The simulation results show that due to the continuous path curvature, the front wheel turning angle of the vehicle will not produce sudden changes, while the path tracking controller can effectively control the vehicle along the planned route with a tracking error of centimeter level, which can realize the automatic parallel parking operation of the vehicle.
