Abstract: Automated parking system (APS) is a hot topic in current research. However, most studies ignore the effect of the initial vehicle heading angle deviation on the overall parking effect, this paper investigates a multi-stage path planning method with a non-zero initial vehicle heading angle in a narrow vertical parking space scenario. Firstly, based on the minimum turning radius of the vehicle, the trajectory equation of the vehicle rear axle center motion is established, and the path planning is optimized by the fifth polynomial to achieve more continuous path curvature, smoother and more comfortable parking. CarSim/Simulink joint simulation validation shows that the five polynomial parking path optimization method is feasible, and the actual width required for the parking space is reduced by 10.4% and the length is reduced by 15.3% on the basis of satisfying the parking safety.

Key words: Vertical parking; Path planning; Quintuple polynomial; Model predictive control; CoSimulation; CarSim/Simulink

摘要： 自动泊车系统（APS）目前受到广泛的关注，但大部分研究忽略了车辆初始航向角偏 差对整体泊车效果的影响，对此，论文研究了狭窄垂直车位场景下车辆初始航向角不为零的 多段式路径规划方法。首先基于车辆最小转弯半径建立了车辆后轴中心运动轨迹方程，通过 五次多项式对路径规划进行优化，实现路径曲率更连续、泊车更平顺和舒适。CarSim/Simulink 联合仿真验证表明，五次多项式泊车路径优化方法可行，在满足泊车安全的基础上，泊车位 所需实际宽度减少 10.4%，长度减少 15.3%。

关键词: 垂直泊车；路径规划；五次多项式；模型预测控制；联合仿真；CarSim/Simulink