一种基于轨迹预判的垂直泊车路径规划算法研究

Research on Vertical Parking Path Planning Algorithm Based on Trajectory Prediction

为解决目前量产自动泊车系统泊车成功率低、泊车完成后车辆姿态偏斜等问题,研究搭建基于阿克曼转向几何学和车辆运动学的车辆运动模型,构建基于可行驶区域的栅格电子地图,在几何路径规划方法中融入基于轨迹预判的碰撞约束方法和路径居中算法,采用车辆膨胀轮廓模型,最大程度利用电子地图可行驶区域和保证路径安全性,设计出一种基于轨迹预判的垂直泊车路径规划算法,并通过仿真测试和实车测试验证算法的可行性。该算法可大大提高泊车成功率,能帮助解决泊车完成后车辆偏斜不居中的问题。

In order to solve the problems of the current mass-produced automatic parking system,such as the low parking success rate and the vehicle attitude skew after parking,a vehicle motion model based on Ackermann steering geometry and vehicle kinematics was built,and a network based on the drivable area was constructed.The grid electronic map integrates the collision constraint method based on trajectory prediction and the path centering algorithm into the geometric path planning method,and adopts the vehicle expansion contour model to maximize the use of the drivable area of the electronic map and ensure the safety of the path.Predictive vertical parking path planning algorithm.Simulation test and real vehicle test verify the feasibility of the algorithm,the algorithm greatly improves the success rate of parking,and solves the problem of vehicle deflection and non-centering after parking is completed.