基于微分平坦的平行泊车路径规划

Parallel Parking Trajectory Planning Based on Differential Flatness

为提高泊车成功率,降低泊车路径规划耗时,提出基于微分平坦理论的平行泊车路径规划方法。泊车路径满足三类约束:为保证行驶安全,综合考虑可能碰撞情况,建立避障约束函数;为满足停放要求,建立了终点状态的约束函数;为保证路径可跟踪,建立了方向盘角度,转角速度和车速的约束函数。利用Matlab非线性约束优化函数求得路径参数。仿真结果表明:该方法鲁棒性强,对车辆初始位置和方位角要求不高,解决了必须从特定位姿开始泊车的问题,增加灵活性和成功率;对于一般泊车环境该方法能得到曲率和车速缓慢变化的轨迹,有效解决了中途停车转向的问题;规划的轨迹满足避撞约束、车辆自身的约束、泊车停放要求、方向盘转角和转角速度约束;基于微分平坦的路径规划方法,降低了计算复杂度,缩短了规划时间,提高了泊车成功率。

In order to improve the parking success rate and reduce the time-consuming of planning of parking path, a parallel parking path planning method based on differential flat theory was proposed. The parking path satisfies three types of constraints: in order to ensure driving safety, possible collision situations was considered. then obstacle avoidance constraint function was established;in order to meet the parking requirements, a constraint function of the end state was established;to ensure that the path can be tracked, the constraint function of steering wheel angle and speed was established.. The path parameters was obtained by using the matlab nonlinear constrained optimization function. The simulation results show that the method is robust and has low requirements on the initial position and azimuth of the vehicle. It solves the problem that parking must start from a specific pose, increasing flexibility and success rate. For the general parking environment, a trajectory with small curvature and slow change of vehicle speed can be obtained, which effectively solves the problem of midway stop with steering;the planned trajectory satisfies collision avoidance constraint, vehicle’s own constraints, parking parking requirements, steering wheel angle and corner speed constraint;The path planning method based on differential flatness can reduce computational complexity, shorten planning time, and improve parking success rate.