凸近似避障及采样区优化的智能车辆轨迹规划

Trajectory Planning Method of Intelligent Vehicle Based on Convex Approximate Obstacle Avoidance Principle and Sampling Area Optimization

针对结构化道路下作匀速运动的智能车辆避障轨迹规划问题,提出一种基于凸近似避障原理及采样区域优化的智能车辆轨迹规划方法。引入凸近似避障原理,得到轨迹可行域范围;将采样区域分为静态采样区、动态采样区两部分,并根据障碍物运动状态,另外划分动态、静态障碍物采样区;采用“动态规划(DP)+二次规划(QP)”思想求解轨迹:利用五次多项式对采样点依次连接,建立动态规划代价函数并筛选得到粗略轨迹;通过二次规划及约束条件的构造,对粗略轨迹进行平滑,最终得到最优轨迹。仿真结果表明:对于静态、低速、动态三种障碍物,该车能够有效地得到平滑轨迹并避开障碍物。

Aiming at the problem of obstacle avoidance trajectory tracking for intelligent vehicles moving at constant speed on structured roads,a trajectory tracking method of intelligent vehicles based on convex approximate obstacle avoidance principle and sampling area optimization is proposed.The convex approximate obstacle avoidance principle is introduced to obtain the feasible range of trajectory.The sampling area is divided into static sampling area and dynamic sampling area,according to the motion state of the obstacle,the dynamic obstacle and static obstacle sampling area are also divided.The idea of“dynamic programming(DP)+quadratic programming(QP)”is used to solve the trajectory.The sampling points are connected successively by the quintic polynomial,and the dynamic programming cost function is established and the rough trajectory is obtained by searching.Through the quadratic programming and the construction of constraints,the rough trajectory is smoothed and the optimal trajectory is obtained.The simulation results show that:the vehicle can effectively obtain smooth trajectories and avoid obstacles for static,low-speed and dynamic obstacles.