基于改进快速探索随机树的复杂环境垂直泊车路径规划

Vertical Parking Path Planning in Complex Environment Based on Improved RRT Algorithm

针对车位狭窄、不规范停车带来的泊车困难问题,提出了一种基于改进快速探索随机树(RRT)的垂直泊车路径规划算法。为加快路径规划速度、提高规划质量,引入逆向树调整RRT目标点,使用高斯分布采样法及偏向性采样法进行融合采样,使用Reeds-Shepp(RS)曲线加快路径规划速度,并基于RS曲线进行了路径平滑优化。最后通过MATLAB仿真并与RRT及基于目标偏好的RRT(Goal-biasing RRT)进行比较,平均规划时间分别缩短52.3%与41.7%,路径代价分别减小17.7%与13.9%,证明了算法的有效性。

For the difficulty of parking caused by narrow parking space and irregular parking,this article proposes a vertical parking path planning algorithm based on improved Rapidly-exploring Randomized Tree(RRT).In order to speed up path planning and improve planning quality,the article introduces a reverse tree to adjust the RRT target point,uses Gaussian distribution sampling method and bias sampling method for fusion sampling,and uses Reeds-Shepp(RS)curve to speed up path planning,and the path is smoothed and optimized based on the RS curve.Through MATLAB simulation and comparison with RRT and Goal-biasing RRT,the average planning time is shortened by 52.3%and 41.7%respectively,and the path cost is reduced by 17.7%and 13.9%,which proves the effectiveness of the algorithm.