基于多次采样启发式策略的改进RRT路径规划算法

Improved RRT Path Planning Algorithm Based on Multiple Sampling Heuristic Strategy

摘要:

在路径规划领域，快速探索随机树（RRT）算法是机械臂解决复杂环境中的路径规划问题的重要工具，然而其纯随机的采样环节导致大量的无效或低效尝试，浪费了计算资源。为解决此问题，提出一种基于多次采样启发式策略的改进RRT算法（MH-RRT）。利用启发式函数评估多个采样点的代价值，选择代价值最低的采样点，引导路径树更快地向目标点生长；将启发式函数策略类似地有效改进了RRT*算法和双向RRT*算法；深入探讨不同参数对改进算法性能的影响，并确定最优参数组合。实验结果表明，改进算法在路径搜索时间、路径长度以及采样点数量等方面均能取得显著提升，提高了路径规划的有效性。

Abstract:

In the field of path planning, rapid exploration random tree (RRT) algorithms are important tools for robotic arms to solve path planning problems in complex environments. However, their purely random sampling process resulted in a large number of invalid or inefficient attempts, wasting computational resources. To solve this problem, an improved RRT algorithm based on multiple sampling heuristic strategy (MH-RRT) is proposed. Firstly, the heuristic function strategy is used to evaluate the proxy value of multiple sampling points, the sampling point with the lowest proxy value is selected, and the path tree is guided to grow faster towards the target point; Then, the heuristic function strategy is effectively improved similarly to the RRT* algorithm and the bidirectional RRT* algorithm; Finally, the impact of different parameters on the performance of the improved algorithm is thoroughly explored, and the optimal parameter combination is determined. The experimental results show that the improved algorithm can significantly improve the time efficiency, path length, and number of sampling points in path search, thereby enhancing the effectiveness of path planning.