Abstract: An improved Informed-RRT* path planning algorithm based on node optimization is proposed to address the characteristics of the current Informed-RRT* algorithm that is slow, poorly purposeful and unsmooth in path planning. The adaptive t-distribution function is introduced to change the distribution probability of random points in different environments to improve the efficiency of the algorithm. Then an elliptic focal bias strategy is used to extend a single bias point to the whole elliptic focal point to make the random tree grow close to the minimum distance between the starting point and the target point, which increases the purposefulness of the algorithm. Finally, the re-election grandparent node strategy is used to reduce redundancy for the whole path, and the symmetric multipole type curve method is used to smooth for the path turns. The improved Informed-RRT* algorithm has higher search efficiency, more purposefulness, and smoother planned paths by comparing the surface with multiple sets of experiments.

Key words: path planning, informed-RRT*, node optimization, path smoothing

摘要： 针对目前Informed-RRT*算法在路径规划时速度慢，目的性差且所得路径不平滑的特点，提出了一种基于节点优化的改进Informed-RRT*路径规划算法。引入自适应t-分布函数改变随机点在不同环境下的分布概率，提高算法的效率。采用椭圆的焦距偏置策略，将单个偏置点扩展为整个椭圆焦距，使随机树的生长贴近起始点与目标点的最小距离，增加了算法的目的性。采用重选祖辈节点策略对整条路径减冗余，对于路径转折处采用对称多极式曲线法进行平滑处理。通过多组实验对比表明，改进后的Informed-RRT*算法搜索效率更高，目的性更强，且规划所得路径更为平滑。

关键词: 路径规划, Informed-RRT*, 节点优化, 路径平滑