Abstract: Because the basic Rapidly-exploring Random Tree（RRT） path planning is unstable and not optimal in dynamic environment，an improved algorithm of robots RRT path planning based on comparison optimization is proposed.In each circle，a stable path can be obtained by trimming and replanning the tree in last circle，and a new path can be planned via the basic RRT method.Comparing these two paths，the optimal one can be found.From the results of simulations and experiments on mobile robots，it concludes that this algorithm can improve the stability of RRT path planning in dynamic environment，and ensures that the path is almost optimal.

Key words: path planning, Rapidly-exploring Random Tree（RRT）, dynamic environment

摘要： 针对动态环境下机器人RRT路径规划算法缺乏稳定性和偏离最优解的问题，提出一种基于对比优化的RRT路径规划改进算法。算法在新一周期的环境下，通过对上一周期路径树进行剪枝和重新规划得到一条稳定的路径，同时利用基本RRT算法规划出一条新路径，通过对比两条路径得到较优解。仿真和真实机器人实验结果均表明，改进的算法提高了动态复杂环境下RRT路径规划的稳定性，并保证了规划的路径逼近最优解。

关键词: 路径规划, 扩展随机树, 动态环境