煤矿环境下改进Informed-RRT*路径规划算法研究

Research on Improved Informed-RRT*Path Planning Algorithm in Coal Mine Environment

为解决煤矿机器人视觉SLAM地图中传统路径规划Informed-RRT*算法收敛速度慢、生成路径代价高、甚至无法找到路径等问题,对视觉SLAM建图进行了研究,提出了一种基于八叉树地图的路径规划方法。首先构建稠密点云地图并生成八叉树地图,将八叉树地图转化为易于路径规划的二维栅格地图,在此地图基础上结合人工势场和Informed-RRT*路径规划方法,提出了一种改进Informed-RRT*路径规划算法。仿真结果表明,该算法比传统Informed-RRT*算法速度提升了29.74%,具有更高的搜索效率和更快的收敛速度,获得了更好的路径规划效果。

In order to solve the problems of slow convergence speed,high cost of path generation and even inability to find the path of the traditional path planning Informed-RRT*algorithm in the visual SLAM map of coal mine robot,the visual SLAM mapping was studied,and a path planning method based on octree map was proposed.Firstly,a dense point cloud map was constructed,an octree map was generated,and the octree map was transformed into a two-dimensional raster map that is easy to plan the path.Based on this map,combined with artificial potential field and Informed-RRT*path planning algorithm,an improved Informed-RRT*path planning algorithm was proposed.The simulation results show that the proposed algorithm is 29.74%faster than the traditional Informed-RRT*algorithm,with higher search efficiency and faster convergence speed,and better path planning resultsare obtained.