摘要： 从全局优化的角度出发,提出一种基于子区域分割的环境建模方法,将各个子区域缩成节点、子区域重心间距离缩成节点间连线,建立起含障区域的全连通图——广义距离矩阵模型,对Hopfield神经网络求解旅行商问题的动态规划方法进行改进后,将其用于求该模型的最优有向连通图。机器人沿最短路径完成覆盖,就是按最优顺序走过连通图中所有节点并依次覆盖各节点所代表的子区域,从而完成移动机器人在含障区域内的全覆盖路径规划问题的优化解决方案。

关键词: 全覆盖, 路径规划, 神经网络, 子区域, 连通图

Abstract: From the point of view of overall optimization,this paper proposes a subregion based method to build the connected graph-generalized distance matrix model of a region with obstacles.It improves the continuous Hopfield neural network method that has been applied to sovle the traveling salesman problem.It uses the improved CHNN to obtain optimized connected graph.Robot going through the optimized connected graph goes through the whole free space along the shortest path or in the shortest time.

中图分类号: 

• TP242