基于信息熵-动态可增减SOM图的机器人路径规划.docVIP

  基于信息熵—动态可增减 SOM 图的机器人 路径规划# 阮晓钢，高静欣* 5 10 15 20 25 30 35 40 （北京工业大学电子信息与控制工程学院，北京 100124） 摘要：针对传统方法建立环境地图需要大量环境信息从而造成的存储搜索困难的问题，本文 基于自组织特征映射图 SOM 提出了一种动态可增减自组织特征映射图 DGPSOM 算法，旨 在用少量神经元来表征大量的环境信息，克服环境信息的存储问题，并在环境信息采集的过 程中引入了信息熵的概念，通过计算不同方向上的信息熵来引导机器人探索未知环境，提高 机器人遍历环境的效率。然后根据 DGPSOM 算法建立的环境地图设计了一种智能 A*搜索 算法，搜索任意起点到终点的最优路径。仿真结果表明，本文建立的信息熵模型可以有效避 免机器人重复访问同一区域，DGPSOM 建立的环境地图可以实现以少量神经元表征大量环 境信息，智能 A*搜索算法可以完成机器人最优路径的搜索。 关键词：移动机器人；信息熵；DGPSOM 算法；A*搜索；路径规划 中图分类号：TP242 Path Planning for Mobile Robot based on the Information Entropy and the DGPSOM algorithm Ruan Xiaogang, Gao Jingxin (School of Electronic Information and Control Engineering，Beijing University of Technology, Beijing 100124) Abstract: The traditional method to build environmental map needs to record a lot of environmental information, which will cause the problem of storage and search difficulty. In this paper, we propose a new algorithm to create the map: the Dynamic Growing and Pruning Self-Organizing Map. The DGPSOM algorithm uses a small number of neurons to express great amount information of environment topology map, so it can properly overcome the problem of storage and search difficulty in the large environment. In the search process, we introduce the concept of information entropy and calculate the information entropy of different directions to guide the robot to explore the unknown environment. Then according to the map of the environment created by the DGPSOM algorithm to design an intelligent A* search algorithm to find the optimal path from any beginning to the end.The simulation results show that, the information entropy model established in this paper can effectively avoid the robot repeatly visits to the same region, the map of the environment which established by the DGPSOM algorithm can be achieved with a small number of neurons represents a large number of environmental information, the A* search algorithm can complete the search of the optimal path. Keywords: Mobile Robot; Information E