在动态环境中移动机器人导航和避碰的一种新方法

本文提出了基于超声传感器的信息，将改进的栅格和回归预测法结合起来，应用于具有静态和动态障碍物的动态环境中，移动机器人THMR-Ⅱ的导航和避碰的一种新方法．对栅格法的改进就是以障碍物为单位记录信息量，结果比原来以栅格为单位记录的信息量少得多，克服了栅格法中存在环境信息存储量大的问题，提高了实时性．对回归预测法也作了改进，并把它们结合起来，在求得最佳候选扇区后，使移动机器人躲避了静态和动态障碍物，实现了导航，最终到达目标．通过三种仿真实验，结果表明作者提出的方法是正确和有效的．

A NEW METHOD OF REAL-TIME NAVIGATION AND OBSTACLE AVOIDANCE FOR MOBILE ROBOT IN DYNAMIC ENVIRONMENT

The paper proposes a new method of real time navigation and obstacle avoidance for mobile robot in dynamic environment. This method consists of the combination of improved Grids method and recurrent predication method based on the information of ultrasonic sensors. In the improved Grids method, we only record information quantities having obstacles for unit in grids which are much less than one of original record grids. The problem of a large number of information storage of an environment is solved. The application of Grids method will become a possibility. The position of moving obstacles is predicated by using recurrent predication method. After finding a solution of optimum candidate sector with improved Grids method, the mobile robot will avoid static and dynamic obstacles to implement navigation towards the target. Three simulation experiment results demonstrate the effectiveness and correctiveness of the method.