摘要
提出一种用于移动机器人对当前感知环境空间进行识别和分类的新方法．文中对系统进行了描述；分析和研究了自适应谐振理论，即ＡＲＴ神经网络；把多传感器信息作为神经网络的输入来设计分类器，对移动机器人的当前感知环境进行快速识别和分类．在ＴＩＴ－１型移动机器人上实验表明，该方法具有识别速度快、分类可靠等优点，为移动机器人在未知环境下实现安全的快速反应导航奠定了必要的基础．
In this paper,a new method for mobile robot's currently perceptual environment partitioning and identification is presented first.We have described the whole environment perceptual system,and then,studied Adaptive Resonance Theory(ART) neural network and chose it as the classifier of the mobile robot's currently perceptual environment.By use of multisonars information,this method can realize mobile robot's currently perceptual environment partitioning and identification.Experiments show that the partitioning and identification speed is improved to a great extent,and the method lays the foundation for a mobile robot to realize fast reactive navigation.
出处
《天津大学学报》
EI
CAS
CSCD
1998年第4期466-470,共5页
Journal of Tianjin University(Science and Technology)
关键词
移动机器人
多传感器信息
神经网络
当前感知环境
mobile robot,multi sensors's information,neural network,currently perceptual environment
