摘要
Navigation of anutonomous mobile robots in known environments has been studied extensively. But the algorithms for controlling progress through unknown environments have not received much study. In this paper, we put forward a new exploration scheme which is based on learning. While travelling, robots use their range/vision sensors to perceive the external world. Newly acquired information about obstacles is added to the system’s knowledge base through learning. And the updated knowledge base is used in planning future navigation paths, thus the generated paths will improve gradually.
Navigation of anutonomous mobile robots in known cnvironments has been studied extensively. But the algorithms for controlling progress through unknown environments have not received much study. In this paper, we put forward a new exploration scheme which is based on learning. While travelling, robots use their range / vision sensors to perceive the external world. Newly acquired information about obstacles is added to the system's knowledge base through learning. And the updated knowledge base is used in planning future navigation paths, thus the generated paths will improve gradually.
