Perception-action method for mobile robot plan and control based on driving experience

A method of navigation and control for mobile robot is introduced. It combines task planning and path tracking together with the principle of “perception-action” under the guidance of “goal planning”. First, we discuss the behavior of the mobile robot in an outdoor real world for the purpose of setting up a mixed layered architecture with “perception-action” and “goal planning”. Then a simple but effective approach of sensor based navigation and control is described for the implementation of the architecture. Finally, we give some improvements based on the human-driving experience concerning path tracking and control for the mobile robot moving on outdoor semistructured roads. The experiments carried out on our THMR-III (Tsinghua Mobile Robot III) mobile robot navigating in the real world showed the method described was effective and robust