Learning of mobile robots using perception-based genetic algorithm

Abstract

This paper deals with adaptation, evolution, and learning for a mobile robot based on fuzzy controllers. If its facing environment is stable, the behavior of a mobile robot can be optimized by conventional genetic algorithms (GAs). Otherwise, the behavior should be tuned by learning according to the change of its environment. However, it is difficult for a mobile robot to maintain various behaviors suitable to changing environmental states. Therefore, this paper proposes a GA based on the perceiving information about the dynamic environment, which is called a Perception-Based GA (PerGA). We apply the proposed method for acquiring collision avoidance behaviors of a mobile robot in a dynamic environment. Furthermore, we conduct several computer simulations and simple experiments of a mobile robot. Simulation results show that the PerGA can maintain various behaviors according to environmental changes.