Emergent behaviors of a fuzzy sensory-motor controller evolved by genetic algorithm

Recently, there has been extensive work on the construction of fuzzy contro llers for mobile robots by a genetic algorithm (GA); therefore, we can real ize evolutionary optimization as a promising method for developing fuzzy co ntrollers. However, much investigation on the evolutionary fuzzy controller , remains because most of the previous works have not seriously attempted t o analyze the fuzzy controller obtained by evolution. This paper develops a fuzzy logic controller for a mobile robot with a GA in simulation environm ents and analyzes the behaviors of the controller with a state transition d iagram of the internal model. Experimental results show that appropriate co ntrol mechanisms of the fuzzy controller are obtained by evolution. The con troller has evolved well enough to smoothly drive the robot in different en vironments. The robot produces emergent behaviors by the interaction of sev eral fuzzy rules obtained.