Behavior-decision fuzzy algorithm for autonomous mobile robots

Abstract

The ambiguous state recognition when an autonomous mobile robot moves around a complex environment is one of the most important problems in robot control. We propose a construction method of the behavior-decision system using fuzzy algorithms capable of expressing sequence flow that includes a mixture of both crisp and fuzzy processing. We also propose in this paper a method of tuning algorithms for giving robots the autonomous ability to judge purposes of actions like humans. In this method, we try to express ambiguous situations that a robot will encounter and decision algorithm flows by using fuzzy algorithms with fuzzy branch-controlled threshold, which we call the behavior-decision fuzzy algorithm. Finally, we report some results of computer simulations and experiments concerning an evaluation of this method in supposed simple indoor environments.