NEURO-FUZZY CONTROLLER FOR MOBILE ROBOT NAVIGATION WITH AVOIDING OBSTACLES AND REACHING TARGET BEHAVIORS

摘要

The modern neuro-fuzzy systems have characteristics which can be used to control the dynamic processes. However, in a real system, like mobile robots, one may find many input signals. In these conditions, an excessive number of inputs could alter the model transparency and increase the calculations complexity. Therefore, it is necessary to select the relevant inputs and consider the priorities. In this paper the authors implemented by simulation in Matlab language a neuro-fuzzy control structure for mobile robot navigation to avoids the "dead cycle" problem by defining a target whereon the robot must be reached.