A GA-based fuzzy logic approach to mobile robot navigation in unknown dynamic environments with moving obstacles

Abstract

A genetic algorithm (GA)-based fuzzy-interference control system with an accelerate/brake (A/B) module is developed for a mobile robot in unknown environments with moving obstacles. The A/B module of the proposed system is to enable the mobile robot to make human-like decisions as it moves toward a target. Under the control of the proposed fuzzy inference model, the robot can perform well in avoiding both static and moving obstacles, like human beings, along a reasonable short path. In addition, a GA module is employed to tune the membership functions, which improves the performance of the fuzzy-inference system. The GA is an effective auto-tuning technique in optimizing systems without suffering from local minima. The effectiveness of the proposed approach is demonstrated by simulation studies.