Fuzzy-genetic algorithms and mobile robot navigation among static obstacles

The paper describes a fuzzy genetic algorithm in which a fuzzy logic controller (FLC) is used with genetic algorithms (GAs) to find obstacle-free paths in a number of find-path problems of a mobile robot. In this algorithm, an obstacle-free direction for the movement of a robot locally is created using an FLC and the extent of travel along obstacle-free direction is determined by a GA. Here, the fuzzy logic approach is used to create initial population and GA crossover and mutation operators. This algorithm is found to perform better than the popular steepest descent approach. The proposed algorithm also finds solutions close to the best known tangent graph with A* algorithm from the accuracy point of view. However, the proposed algorithm finds a near-optimal solution faster than the tangent graph and A* algorithm. Moreover, the proposed approach shows how genetic operators can be modified with problem-specific information to create a search algorithm which is efficient for the particular application.