An Efficient Path Planner Combining Random Sampling and a Fuzzy System

In order to mimic human behavior in a maze-like environment, we present an efficient and practical path planning method that combines a random-sampling system with a fuzzy system. The fuzzy rule is applied to obstacles to derive the necessary information that will allow a robot to follow a path without colliding with nearby obstacles. Then the random sampling system chooses the forward moving direction based on the information provided by the fuzzy system and checks if the goal is reachable until the goal is reached. Our algorithm remedies the weakness of using RRT or the fuzzy system alone in planning a path through a map with narrow passages or dead ends. Experiments performed in different environments show that the prepared method is efficient and practical in a variety of difficult situations, including narrow passages and dead ends in maze-like maps.