Fuzzy Based Autonomous Robot Navigation System

Abstract

Fuzzy systems have been used in past years, for robot navigation applications because of their ability to capture human mind expertise and to utilize knowledge to develop autonomous navigation strategies. In this paper, the Fuzzy based systems are developed for behavior based control of a mobile robot for reactive navigation. The Fuzzy system transforms sensors' input to yield wheel velocities. The rule bases for Fuzzy systems are generated by "handcrafting" that generates training sequences optimally. Fuzzy membership functions are fine tuned manually based on simulation experiments. With a view to ascertain the efficacy of the Fuzzy based systems; our developed Fuzzy system's basic architecture is compared with architecture based on heuristic approach. Simulation results show effectiveness of proposed system in all kind of obstacle environments.