Mobile robots navigation in unknown environments by using fuzzy logic and learning automata

Autonomously navigation of mobile robots in unknown environments is a basic challenge in robotics. We can use behavior based approach in navigation of mobile robots in environments with obstacles. If actions of robot be taken as behavior, we can design them by fuzzy logic. It decreases the problem states, make the navigation easier and also can be used as initial knowledge for reinforcement learning. In this paper we used learning automata for coordinating behaviors which caused robot to choose the best action in any situation. Using Pioneer robot in V-rep simulator environment showed that fuzzy logic and learning automata for robot navigation had a better performance in convergence and learning speed rather than fuzzy logic and Q-algorithm.