Navigation for Mobile Robots Using Reinforcement Learning and Fuzzy Logic

Abstract

Autonomous navigation is the key technology of mobile robots. Navigation control of autonomous robots in uncertain environments is realized by using the reinforcement learning and fuzzy logic in this paper. First, the principle of reinforcement learning is introduced. And then a framework is proposed which consists of avoidance module, goal-seeking module and behavior selecting module for navigation of autonomous mobile robots in uncertain environments. According to the framework, a learning-planning method is proposed, which utilizes reinforcement learning and fuzzy logic. Two behaviors are independently designed in training stage and then combined by a behavior selector at running stage. According to information acquired by ultrasonic sensors, the behavior selector chooses a behavior at each action step so that the mobile robot can reach the goal position without colliding with obstacles. At last, the effectiveness of the proposed method is verified by a series of simulations.