Autonomous navigation based on a Q-learning algorithm for a robot in a real environment

This paper explores autonomous navigation and obstacle avoidance techniques based on Q-learning for a mobile robot in a real environment. The implemented algorithm focuses on simplicity and efficiency. The learning process takes place in both simulation and real world allowing the combination of a longer learning time in the simulator with a more accurate knowledge from the real world. After learning is completed in simulation and in the real world, the robot was able to navigate without hitting obstacles and able to generate control law for complex situations such as corners and small objects.

[1]  Bo Yin,et al.  Robot path planning in complex environment based on delayed-optimization reinforcement learning , 2002, Proceedings. International Conference on Machine Learning and Cybernetics.

[2]  Peter Dayan,et al.  Technical Note: Q-Learning , 2004, Machine Learning.

[3]  Chen Gang.  GENETIC PATH PLANNING ALGORITHM FOR COMPLEX ENVIRONMENT PATH PLANNING , 2001 .

[4]  N. Peric,et al.  A reinforcement learning approach to obstacle avoidance of mobile robots , 2002, 7th International Workshop on Advanced Motion Control. Proceedings (Cat. No.02TH8623).

[5]  J. Ota,et al.  Acceleration of reinforcement learning by a mobile robot using generalized rules , 2000, Proceedings. 2000 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2000) (Cat. No.00CH37113).

[6]  Min Guo,et al.  Reinforcement Learning Neural Network to the Problem of Autonomous Mobile Robot Obstacle Avoidance , 2005, 2005 International Conference on Machine Learning and Cybernetics.

[7]  N. H. C. Yung,et al.  An intelligent mobile vehicle navigator based on fuzzy logic and reinforcement learning , 1999, IEEE Trans. Syst. Man Cybern. Part B.