A Reinforcement Learning Based ART2 Neural Network: RL-ART2
FAN Jian1,2, FEI Mei-Rui1
1.Key Laboratory of Power Station Automation Technology, Shanghai University, Shanghai 2000722. Operations Research Centre, Nanjing Army Command College, Nanjing 210045
Abstract:A reinforcement learning based ART2 neural network (RLART2) is proposed and its learning algorithm is given. It is capable of online learning without training samples by using the characteristic of alteration with environment of reinforcement learning. In RLART2, the output classified pattern is got by inner competition of ART2, then the running effect of the classified pattern is gained and evaluated through altering with environment. With enough time of being online and interactive learning with environment, a certain recognition ratio of ART2 neural network is attained. The simulation results of path planning for mobile robot indicate that the collision times of robot is effectively decreased by using RLART2 . Moreover, the rationality and validity of RLART2 are also demonstrated by the results.
[1] Carpenter G A, Grossberg S. ART2: Stable SelfOrganization of Category Recognition Codes for Analog Input Patterns. Applied Optics, 1987, 26(23): 4919-4930 [2] Whiteley J R, Davis J F, Mehrotra A, et al. Observations and Problems Applying ART2 for Dynamic Sensor Patter Interpretation. IEEE Trans on Systems, Man and Cybernetics, 1996, 26(4): 423-437 [3] Grossberg S. Adaptive Pattern Classification and Universal Recoding II: Feedback, Expectation, Olfaction, and Illusions. Biological Cybernetics, 1976, 23(1): 187-202 [4] Qian Xiaodong,Wang Zhenou, Wang Yu. A Method of Data Clustering Based on Improved Algorithm of ART2 // Proc of the 4th International Conference on Machine Learning and Cybernetics. Guangzhou, China, 2005: 2021-2026 [5] Burchardt H, Salomon R. Implementation of Path Planning Using Genetic Algorithms on Mobile Robots // Proc of the 5th IEEE Congress on Evolutionary Computation. Vancouver, Canada, 2006: 1831-1836 [6] Surmann H, Huser J, Wehking J. Path Planning for a Fuzzy Controlled Autonomous Mobile Robot // Proc of the 5th IEEE International Conference on Fuzzy Systems. New Orleans, USA, 1996, Ⅲ: 1610-1665 [7] Fujimori A, Tani S. A Navigation of Mobile Robot with Collision Avoidance for Moving Obstacles // Proc of the IEEE International Conference on Industrial Technology. Bangkok, Thailand, 2002, Ⅰ: 1-6 [8] Grossberg S. Adaptive Pattern Classification and Universal Recoding I: Parallel Development and Coding of Neural Feature Detectors. Biological Cybernetics, 1976, 23(1): 121-134 [9] Sutton R S, Barto A G. Reinforcement Learning: An Introduction. Cambridge, USA: MIT Press, 1998 [10] Xiao Nanfeng, Nahavandi S. A Reinforcement Learning Approach for Robot Control in an Unknown Environment // Proc of the IEEE International Conference on Industrial Technology. Bangkok, Thailand, 2002, Ⅱ: 1096-1099