Abstract：This paper introduces a new kind of backward self-learning control strategy based on the neural networks joined with functional relationship. This control strategy is mainly used to solve problems of the movement control of robot arm with two joints. The simulation and test show that this control strategy is the able to increase the speed of self-adaptation and the starting property of the control system. In order to increase the stability of the control system, ation a series learning scheme is added into the control system and the control system then becomes more stable.
1 Mukhopadhyay S.Disturbance Rejection in Nonlinear System Using Neural Network. IEEE Trans,on Neural Net work,1993, 4(1) 2Musavi M T. On the Training of Radial Basis Functi on class ifiers. Neural Net work,1992, 5:595-603 3 焦李成. 神经网络的应用与实现. 西安电子科技大学出版社, 1993 4Giles C L, Maxwell T. Learning, Invariance and Generalization in High-order Neural Net works. A pplied Optics,1987,26(23):4972-4978 5 Kohonen T.The Self-organizing Map. Proc of IEEE, 1990, 78(9):1464-1479 6 Albus J S.A New Approach to Manipulator Control:The Cerebellar Model Articulation Controller (CMAC).ASME. Dynam Syst Meas Contr, 1975,97(3):220-227
