Identification and control of dynamical system by one neural network

The paper proposes a new neural control scheme that can perform identification and control for a dynamical system with linear and nonlinear uncertainties. This scheme uses a single neural network for both the identification and the control. By using the Lyapunov stability technique, stability of the proposed scheme is analyzed and a sufficient condition of the local asymptotic stability is derived. Then, a computer simulation is performed in order to illustrate the effectiveness and the applicability of the proposed scheme.

[1]  Eduardo F. Camacho,et al.  A neural network based feedforward adaptive controller for robots , 1995, IEEE Trans. Syst. Man Cybern..

[2]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1992, Math. Control. Signals Syst..

[3]  Takayuki Yamada,et al.  Neural network controller characteristics with regard to adaptive control , 1992, IEEE Trans. Syst. Man Cybern..

[4]  K S Narendra,et al.  Control of nonlinear dynamical systems using neural networks. II. Observability, identification, and control , 1996, IEEE Trans. Neural Networks.

[5]  Makoto Kaneko,et al.  Adaptive neural controller for a class of plant with nonlinear uncertainties , 1996, Proceedings of 4th IEEE International Workshop on Advanced Motion Control - AMC '96 - MIE.

[6]  Marzuki Khalid,et al.  Temperature regulation with neural networks and alternative control schemes , 1995, IEEE Trans. Neural Networks.

[7]  Kwang Y. Lee,et al.  Diagonal recurrent neural networks for dynamic systems control , 1995, IEEE Trans. Neural Networks.