论文信息 - Adaptive feedforward cancellation for nonlinear systems using RLLM networks

Adaptive feedforward cancellation for nonlinear systems using RLLM networks

An algorithm for adaptive feedforward cancellation (AFC) of high frequency measurable disturbances for a class of nonlinear dynamical systems is developed. A rectangular local linear model (RLLM) network is used to model nonlinear dynamical behaviour of the system and to predict the effect of the disturbances. Another network of the same kind containing local linear compensation units is used as a feedforward controller based on the prediction mentioned above. This scheme is appended with a feedback controller for the sake of better tracking performance.

H. Unbehauen | T. F. Junge | A. Ali

[1] Frank W. Paul,et al. Nonlinear System Identification with Fourier Series Neural Networks , 1997 .

[2] M. Ashfaq,et al. Model reference adaptive control of nonlinear systems using RLLM networks , 2000, Proceedings of the 2000 American Control Conference. ACC (IEEE Cat. No.00CH36334).

[3] Heinz Unbehauen,et al. Real time learning control of an emergency turbo-generator plant using structurally adaptive neural networks , 1998, IECON '98. Proceedings of the 24th Annual Conference of the IEEE Industrial Electronics Society (Cat. No.98CH36200).

[4] H. S. Ranganath,et al. On-line identification of linear systems , 1971 .

[5] Tor Arne Johansen,et al. Design and analysis of gain-scheduled control using local controller networks , 1997 .

[6] Kumpati S. Narendra,et al. Identification and control of dynamical systems using neural networks , 1990, IEEE Trans. Neural Networks.

[7] Thomas Fernando Junge. "On-line"-Identifikation und lernende Regelung nichtlinearer Regelstrecken mittels neuronaler Netze , 1999 .