Smith predictive control based on NN

Author

Wang, Pei-Guang ; Feng, He-Ping ; Zong, Xiao-ping

Author_Institution

Coll. of Electron. & Inf. Eng., Hebei Univ., Baoding, China

Volume

7

fYear

2005

fDate

18-21 Aug. 2005

Firstpage

4179

Abstract

Researching the application of neural networks and the Smith predictor in the system with time-delay, considering the delayed problem of the system, a new control approach is introduced, which can realize the predictive control by combining the Smith predictor with the NN and PID control which is tuned by BP neural network through the predictive errors effectively. We use the RBFNN in the Smith predictor, which is better than BPNN. The simulation shows that Smith predictive control based on NN has a strong robustness and better controlling character. A satisfactory controlling effect is obtained.

Keywords

backpropagation; compensation; delays; neurocontrollers; nonlinear control systems; predictive control; radial basis function networks; robust control; three-term control; tuning; PID control tuning; Smith predictive control; backpropagation neural network; predictive error; radial basis function neural network; robustness; time delay; Control systems; Educational institutions; Electrical equipment industry; Information analysis; Mathematical model; Neural networks; Optimal control; Predictive control; Predictive models; Three-term control; BPNN; PID control; RBFNN; Smith predictor;

fLanguage

English

Publisher

ieee

Conference_Titel

Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on

Conference_Location

Guangzhou, China

Print_ISBN

0-7803-9091-1

Type

conf

DOI

10.1109/ICMLC.2005.1527670

Filename

1527670