Title :
Real-time neurocontrol of a pendulum system
Author :
Yang, Wei-Chung ; Hagan, Martin T.
Author_Institution :
18599 Mesa Verde Way, Castro Valley, CA, USA
Abstract :
This research investigates the use of neural networks for system identification and control of nonlinear dynamic systems. Supervised learning algorithms are used to train both feedforward and recurrent neural networks. The trained networks are implemented in real-time for the control of an experimental pendulum system
Keywords :
digital control; feedforward neural nets; learning (artificial intelligence); model reference adaptive control systems; neurocontrollers; pendulums; real-time systems; recurrent neural nets; feedforward neural networks; nonlinear dynamic systems; pendulum system; real-time neurocontrol; recurrent neural networks; supervised learning algorithms; system identification; training; Backpropagation algorithms; Control systems; Neural networks; Neurocontrollers; Nonhomogeneous media; Real time systems; Recurrent neural networks; Signal processing algorithms; System identification; System testing;
Conference_Titel :
Industry Applications Society Annual Meeting, 1994., Conference Record of the 1994 IEEE
Conference_Location :
Denver, CO
Print_ISBN :
0-7803-1993-1
DOI :
10.1109/IAS.1994.377661
