제어로봇시스템학회:학술대회논문집
- 1992.10b
- /
- Pages.56-60
- /
- 1992
A tracking controller using multi-layered neural networks
- Bae, Byeong-Woo (Department of Electronics Kyungpook National university Taegu) ;
- Jeon, Gi-Joon (Department of Electronics Kyungpook National university Taegu) ;
-
Kim, Kyung-Youn
(Department of Electronic Cheju national University)
- Published : 1992.10.01
Abstract
This paper addresses the problem of designing a neural network based controller for a discrete-time nonlinear dynamical system. Using two multi-layered neural networks we first design an indirect controller the weights of which are updated by the informations obtained from system identification. The weight update is executed by parameter optimization method under Lagrangian formulation. For the nonlinear dynamical system, we define several cost functions and by computer simulations analyze the control performances of them and the effects of penalty-weighting values.
Keywords