engleski

On-line Approximation of Dynamical System Nonlinearities with Neural Networks

The paper deals with the possibilities of using on-line nonlinear function approximation with neural networks in control system design procedures. Used it this way, a neural network can externally linearize a dynamical system or can make better tracking performance. This approach also opens possibilities of using linear control design procedures in nonlinear system controller design. Neural networks are trained on line as in the adaptive control schemes.

on-line approximation; adaptive control; neural networks

nije evidentirano