Title :
Passivity properties of neuro-identifier
Author :
Yu, Wen ; Li, XiaoOu
Author_Institution :
Dept. de Control Autom., CINVESTAV-IPN, Mexico City, Mexico
Abstract :
In this paper the passivity approach is applied to access several stability properties of neuro-identifier. A dynamic neural network is used for nonlinear system online identification. By using a simple gradient learning law, the conditions for passivity stability, asymptotic stability and input-to-state stability are established. The result obtained shows that the gradient algorithm is robust with respect to all kinds of bounded uncertainties for the neuro-identifier
Keywords :
gradient methods; identification; learning (artificial intelligence); neural nets; nonlinear systems; stability; dynamic neural network; gradient algorithm; gradient learning; identification; nonlinear systems; stability; Automatic control; Circuit stability; Error correction; Multi-layer neural network; Multilayer perceptrons; Neural networks; Nonlinear dynamical systems; Nonlinear systems; Stability analysis; Vehicle dynamics;
Conference_Titel :
Decision and Control, 2000. Proceedings of the 39th IEEE Conference on
Conference_Location :
Sydney, NSW
Print_ISBN :
0-7803-6638-7
DOI :
10.1109/CDC.2000.912312
