Title :
Adaptive prediction using fuzzy systems and neural networks
Author :
Spooner, Jeffrey T. ; Passino, Kevin M.
Author_Institution :
Control Subsyst. Dept., Sandia Nat. Labs., Albuquerque, NM, USA
Abstract :
A collection of adaptive prediction schemes which use the functional approximation properties of fuzzy systems is presented. Both direct and indirect approaches are developed using gradient and least squares update laws. It is proven that the prediction error converges asymptotically to zero for each case provided some minor assumptions hold
Keywords :
discrete time systems; feedforward neural nets; function approximation; fuzzy systems; least squares approximations; multilayer perceptrons; neurocontrollers; prediction theory; adaptive prediction; direct approach; functional approximation; fuzzy systems; gradient laws; indirect approach; least squares update laws; neural networks; prediction error; Adaptive control; Biological neural networks; Control systems; Fuzzy sets; Fuzzy systems; Laboratories; Least squares approximation; Least squares methods; Neural networks; Programmable control;
Conference_Titel :
American Control Conference, 1997. Proceedings of the 1997
Conference_Location :
Albuquerque, NM
Print_ISBN :
0-7803-3832-4
DOI :
10.1109/ACC.1997.609738
