Abstract
New developments are presented on the parameter identification of general discrete non-linear systems with linear parameters in both the deterministic and stochastic cases. The algorithms presented are based on the least-squares method. Parameter convergence and output error are studied and preliminary persistently exciting conditions are given. Under some conditions, it is proved that the identified parameters are guaranteed to converge to their original values. Three computer simulation experiments are carried out and simulation results show good performance of the algorithms for the non-linear systems described.

This publication has 13 references indexed in Scilit: