Some results of computer research of basic algorithm of adaptive identification

Abstract

Qualitative (graphic) analysis of the common properties of Basic Algorithm, when presence of Gaussian noises has place, is given. On the basis of its computer simulation with use of computing tools of MatLab v.6.5 it is shown, that influence of noise complicates dynamics of convergence. If there is no noise, the quality identification increases for move multistage realizations. The realizations, which have the amount of computing stages, numerically less or equal to dimension of control object input vector, are making worse an identification quality. After further increasing of amount of computing stages basic algorithm multistage realizations improve an identification quality in comparison with realizations, which have less amount of computing stages.