Adaptive control with stochastic gradient algorithm: rate of convergence

Author

Radenkovic, Miloje S.

Author_Institution

Dept. of Electr. Eng., Colorado Univ., Denver, CO, USA

Volume

3

fYear

1995

fDate

21-23 Jun 1995

Firstpage

1585

Abstract

In this paper, we consider the rate of convergence of the parameter estimation error and the cost function for the stochastic approximation type algorithm. The problem is solve in the case of the minimum-variance stochastic adaptive control. It is proven that stochastic approximation algorithm has the same rate of convergence as the one established for the least-squares algorithm. Comparison of the two algorithms is made under the same conditions related to the process noise and the structure of the system model

Keywords

adaptive control; convergence of numerical methods; discrete time systems; least squares approximations; parameter estimation; stability; stochastic processes; SISO systems; adaptive control; convergence; cost function; discrete time systems; least-squares algorithm; minimum-variance; parameter estimation error; process noise; stability; stochastic approximation; stochastic gradient algorithm; system model; Adaptive control; Adaptive systems; Approximation algorithms; Convergence; Cost function; Parameter estimation; Polynomials; Stochastic processes; Stochastic resonance; Stochastic systems;

fLanguage

English

Publisher

ieee

Conference_Titel

American Control Conference, Proceedings of the 1995

Conference_Location

Seattle, WA

Print_ISBN

0-7803-2445-5

Type

conf

DOI

10.1109/ACC.1995.529774

Filename

529774