Step-Size Rules for Lie Bracket-Based Extremum Seeking With Asymptotic Convergence Guarantees

IEEE Control Systems Letters, vol. 8, pp. 1967–1972

Abstract

In this paper, we derive a class of step-size rules (time-varying gains) for gradient-based extremum seeking algorithms that guarantee classical asymptotic convergence rather than practical convergence. The obtained step-size rule conditions are similar to the classical step-size rules known in stochastic approximation theory.

Authors 2

1. Victoria Grushkovskaya

   University of Klagenfurt

   Affiliation as printed

   Department of Mathematics, University of Klagenfurt, Klagenfurt am Wörthersee, Austria

2. Christian Ebenbauer Aachen Chair of Intelligent Control Systems

   RWTH Aachen University

   Affiliation as printed

   Chair of Intelligent Control Systems, RWTH Aachen University, Aachen, Germany