Convergence and synthesis issues in extremum seeking control

Abstract

Convergence and synthesis issues in an extremum seeking (also peak seeking) control scheme are addressed. The goal of an extremum seeking controller is to operate at a set-point that represents the optimal value of a functional being optimized in the control loop. The results presented here are based solely on assumptions on the curvature of the function being optimized in the loop. Other than pure integrators, we present a general class of extremum seeking compensators.