Robust and adaptive design of numerical optimization-based extremum seeking control

We consider the employment of numerical optimization and state regulation to solve the extremum seeking control (ESC) problem, which does not assume the time scale separation between the plant dynamics and the extremum seeking loop. Extremum seeking is realized via a state regulator that drives the state traveling along a convergent set point sequence generated by a numerical optimization algorithm. In this paper, we propose a novel design of an asymptotic state regulator via output tracking for state feedback linearizable systems, where we trade off finite time state regulation to obtain flexibility in designing a robust extremum seeking controller. Existing techniques such as nonlinear damping and nonlinear adaptive control are then used to deal with input disturbance and unmodeled plant dynamics. Simulation examples illustrate the effectiveness of the basic and robust extremum seeking schemes, and some design guidelines are provided for engineering applications.