On a Parameter Adaptive Extremum Controller

An extremum controller based on parameter adaptive control is investigated in this contribution. The objective function to be minimized is parameterized using a simple nonlinear model, and the control is based on real-time estimation of the parameters of this model. Asymptotic analysis of the extremum controller gives insight concerning the compromise between two contradicting goals, namely accurate control and accurate estimation. This is similar in spirit to stochastic dual control. One motivation of this study is an application to automotive engine control, which is briefly described.