Adaptive control of nonlinear servo system: Comparison of standard and dual approaches

Abstract

The general task of optimal adaptive control with recursive identification (self-tuning control) is very complicated problem. This problem is solved usually by the separation of identification and control - the Certainty Equivalency (CE) Principle. The aim of this paper is to present the solution of this problem using the Dual Adaptive Control (Bicriterial Approach). The main idea of this approach involves two cost functions: (1) the system output should track cautiously the desired reference signal; (2) the control signal should excite the controlled process sufficiently for accelerating the parameter estimates. This approach was verified and compared with some other adaptive control approaches by a real-time control of highly nonlinear laboratory model DR300 Speed Control with Variable Load.