Indirect Adaptive Control: An Integrated Approach

The aim of this paper is to present the basis of the adaptive control approach being pursued in the Centre for Industrial Control Science at the University of Newcastle. A key feature of this approach is the separate optimization of procedures for estimation and on-line control law synthesis. Regarding the estimation module, attention is given to such features as data filtering, the response of the estimator, techniques for handling unmodelled dynamics and the question of numerical accuracy. Regarding the control law synthesis module, attention is given to robustness to unstructured modelling errors, disturbance and set-point feedforward techniques and implementation issues. In the paper, these various aspects are illustrated by reference to a case study problem proposed by Mike Master and Herb Cohn from Texas Instruments.