Robustness enhancement of adaptive controllers by incorporation of process a priori knowledge

This paper addresses the problem of robustness improvement of adaptive controllers by incorporation into the design of the prior knowledge about the transfer function attainable with the ‘ideal’ fixed gain regulator, i.e. the tuned transfer function [2]. A modification to the adaptive control scheme, which has proven useful in practice [12], is proposed. It consists in the design of a fixed gain regulation loop around the tracking error. The additional compensator is chosen based on the estimated tuned transfer function. This approach, in contrast to other modified adaptive schemes [3–6,10], fully exploits the capability of the adaptive controller to tune to the unknown system due to the outer-loop nature of the compensation scheme.