An approach to stable indirect adaptive control (Q1064300)

The stable indirect adaptive control of unknown linear continuous-time systems is treated subject to partial prior knowledge of the plant. It is assumed that the plant is of known order n, that bounds on the plant parameters are known, and that for each set of parameter values within these bounds the corresponding plant does not show an unstable pole-zero cancellation. Except for these assumptions, the plant is arbitrary and hence both minimum and nonminimum phase plants are included. The uniform boundedness of all states of the overall adaptive system is also proved. For this purpose, it is shown that, by using a special nonminimal representation of the plant, the closed loop system can be rewritten as an exponentially stable system with a feedback loop around it, where the feedback signal is exactly the identification error. From this argument, a direct link between the identification error and the closed loop stability is established.