Search

This paper adresses the important issue of robustness for direct multivariable adaptive control. Our interest is focused on minimum phase processes which are not necessarily decouplable by static state feedback. This paper presents a multivariable version of a robustness result given by Ortega Praly and Landau (1984) where the analysis is carried out, using the sector stability and passivity theorems. An application of this result to a cement kiln model is given. Some simulations emphasize the fact that our sufficient conditions are far from being necessary.

In this paper, we are interested by the adaptive control of discrete multivariable minimum phase systems not necessarily characterized by a diagonal interactor matrix. We present some simulation results which emphasize the behaviour of different algorithms. Constant trace gain algorithms, projection algorithms and variable gain algorithms will be treated. The adaptive control design presented here allows to specify independently deterministic regulation and tracking objectives. The global convergence of the overall scheme is ensured.

This paper describes algorithms for model reference adaptive control of linear multivariable systems. A survey of recent work on MIMO direct adaptive control is presented, focusing on the general case of processes which are not decouplable by static state feedback. The principal aim of this paper is to examine parametrizations of linear system models for use in adaptive control. More precisely, comparison is made of the linear control schemes as well as the necessary a priori knowledge on the process in order to achieve some specified objectives.