Showing papers by "Jamal Daafouz published in 2003"
TL;DR: It is shown that IIGS may result from both a special partial-pole placement and the notion of polyquadratic stability, and an input-independent global synchronization (IIGS) observer-based method is proposed that differs from the standard unknown input observers framework.
Abstract: The work presented in this paper deals with chaos synchronization of discrete-time switched systems for communication purposes. In a chaotic masking scheme, the main problem arising lies in the fact that the state-reconstruction-error equation involves a term depending on the embedded information which may prevent the convergence. To tackle such a situation, an input-independent global synchronization (IIGS) observer-based method is proposed. It differs from the standard unknown input observers framework, which cannot be applied for the class of systems considered in this paper. It is shown that IIGS may result from both a special partial-pole placement and the notion of polyquadratic stability. Two configurations are distinguished: a subsystem configuration and an observer one. The analysis in the subsystem configuration or the design in the observer configuration is performed by solving a tractable linear-matrix inequalities setting. Two illustrative examples are given.
62 citations
09 Dec 2003
TL;DR: It is shown in context of switched systems that there is some possibility to take advantage of the study of a convex embedding problem for which a solution exists and how suboptimal chattering solutions can be found.
Abstract: In this paper suboptimal control in the case of hybrid systems is addressed. After a brief recall of necessary conditions on the optimal hybrid trajectory, it is shown that many hybrid optimal problems cannot have an hybrid solution. Nevertheless, we show in context of switched systems that there is some possibility to take advantage of the study of a convex embedding problem for which a solution exists. Indeed the presence of singular arcs in this solution explains why the original switched system has no solution and how suboptimal chattering solutions can be found.
23 citations
01 Sep 2003
TL;DR: The main result consists in proving a separation principle for linear discrete-time switched systems which means the design of the switched state feedback control and the switched observer can be carried out independently.
Abstract: In this paper, we consider switched linear discrete-time systems. We propose a method to design an observer-based switched control which guarantees that the switched system is asymptotically stable. The main result consists in proving a separation principle for linear discrete-time switched systems. Hence, the design of the switched state feedback control and the switched observer can be carried out independently. Such a design is formulated in terms of Linear Matrix Inequalities (LMI).
11 citations
TL;DR: A method to design a switched state feedback control which allows to place the poles of each subsystem in desired locations is proposed and guarantees that the switched system is asymptotically stable using Lyapunov arguments.
5 citations