On observer design for interconnected systems

From MaRDI portal





The authors develop decomposition methods for constructing observers for nonlinear systems. Taking account of the measured signal, an initial nonlinear system is considered as an interconnection between several subsystems, each of them satisfying some properties required for an observer to be easily computable. The aim is then to design an observer for the whole system with the aid of separate designs for each of the subsystems. The observer resulting from such a design is not always stable, as illustrated by examples in the paper. Two situations are considered: either the convergence rate of the resulting observer can be arbitrarily fast, which happens when the separate synthesis allows arbitrary rate of convergence or the rate of convergence, of the complete observer admits an upper bound, which occurs when one of the separate designs has a convergence rate imposed by the subsystem. With the aid of Lyapunov's second method sufficient conditions for separate designs to be possible are presented, highlighting the two situations which can occur. Examples showing the importance of the conditions derived are given. To illustrate the approach presented in the paper, observer designs for an induction motor are described under stator currents as well as both stator currents and rotor positions measurements.











This page was built for publication: On observer design for interconnected systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4398703)