Design of observers for time-varying discrete-time descriptor systems
- 1 September 1987
- journal article
- Published by Taylor & Francis in International Journal of Control
- Vol. 46 (3) , 841-848
- https://doi.org/10.1080/00207178708547397
Abstract
In this note, linear discrete-time time-varying ‘descriptor’ systems, i.e. control systems of the form Ek+1 Xk+1 = AkXk + Bkuk, yk = Ckxk are considered where Ek+1 and Ak+1 are constant, not necessarily square, matrices. The problem is to design a reduced-order observer for the system. Using matrix generalized inverses and singular-value decomposition, a procedure is developed which either leads to a routine observer-design problem involving an ordinary discrete-time state equation or proves that no observer exists for the given ‘descriptor’ system. The method of this note is partly based on an earlier paper of the authors on observer design for continuous-time time invariant systems of the form E dx/dt=Ax + Bu, y=Cx where E and A are square matrices (El-Tohami et al. 1983) and makes use of some results of Luenberger on discrete-time time-varying descriptor systems (Luenberger 1977).Keywords
This publication has 5 references indexed in Scilit:
- On the design of observers for generalized state space systems using singular value decompositionInternational Journal of Control, 1983
- On the application of matrix generalized inverses to the design of observers for time-varying and time-invariant linear systemsIEEE Transactions on Automatic Control, 1980
- The singular value decomposition: Its computation and some applicationsIEEE Transactions on Automatic Control, 1980
- An introduction to the application of the simplest matrix-generalized inverse in systems scienceIEEE Transactions on Circuits and Systems, 1978
- Dynamic equations in descriptor formIEEE Transactions on Automatic Control, 1977