Abstract
This paper studies the decentralised filtering problem for interconnected discrete-time systems. In the developed decentralised filter scheme, only local output information is available for each local filter. The dynamics of the original system and the dynamics of the proposed filter are expressed as an augmented system. The system matrices of the augmented system consist of matrix inversion terms, which are computationally expensive and even meet numerical stability problems. In order to avoid the matrix inversion computation, this paper derives novel computationally attractive sufficient conditions without inversion terms to guarantee the
performance of the augmented system. Then, these conditions are transformed into linear matrix inequalities. The decentralised filter parameters can be obtained by solving linear matrix inequalities. Finally, two simulation examples are given to demonstrate the effectiveness and the advantages of the proposed approach.
Disclosure statement
No potential conflict of interest was reported by the authors.