Input design for asymptotic robust H2-filtering

In this paper we study the problem of robust discrete-time H₂ filtering using a Linear Matrix Inequality approach. By assuming that the number of samples available for the identification of the system is large enough, we describe the filter design problem as a semidefinite program. Afterwards, the problem of designing an input signal for the identification of the system, to improve the performance of the conceived filter, is examined, and we show how to solve this problem using convex optimization.