Mixed H_2/H_∞ filtering with estimation error variance constraints

The mixed H2/H∞ filtering problem for linear systems with steady-state error variance constraints is considered in this paper. The filter is directly obtained by solving a linear matrix inequality and the following requirements are satisfied: The filter is asymptotically stable and the H∞ norm is minimized with the steady-state variance of the estimation error of each state being not more than the prespecified upper bound. An numerical example is provided to demonstrate the flexibility of the proposed approach.