Robust H∞, filtering for uncertain linear systems with unstable modes

Abstract This paper presents a robust H ∞ filter design method for a class of uncertain linear systems with unstable modes. The uncertain parameters are time-varying and their admissible values and rates of variation lie in a given polytope. An LMI method based on a parameter-dependent Lyapunov function is proposed for designing linear asymptotically stable filters which ensure a prescribed upper-bound on the L 2 -gain from the noise signals to the estimation error, for all admissible values and rates of variation of the uncertain parameters.