Guaranteed Cost Robust Centralized Fusion Steady-State Kalman Predictors with Uncertain Noise Variances and Missing Measurements

The guaranteed cost prediction problem for discrete-time systems with uncertain noise variances and missing measurements is considered where missing measurement is described as Bernoulli random variables. The system under consideration can he converted into that only with uncertain noise variances by introducing fictitious measurement noises, then two classes of guaranteed cost robust centralized fusion Kalman predictors are presented by Lyapunov equation approach based on minimax robust estimation principle and parameterization representation of uncertain noise variances. One class is to find the maximal perturbation region of uncertain noise variances so as to guarantee that the accuracy deviations remain within a given free degree. The other class is to find the maximal lower hound and the minimal upper hound of the accuracy deviations over the prescribed hounded perturbation region of noise variances. The proposed guaranteed cost predictors can give minimal upper and maximal lower hound of accuracy deviations. A simulation example shows the correctness and effectiveness of the proposed results.