Robust Kalman Predictor under Linearly Correlated Noise and Mixed Uncertainties of Noise Variances and Multiple Networked-inducements

The paper solves the robust Kalman prediction problem for systems with linearly correlated noise and mixed uncertainties of noise variances, multiplicative noises and multiple networked-inducements including missing measurements, packets dropouts and two-step random measurement delays. The original system with mixed uncertainties is transformed into one with only uncertain fictitious noise variance by the proposed model-transformation method. Then the robust steady-state Kalman predictor is presented by minimax robust estimation principle. The robustness of robust Kalman predictor is proved by the extended Lyapunov equation approach and matrix factorization. Finally, a simulation study applied to tracking system is provided to examine effectiveness and applicability of the proposed algorithm.