State estimation algorithms for linear systems with strong disturbance

This paper considers state estimation algorithms for linear systems contaminated by complex strong disturbance. The strong disturbance is characterized as a general random matrix, which can be non-diagonal, and its components (the parameter of each observation channel) can be simultaneously corrected. For this strong disturbance, we propose an optimal filtering algorithm to filter it in order to estimate the state by using projection theorem. We also present an optimal fixed-interval smoothing and a deconvolution algorithm to address it. Simulation results verify the presented algorithms.