Linear minimum variance estimators for systems with bounded random measurement delays and packet dropouts

For discrete-time stochastic linear systems with bounded random measurement delays and packet dropouts, the optimal estimators including filter, predictor and smoother are developed in the linear minimum variance sense based on the innovation analysis approach. Some binary distributed random variables with known distributions are employed to describe the phenomenon of random delays and packet dropouts. Compared with the augmented approach, the computational cost is reduced. Furthermore, the proposed algorithm also gives a suboptimal estimate for systems with unbounded delays and packet dropouts by selecting a sufficient large upper bound. A simulation shows the effectiveness of the proposed algorithms.