OPTIMAL FUSION REDUCED-ORDER KALMAN ESTIMATORS FOR DISCRETE-TIME STOCHASTIC SINGULAR SYSTEMS

摘要

Based on the optimal fusion algorithm weighted by scalars in the linear minimum variance (LMV) sense, the distributed optimal fusion reduced-order Kalman estimators including predictor, filter and smoother are presented for discrete-time stochastic singular linear systems with multiple sensors and correlated noises. The fusion estimation problem of original high-order singular system is transferred to that of two reduced-order subsystems. They have better precision than any local estimators from every sensor do. The estimation error cross-covariance matrices between any two sensor subsystems are derived for two reduced-order subsystems, respectively. Furthermore, the steady-state fusion estimators are also investigated, which have the reduced online computational burden. A simulation example with three sensors shows the effectiveness.