Multi-Sensor Weighted Fusion Suboptimal Filtering for Systems with Multiple Time Delayed Measurements

This paper is concerned with distributed fusion estimation for discrete-time stochastic linear systems with multiple sensors having multiple time delayed measurements. A distributed weighted fusion suboptimal Kalman filter is given based on the local suboptimal Kalman filters and the optimal fusion algorithm weighted by matrices in the linear minimum variance sense. Compared with the augmented Kalman filter and the fusion optimal filter, it avoids the expensive high-dimension computation and the complicated smoothing computation. So it has the reduced computation burden. The suboptimal filtering error cross-covariance matrix between any two subsystems is derived. Applying it to a tracking system with three sensors demonstrates its effectiveness