Multi-channel networked data fusion with intermittent observations

In this paper, the problem of networked data fusion is considered. Based on the state estimation algorithm with packet delay in single channel, two multi-channel state estimation methods are presented. A centralized architecture and a distributed architecture with buffer have been proposed. Modified Kalman Filter is adopted to solve the state estimation with measurement delay and loss. The stability of the system is analyzed and it shows that the trace of the error covariance after fusing is much smaller and the performance of the filter is much better than that without buffer. Two numerical examples are given to show the validity of the proposed techniques.