A Class of Suboptimal Distributed Consensus Filter for Sensor Networks with Correlated Noises

In this paper, the problem of distributed consensus filtering in sensor networks with correlated noises is studied. In the sensor networks, each sensor is equivalent to a local fusion center. A class of suboptimal distributed consensus filter is proposed based on a matrix inequality. First, each sensor node produces a local filter based on its own measurement data. Then, each node iteratively performs fusion estimation based on the estimates of itself and neighbor nodes through multiple data exchanges between nodes. Finally, the fusion estimate of each node trends to be consensus. The effectiveness of the algorithm is verified by simulation experiments.