Implementation of Distributed Kalman Filter Based on Mutual Coupled Oscillators in Sensor Networks

The problem of distributed Kalman filter is one of the most fundamental collaborative information process problems for data fusion in sensor networks. This paper models the nodes of sensor networks as coupled oscillators and propose a novel implementation of distributed Kalman filter based on mutually coupled oscillators. This paper firstly demonstrates that the structure of distributed Kalman filter algorithm is equivalent to the local Kalman filter with the inputs from optimal ML estimator. Then mutually coupled oscillators are used for distributed ML estimate and thus implementing the distributed Kalman filter. This method not only provides a simple and convenient implementation of collaborative information process in sensor networks, but also maintains the robustness and reliability.