Switching adaptive filter design using Bayesian classification approach for multi-sensor data fusion

The aim of this paper is to present an adaptive filtering fusion approach for tracking the same maneuvering target in a multi-sensor environment. The hierarchical estimation fusion consists of several local nodes and a global node. A linear Kalman filter is employed by each local node to perform the tracking function and the resulting track file communicates to the global node. In the global node, an algorithm, which consists of dual-band Information Matrix Filter (IMF) and a two-category Bayesian classifier, is employed to generate an appropriately global estimate. By incorporating Bayesian decision rule into a classification scheme, a Bayesian classifier is developed which involves switching between high-level-band IMF and low-level-band IMF against the rapid variation of target dynamics. The proposed filter, so-called switching adaptive filter, has better estimation accuracy than each individual IMF. Computer simulation results are included to demonstrate the effectiveness of proposed algorithm.