The Research Of Information Fusion Algorithm Based On The Estimation Of Target State

The problem of target state estimation using information fusion technology is concerned in the thesis. And two improved fusion algorithms were proposed. One is the measurement fusion algorithm based on fuzzy Kalman filter. By using the algorithm the innovation process can be efficiently controlled and the divergence is also prevented when innovation process isn't Gauss white noise through weighting the measurement noise and system noise by adaptive fuzzy logic system. The second algorithm is fully decentralized state-vector fusion algorithm in which the communicate structure of Kalman filters and particle filters is used. In the algorithm each sensing node arrives at a partial decision and then broadcasts the information of state error and variance error to every other node. Each node then assimilates this received information to arrive at last estimation. Using the algorithm, potential computation bottlenecks can be removed. Therefore, the algorithm implement faster than measurement fusion method because of the ability of parallel processing.The performance of the two fusion algorithms presented here is researched by numerical simulation. The results show that better estimation accuracy and robustness can be obtained comparing to the traditional method.