Distributed and parallel multisensor fusion in the wavelet domain

By combining the wavelet domain filtering and multisensor linear information fusion rule, we propose a new distributed and parallel algorithm for multi-sensor information fusion using the theory of wavelet domain multi-scale analysis. The proposed algorithm decomposes the single sensor's information into different scale, and then estimates the target state in different scale; then the local state estimates are transmitted to the fusion-center to get the optimal information fusion estimate. The posterior Cramer-Rao lower bound (PCRLB) of the proposed algorithm is also given. An example of radar tracking system with three sensors is provided to illuminate the effectiveness of the method.