A fusion algorithm for cooperative reconnaissance based on the correlation of sparse representation

Information fusion is an important step in cooperative reconnaissance signal processing, which can be divided into three types: data-level, feature-level and decision-level fusion. Data-level fusion has the advantage of information completeness but the large data volume makes it hard to apply practically. Based on the sparse representation of intercepted signal and the correlations between signals, a data-level fusion algorithm is proposed. In the algorithm, the intercepted signal is expressed in sparse representation and compressed for communication. Then fuse the sparse coefficients based on the correlations between them. The experiments indicate that the algorithm can reduce the amount of data effectively with the same information and compared with the existing sparse representation fusion algorithm based on activeness of atom, the algorithm can reach better fusion effect.