A Gaussian mixture probability hypothesis density smoothing algorithm for multi-target track-before-detect

摘要

When the signal-to-noise ratio (SNR) is reduced in case of track-before-detect (TBD) for weak target detection, the TBD algorithm based on Gaussian mixture probability hypothesis density (GM-PHD) filter cannot estimate the number or status of targets accurately. In order to solve this problem, a TBD algorithm based on GM-PHD smoothing filter (SGM-PHD-TBD) is proposed. Within the framework of TBD standard observation model, the algorithm employs smooth recursive method, using quantities of measurement data to smooth the filtering results. The simulation result shows that the proposed algorithm is better than the GM-PHD-TBD algorithm under low SNR, especially in the accuracy of target number estimation and the precision of target status estimation.