To separate the target group siganl in midcourse, a method based on time-frequency filtering is proposed in this paper. Firstly, the micro-motion period of one target is estimated based on the auto correlation method. Then, the signal is divided to several segments based on the estimated period. Also, the strong energy area in time-frequency domain for each segment signal is obtained by S-transform. The intersection of strong energy areas for different period can be seem as the support area of target. So, the signal attribute to one target can be obtained by time-frequency filtering based on the estimated support area. Simulation results verify the effectiveness of the proposed algorithm.

The National Natural Science Foundation of China (61372166)