Non-parametric radar signal sorting algorithm based on improved DSets

College of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, China Abstract: ［Objective］With the increasingly complex electromagnetic environment, it is more and more difficult to separate signals. In order to solve the problem that the performance of the existing efficient sorting algorithm depends heavily on external input parameters, a non-parametric radar signal clustering algorithm based on Dominant Sets (DSets) and Dempster-Shafer（D-S） evidence theory is proposed.［Method］The algorithm is applied to radar signal sorting; a non-parametric radar signal pulse clustering algorithm is then given and combined with D-S evidence theory to solve the problem of excessive DSets algorithm segmentation. ［Results ］ In the absence of any prior information, this algorithm can complete the precise clustering of mixed pulse radar signals. In addition, its sorting accuracy rate is more than 93.13% when the false pulse ratio is less than 50%. ［Conclusion］The combination of the DSets algorithm and D-S evidence theory can effectively complete the clustering of radar signal pulses without prior information, and achieves good clustering performance.

Key words: non-parametric clustering；dominant sets algorithm；histogram equalization；D-S evidence theory