A Non-Parametric Model-Based Intelligent Recognition Method for Radar Emitters
REN Haohao1， MA Hanfei2， WEI Yulu1， YU Sihui1， ZHOU Yun1
1. School of Information and Communication Engineering， University of Electronic Science and Technology of China， Chengdu 611731， Sichuan， China； 2. Shanghai Academy of Spaceflight Technology， Shanghai 201109， China
Abstract：Given the problem of increasing the number of radar emitters and pulse density in a current complex electromagnetic environment, this study proposed an intelligent recognition algorithm using a non-parametric model to achieve high-precision radar emitters recognition. The proposed method was composed of three vital steps: static clustering of radar emitters, selection of representative points for radar emitter classes, and oriented-data stream incremental recognition. Compared to conventional clustering analysis methods, the proposed method utilized the clustering idea of Dirichlet Process Mixture Models (DPMM), which adaptively realized radar emitters’ class sorting. To reduce the data storage burden while maximizing the preservation of class distribution, a gradient-based representative point selection strategy for emitter classes was established. Besides, to satisfy the requirement of incremental processing of the data stream, a label propagation-based classification inference method was applied to achieve incremental radar emitters recognition. Finally, the proposed algorithm was validated using simulated radar emitters’ data in two scenarios. The evaluation results show that the proposed method is more robust and accurate than the other seven advanced emitters recognition clustering algorithms.
