Radar Emitter Recognition in Airborne RWR/ESM Based on Improved K Nearest Neighbor Algorithm

The electromagnetic environments of war-field are intricate and complex, there exists a significant challenge to RWR/ESM (Radar Warning Receiver / Electronic Support Measure) for recognizing radar emitter signals. Because of the computing ability of RWR/ESM is limited by the hardware, the traditional liner algorithm of radar emitter recognition need a long time, which cannot adapt to the electromagnetic environments. In this paper, considering the front-end of RWR system is frequency dividing receiver, then the frequency dividing method can also be used to the radar sample database. The radar sample database is divided into some radar sample sub database by the front-end frequency of RWR, and using K nearest neighbor algorithm in radar sample sub database to decrease the computing time. We compared the computing time of the tradition liner algorithm and of the K Nearest Neighbor algorithm in radar sample sub database, and the results of simulation showed that the improved K Nearest Neighbor mating algorithm is efficient.