Radar Emitter Recognition Method Based On Pattern Recognition Technology

The modern warfare is the high-tech war. The electronic warfare plays the important role in the modern warfare. Reconnaissance of radar emitter is key link in electronic warfare process. Technology about radar emitter recognition therefore becomes the important subject in domestic and foreign military science and technology domain.Database matching and search is the primary method in traditional radar emitter recognition. Firstly, a database of radar parameters is bulit. These parameters is acquired by any means. Then by means of database matching a new radar emitter whose parameters is acquired by reconnaissance radar can be recognised. Later, expert system which include some new method such as fuzzy matching and evidence theory and inference engine is used in radar emitter identification. However, these methods can not identify a new unknown radar emitter. In order to identify new unknown radar emitter a new method must be researched.In this thesis a new method which applies pattern recognition technology is proposed to identify a new radar emitter. Firstly, this paper introduces some knowledge about the radar classification and radar emitter identification and also analyzes the relation between the radar parameter and the radar use. Secondly, several typical pattern recongition classification techniques including Bayesian Classification, k-Nearest-Neighbor classification, Cluster Analysis method and Artifical Neural Networks classification is introduced in detail. Finally, a new classification based on support vector machine method combine whit principal component analysis technique is proposed. The results of experiments indicates that this classification can acquire a better effect in radar emitter recognition.