Research On Emitter Signal Sorting Techniques In Complicated Electromagnetic Environments

Radar emitter signal sorting refers to the signal processing of interleaved and overlapped emitter signals received and intercepted by the receiver,so that the signals emitted by the same emitter can be divided into the same category.In the current information era,emitter signal sorting has become an important part of electronic warfare,which has an important impact on electronic warfare.At present,with the rapid development of radar technology,there are many new radar systems.The electromagnetic environment in the battlefield is becoming more and more complex.How to avoid the signal overlapping and loss in the complex electromagnetic environment,and how to select the accurate emitter signal has become an urgent problem to be solved.This thesis first introduces the research background and significance of this topic,and then studies the research status at home and abroad.Through the analysis of the research status at home and abroad,it is found that the separation algorithm of radiation sources faces the problem of complex electromagnetic environment interference in practical application,and the separation effect still needs to be improved.Then it introduces the basic process of emitter signal sorting,analyzes the characteristic parameters of each dimension and the change mode of each parameter in the current new radar,and makes the data basis for emitter signal sorting.Then it introduces two classic PRI interlacing estimation algorithms and a sequence search algorithm for main sorting,and analyzes the original performance degradation of traditional algorithms in the face of complex electromagnetic environment At last,the thesis introduces the machine learning technology and its application in the radiation source selection.In order to solve the problem of poor sorting performance of traditional sorting method in complex electromagnetic environment,a new sorting method is proposed by combining machine learning algorithm with emitter sorting.The main work and research results are as follows:In view of the characteristics of traditional unsupervised learning clustering and sorting methods that need to know the types of radiation sources in advance,and the selection of clustering centers has a great influence on the sorting effect,this thesis first studies the clustering and sorting algorithm of FCM(fuzzy c-means)radar radiation source signals,and then proposes the AP(affinity propagation)clustering algorithm and introduces the AP clustering technology into radar signal sorting,which greatly improves the reliability of the algorithm by using its application of attraction propagation characteristics,and introduces the ADAP algorithm for the vibration elimination problem in the AP algorithm.The reliability of the three clustering and sorting algorithms is verified by experiments,and the application scenarios of the three algorithms are analyzed.Aiming at the problem that AP unsupervised learning algorithm can't use the known information,the semi supervised learning algorithm sap(semi supervised affinity propagation)based on AP algorithm is introduced into the radiation source sub selection,and seed AP algorithm is used and some improvements are made for the possible violation of constraint information in SAP algorithm results.Experimental results show that the improved algorithm can make full use of constraint information and prior information,so that the clustering effect is better.