Research On Radar Emitter Signal Sorting Technology

Signal sorting of radar emitter is the key technology of reconnaissance system and the key link of radar countermeasure information processing in modern electronic warfare.The signal sorting system can de-interleave the complex radar signal,extract the characteristic information,identify of radar emitter,assess threat level and provide important military intelligence.In view of the problem of signal sorting of radar reconnaissance equipment in complex electromagnetic environment,The following aspects are studied.The complex electromagnetic environment of the battlefield is analyzed.The common radar signals are explained and their mathematical models are established.On this basis,the time domain,frequency domain and spatial domain characteristics of the signal are extracted.The radar reconnaissance system and its signal processing system are studied,and the key factors affecting the signal sorting effect are analyzed.These factors will be considered in the subsequent simulation experiments.Aiming at the problem that signal pulse density intercepted by radar reconnaissance receiver is high.A method to mine the intra-pulse resemblance coefficient features of signals and then cluster them is proposed.The dilution of signal flow is realized and the pressure of separation is reduced.As a tool,the similarity coefficient can measure the degree of similarity between two signals.In this paper,the resemblance coefficient of radar signal are calculated with rectangular signal and triangular signal respectively,forming a joint eigenvector.Then the joint eigenvector are clustered by k-means to classify radar signals with different modulation modes.In addition,the classical k-means algorithm needs to be given the number of clustering in advance is improved.The idea of incremental clustering is applied to signal flow dilution.Finally,the feasibility and effectiveness of signal dilution using similarity coefficient are verified by simulation and actual data experiment.As a data mining method,clustering can mine the internal characteristics and distribution rules contained in the data and has been applied to the radar signal sorting field for signal processing.Aiming at the problem that the performance of traditional FCM clustering is degraded when the signals overlap seriously,wavecluster algorithm is applied to signal sorting.Wavecluster algorithm is formed by the combination of wavelet transform and grid clustering.Based on the wavecluster model of radar signal,the workflow of signal sorting using wavecluster is presented and the time complexity of the sorting method is analyzed.The algorithm has the advantages of insensitivity to input data volume and input data order,and can find clusters of arbitrary shapes.At the same time,wavelet transform is used to realize fast multi-scale and multi-resolution clustering and noise removal.On the other hand,the algorithm does not need to give the number of clustering in advance,which increases its practicability.Computer simulation results show that the proposed algorithm can achieve high sorting accuracy in the case of serious overlap of signal parameters and large jitter,which verifies the effectiveness of the proposed algorithm.