The Research On Sorting Of Various Waveform Radar Signals

Radar signal sorting is a key link of the electronic warfare equipment, Countries havepaid highly attention to its theory and technique. With the intense resistance of the modernelectronic warfare, the number of the electronic warfare equipment is greatly increasing. Thesignal environment becomes highly dense and the signal density of electromagnetic threatenvironment could reach millions. At the same time the signal forms of radar becomes morecomplex, such as the technology of various waveform radar signals. It makes the successprobability of the electronic warfare equipment have been enormously influenced. All theselead to the bad sorting effect of traditional sorting methods. In order to make the sortingtechnology adapt the complex signal environment. there has been a strong push for theresearch on the sorting technology and the new signal sorting processor structure which canadapt the complex signal environment. This article makes a thorough theoretical study ofcorrelation technique of signal sorting and proved by using the method of computersimulation.Firstly the complex electromagnetic environment of radar signal sorting is analyzed inthis paper and some common signal sorting technology is introduced here. Then we study theblind source separation(BSS) algorithm for radar signal sorting. Blind source separationtechnology can reconstruct original signals just through the observed mixed signals. It cansolve radar signal sorting under the complex environment well, and has many advantages insimultaneous arriving signals processing. We have studied FASTICA algorithm BSSalgorithm base on maximum Signal Noise Ratio the joint approximative diagonalization ofeigenmatrix(JADE) and the BSS switching algorithm in this paper. As the separationefficiency of original switching algorithm is not good when the mixed original signals is largein number. An improved algorithm is proposed, it has improved the performance of the BSSby loosening the selection of activation functions and using kurtosis value to learn activationfunction adaptively.In order to improve the performance of sorting and making up for the shortage that thetraditional parameters couldn’t describe the intra-pulse modulated features, the intra-pulsemodulated features of radar signals is studied in this paper. And then the feature parametersare extracted to sorting signals. It can reduced the probability of the overlap in multidimensional parameters space and improves the probability of sorting and recognizing.Main research topics include: Calculate the instantaneous frequency(IF) of differentmodulated signals by the method of self-correlation. And then extract the further feature ofthe instantaneous frequency as the complement of parameters. On the other hand, thefractional Fourier transform(FRFT) is studied, the different modulated signals are processedby fractional Fourier transform with the order p from0to4. Then curve of amplitude vvarying with order p is obtained. Make cluster analysis for the curve distributioncharacteristics parameters and the effectiveness of extracted characteristics parameters forsignal sorting is confirmed by computer simulation.