Detection And Parameters Estimation Of LFM Signal Based On Time-Frequency Analysis

Linear Frequency Modulated(LFM) Signal is a kind of nonstationary signal. As awide bandwidth signal with Generalized Time-Bandwidth Product, it has high rangeresolution. Thus, LFM signal has been well applied to many fields in radar informationsystem, such as target detection, precise position, target identification and imaging etc.The detection and parameters estimation of LFM signal have been one of the main tasksin Electronic Warfare signal analysis.This paper is mainly concerned with researching the detection and parametersestimation methods of LFM signal, mainly as follows:Firstly, the basic characteristics of LFM signal are introduced, and then theshort-time Fourier transform(STFT), the Wigner-Ville distribution(WVD) and theCohen distribution are studied, and then, the phenomenon of Cross-terms and thetechnology of Kernel design for reducing interference distributions are analyzed. Theconclusion that there is a compromised relationship between the Time-Frequencyconcentration and the Cross-terms is obtained. And then, the algorithm with rapidcorrelative dechirp estimation is researched. It draws the conclusion that the precision ofparameters estimation is related to the delay time. Then, to overcome the disadvantagesof the above algorithms, a Time-Frequency analysis method for detecting andestimating the multicomponent LFM signal based on the Radon-ambiguitytransform(RAT) is studied. Because of the linearity of LFM signal in Time-Frequencyplane and the feature of detecting straight line about Radon transform, the RATalgorithm can detect LFM signal in low SNR. The simulation results on MATLABindicate that the performance of the modified algorithm is superior to other algorithms.Lastly, the thesis concludes with a summary of the overall research and a prospect forthe following research.