Research On The Modulation Recognition Algorithms And Parameter Estimation Algorithms Of Radar Signal

Combined with the project of “Signal reconnaissance and processing technology of×××which is based on broadband real time spectrum analysis” closely, technologyinvestigation and research are undertaken to be used for reconnaissance and processingof radar signals. It includes time-domain and frequency-domain characteristics’ analysisof radar signals, as well as some algorithms that are basic for modulation recognitionand parameters estimation of radar signals. This thesis is composed of three parts asfollows:In part Ⅰ, time-domain and frequency-domain features of Liner FrequencyModulation(LFM) signal and Buck code signal are analyzed. Firstly the time-domainfeatures of radar signals studied by time-domain analysis methods are given.Subsequently, features are analysed in frequency-domain as well, and finally, thetime-frequency features are studied by Liner Time-frequency Distribution andWigner-Ville Distribution.In part Ⅱ, modulation recognition algorithms of radar signals are studied. And theresearch objects are constant carrier-frequency, LFM and the Buck Code. Firstly thealgorithm based on frequency’s offset is given. It distinguishes different modulations byrecognizing different instantaneous error characters of frequency of signals in differentmodulation signal. Then the algorithm based on instantaneous autocorrelation function’sspectrum is given. The spectrum of instantaneous autocorrelation function is applied tomodulation recognition. Finally the algorithm based on Morlet wavelet ridge isoptimized. It is a time-frequency analysis method based on Morlet wavelet transform.The calculation quantity is reduced after optimizing the algorithm.In part Ⅲ, parameter estimation algorithms of LFM are studied. The parametersinclude pulse time of arrival(TOA), pulse width(PW), pulse recurrence interval(PRI),starting frequency, cut-off frequency, bandwidth, frequency modulation slope andamplitude. Firstly a pulse detection algorithm based on the frequency-domain MNmethod is gived. It is the same thinking with the MN method, but it is in thefrequency-domain. Then the frequency parameters estimation algorithm based on STFTis optimized. Then the signal amplitude is estimated directly by method based on STFT.The idea comes from the amplitude estimation algorithm based on FFT which is usedfor point frequency signal only. Finally combined the frequency-domain MN methodand STFT, parameter estimation algorithm is given. A soft according to this algorithmwhich is used to measure these parameters is developed, and it has achived to processthe measured data.