| In modern signal processing domain, frequency-modulation signal is one of the representative signals of non-stationary signals. It can be divided into linear frequency modulation (LFM) signal and non-linear frequency modulation (NLFM) signal. LFM signal is wildly used in radar, sonar, speech and biological signal analysis areas. Time-frequency analysis is an effective tool to analyze and process non-stationary signals. This paper studies frequency-modulation signal processing with Wigner-Ville distribution.WVD and Radon-WVD are used to estimate parameters of mono-component LFM signal and the reference signal is established according to local signal to correct the rough estimated frequency. Variable step-length optimization method is used to reduce the amount of computation. The result of experimentations in the paper indicates that the modified WVD method at SNR (Signal-to-noise) of -6dB and Radon-WVD at -15dB still has higher frequency estimation precision than original algorithm.For multi-component LFM, using Radon-WVD combined with"CLEAN"to detect and estimates its parameters one by one. The parameter estimated performance is also discussed in the paper. Theoretical analysis and experiment result indicate that this method can separate the signal and restrain the noise.In order to get the better result of parameter estimation, based on frequency correction WVD is used in NLFM signal processing. Using delay and correlation to reduce the order of NLFM into LFM, estimate the parameters of NLFM one by one.In consideration of instantaneous frequency estimation of the radar echo signal in low frequency were not very accuracy, frequency resolution were discussed in this paper. Combined multi-ratio processing with WVD method is proposed, which combines a frequency division in order to improve the precision of the echo signal. The result of simulation and actual measurement indicate that, this algorithm can effectively improve the accuracy of low frequency. |
PDF Full Text Request