Component Parameter Estimation Of Lfm Signal And Fractional Filter Design

With the linear FM signal further applied in radar, seismic, speech, medical electronic and other new areas, the research on the detection of the linear FM signal and its parameter estimation becomes more popular. This study of this article is carried out as following:At first, this paper studies multi-component LFM signal detection and parameter estimation algorithm based on the Cubic Phase Function-Fractional Fourier Transform (CPF-FrFT). The algorithm first uses CPF, a quadratic function, to obtain the parameter estimation for instantaneous frequency rate, and takes it as the optimal order of the fractional Fourier transform(FrFT) of the remaining multi-component signal, which transforms the signal to the strongest signal component of the optimal fractional transform domain. Afterwards, the estimation for the parameter of the starting frequency and amplitude of the components is derived. while the corresponding signal components is filtered. The algorithm takes full advantage of the high energy convergence in the optimal fractional transform domain, and can be used for detection and separation of multi-component signal, and effectively enhance the estimation accuracy of the parameter of instantaneous frequency rate, initial frequency and amplitude.In the above study, if the difference between the signal magnitude of each components is small, pseudo-peak interference is generated because of cross-terms of the CPF method, this paper gives the corresponding solution-PCPF (Procduct CPF). After further analysis of the reasons for the failure of traditional CPF methods in case that the instantaneous frequency rate of multi-component signal is close to each other, this paper puts forward the FrFT neighborhood scanning methods to improve the estimation accuracy of parameter IFR, and accordingly proposed detection and parameter estimation algorithm of same IFR multi-component linear FM signal PCPF-FrFT, which first obtains by PCPF coarse eastimation of parameter IFR, and then adjusts it using FrFT. And then in the optimal fractional domain the parameters of starting frequency and amplitude is estimated. The algorithm solves the pseudo-peak interference and increases the estimation accuracy of IFR in the same multi-component linear FM signal.Finally, in combination of linear S transform and quadratic wigner distribution, the proposed new time-frequency transform Stockwell-Wigner transform is derived. The transform not only eliminates the cross terms of the quadratic transformation, but also increases the signal energy aggregation in the two-dimensional time-frequency plane, therefore, this new transform is applied to the fractional filter design, and behaves well in simulation experiments.