Research Of Fast Parameter Estimation Algorithm Of LFM Signal

Linear Frequency Modulated Signal is a kind of signal widely used in multitudinous signal processing areas, such as radar, audio, sonar and so on. Therefore parameter estimation of LFM signals is deeply researched worldwide. As the non-stationary feature of LFM signals, parameters are harder to detect than stable signals. Methods used now are based on two-dimension search in time-frequency plane which are always complex. It's urgent to find simple and effective algorithm with low computational complexity.This paper is mainly concerned with looking for fast estimation methods of LFM signals. First of all, LFM signal model is introduced, and then short-time Fourier transform is studied. A new kind of LFM signal estimation method based on STFT, Maximum likelihood estimation is proposed. The performance of this method is analyzed according to MATLAB simulation result and it's approved to be an fast method to wide band signals.When narrow band LFM signal is concerned, we discuss the method of Wigner-Ville distribution and ambiguity function. Because of the linearity of LFM signal in time-frequency plane and the feature of detecting straight line of Radon transform, Radon-Wigner transform and Radon-Ambiguity transform can detect LFM signal in low SNR. The performance and computational complexity are studied.Then a new parameter estimation algorithm based on multi-class search, Radon-Ambiguity, fractional Fourier transform is presented. Computational complexity is reduced greatly compared with normal Radon-Wigner transform. A lot of simulation results are given which indicate that it is a good method. At last, the conclusions are drawn, and some further research about LFM signal detection and parameter estimation is pointed out.