FRFT’s Application In Signal Separation And Parameters Estimation

In the field of modern electronic surveillance and confrontation, with the development of complex electromagnetic environment and electronic defense measures, signal detection and recognition become more difficult. The receiver has great probability of receiving many signals at the same time, therefore, separating each signal and its characteristic parameters from the mixed signals is the key to electronic surveillance and confrontation. In this paper, it is focused on separating the Linear Frequency Modulated (LFM) signals and parameters recognition that have important application in radar system, and the time-frequency analysis methods based on fractional Fourier transform (FRFT) are researched.The theory of FRFT as well as other time-frequency analysis methods, including short time Fourier transform (STFT)、Wigner-Ville Distribution (WVD), is studied and their relationships are analyzed.WVD and FRFT are used to detect signal and estimate parameters of mono-component LFM signal. For the complexity of the2D peak searching algorithm in the FRFT, it presents an FRFT modulus detector via ID curve-fitting optimization technique that can greatly reduce the computational complexity. The result of experimentations indicates that FRFT can get better performance in presence of noise compared to WVD and correlation demodulation methods.For multi-component LFM signals, FRFT combined with the "CLEAN" algorithm separates each signal from the mixed signal and estimates its parameters one by one. The simulation results show that the FRFT separation algorithm overcomes the cross-term interference in WVD method, has better ability to signals separation and noise suppression than Wigner-Hough algorithm.For single-frequency signal and LFM signals, and the power of single-frequency signal is bigger than LFM signals, that is to say, LFM signals are submerged in the single-frequency signal, according to the different time shift characteristics between the LFM signal FRFT module function and single-frequency signal’s, it presents an separation algorithm based on difference of FRFT module that overcomes the shortcoming of frequency domain masking method, the simulation results show that this method has better signal detection and parameters estimation ability, and the noise can be suppressed partly.