Linear Frequency Modulation Signal Adaptive Processing Algorithm Research

Chirp (Linear Frequency Modulation, LFM) is a kind of non-stationary signal, which applied widely in radar, medical science, communications, and many other fields in recent years, so the corresponding digital signal processing methods have been paid great attention and researched by people.In practical application, the parameters estimation of LFM is an important research orientation. The existing signal process methods cannot meet the requirement because of their shortcomings. For example, short-time Fourier Transform (STFT) has problem in selecting window function and is limited by uncertainty principle; Wigner-Ville transformation exits cross-term interference while dealing with multiple LFM signals. In contract, the Fractional Fourier Transform (FRFT) does not exit the limit. This paper first use FRFT to estimate the parameters of LFM, but this conversion calculation is very complicated, in order to reduce the amount of the calculation, a new method is put in use which brings the convenience for different frequency LFM analysis in this paper.In the LFM signal processing, the noise almost submerged due to nonlinearity, so the paper puts forwards the appropriate filtering method to reduce the noise, which improves the accuracy of processing results. The traditional adaptive algorithm, minimum mean square (LMS) algorithm and recursive least square (RLS) algorithm have a good effect in dealing with stable signals. When dealing with the unknown environment or non-stationary signals, the adaptive filter shows its special advantages. But LFM is a kind of non-stationary signal, there is a new adaptive algorithm basing on the FRFT to handle the chirp signal, and simulate by the MATLAB. The result shows the algorithm can estimate the parameters of LFM after denosing.According to the simulation algorithm and results, In paper, adaptive hardware simulation results of the LFM is base on TMS320F2812platform, the result has the same effect with MATLAB, which proves the method has a certain application value.