一种基于MT-FCRA的高分辨率时频增强算法OA

In recent years,Doppler through-wall radar(TWR)technology has been widely applied in non-contact target localization and tracking.However,in complex multi-target scenarios,the received signals often contain crossing instantaneous frequency(FI)components,which interferes with accurate time-frequency feature extraction and reduces localization accuracy.To address this issue,a high-resolution and high-energy-concentration time-frequency enhancement method,i.e.,multi-target frequency-chirp rate analysis(MT-FCRA),was proposed to improve frequency extraction accuracy and component separability.The proposed method firstly utilized Bezier curves model to flexibly fit the complex time-frequency ridges of non-stationary signals to achieve high-precision FI estimation,and thereby provided crucial a priori information for subsequent signal processing.In addition,a novel analysis framework based on the frequency-chirp rate domain was constructed,where a local maximum synchrosqueezing operator was designed to redistribute the time-frequency coefficients in an orderly manner,forming the high time-frequency resolution time-frequency distribution.The results demonstrate that the proposed algorithm reduces the frequency estimation error by 61%,63%and 69%compared to short time Fourier transform(STFT),synchrosqueezing transform(SST)and multiple synchrosqueezing transform(MSST),respectively,and the localization accuracy is improved by 93%,88%and 93%,respectively.