Abstract: Based on short-time fractional Fourier transform, a new pseudo Wigner-Ville distribution is proposed to analyze and separate multi-component Chirp signal with additive white Gaussian noise. The appropriate fractional domain is found from the knowledge of the forth-order fractional Fourier transform moments. Based on rotated short-time Fourier transform, the proposed pseudo Wigner-Ville distribution preserves the WVD auto-terms and cancels the cross-terms and the noise. The impulse signal is transformed from time-frequency domain to time domain and using inverse fractional Fourier transform to separate multi-component chirp signal. Simulation results show the method can separate multi-component chirp signal with additive white Gaussian noise and is beneficial for parameter estimation.

Key words: multi-component Chirp signal, short-time fractional Fourier transform, pseudo Wigner-Ville distribution, signal separation

摘要： 针对加性高斯白噪声背景下多分量Chirp信号的分离问题，采用一种基于短时分数阶傅里叶的伪魏格纳变换来实现对多分量Chirp信号的分离。该方法利用分数阶傅里叶变换四阶中心矩寻找极值点来确定最佳变换域，在最佳变换域对信号进行旋转的短时傅里叶变换，并进行伪魏格纳变换，最后把在时频面得到的冲激信号变换到时域再进行分数阶傅里叶逆变换，实现了多分量Chirp信号的分离。仿真实验证明该方法可有效地实现多分量Chirp信号分离，有助于后续对各分量的参数估计。