多线性调频信号瞬时频率估计快速算法

Fast Estimator of Instantaneous Frequency of Multiple Linear Frequency Modulation Signals

对多线性调频信号的瞬时频率估计问题提出一种快速算法,该算法以特征子空间跟踪算法为基础,结合矩阵线性变换和多项式方程求根得到参数估计。该算法的优点是计算量小,其计算量仅与短时傅里叶变换相当;频率分辨力较高;多信号情况下不存在交叉项问题;当多个分量的功率差异达到14 d B时仍能有效估计瞬时频率。但是,由于采用了矩阵求逆的步骤,该算法在低信噪比环境下性能较差。仿真实验显示在信噪比不低于6 d B时,算法具有明显的优越性。

A fast algorithm is proposed aiming at the problem of instantaneous frequency estimation of multiple linear frequency modulation signals.The algorithm is based on subspace tracking,and the matrix linear transformation and the polynomial rooting are adopted to estimate the parameters.The advantages of the proposed algorithm are as follows：lower computational complexity,which is comparable to that of the short time Fourier transformation; higher frequency resolution; no cross-terms problem in the multiple signals situation; still effective when the differences between the signals＇ power are up to 14 d B.Unfortunately,since the matrix inverse is adopted,the performance of the novel algorithm will be lost in a lower signal-noise-ratio（SNR） situation.Simulation results show that the proposed algorithm possesses obvious superiority when the SNR is not lower than 6 d B.