基于S变换的线性调频信号时频滤波

Time-Frequency Filtering for LFM Based on S Transform

线性调频(LFM)信号是一种典型的非平稳信号。对非平稳信号的分析和处理不能仅依靠传统的基于傅里叶变换的分析方法,必须采用时域和频域联合的时频分析方法。将S变换(ST)的时频分析方法应用到LFM信号的滤波中,用tST-tIST和fST-fIST 2种算法组合分别实现了对LFM信号的滤波,然后采用均方误差(MSE)的衡量标准,仿真对比了以上2种算法组合的滤波性能。仿真结果表明:fST-fIST算法的滤波误差较大,不适合对LFM信号的滤波;tST-tIST算法的滤波性能较好,适合对LFM信号的滤波。

The linear frequency modulation（LFM） is a classical non-stationary signal. The traditional analysis method based on the Fourier transform is not suitable to the processing of the non-stationary signal,while the time-frequency analysis method is suitable. The time-frequency analysis method named S transform is applied to the LFM filtering. Two algorithms are used to the LFM filtering. The simulation is carried out to compare the filtering performance of the two algorithms using the mean square error criterion. T^ae simulation results show that the fST-flST algorithm is not suitable to the LFM filtering,while the tST-tIST algorithm is suitable.