复杂噪声下基于同步压缩Chirplet变换的LFM信号参数估计

Parameter Estimation of LFM Signals Based on Synchrosqueezing Chirplet Transform in Complicated Noise

同步压缩变换建立在小波变换的基础上,通过在较小频域范围内压缩小波系数,可有效改善信号的能量分布,提高时频聚集性。该文针对线性调频(LFM)信号的参数估计问题,根据适用于LFM信号的Chirplet变换,在同步压缩理论的框架下,提出一种同步压缩Chirplet变换方法(SSCT)。由于充分利用了LFM信号时间与频率的线性关系,SSCT方法在提高Chirplet变换时频平面能量聚集性的同时,可实现信号参数的精确估计,且保留了Chirplet变换窗函数选取灵活,无交叉项干扰等优点。针对复杂噪声环境下的参数估计问题,进一步提出分数低阶SSCT方法(FLOSSCT)。仿真结果表明,在高斯噪声以及脉冲性更强的α稳定分布噪声背景下,该方法可有效实现LFM信号的参数提取,具有较好的鲁棒性。

SynchroSqueezing Transform （SST）, based on the wavelet transform, can effectively improve the energy distribution and time-frequency aggregation of a signal by compressing the wavelet coefficients in a short frequency domain. To solve the parameter estimation problem of Linear Frequency Modulation （LFM） signals, a new SynchroSqueezing Chirplet Transform （SSCT） is proposed within the framework of synchrosqueezing. Taking full use of the linear relationship between the time and the frequency of an LFM signal, the SSCT method can improve the energy density on the time-frequency plane and estimate the signal parameters accurately, which at the same time keeps the advantages of the chirplet transform, such as flexible window function selecting and no cross-term interfering. Then a Fractional Lower Order SSCT （FLOSSCT） method is proposed in order to estimate the parameters of an LFM signal in the complex noise environment. The simulation results show that the SSCT and the FLOSSCT have good performance under the background of Gaussian and impulsive noise, respectively.