一种α稳定分布噪声下的跳频参数盲估计方法  

A Blind Parameters Estimation Method for Frequency-hopping Parameters with α Stably Distributed Noise

作　　者：杨芸丞 孙雪丽 钟兆根 YANG Yun-cheng;SUN Xue-li;ZHONG Zhao-gen(Naval Aviation University,Yantai 264001,China)

机构地区：[1]海军航空大学,山东烟台264001

出　　处：《火力与指挥控制》2020年第2期16-22,共7页Fire Control & Command Control

基　　金：国家自然科学基金(61179016);国家自然科学基金重大研究计划(91538201);泰山学者工程专项基金资助项目(ts201511020)。

摘　　要：针对目前以高斯白噪声为模型的大部分跳频参数估计方法在α稳定分布噪声背景下,性能急剧下降的缺点,对跳频信号进行两次窗函数长短不同的分数低阶STFT,从而得到两组时频数据,,将两组时频数据点乘,得到新的时频表示,基于时频分析的跳频参数估计方法,实现跳频参数的估计。仿真实验表明,提出的方法有效抑制了α噪声,在α=0.8,GSNR≥1 d B;α=1.5,GSNR≥0 d B时,可以实现跳频周期的准确估计。在α=1.5,GSNR=3 d B时,该算法跳变时刻估计值最大相对误差比STFT低3%、比分数低阶STFT低1.6%,跳变频率估计值更加精确。Aiming at the problem that most frequency-hopping parameters estimation methods based on Gaussian white noise are currently under the stably distributed noise background,the performance is drastically reduced. In this paper,two fractional low-order STFT with different window functions are firstly applied to the frequency-hopping signal to obtain two sets of time-frequency data.Then,two sets of time-frequency data are multiplied to obtain a new time-frequency representation,and finally based on time-frequency analysis. Frequency-hopping parameter estimation method realizes accurate estimation of frequency-hopping parameters. Simulation experiments show that the proposed method effectively suppresses the α stable distribution noise. When α =0.8,GSNR ≥1 dB;α =1.5,GSNR≥0 dB,an accurate estimation of the frequency-hopping period can be realized. At the time ofα=1.5,the maximum relative error of the estimation time of the algorithm is 3 % lower than the STFT,1.6 % lower than the fractional low-order STFT,and the estimated hopping-frequency is more accurate.

关 键 词：稳定分布噪声 组合窗函数 分数低阶STFT 跳频 参数估计 

分 类 号：TN974[电子电信—信号与信息处理]