基于分数低阶协方差的AR SαS模型α谱估计  

α-Spectrum Estimation for AR SαS Processes Based on FLOC

作　　者：王首勇[1] 朱晓波[2] 李旭涛[3] 王永良[1] 

机构地区：[1]空军雷达学院军队重点实验室,湖北武汉430019 [2]空军雷达学院研究生管理大队,湖北武汉430019 [3]华中科技大学电子与信息工程系,湖北武汉430074

出　　处：《电子学报》2007年第9期1637-1641,共5页Acta Electronica Sinica

摘　　要：根据自回归(AR)SαS模型的α谱,分析了基于分数低阶矩(FLOM)法估计AR SαS模型参数的不足.提出了一种基于分数低阶协方差(FLOC)的AR SαS模型参数估计方法,并给出了基于FLOC的AR SαS模型α谱方法.分别对AR SαS模型参数的估计、α稳定分布噪声中单一正弦信号的估计和两个正弦信号的分辨进行了仿真.仿真结果表明,基于FLOC的AR SαS模型α谱估计方法对于不同的α值均具有较好的韧性.特别是在α值较小,即α稳定分布噪声概率密度函数(PDF)拖尾比较严重时,本文所提出的基于FLOC的AR SαS模型α谱估计方法,其性能明显优于基于FLOM的AR SαS模型α谱估计方法.Based on the α-spectrum of AR SαS processes, the drawback of the parameter estimation of AR SαS processes based on Fractional Lower Order Moments （FLOM） was analyzed. Then a parameter estimation method of AR SαS processes based on Fractional Lower Order Covariance （FLOC） was proposed, and the α-spectrum method of AR SαS processes based on FLOC was presented. The parameter estimation of AR SetS processes, α-spectrum estimation, the estimation of single sine signal and the resolving ability of two sine signals embedded in the α-stable noise were simulated respectively. Simulation results showed that the α- spectrum estimation method of the AR SαS processes based on FLOC were robust for different values of α（1 〈 α≤ 2）, especially when α was small,i.e. ,when the probability density function （PDF） of α-stable noise had heavier tails, the proposed α-spectrum estimation method of AR SαS processes based on FLOC provided better performance than the method based on FLOM.

关 键 词：Α谱 自回归SαS模型 Α稳定分布 分数低阶协方差 分数低阶矩 

分 类 号：TN911.23[电子电信—通信与信息系统]