基于高阶统计量的对称分布信号检测  

Detection of Signals with Symmetry Distribution Based on Higher Order Statistics

作　　者：赵倩[1] 李宏伟[1] 余绍权[1] 

机构地区：[1]中国地质大学数理系,武汉430074

出　　处：《信号处理》2005年第6期565-569,共5页Journal of Signal Processing

基　　金：湖北省自然科学基金资助项目(No.2004ABA038)国家自然科学基金资助项目(No.60472062)

摘　　要：本文研究高斯噪声中具有对称分布的随机信号检测问题,提出了一种基于偶数阶高阶统计量的信号检测方 法。该方法利用高阶累积量估计的渐近正态性,采用随机变量Taylor展开的方法估计高阶累积量估计的方差,再对偶数阶 的高阶累积量进行归一化处理,将信号的检测转化为一般的X2检验问题,并通过仿真实验以广义高斯褶积信号为例对其 检测性能进行分析。In this paper, a new detection algorithm based on higher order statistics （HOS） is proposed, which can be used to detect signals with symmetry distribution in additive Gaussian noise. We estimate the variance of higher-order cumulant estimate with Taylor expansion of random variable, and normalize the even-order cumulant to translate this kind of signal detection to the 2 general χ^2 test with asymptotic normality of higher-order cumulant estimate. In addition, asymptotic performance analysis for the algorithm is also given with generalized Gaussian convolution signals in simulation.

关 键 词：高阶统计量 广义高斯分布 信号检测 Χ^2检验 

分 类 号：TN911.23[电子电信—通信与信息系统] O211.3[电子电信—信息与通信工程]