本文以α稳定分布作为噪声模型,研究了非高斯噪声对传统的二阶循环统计量的影响,提出了分数低阶循环统计量的概念,研究并证明了其性质。在此基础上提出基于分数低阶循环统计量的新的时延估计方法-RCCC(Robust Correlated Cyclic Covariation)。计算机模拟表明,这种算法是一种在高斯和α稳定分布噪声条件下具有良好韧性的时延估计方法。
Cyclostationarity is an important statistical property of signals in many areas. However it has never been considered so far under impulsive noises which exist almost anywhere in real world. In this paper, the degradation of a signal' s cyclic property is evalu- ated in the presence ofα stable distributed noises, novel definitions referred to as fractional lower order cyclic statistics are proposed and some properties of the proposed definitions are developed and proven. Based on the proposed definition, a novel time delay estimation method named RCCC (Robust Correlated Cyclic Covariation) is then introduced. Simulations show that the proposed algorithm enhances the robustness to the noise no matter it is either Gaussian or αstable distributed noises.