基于Durbin与Burg算法的AR SαS有色噪声α谱估计

Alpha Spectrum Estimation of AR SαS Colored Noise Based on Durbin and Burg Algorithms

稳定分布可更好地描述实际中所遇到的具有显著脉冲特性的随机噪声。在简要介绍稳定分布统计特性的基础上,描述了稳定分布的谱表示。为了对AR SαS过程进行α谱估值,通过对广义Yule-Walker方程进行分析,借鉴杜宾推算法和Burg算法推导了一种求解广义Yule-Walker方程的新方法。仿真结果表明,该方法能够较为准确地估计出AR模型参数,且能准确地估计出AR SαS过程的α谱。

Stable distribution could better model better the impulsive random signals and noises in physical observation,and has no close form of probability density function and finite second moment.The paper briefly describes the statistical characteristics of stable distribution,tells of its spectral representation.In order to estimate the α -spectrum of AR Sα Sprocesses,a new solution algorithm to the generalized Yule-Walker（GYW） equation is deduced through the analysis of generalized Yule-Walker（GYW） Equation and reference to Durbin and Burg algorithms.The simulation results show that the new method could accurately estimate the parameter and α -spectrum of AR model.