Alpha Spectrum Estimation of AR SαS Colored Noise Based on Durbin and Burg Algorithms
WU Liang-dan · Communications technology · 2010
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.