Research on methods of higher-order statistics for phase difference detection and frequency estimation

Higher-order statistics is widely used in radar, communication, biomedicine and fault diagnosis. Higher-order statistics not only reveal phase properties of random process but also resist Gaussian white and colored noise. Phase difference detection methods based on higher-order spectra (bispectrum, cross bispectrum and trispectrum) as well as frequency estimation algorithms based on fourth-order cumulant diagonal slices are researched. Experimental results under different conditions significantly demonstrate the precision, effectiveness and perfect filtering effect to Gaussian white or colored noise of the methods.