Periodogram Processing:Wavelet and Cepstral Thresholding Approaches

The book provides an introduction to Nonparametric spectral estimation. The periodogram is the basic nonparametric PSD estimation method. It is not consistent due to erratic fluctuations in its spectrum. The technique of averaging or smoothing has been proposed to reduce variance of the periodogram. However, there is a tradeoff between the bias and the variance. In this book, wavelet thresholding and cepstrum thresholding approaches are utilized for processing the periodogram. Several data independent and dependent thresholding schemes are discussed to achieve variance reduction in power spectrum. Modified cepstrum is presented and applied to MST radar data to improve visibility in dopper profile. its performance is in agreement with that of NARL Atmosperic Data Processor.