ABSTRACT
This chapter provides a review of some of the existing work in the area of nonparametric spectral estimation, including fast implementations of the most successful estimators as well as various extensions to spectral analysis of incomplete data. It discusses several spectral estimators that can be obtained via a weighted least-squares (WLS) fitting criterion. The chapter provides a filterbank interpretation of the so-obtained estimators and shows how the estimators can be extended to two-dimensional data. It discusses how the so-called forward-backward averaging can be used to improve the statistical properties of the estimators. The chapter summarizes the available results on statistical performance of adaptive filterbank estimators, and describes fast techniques for the evaluation of the Capon and amplitude and phase estimation spectra (APES). It collects various numerical examples of the performance of these estimators, and treats an important problem where the filterbank estimators have proven to be successful: spectral estimation of data with missing samples.
