Fundamentals of Statistical Signal Processing, Volume III

Chapter 11. Spectral Estimation

11.1. Introduction

The power spectral density (PSD) of a random process is a function that quantifies the distribution of power with frequency. The estimation of this function is commonly referred to as spectral estimation. How this is accomplished is the subject of this chapter. Unlike estimation or detection problems for which some general theory exists, leading to optimal approaches, spectral estimation has no such theory to help guide us in developing practical methods. This is because it is inherently an ill-posed problem, requiring the estimation of a general curve based on only a finite number of data samples. Its principal and very important utility, however, is as a preliminary data analysis tool. Depending ...