Entropy; Signals and signaling; Spectral energy distribution
Presented is a generalized derivation of the Tzannes et al. algorithm (1985) to accommodate multiple signals and also a new algorithm for implementing the Relative Entropy method for estimating the spectral densities of multisignal input data.
Chapter 2 presents several of the current methods of Relative Entropy spectral estimation which involve solving N non-linear equations simultaneously. Chapter 3 presents the multisignal derivation for the Tzannes et al. method and develops the new algorithm which solves for one unknown at a time in an iterative fashion.
The performance of this new method is compared with the Maximum Entropy and Minimum Relative Entropy methods in terms of precision and computation time. Some classical examples are repeated and compared with results of these methods. An example demonstrating the power of the multisignal method is presented in Chapter 4 which estimates the voice and noise spectral densities from an input of voice in the presence of additive background noise.
Master of Science (M.S.)
College of Engineering
Length of Campus-only Access
Masters Thesis (Open Access)
Gregory, David Martin, "Multiple Signal Relative Entropy Spectral Estimation" (1988). Retrospective Theses and Dissertations. 5138.