Algebraic signal processing theory: 1-D nearest neighbor models

Aliaksei Sandryhaila, Jelena Kovačević, Markus Püschel

Research output: Contribution to journalArticlepeer-review

Abstract

We present a signal processing framework for the analysis of discrete signals represented as linear combinations of orthogonal polynomials. We demonstrate that this representation implicitly changes the associated shift operation from the standard time shift to the nearest neighbor shift introduced in this paper. Using the algebraic signal processing theory, we construct signal models based on this shift and derive their corresponding signal processing concepts, including the proper notions of signal and filter spaces, z-transform, convolution, spectrum, and Fourier transform. The presented results extend the algebraic signal processing theory and provide a general theoretical framework for signal analysis using orthogonal polynomials.

Original languageEnglish (US)
Article number6140984
Pages (from-to)2247-2259
Number of pages13
JournalIEEE Transactions on Signal Processing
Volume60
Issue number5
DOIs
StatePublished - May 2012

Keywords

  • Algebra
  • Fourier transform
  • Hermite polynomials
  • Laguerre polynomials
  • Legendre polynomials
  • convolution
  • filter
  • module
  • orthogonal polynomials
  • shift
  • signal model
  • signal representation

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Algebraic signal processing theory: 1-D nearest neighbor models'. Together they form a unique fingerprint.

Cite this