Dynamic, noisy channel deconvolution: A model based approach

Michael J. Roan, Mark Gramann

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Blind Deconvolution (BDC) algorithms typically assume a noiseless signal model, and a stationary signal propagating through a static channel. In real world systems this is almost never the case (e.g. the problem of interest in this paper: a noisy multipath propagation environment where the source and receiver are moving). In such cases, it is proposed that model-based techniques be applied to incorporate further a priori information about the system into the existing blind processing framework. The significant original contributions of this work are as follows: First, a modified formulation of the extended Kalman filter (EKF) is developed that allows incorporation of a priori information into gradient-based blind processing algorithms. This formulation is then applied to the existing Natural Gradient BDC algorithm. Finally, simulation results are presented that suggest significant improvement in signal recovery performance through application of the modified EKF to the NG BDC algorithm for dynamic noisy channels. Copyright

Original languageEnglish (US)
Title of host publication13th International Congress on Sound and Vibration 2006, ICSV 2006
Pages4697-4704
Number of pages8
StatePublished - 2006
Event13th International Congress on Sound and Vibration 2006, ICSV 2006 - Vienna, Austria
Duration: Jul 2 2006Jul 6 2006

Publication series

Name13th International Congress on Sound and Vibration 2006, ICSV 2006
Volume6

Other

Other13th International Congress on Sound and Vibration 2006, ICSV 2006
Country/TerritoryAustria
CityVienna
Period7/2/067/6/06

All Science Journal Classification (ASJC) codes

  • Acoustics and Ultrasonics

Fingerprint

Dive into the research topics of 'Dynamic, noisy channel deconvolution: A model based approach'. Together they form a unique fingerprint.

Cite this