Ad Hoc Microphone Array Beamforming Using the Primal-Dual Method of Multipliers

Vincent Mohammad Tavakoli, Jesper Rindom Jensen, Richard Heusdens, Jacob Benesty, Mads Græsbøll Christensen

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

9 Citations (Scopus)
436 Downloads (Pure)

Abstract

In the recent years, there have been increasing amount of researches aiming at optimal beamforming with ad hoc microphone arrays, mostly with fusion-based schemes. However, huge amount of computational complexity and communication overhead impede many of these algorithms from being useful in practice. In this paper, we propose a low-footprint optimization approach to reduce the convergence time and overheads for the convex beamforming problem. We transcribe the beamforming with pseudo-coherence-based formulation which is insightful for taking into account the nature of speech. We formulate the distributed linearly-constrained minimum variance beamformer using the the state of the art primal-dual method of multipliers. We study the proposed algorithm with an experiment.
Original languageEnglish
Title of host publicationProceedings of the 2016 24th European Signal Processing Conference (EUSIPCO)
PublisherIEEE
Publication date29 Aug 2016
Pages1088-1092
ISBN (Print)978-0-9928-6265-7
DOIs
Publication statusPublished - 29 Aug 2016
Event European Signal Processing Conference - Hotel Hilton Budapest, Budapest, Hungary
Duration: 29 Aug 20162 Sep 2016
http://www.eusipco2016.org/

Conference

Conference European Signal Processing Conference
LocationHotel Hilton Budapest
Country/TerritoryHungary
CityBudapest
Period29/08/201602/09/2016
Internet address
SeriesProceedings of the European Signal Processing Conference (EUSIPCO)
ISSN2076-1465

Keywords

  • speech enhancement
  • ad hoc microphone array
  • distributed beamforming
  • primal-dual method of multipliers
  • PDMM

Fingerprint

Dive into the research topics of 'Ad Hoc Microphone Array Beamforming Using the Primal-Dual Method of Multipliers'. Together they form a unique fingerprint.

Cite this