Distributed LCMV beamformer design by randomly permuted ADMM

Zhibao Li, Ka Fai Cedric Yiu, Yu Hong Dai, Sven Nordholm

Research output: Journal article publicationJournal articleAcademic researchpeer-review

Abstract

In recent years, distributed beamforming has attracted a lot of attention. Since each node has its own processing power, one significant advantage is the capability of distributed computing. In general, almost all distributed beamforming approaches are solving certain multi-block optimization problems. However, additional conditions are usually required to ensure convergence. In this paper, a new distributed beamforming algorithm is proposed. We first introduce the augmented Lagrangian method to implement the centralized LCMV beamformer design. Then, we propose an effective blockwise optimization method for the design of distributed LCMV beamformer based on the randomly permuted alternating direction method of multiplier (RP-ADMM). The expected convergence is obtained for distributed LCMV beamformer design without additional conditions. Numerical experiments are conducted to illustrate the performance of the proposed method.

Original languageEnglish
Article number102820
Pages (from-to)1-9
Number of pages9
JournalDigital Signal Processing: A Review Journal
Volume106
DOIs
Publication statusPublished - Nov 2020

Keywords

  • ADMM
  • Blockwise optimization
  • Distributed LCMV beamformer
  • Random permutation
  • Speech enhancement

ASJC Scopus subject areas

  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Statistics, Probability and Uncertainty
  • Computational Theory and Mathematics
  • Electrical and Electronic Engineering
  • Artificial Intelligence
  • Applied Mathematics

Cite this