GPU Implementation of a Fast Multichannel Wiener Filter Algorithm for Active Noise Control

Junlin Liu, Hongling Sun*, Ming Wu, Jun Yang

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)

Abstract

Most of the traditional active noise control (ANC) systems are implemented using digital signal processor (DSP), but the lack of computational power of DSP has been a limitation to the application and development of ANC technology. Currently, it is a feasible approach to utilize Graphics Processing Unit (GPU) to enhance the computational power of the ANC system. The advantage of GPU is parallel computation, but most of the traditional ANC algorithms can not be efficiently executed in parallel. In this letter, a parallelizable Fast Multi-channel Wiener Filter (FMWF) algorithm is proposed, and the feasibility of implementing the FMWF algorithm on GPU is verified through experiments, which show that the FMWF algorithm has obvious advantages in parallel execution on GPU. In addition, a DSP-CPU-GPU architecture for ANC systems is designed. In this architecture, each processor can make full use of its own advantages to enhance the computational capability of the system and guarantee the real-time processing of the signals at the same time.

Original languageEnglish
Pages (from-to)1009-1013
Number of pages5
JournalIEEE Signal Processing Letters
Volume31
DOIs
Publication statusPublished - 2024
Externally publishedYes

Bibliographical note

Publisher Copyright:
© 1994-2012 IEEE.

ASJC Scopus Subject Areas

  • Signal Processing
  • Electrical and Electronic Engineering
  • Applied Mathematics

Keywords

  • Active noise control
  • FMWF algorithm
  • Graphics processing units

Cite this