Optimal and Adaptive Microphone Arrays for Speech Input in Automobiles

Sven Nordholm, Ingvar Claesson, Nedelko Grbic

Research output: Chapter in Book/Report/Conference proceedingBook chapterResearchpeer-review

Abstract

In this chapter, speech enhancement and echo cancellation for hands-free mobile telephony are discussed. A number of microphone array methods have been tested and results from car measurements are given. Traditional methods such as linearly constrained beamforming and generalized sidelobe cancelers are discussed as well as array gain optimization methods. An in situ calibrated method which gives an overall improved performance is also presented. Algorithms such as Least-Squares (LS) and Normalized-Least-Mean-Squares (NLMS) are used to find optimal weights. Improved performance using an LS-method is shown, but at the cost of increased numerical complexity limiting its implementation in real-time applications. By introducing subband processing this issue can be avoided. The results show a noise suppression of approximately 18 dB and hands-free loudspeaker suppression of the same order.
Original languageEnglish
Title of host publicationMicrophone Arrays
Subtitle of host publicationSignal Processing Techniques and Applications
EditorsMichael Brandstein, Darren Ward
PublisherSpringer
Pages307-329
Number of pages23
ISBN (Electronic)978-3-662-04619-7
ISBN (Print)978-3-540-41953-2, 978-3-642-07547-6
DOIs
Publication statusPublished - 2001
Externally publishedYes

Publication series

NameDigital Signal Processing
PublisherSpringer
ISSN (Print)1612-1457

Subject classification (UKÄ)

  • Signal Processing

Fingerprint

Dive into the research topics of 'Optimal and Adaptive Microphone Arrays for Speech Input in Automobiles'. Together they form a unique fingerprint.

Cite this