FXLMS with Multiple Error Switching for Active Noise-Cancelling Casings

Passive noise reduction methods require thick and heavy barriers to be effective for low frequencies and those clasical ones are thus not suitable for reduction of low frequency noise generated by devices. Active noise-cancelling casings, where casing walls vibrations are actively controlled, are an...

Full description

Saved in:
Bibliographic Details
Published in:Archives of acoustics 2019-01, Vol.44 (4), p.775
Main Authors: Mazur, Krzysztof, Wrona, Stanislaw, Chraponska, Anna, Rzepecki, Jaroslaw, Pawelczyk, Marek
Format: Article
Language:English
Subjects:
Actuators
Algorithms
Error analysis
Error signals
LF noise
Microphones
Noise control
Noise measurement
Noise reduction
Sensors
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Passive noise reduction methods require thick and heavy barriers to be effective for low frequencies and those clasical ones are thus not suitable for reduction of low frequency noise generated by devices. Active noise-cancelling casings, where casing walls vibrations are actively controlled, are an interesting alternative that can provide much higher low-frequency noise reduction. Such systems, compared to classical ANC systems, can provide not only local, but also global noise reduction, which is highly expected for most applications. For effective control of casing vibrations a large number of actuators is required. Additionally, a high number of error sensors, usually microphones that measure noise emission from the device, is also required. All actuators have an effect on all error sensors, and the control system must take into account all paths, from each actuator to each error sensor. The Multiple Error FXLMS has very high computational requirements. To reduce it a Switched-Error FXLMS, where only one error signal is used at the given time, have been proposed. This, however, significantly reduces convergence rate. In this paper an algorithm that uses multiple errors at once, but not all, is proposed. The performance of various algorithm variants is compared using simulations with the models obtained from real active-noise cancelling casing.
ISSN:0137-5075
2300-262X
DOI:10.24425/aoa.2019.129732