Back

TUTORIALS

AFTERNOON SESSION: 1.00 PM to 4.30 PM

T6: ACTIVE NOISE CONTROL: A SIGNAL PROCESSING APPROACH
Dr. S. V. Narasimhan (about the Speaker)
NAL, Bangalore, India

Abstract: Acoustic noise of high level not only impedes human oral communication but also prolonged exposure to it results in health hazards. Acoustic noise reduction has become important in many applications like aeronautical, automobile, industrial (exhaust ducts, high noise level manufacturing sites) and commercial applications. Even in daily life, increased use of automobile engines, blowers, fans, compressors and transformers has made the noise reduction mandatory. Further, use of fuel-efficient lightweight structures to overcome cost constraints aggravates the noise problem. The passive absorbers reduce the noise due to high frequency. However, these are not suitable for the low frequency noise, predominant in industries and mobile vehicles due to their size, weight and hence the cost. The new technology of active control that attenuates the noise by superimposing an antinoise on the existing noise field appears to be promising. Low cost DSPs are readily available since 1990, making active noise control emerge as a major technology in all walks of life.

This tutorial will present a signal processing approach to active noise control (ANC). It will introduce adaptive signal processing concepts, including system identification, mean square error, linear minimum mean square filtering, LMS algorithm, theory of transform domain analysis. Single channel ANC in a duct and multichannel ANC will be described. Filtered-X LMS (FXLMS) algorithm, effect of measurement noise, secondary path modeling, feedback effects and solutions, feedback ANC, frequency domain implementation of FXLMS algorithm, and complexity of FXLMS algorithm will be covered. Implementation on DSP processors, and practical results for ducts and enclosures will be presented.

Back