Multichannel active noise control system for local spectral reshaping of multifrequency noise

This paper discusses the development of a multichannel active system for local spectral reshaping of multitone noise. The aim of this work is to design a real practical system that performs well in local active noise control (ANC) applications, and so improving the comfort sensation produced by enclosed sound fields. The adaptive algorithm implemented in the controller is a multichannel extension of the multifrequency adaptive equalizer developed by Kuo. The philosophy behind these equalizers lies in independently controlling some given frequencies of a primary signal. Moreover, the algorithm should manage to generate usefully sized zones of equalization in order to allow for the head motion of somebody with restricted mobility; for example, a passenger seated in a car. To verify the successful implementation of the multichannel system, experiments were carried out under listening room conditions. The developed prototype consists of an array of up to four microphones used as error sensors and two secondary sources. A 4100-type Bruel and Kjaer mannequin with two calibrated microphones at the ear canals was used to measure sound levels in a hypothetical listenerʹs head. Different synthesized repetitive noises were used, as reference signals, specifically repetitive noise with harmonics of 15, 20, and 28 Hz, as well as an 80 Hz single tone. The equalized points include an area around the error sensor positions, and these are measured using the mannequin and an x–y moving platform. The extent to which the experimental equalization zones obtained favourable results validates the multichannel local ANC equalization system. Different error sensor positions around the listener head were also tested. The residual field inside the equalization zone was measured in all cases.