Multichannel active sound control using adaptive filtering

Abstract

The active control of harmonic sound using multiple secondary sources and multiple error microphones is formulated in an adaptive digital filtering framework. A stochastic gradient algorithm is presented to minimize the sum of the squares of the error microphone outputs by adapting the coefficients of an array of FIR (finite-impulse response) filters driving the secondary sources. The properties and implementation of the algorithm are discussed and some results are presented which demonstrate that the noise at the engine firing frequency inside an automobile can be reduced by 10-15 dB using a practical implementation of such an active sound control system, which had a response time of 0.1 s in this application.