Active noise control system using the simultaneous equation method without the estimation of error path filter coefficients

Abstract

In an active noise control system that uses the Filtered‐x method as an adaptive algorithm, the impulse response on the error path is observed and the result is assigned as a coefficient of an error path filter before the system is started. Such an impulse response can obviously change after the system is started. The change can increase the difference between the assigned coefficient and the inherent coefficient of the error path filter, and thus render the operation of noise control unstable. This paper proposes a noise control filter coefficient renewal method that does not require the calculation of an error path filter coefficient by focusing on the fact that, when a set of two different coefficients is assigned to a noise control filter, the system consisting of components from a noise detection microphone to an error detection microphone establishes two independent equations whose variables are impulse responses for the feedforward path and the error path system. The coefficient of the noise control filer can be renewed by solving the simultaneous equations by iteration. © 2002 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 85(12): 101–108, 2002; Published online in Wiley InterScience (www.interscience.wiley. com). DOI 10.1002/ecjc.1132

Verification of extension of simultaneous equations method extended to multiple-channel active noise control systems

“…Estimate the coefficient vectors of the auxiliary filters to yield (6). By repeating the above procedure, the simultaneous equations method can continuously adjust the coefficient vectors of the noise control filters so that the outputs of the error microphones are minimized, and thereby the noise reduction effect degraded by path changes can be automatically recovered.…”

“…In practical systems, however, the feeding of the extra noise is undesirable. The simultaneous equations method is well known as a technique capable of automatically recovering the noise reduction effect degraded by the path change without feeding the extra noise [6][7][8]. The method is characterized by an auxiliary filter substituted for the secondary path filter forming the core of the filtered-x algorithm.…”

Frequency domain simultaneous equations method for active noise control systems

“…A few method capable of automatically recovering the noise reduction effect without feeding the extra noise hence have proposed [6]- [8]. However, [6] and [7] neglect the feedback path from the loudspeaker to the noise detection microphone.…”

“…However, [6] and [7] neglect the feedback path from the loudspeaker to the noise detection microphone. In addition, the noise reduction speed of [6] and [7] is slower than that of the filtered-x algorithm, and the processing cost of [7] and [8] is high. On the other hand, [9] shows that the simultaneous equations method proposed in [8] can successfully work on condition that the feedback path causes no howling.…”

“…In addition, the noise reduction speed of [6] and [7] is slower than that of the filtered-x algorithm, and the processing cost of [7] and [8] is high. On the other hand, [9] shows that the simultaneous equations method proposed in [8] can successfully work on condition that the feedback path causes no howling. This paper reduces the processing cost by applying a frequency domain technique to the simultaneous equations method.…”