Near-field error sensing of multi-channel active noise control using virtual sensing technique

Nan Jiang,Chuang Shi

2018-01-01

Abstract:This paper proposes a near-field error sensing method for multi-channel feedforward active noise control (MCFFANC). The proposed method is particularly useful in applications such as the active noise canceling window that can mitigate noise passing through an open window. In MCFFANC, the classic multiple-error LMS algorithm is derived for minimizing the sum of the squared errors, which is an appropriate approximation of the noise power when the error microphones are placed in the far field. However, owing to the room reverberation, the error microphones are preferably placed in the near field. Therefore, there is a mismatch between the algorithms and applications of MCFFANC. Near-field error sensing methods are desired in MCFFANC. For such purposes, this paper extends the single-channel virtual sensing technique to the multiple-channel case. The effectiveness of the proposed method is validated by an experimental setup of the active noise canceling window. An improved global reduction of noise is confirmed at far-field observation points.