A scalable hybrid analog-digital architecture for multi-channel feedforward active noise control

Abstract

This paper proposes a hybrid analog-digital architecture, aimed at reducing hardware resource consumption and improving the scalability of a multi-channel feedforward active noise control system to accommodate a large number of error microphones. This hybrid architecture employs a partial-update filtered-x least mean squares algorithm that is lightweight and suitable for real-time execution, as it updates the control filters in each iteration using only two selected error signals. The selection of these two error signals is achieved through a dedicated analog circuit comprising analog comparators and multiplexers. Thus, the hybrid architecture requires only two analog-to-digital converters for error signals, regardless of the number of error microphones. Experiments were conducted on two active noise control casings, demonstrating that, with a specific computational capacity, the proposed hybrid architecture can accommodate significantly longer control filters, resulting in higher steady-state noise reduction. Moreover, additional error microphones were introduced to demonstrate the scalability. The hybrid architecture simplifies the implementation of the multi-channel feedforward active noise control system when additional error microphones are needed for more evenly distributed residual noise levels or a larger zone of quiet.

More information

Identifiers

ORCID for Chuang Shi: orcid.org/0000-0002-1517-2775

Catalogue record

Export record

Altmetrics