Basic study on active acoustic shielding (Improving the method to enlarge the AAS window)

Abstract

In this paper, the M[(1-1)-L’] FXLMS algorithm is shown to be useful in enlarging the window size for Active Acoustic Shielding (AAS) window. AAS is a system that can attenuate sound passing through an open window. In our previous study, an AAS window was fabricated with 4 AAS cells, each of which consisted of an collocated microphone and speaker. This AAS window was proved to be effective for not only single stable noise sources but also multiple and moving noise sources. The AAS system consists of many AAS cells set in an array. However, the size of the existing type AAS window is only 250mm square. Therefore in this paper, we propose the M[(1-1)-L’] FXLMS algorithm (M: number of AAS systems, L’ : number of error signals used for controlling 1 AAS cell) for controlling a larger window and a larger number of AAS cells. This algorithm is a similar to FXLMS algorithm, as each AAS cell is individually controlled by its own reference and neighboring error sensors. Six AAS cells are fixed on a rectangular window (125 × 750mm square) and controlled using 6 [(1-1)-3’ ] FXLMS. This system proved useful in controlling large number of AAS cells. In this study a simulator of the M[(1-1)-L’] FXLMS algorithm is constructed. The simulation results from the 6 [(1-1)-3’] FXLMS algorithms coincided well with the experimental results of 6 [(1-1)-3’ ] FXLMS, which suggests that a large AAS window is feasible.

Keywords

• active noise control
• active acoustic shielding
• multiple filtered-x-lms
• decentralized system