騒音源情報推定によるVirtual Sensing ANC

概要

論文の詳細を見る
Virtual sensing emancipates the error sensor placement for active noise control (ANC) from the constraint such that the location of error sensor needs to coincide with that of a control point. When summarizing the virtual sensing methods reported so far, several shortcomings emerge, which disable a virtual sensing method to be viable: the configuration of an ANC system using the virtual sensing is limited to a one-dimensional case; the location of virtual sensor needs to be close enough to that of real sensor, otherwise no effective. To overcome the problems, this paper presents novel virtual sensing, the fundamental concept of which is not so much extrapolating the virtual sound pressure level as conventionally employed as estimating the properties of noise source information, thereby enabling one to deal with a two-dimensional case, enhancing dramatically the estimation precision. Furthermore, the amplitude and phase at the control point are estimated based on the assumption that the sound decreases in inverse proportion to the distance from a noise source. Even if the sound decreases in inverse proportion to the bth power, where b varies from 0 to 1, of the distance from a noise source, the assumption permits one to cope with the actual sound propagation. Finally, a numerical analysis followed by an experiment is conducted, demonstrating the validity of the virtual sensing method proposed.