An inverse method for in-situ estimation of acoustic surface impedance targeting inverse sound rendering in rooms

摘要

We propose two methods based on the inverse boundary element method (IBEM) for theestimation of acoustic impedances of the interior surfaces of a room. The algorithms take asinput the geometrical model of the space, a set of sound field pressures measured at randompoints and the vibration strength of the sound source. A first approach is a constrained leastsquaresmethod that attempts to find simultaneously surface sound pressures and particlevelocities, parameters that define the acoustic impedance. This is in contrast with previouswork on similar acoustic inverse problems that solve for one parameter (typically the strengthof the vibration velocity) and use regularization techniques to reduce the effects of the illconditioningand obtain smooth solutions. Our second approach is an iterative optimizationprocess that estimates directly the sought impedances assuming that the interior surfaces havehomogeneous impedance values. This assumption allows in addition a significant reduction ofthe dimensionality of the optimization problem. A numerical example of the performance ofboth methods is presented. Preliminary experiments in a reverberation chamber are also shown.