SENSITIVITY ANALYSIS OF ACOUSTIC INVERSE PROBLEMS
L. Gilquin1, S. Bouley1, J. Antoni1, C. Marteau2 and T. Le Magueresse3
1 Univ Lyon, INSA-Lyon, Laboratoire Vibrations Acoustique, F-69621 Villeurbanne, France
2 Univ Lyon, Université Claude Bernard Lyon 1, CNRS UMR 5208, Institut Camille Jordan, F-69622 Villeurbanne, France
3 MicrodB, 28 Chemin du Petit Bois, Ecully, FRANCE
Abstract :
The characterisation of an acoustic source is achieved by retro-propagating the acoustic field measured with a microphone array to a mesh of the surface of interest. This latter step is performed by solving an inverse problem. Such inverse problems are built on the basis of a forward model prone to uncertainties arising from the physics of the experiment.
Assessing the effects of these uncertainties on the resolution of the inverse problem represents a challenge. This paper introduces a practical solution to measure these effects by conducting a sensitivity analysis. The sensitivity analysis method applied here provides a mean to identify and rank the main sources of uncertainty through the estimation of sensitivity indices. Two inverse methods are investigated for the reconstruction of a monopole source: Beamforming and Bayesian focusing. The propagation of uncertainties is carried on numerically. The consistency between the real experiment and its numerical approximation is assessed by means of a small batch of measurements performed in a semi-anechoic chamber.
Distribution of the numerical outputs when the acoustic source is reconstructed either by Beamforming or by Bayesian focusing. The purple curve represents the mean of the numerical outputs, the orange region is delimited by the fifth and the 95-th percentiles, the blue region contains the remaining values. The real outputs are represented by the black curves.
Partenaires
MicrodB, Terabee and Rtone are members of the Minalogic cluster
www.microdb.fr www.terabee.com www.rtone.fr
The LUG2 project is funded by the Auvergne Rhône-Alpes Region and BPI France
