Using Global Optimization Methods for Acoustic Source Localization

Conventional beamforming is a common method to localize sound sources with a microphone array. The method, which is based on the delay-and-sum beamforming, provides an estimate value for the source strength at a given spatial position. It suffers from low spatial resolution at low frequencies, high side lobe levels and requires the user to initialize a two dimensional scan area at a certain distance from the array which can trouble source identification. In this work we use the global optimization method Differential Evolution to efficiently search for source locations. The source locations maximizes the agreement between the modelled signal and measurement. This method also allows for inclusion of more unknowns, such as environmental parameters or a search in three dimensions. Using simulated data, results show that the acoustic source can be identified very accurately with good spatial resolution