Direction-finding accuracy of an air acoustic vector sensor in correlated noise field

Abstract

The direction-of-arrival (DoA) of an acoustic source has been estimated using an air acoustic vector sensor (AVS) in the presence of spatially correlated noise field. T he DoA of an acoustic source is extracted from the acoustic intensity parameter and the acoustic intensity is calculated using an AVS (implementated with omni-directional microphones), where acoustic intensity is derived using pressure signal differences. In practice, the noise field measured at the sensors of an AVS will be correlated due to the spatial closeness of the microphones that is necessary to derive the acoustic intensity from the finite difference (FD) approximated pressure-gradient. The FD approximated pressure-gradient based acoustic intensity gives better estimate of DoA in correlated noise field than t he uncorrelated noise field. The experimental environment have been setup using finite element m ethod (FEM) simulation tool f or generating source signal and received signals at the AVS. It has been observed for delta configuration A VS, t hat t he average root mean square angular error (RMSAE) is less than 0.4° at 10 dB SNR for the correlated noise field, whereas f or the uncorrelated noise field t he average RMSAE is increased by a factor of more than five.