Sound source localization with DAS beamforming method using small number of microphones

Abstract

Localization of the sound source using microphone arrays is largely done by Acoustical Holography or Beamforming method. In this paper we present an implementation of DelayAnd-Sum algorithm for beamforming computation using various microphone arrays. A MATLAB code was implemented for postprocessing of obtained acoustic signals. There is also included a description of the fractional delay filter implementation to delay the signal samples by racional values of sample period. The goal of this paper was comparing the accuracy of the source localization with the microphone arrays using small number of microphones. The results of the localization of several noise sources by the different types of arrays is presented.

Microphone Array for Speaker Localization and Identification in Shared Autonomous Vehicles

“…Outside the automotive context, current audio-only sensor systems use microphone arrays to localize different sound sources in a wide range of applications, for example, robot and human-robot interactions [19,20], drones direction calculation [21], audio recording for multi-channel reproduction [22], and multi-speaker voice and speech recognition [23]. In such solutions, the accuracy and detection performance is affected by the array geometry, where linear arrays are only able to localize sound sources in a 2D range [24], and circular [19,22,25], spherical [20,26], or other geometries [27,28] allow the system to localize in a 3D space. Besides the geometry, the number of microphones also affects the localization accuracy [27].…”

“…In such solutions, the accuracy and detection performance is affected by the array geometry, where linear arrays are only able to localize sound sources in a 2D range [24], and circular [19,22,25], spherical [20,26], or other geometries [27,28] allow the system to localize in a 3D space. Besides the geometry, the number of microphones also affects the localization accuracy [27]. Other hybrid approaches combine microphone arrays with video cameras combined with facial recognition techniques to localize and detect audio sources, which can be used to monitor the SAV [29,30].…”