Broadband Source Angle Estimation Using a Sparse Uniform Linear Acoustic Array

Abstract

A microphone array can be used to estimate the angular positions of acoustic sources by beamforming. When the array is sparse and periodic, and the source is narrowband, the presence of grating lobes may result in anomalous angle estimates. However, for broadband sources, the effect of grating lobes can be mitigated using the wide spectra of the received signals. In this paper, two methods based on the Fourier and minimum variance distortionless response (MVDR) beamformers are proposed for broadband source angle estimation using a sparse uniform linear acoustic array. The feasibilities of the two angle estimation methods are demonstrated and their angular resolution capabilities compared using simulated data for fixed far-field point sources. The methods are applied to estimating the elevation angle trajectories of jet aircraft. Real acoustic data were collected for several transits of a jet aircraft flying directly over and along the array axis. Scenarios of two jet aircraft in transit are emulated by shifting in time the real data recorded for a particular transit of the single aircraft and adding the time-shifted data to the original data, or by combining two different sets of real data each corresponding to a different transit of the single aircraft. The emulated data are processed using the proposed methods, and the results demonstrate their potential for estimating the angular trajectories of multiple broadband sources using a sparse uniform linear array.