Parametric direction-of-arrival estimation of sound sources using a spherical microphone array

Abstract

This dissertation presents theories and methodologies for parametric direction-of-arrival (DoA) estimation using a spherical microphone array. There are known limitations of the conventional parametric DoA estimation method including a singularity for a specific direction, ambiguity issue, and the small number of detectable sources. The theories developed in this work resolve these limitations by introducing novel recurrence relations of spherical harmonics and vector-based directional parameters. Moreover, it is shown that the number of detectable sources can be increased beyond the formerly known limit, by deriving additional recurrence relations involving high order spherical harmonic coefficients. This work delivers the mathematical derivation for the different recurrence relations and the methodology for solving those equations simultaneously to detect DoAs through the estimation of directional parameters. The superior estimation performance and robustness against noises of the proposed methods are demonstrated and verified through numerical simulations and experiments under various signal-to-noise ratios, the number of sources, and room conditions.