One-Dimensional MUSIC-Type Algorithm for Spherical Microphone Arrays

Spherical microphone arrays have been widely used in direction-of-arrival (DOA) estimation in recent years. The conventional multiple signal classification method formulated in spherical harmonic domain (SHMUSIC) is computationally expensive, particularly in the application of joint estimation of the azimuth and elevation. In this paper, one-dimensional MUSIC-type approach (ODMUSIC) is proposed to estimate the elevation and azimuth. We construct a mapping matrix between the spherical harmonic function and Fourier series, then utilize the Vandermonde structure of Fourier series to develop the polynomial rooting method. The mapping matrix only depends on the array configuration and can be calculated prior to DOA estimation. The two-dimensional angle search can be transformed into two one-dimensional angle findings based on the mapping matrix. By using the characteristic of Fourier series, we can construct two root polynomials to estimate the azimuth and elevation, respectively. The proposed method can avoid the two-dimensional spectral search and the angles are automatically paired. Computer simulations validate the effectiveness of the proposed method.