DFT and PSD for estimating DOA with an active acoustic array

Direction of arrival (DOA) estimation of narrowband wavefronts impinging on an array of sensors has long been of great research interest. In this paper, it is shown that for a uniform linear array (ULA), discrete Fourier transform (DFT) domain-based DOA estimation rely on knowledge of the power spectral density (PSD). We present a novel approach for signal tracking, which updates the signal PSD for each DFT coefficient in the presence of both signal and noise. This method is based on the eigenvalue decomposition of correlation matrices that are constructed from snapshots of time series of array sensors. The presented method is very well capable of estimating more than five signals from different spatial directions. In comparison to the well-known multiple signal classification (MUSIC) algorithms the proposed method reduces the estimation error between the estimated and the true signal direction.