Signal-subspace high-resolution direction finding based on orthogonal projections

In this paper, an orthogonal projection algorithm based on subspace estimation, referred as OPS, is presented for direction finding of narrowband plane waves in temporally and spatially white noise. Firstly, a base of signal-subspace is determined by a sequence of orthogonal projections which reach a linear complexity and guarantee maximum signal energy in the spanned signal-subspace. Then, the direction-of-arrivals (DOAs) of signals are estimated by detecting intersections between the estimated signal-subspace and the scanning vector. And the QR decomposition of a special matrix comprising the signal-subspace base vectors and the scanning vector is exploited. Simulation results show that the presented algorithm outperforms the classical subspace-based high-resolution MUSIC method, especially in the case of low signal-to-noise ratio (SNR).