Real-Valued DOA Estimation for Impulsive Noise

Aiming at DOA estimation under impulsive noise, this paper propose a real-valued sparse Bayesian learning (SBL) method. A unitary transformation is utilized to convert complex-valued direction-of-arrival (DOA) estimation into real ones. The variational Bayesian inference (VBI) technique is then adopted to perform the Bayesian inference with such real-valued prior. Consequently, the computational complexity of this Bayesian inference is significantly reduced. Simulation outcomes demonstrate the great robust performance and low computational load of the new method.