Partial Noise Subspace Method for DOA Estimation Applications

This paper presents a new Angle of Arrival (AoA) method to overcome the problem of spurious peaks that appear with Pisarenko Harmonic Decomposition (PHD) algorithm. The proposed method is called Partial Noise Subspace (PNS); it picks k-subsets of the rows/columns of the noise subspace matrix. The PNS algorithm is computational load than the Multiple Signal Classification (MUSIC) in the grid searching stage, while not debasing the performance of angle estimation. The idea and working principle of the proposed method are presented and the mathematical model is derived. Numerical simulations are achieved with a different signal to noise ratio to show the effect of the false peaks. A Monte Carlo simulation with a different number of antenna elements is implemented and results verified that the detection performance and estimation accuracy of the PNS and MUSIC are comparable and both are much better than the PHD algorithm. It is also demonstrated that PNS presents a lower computational load than the MUSIC method.