DOA Estimation via Manifold Separation Technique

The manifold separation technique (MST) models the steering vector of antenna arrays with arbitrary geometry, which can be expressed as product of a sampling matrix (dependent on the antenna array only) and a Vandermonde structured vector (dependent on the wavefield only). This allows fast direction-of-arrival (DOA) algorithms designed for linear arrays to be used on arrays with arbitrary configuration, such as root-MUSIC. The calibration measurements are corrupted by noise in real-world applications, which impairs the performance of MST-based algorithms. The effect of noisy measurements on subspace-based DOA algorithms is analyzed. A link between the optimal number of selected modes and the statistics of noise is established. Simulation results show that performance is satisfactory when MST is used for 1-D (azimuth) DOA estimation.