DOA estimation for coprime EMVS arrays via minimum distance criterion based on PARAFAC analysis

This study presents a minimum distance‐based direction‐of‐arrival (DOA) estimation algorithm for coprime electromagnetic vector sensor (EMVS) arrays. The idea is to split‐up the coprime array into two uniform linear arrays (ULAs) of vector sensors and arrange the received ULA data in the form of a three‐way array suitable for parallel factor (PARAFAC) analysis, which fits least‐square models to the received source signal mixtures of ULAs and thus enables to retrieve the model matrices corresponding to each ULA. Nevertheless, because of the array splitting the estimated DOAs from these matrices are not unique. To uniquely determine the DOA, the authors state and prove a theorem which is fundamental to the proposed algorithm and provides a means to find an estimate based on the minimum distance criterion. Efficacy of the proposed algorithm is demonstrated through performance comparison with other existing algorithms such as Estimation of Signal Parameters via Rotational Invariance Technique (ESPRIT), long vector MUltiple SIgnal Classification (MUSIC), conventional PARAFAC and the propagator method being simulated for an equivalent element ULA of EMVS and spaced half a wavelength apart. Numerical simulations reveal that the proposed algorithm outperforms the others.