High-Resolution Quaternion-Based Algorithm for Coherent Underwater Sources with Linear Vector-Hydrophone Array

For scenarios of coherent underwater signals at a low signal-to-noise ratio (SNR), a novel high-resolution DOA algorithm based on quaternion algebra has been pro-posed to improve the poor performance of the subspace-based method without eigendecomposition (SUMWE). We construct four quaternion-based models by judiciously arranging the received data and then obtain the signals' statistical characteristics by performing cross-correlations between the models. In the quaternion algebra framework, we statistically eliminate the autocorrelation noise by using the properties of circular signals. The noise elimination provides a high-resolution performance for the algorithm at a low SNR. Moreover, both the cross-correlations and autocorrelations are utilized to enhance DOA estimation accuracy. Compared with the SUMWE-like methods, the proposed algorithm shows superiority in terms of estimation accuracy and angular resolution.