True-angle sparse Bayesian learning for horizontal line arrays in a multipath shadow-zone environment

Horizontal line arrays may exhibit large bearing-estimation errors in deep-ocean shadow zones, particularly near the endfire direction, because the received field is dominated by vertical-plane multipath rather than a direct arrival. To address this problem, this paper analyzes the geometric relationship among the ray arrival angle, the apparent bearing measured by a horizontal line array, and the true target bearing. Based on this relationship, a true-angle sparse Bayesian learning (TA-SBL) method is proposed. In TA-SBL, several apparent-angle branches associated with the same candidate true bearing are represented by one sparse block and controlled by a shared hyperparameter so that the bearing estimate is obtained directly in the true-angle domain. Simulations show reduced multipath-induced bearing bias in the examined cases.