A Low-Complexity 3-D Source Localization Method Using 1-D AOAs of Multiple Linear Arrays

Traditional Angle of Arrival (AOA) localization in three-dimensional space typically requires sensors equipped with planar arrays, which incurs additional hardware costs. This limitation restricts its application in systems such as the Internet of Underwater Things (IoUT). Recent studies have shown that localization can also be achieved using sensor networks composed solely of linear arrays. However, most existing methods impose strict constraints on the orientation or placement of sensor arrays. Although some studies have proposed localization solutions free from these two limitations, such methods still exhibit two notable shortcomings: (1) high computational complexity that scales with network size, making them unsuitable for resource-constrained scenarios or large-scale sensor network deployments; and (2) poor robustness to sensor position errors, where nodal deviations can significantly degrade localization accuracy. To address these challenges, this study proposes a novel computationally-efficient 3-D source localization method based on 1-D AOA measurements from multiple linear arrays. Significantly, we innovatively incorporate a Weighted Least Squares (WLS) compensation model that effectively enhances the method’s robustness against sensor position errors. Experimental results demonstrate that: (1) while achieving the theoretical optimum localization accuracy as defined by the Cramér-Rao Lower Bound (CRLB), the proposed method shows significantly lower computational complexity than existing methods, with complexity independent of sensor network scale; (2) in practical scenarios with node position errors, our method outperforms other state-of-the-art methods in localization accuracy.