The first calibration model for bluetooth angle of arrival: Enhancing positioning accuracy in indoor environments

Internet of Things (IoT) applications are increasingly reliant on indoor positioning systems to deliver precise and reliable navigation in GNSS-denied environments, including urban areas, smart warehouses, hospitals, and underground or multi-level parking systems. Bluetooth Angle of Arrival (AoA) positioning offers cost-effective solutions with the potential to provide users with sub-meter position accuracy, which is crucial for applications such as underground navigation, firefighters, and robotic navigation. Bluetooth AoA positioning uses angles to determine the position of Bluetooth tags; these angles, measured in the anchor coordinate system, need to be transferred to the user's coordinate system. This requires models or techniques to compute 3D rotation matrices between the anchor and user coordinate system. Until now, no model or technique has been developed to compute these rotation matrices. Therefore, the development of the AoA positioning model focuses on simulated scenarios. This paper introduces the first model, named the AoA calibration model, capable of estimating these rotation matrices, thereby facilitating the practical application of this technology. In addition, this paper tests the Bluetooth AoA calibration and positioning model on a real dataset and presents end-toend functional architectures for AoA positioning. The results demonstrate that the proposed calibration model can estimate the 3D transformation rotation angles with a standard deviation better than 2.5 degrees. The findings also reveal that AoA positioning can achieve sub-meter accuracy in both static and kinematic modes, with accuracy significantly influenced by the distance to the anchors and the geometry factor.