Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/0423BAC1-64B4-4C60-A934-3CC0858AFE1A | ||
| Year | 2024 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Georgios Andreadis, "Phase-based localization of low power Bluetooth tags with multi-antenna receivers and comparison with RFID technology", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2024 https://doi.org/10.26233/heallink.tuc.100609 | ||
| Appears in Collections |
Localization technologies are essential for applications ranging from asset tracking to navigation systems. This study explores minimizing Bluetooth Low Energy (BLE) multi-antenna receivers (locators) to reduce energy consumption, hardware usage, and installation costs. It introduces a hyperbolic localization technique based on phase differences, using linear approximations of hyperbolas to calculate 3D direction of arrival (DoA) and estimate BLE tag positions. On a single locator, DoA estimation achieves a Mean Absolute Error (MAE) under 10° for azimuth and under 7° for elevation. However, a single locator cannot determine the tag's position due to system limitations. Therefore, the study explores using multiple locators. On multiple locators, the proposed localization method can enhance the prior art localization method, reducing the MAE of localization error by 10%.The work also compares Radio Frequency Identification (RFID) and BLE for tag localization accuracy under static conditions. While the hyperbolic localization technique is effective, it is sensitive to multipath noise. BLE's spatial and frequency diversity mitigate multipath issues, achieving a 3D localization MAE of 30 cm for the topology with high coverage and 80 cm for the topology with low coverage. In contrast, RFID, lacking such diversity, results in a 3D localization MAE of 1.2 m for the second topology. This comparison suggests developing new algorithms for consistent results across both technologies and the potential for merging BLE and RFID into a single tag to minimize energy consumption and enhance localization accuracy.To further minimize BLE locators, neural networks were employed. Recognizing BLE single position measurements as sequential data, Recurrent Neural Networks (RNNs) were utilized, achieving a 2D localization MAE of 30 cm with a single locator, a challenge unsolvable by deterministic methods.
Copyright © DIAS 2013
