|
|
Enhancing Localization Accuracy in Industrial Wearables with LoRaWAN
Svertoka, Ekaterina ; Martian, Alexandru (oponent) ; Digulescu-Popescu,, Angela (oponent) ; Lohan, Elena Simona (oponent) ; Hošek,, Jiří (oponent) ; Burget, Radim (vedoucí práce)
This research combines theoretical insights, simulation studies, and practical experimentation to explore the field of industrial wearables, with a focus on enhancing their location accuracy through LoRaWAN technology. The work led to creating two classifications on functions and metrics of industrial wearables and 7 open access LoRaWAN datasets in diverse environments (indoor, outdoor, and underground). Moreover, the paper conducts a comprehensive assessment of localization accuracy through multiple approaches, analyzing the influence of variables from the measurement campaign and data processing techniques. Furthermore, it proposes modifications to the k-NN algorithm, which, alongside preprocessing methods, results in a 17.2% increase in accuracy compared to the original benchmark. Validated on LoRaWAN datasets, the proposed algorithms offer potential applications across various fields. The study concludes validation of LoRaWAN-based localization with an accuracy of 2.6m indoors and 4m outdoors, suggesting that while LoRaWAN-based localisation is not as precise as leading technologies, it can be used in sectors such as logistics, agriculture, and smart manufacturing where absolute precision is not essential.
|
host ::
RSS.