Real time long range (LoRa) based indoor positioning system using Deep Gaussian Process (DGP) algorithm / Ng Tarng Jian
This thesis explores the development of a real-time LoRa-based indoor positioning system in industrial production lines. Recognizing the limitations of traditional GPS and other indoor positioning technologies, this research investigates the feasibility of LoRa and proposes a hybrid machine learning...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Published: |
2025
|
| Subjects: | |
| Online Access: | http://studentsrepo.um.edu.my/15999/1/Ng_Tarng_Jian.pdf http://studentsrepo.um.edu.my/15999/2/Ng_Tarng_Jian.pdf http://studentsrepo.um.edu.my/15999/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This thesis explores the development of a real-time LoRa-based indoor positioning system in industrial production lines. Recognizing the limitations of traditional GPS and other indoor positioning technologies, this research investigates the feasibility of LoRa and proposes a hybrid machine learning approach for accurate and reliable positioning. The study addresses challenges posed by signal fluctuations, non-line-of-sight propagation, and the need for continuous positioning estimation in dynamic environments. Through experimental evaluation and comparison of various machine learning algorithms, including Deep Gaussian Process (DGP) regression, the research demonstrates the effectiveness of DGPs in achieving precise single-point estimation, by keeping the mean absolute error to below 5 meters. Furthermore, the thesis introduces enhancement techniques such as Temporal-Weighted RSSI averaging, Kalman filtering, and lane constraints to improve the system's performance further. The experimental results, conducted in a real industrial environment, demonstrate that the proposed system achieves a mean absolute error of 1.58 meters and a root mean square error of 1.90 meters. These findings highlight the potential of combining LoRa technology with advanced machine learning algorithms and filtering techniques to achieve precise and reliable indoor tracking.
|
|---|