Research Info

Home \A cost-effective ...
Title
A cost-effective LoRaWAN-based IoT localization method using fixed reference nodes and dual-slope path-loss modeling
Type Article
Keywords
GNSS-free, IoT localization, RSSI, Path-loss modeling, LoRaWAN
Abstract
Recently, RSSI-based localization has gained popularity for outdoor localization within the IoT ecosystem due to its cost-effectiveness, low-power and low-cost deployment, and ability to operate loT nodes for years on a single battery. However, this approach typically sacrifices accuracy, resulting in location estimates in the tens or hundreds of meters. For example, LoRaWAN communications have been widely used in IoT applications, being an attractive solution due to their long-range coverage, low implementation cost, and higher autonomy. However, RSSI-based LoRaWAN location methods suffer from multipath and fading interference which results in the near-far problem, impacting ranging accuracy for short and long distances, and thus leading to an overall decrease in the localization accuracy. In this article, we address the low accuracy of RSSI based localization challenge for outdoor IoT node localization by computing the Path Loss (PL) parameters of LoRaWAN, for short and long distances separately, thus providing a dual-slope PL model. In addition, some IoT nodes in specific locations are adopted as Reference Nodes (RNs), whose main task is to estimate the interference effect on the transmitted signals dynamically. The proposed model has been evaluated using a publicly available LoRaWAN dataset collected in urban areas in the city of Antwerp, Belgium, which serves as a benchmark for the evaluation of the results. Its effectiveness is assessed by simulation and comparison to the state-of-the-art. In addition, results are compared with the derived Cramer–Rao Lower Bound (CRLB). The localization error achieves a median error of 117 m and a mean error of 236 m.
Researchers Azin Moradbeiki (First researcher) , Ahmad Keshavarz (Second researcher) , Habib Rostami (Third researcher) , Sara Paiva (Fourth researcher) , Serjio Lopez (Fifth researcher)