Long-range, Low-Power Wide Area Network (LoRaWAN) is a very scalable solution for the Internet of Things (IoT). Due to the air channel environment’s complexity, connectivity is a crucial parameter for successfully planning and deploying the IoT networks. Measurements and simulations have been used to evaluate LoRaWAN propagation models in the Urban environment, but it is a challenging task. While practical propagation evaluation has been prohibitively expensive, the theoretical modeling results have been less accurate. This paper uses real-world measurements and a trace-driven simulation technique to evaluate the RF propagation models’ prediction performance for LoRaWAN 868 MHz propagation. First, a novel LoRaWAN trace- driven simulation of Glasgow city centre has been performed. Second, LoRaWAN 868 MHz measurements have been used to perform a critical analysis of LoRaWAN trace-driven Radio Frequency (RF) propagation models and validation. The processed trace dataset is composed of GPS coordinates, and the corresponding LoRaWAN received signal strength. The dataset has been extracted from 5017 datasets of LoRaWAN measurements taken from Glasgow city centre. A trace simulation program built-in ICS-Telecom was used to simulate LoRaWAN propagation in the real-world urban environment. Comparison of LoRaWAN simulation traces and the real-world data was performed to evaluate the prediction performance accuracy of Deygout 94, ITU-R 525/526, and COST-Walfish Ikegami (COST-WI) propagation models. All models over-estimated LoRaWAN trace-simulated RSS levels in comparison to collected measurement samples. While Deygout 94 prediction accuracy was higher with mean absolute error (MAE) at 0.83 dBm and standard deviation (SD) at 4.17 dBm, COST-WI performed poorly with MAE and SD at 2.87 dBm and 10.96 dBm respectively.
|Number of pages||10|
|Journal||Advances in Science, Technology and Engineering Systems|
|Publication status||Published - 8 Nov 2020|
- LoRaWAN, trace-simulation, received signal strength, deterministic models, empirical models, urban environment