Long Range Wide Area Network (LoRaWAN) propagation model performance in Botswana

Abstract

Long Range (LoRa) is a leading low power wide area network under the IEEE802.15.4 designation. The technology has been used extensively in wireless sensor network applications (WSN) and studies are shifting to investigate the technology’s viability when used for navigation. However, the problem of planning and deployment of LoRa in navigation as well as the repeatable accuracy of such a system based on known timing errors has not been done. The first chapter of this thesis presents a background into the LoRa technology and introduce some of the metrics of interest in navigation as well as in WSN applications, in which LoRa has traditionally been used. Additionally, the objectives and problem statement are presented. In the second chapter of the thesis, a closed form of the variance in time of arrival(TOA),away to determine pseudo-range for LoRa and introduce the sources of error in the ranging function, namely path-loss and noise are derived. Furthermore, the geometry in the navigation solution is determined. By investigating the contribution of geometry, assessments of the ideal gateway placement technique are performed. The gateway placement techniques under assessment are random placement, naive placement and imaginary triangular tessellation. These gateway placement techniques are then validated statistically against one another using effect size, statistical significance and the Z-statistic. The findings of this research show that LoRa is viable to use in navigation applications. It was observed through simulations that LoRa is capable of attaining repeatable accuracy below 20m. Moreover, the imaginary tessellation technique is the best placement technique based on statistical analysis as well as practical analysis of the results. It was also determined that the Ericsson 9999 model is able to produce simulation results which are more consistent with experimental findings compared to the COST 231- Walfish Ikagami path-loss model and the free space path-loss model. In conclusion, this work investigated the use of LoRa in a navigation application and found the technology viable. A gateway placement technique which is ideal for navigation, reduces geometry based errors and assured a low horizontal dilution of precision (HDOP) in a desired coverage area therefore increasing the accuracy of the position solution for LoRa was also determined. The work also proposes that LoRa be considered as a signal of opportunity (SoOP) in future works given the documented performance.