Towards scalable steady LoRa networks: performance prospects from link level to system level
File(s)
Author(s)
Bouazizi, Yathreb
Type
Thesis
Abstract
With a plethora of applications perpetually expanding to shore up different business use cases, the IoT ecosystem is continuously growing to underpin a wide variety of "Things". In response to this growth, the Low Power Wide Area Networks (LPWAN) paradigm has emerged, encompassing an array of technologies designed to extend network coverage, support low data rates, and consume less power. Among the LPWAN, LoRa has particularly attracted the attention of research and industrial communities thanks to the opportunity it offers to stakeholders to build private IoT infrastructures while reconciling range, power, and cost. However, with an unplanned deployment model and uncoordinated communications in the license-free spectrum, LoRa could face scalability issues. Characterising the scalability of LoRa networks and revealing what impacts it and how it trades with other performance KPIs are fundamental steps in engineering well-provisioned LoRa networks.
Unlike its licensed counterparts, LoRa is a new technology with a patented physical layer that relies on several tuneable parameters. While beneficial to answer the requirements of different IoT scenarios, these parameters bring more challenges to LoRa modelling, performance estimation, and scalability characterisation. This thesis explores LoRa and provides a comprehensive picture of its link-level and system-level performance. From a link-level perspective, we deeply investigate LoRa signal theory and modulation scheme. Our study reveals and quantifies the non-orthogonality of LoRa modulation. From a network-level standpoint, we use the outcome of the link-level analysis to devise meticulous, more accurate analytical frameworks to evaluate the performance of large-scale LoRa networks. The proposed analytical frameworks uniquely join stochastic geometry and queueing theory to reflect both device-level and network-level, account for the access restriction to the shared medium, which is a significant challenge for license-free technologies, and comprehensively evaluate the interplay between different performance indicators in LoRa-based IoT networks under both static and dynamic SF allocations.
Unlike its licensed counterparts, LoRa is a new technology with a patented physical layer that relies on several tuneable parameters. While beneficial to answer the requirements of different IoT scenarios, these parameters bring more challenges to LoRa modelling, performance estimation, and scalability characterisation. This thesis explores LoRa and provides a comprehensive picture of its link-level and system-level performance. From a link-level perspective, we deeply investigate LoRa signal theory and modulation scheme. Our study reveals and quantifies the non-orthogonality of LoRa modulation. From a network-level standpoint, we use the outcome of the link-level analysis to devise meticulous, more accurate analytical frameworks to evaluate the performance of large-scale LoRa networks. The proposed analytical frameworks uniquely join stochastic geometry and queueing theory to reflect both device-level and network-level, account for the access restriction to the shared medium, which is a significant challenge for license-free technologies, and comprehensively evaluate the interplay between different performance indicators in LoRa-based IoT networks under both static and dynamic SF allocations.
Version
Open Access
Date Issued
2023-07
Date Awarded
2024-02
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
McCann, Julie
Benkhelifa, Fatma
Sponsor
Singapore Ministry of National Development
Grant Number
L2NICTDF1-2017-3
Publisher Department
Computing
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)