Machine learning-driven improvements for efficient routing in advanced metering infrastructure for Smart Grids
| dc.contributor.advisor | Meng, Julian | |
| dc.contributor.advisor | Mohamad Mezher, Ahmad | |
| dc.contributor.author | Duenas Santos, Carlos Lester | |
| dc.date.accessioned | 2024-10-10T13:06:03Z | |
| dc.date.available | 2024-10-10T13:06:03Z | |
| dc.date.issued | 2024-08 | |
| dc.description.abstract | Smart Grids envisioned as the next-generation power grid, aim to revolutionize the way electricity is generated, distributed, and consumed. They promise enhanced efficiency, reliability, and greater integration of renewable energy sources. However, the achievement of these ambitious objectives is deeply dependent on the robustness and sophistication of Advanced Metering Infrastructure (AMI) systems. AMI is a critical component of Smart Grids, providing essential real-time data and two-way communication capabilities that are fundamental to the smart management of energy resources. Through AMI, Smart Grids gain the ability to dynamically respond to changing energy demands, implement efficient demand-response strategies, and offer consumers greater control over their energy usage. AMI systems rely heavily on robust and reliable communication networks. However, several critical challenges and limitations persist in the current communication networks used within AMIs, necessitating further research and innovation in this domain. One of those challenges is the way the data packets are routed so that they reach their intended destinations. Thus, this PhD thesis focuses on addressing the inherent limitation of the current routing protocols employed in AMI deployments. In pursuing this goal, three routing algorithms are proposed, RPL+, ML-RPL, and Q-RPL. They take as a base the Routing Protocol for Low Power and Lossy Networks (RPL) and on top of this implement Machine Learning-driven mechanisms to improve the routing process. The developed routing algorithms are validated through extensive simulations and experiments in representative AMI deployments. Evaluation criteria include typical metrics to measure the performance of communication networks such as packet losses, and latency. When the proposed routing algorithms are compared to the standard RPL implementations and other benchmark algorithms found in the literature, it is observed a significant improvement in network performance, which underscores their potential. These advancements represent a more robust, efficient, and reliable communication system within Smart Grids and lay the groundwork for future innovations in communication technologies in this context. | |
| dc.description.copyright | © Carlos Lester Duenas Santos, 2024 | |
| dc.format.extent | xv, 140 | |
| dc.format.medium | electronic | |
| dc.identifier.uri | https://unbscholar.lib.unb.ca/handle/1882/38151 | |
| dc.language.iso | en | |
| dc.publisher | University of New Brunswick | |
| dc.rights | http://purl.org/coar/access_right/c_abf2 | |
| dc.subject.discipline | Electrical and Computer Engineering | |
| dc.title | Machine learning-driven improvements for efficient routing in advanced metering infrastructure for Smart Grids | |
| dc.type | doctoral thesis | |
| oaire.license.condition | other | |
| thesis.degree.discipline | Electrical and Computer Engineering | |
| thesis.degree.grantor | University of New Brunswick | |
| thesis.degree.level | doctorate | |
| thesis.degree.name | Ph.D. |