Machine learning-based energy-efficient routing and dynamic reconfiguration framework for smart city IoT networks

Abstract

Energy efficiency is critical in the sustainable operation of Internet of Things (IoT) networks, particularly in resource-constrained smart city environments. This paper delves into the challenges and opportunities for optimizing energy consumption in IoT routing protocols. We explore the limitations of traditional routing protocols and highlight the need for innovative approaches that can adapt to dynamic network conditions and device energy constraints. We propose a novel energy-efficient routing protocol that leverages advanced techniques such as machine learning and reinforcement learning to optimize routing decisions dynamically. Our protocol considers factors like node energy levels, link quality, and traffic load to select energy-efficient paths for data transmission. Additionally, we incorporate sleep scheduling mechanisms to minimize idle power consumption and prolong the network lifetime. Through rigorous simulations and evaluations, we demonstrate the significant energy savings and performance improvements achieved by our proposed protocol compared to existing solutions. Our findings provide valuable insights into designing and deploying energy-efficient IoT networks in smart cities, contributing to realizing sustainable and resilient urban environments.