ReLeC: A Reinforcement Learning-Based Clustering-Enhanced Protocol for Efficient Energy Optimization in Wireless Sensor Networks

Abstract

Wireless sensor networks (WSNs) are a widely studied area in the field of networked embedded computing. They are made up of several sensor nodes, which keep track of a variety of physical and environmental parameters, like temperature and humidity. The nodes are autonomous, self-configuring, and wireless. A significant problem in WSNs is that sensors in these networks consume a lot of energy. Energy consumption is a big issue when it comes to the deployment of sensor networks. The reason for this is the cost of operating a sensor node and the cost incurred due to energy consumption. Energy optimization is based on intelligent energy management. This paper presents a reinforcement learning-based and clustering-enhanced method. Reinforcement learning is a set of algorithms inspired by operant conditioning in animal behavior, and clustering-based methods have been extensively used for devising energy-efficient protocols. The proposed method is able to plan and schedule the nodes to ensure an extended network lifetime. In this work, we aim to assess and increase the efficiency of power consumption and reduce sensor node energy loss. The simulation results prove that the presented protocol effectively reduces the energy consumption of sensor nodes and ensures a prolonged lifetime of the sensor network.