Deep Learning Based Energy Efficiency in Wireless Sensor Network

Abstract

Wireless Sensor Network (WSN) comprise of huge amount of sensor nodes. These nodes sense the data from their surroundings and pass this information to the sink node using cluster head. Due to the emergence of new technology, it is widely used in distinct applications such as habitat monitoring, health science, border surveillance etc. There are several issues in WSN such as Quality of Service (QoS), localization, routing and data aggregation. Sensor nodes have limited energy, so there is a need to enhance the energy efficiency across the network. This paper focuses on two mechanisms of energy efficiency: energy consumption and energy harvesting. Energy consumption can be minimized by using different Machine Learning (ML) approaches. The other mechanism is energy harvesting. It provides the further two sources: ambient source and external source. Ambient source consists of renewable resources such as radio frequency, solar, thermal and flow-based energy harvesting. Radio frequency converts the radio waves into electric signal, solar mechanism converts solar rays to electric signals, thermal mechanism converts heat energy to electric energy and flow technique convert the rotatory movement to electric signal. External source includes mechanical and human based energy harvesting. Further, the proposed statistical analysis of eight years (2014-2021) illustrated the fact that different ML techniques applied in energy efficient parameter reduces the consumption of energy across the network. These two mechanisms enhance energy efficiency parameter and network lifetime by using ML. Author vision is discussed as an open issue in the last.