Fuzzy Based Sleep Scheduling Algorithm with Machine Learning Techniques to Enhance Energy Efficiency in Wireless Sensor Networks

Abstract

Wireless sensor networks, generally, are grouped into clusters to collect information effectively. Such grouping of nodes helps immensely to elongate the life of Wireless Sensor Networks. Message exchanges between nodes for consecutive and periodic clustering overload the sensor nodes and cause a shortfall of energy. Additional overhead during cluster formation, instability in energy use and the difficulty of information sharing during clustering, uncertain network structure, etc. are the current clustering problems. There is also a need to enhance intra-cluster transmission and to find effective methods to extend the network's lifespan. This paper aims to reduce the energy loss of nodes by reducing the message transmission overhead and simplifying the creation and upgrading of clusters to improve the lifespan of the network. A clustering strategy where the cluster is regularly restructured to decrease the overhead on cluster head nodes is also proposed in the paper. The suggested approach reduces data transmission using machine learning by the cluster member nodes and reduces the energy consumption of individual sensor nodes by implementing a suitable active/sleep schedule. To calculate the cluster update cycle and sleep cycle, it also makes use of the advantages of fuzzy logic by selecting appropriate fuzzy descriptors such as average data rate, distance from the head node to the sink and the remaining energy. The proposed approach optimizes the energy utilization of cluster heads and node members thereby enhancing the lifespan of the sensor network.