A Machine Learning Dataset for Enhancing Energy Efficiency in WSN

Abstract

WSNs are constructed of a large number of tiny energy-constrained nodes and have low capacity. Sensor nodes are skilled to carry the functioning of sense, aggregating, and transmitting information. In this paper, the use of machine learning is suggested in order to enhance the energy efficiency of WSNs. The proposed method aims at establishing a dataset that is used by a machine learning model to choose the best Cluster Head (CH) in WSN. Forming a sufficient dataset is primarily based on assuming several network parameters. For each combination of these parameters, the node which leads to the least energy consumption will be selected as CH. The system parameters used to build this dataset are inter-cluster distance, node residual energies, and how often each node is selected as a CH. As a result, a dataset for choosing the best cluster head in the WSN is created and would be trained by a machine learning model, where the dataset labels the best node to be chosen as a cluster head compared with the physical location of the node on the network.