DAAG-SNP: Energy Efficient Distance and Angulation-Based Agglomerative Clustering for Sink Node Placement

Abstract

Wireless Body Area Networks (WBANs) have significantly enhanced various aspects of human life, particularly in healthcare, fitness, entertainment, sports, and etc. In WBANs, the sensor nodes are placed in and around the body along with the sink node, which collects the physiological data from these sensors and forwards it for further processing. The placement of the sink node is one of the critical aspects in the design of WABNs as it affects both the energy efficiency and connectivity. To this end, this paper introduces a hybrid method called Distance and Angulation based AGglomerative Clustering (DAAG). DAAG, initially clusters the WBAN sensors using Distance and Angulation based k-Mean clustering. Afterward, Agglomerative Clustering is applied to determine the optimal placement of the sink node. The results of DAAG are compared with various machine learning and optimization approaches, including D-RMS (Distance based Random mean shift clustering), Reinforcement Q-Learning Approach (QL), Humpback Whale optimization (HWOA), Multi-Angulation (MA) and Closeness Centrality (CC). Given an initial energy, the results show that the DAAG exhibits superior performance in terms of latency, packet error rate (PER), and energy consumption. DAAG shows an energy consumption of only 1.51% outperforming QL, HWOA, MA, CC, and D-RMS along with an improved localization accuracy of 0.36 m.