DEAR-SFCM: Dynamic Energy Aware Radius Adaptive Subtractive Fuzzy C-Means Clustering Framework for Efficient and Sustainable Wireless Sensor Networks

Managing energy efficiently in Wireless Sensor Networks (WSNs) is challenging because uneven energy consumption and early node failures degrade network performance. To address this issue, the dynamic energy-aware and radius-adaptive subtractive fuzzy C-means clustering method (DEAR-SFCM) is proposed to create more balanced clusters and improve both network stability and lifetime. The proposed method combines adaptive subtractive clustering, an energy-distance weighted measure, and entropy-regularized fuzzy membership updates to generate clusters that better reflect both the energy levels and spatial locations of sensor nodes. It incorporates an adaptive radius control mechanism that adjusts the clustering radius according to node density, thereby preventing overcrowding in dense regions and excessive cluster formation in sparse areas. In addition, optimal nodes are selected as cluster heads (CHs) using a multi-criteria CH selection strategy. Compared with existing state-of-the-art methods, DEAR-SFCM demonstrates superior performance in terms of energy balancing, network lifetime, alive-node ratio, and packet delivery rate. The results show that DEAR-SFCM reduces hotspot formation, distributes communication tasks more evenly, and extends both the stable operation period and the overall network lifetime. These improvements make DEAR-SFCM a robust and adaptable solution for energy-constrained WSNs and future large-scale Internet of Things (IoT) monitoring applications. Keywords: Cluster head selection, energy efficiency, fuzzy C-means, network lifetime, wireless sensor networks