IEEE Access, cilt.13, ss.127868-127884, 2025 (SCI-Expanded)
Wireless sensor networks (WSNs) play a vital role in modern applications such as healthcare, smart cities, and environmental monitoring. However, their potential is often limited by energy constraints, which reduce the lifetime of the network and the data collection capabilities. To address this challenge, we propose KDL, a novel hybrid clustering algorithm that combines K-Nearest Neighbor (KNN) and Density-Based Spatial Clustering (DBSCAN) to optimize energy efficiency in WSNs. KDL first uses KNN to analyze internode distances and determine optimal clustering parameters, which guide DBSCAN in forming robust clusters. After clustering, KNN reassigns noise points to appropriate clusters, improving coverage. Next, an energy-aware cluster head (CH) selection mechanism, inspired by LEACH but enhanced with node energy levels and cluster centroid distances, ensures balanced energy consumption. In addition, a relay-assisted communication strategy optimizes data transmission by strategically placing relay nodes between the CH and the base station. Through extensive simulations, KDL shows significant performance improvements over existing approaches, substantially enhancing network lifetime while maintaining energy efficiency. These advancements position KDL as a promising solution for real-world energy-efficient WSN deployments.