Enhancing Wireless Sensor Networks Performance by Integrating Particle Swarm Optimization with Intelligent Clustering Techniques

Abstract

Wireless sensor networks play a crucial role in various applications, including industrial automation, healthcare, and environmental monitoring. Energy consumption remains a major challenge due to the limited energy of sensor nodes. This study utiliz-es FCM, Heuristic, K-Means, DBSCAN, and PSO to enhance network performance and longevity. K-Means balances clusters, DBSCAN detects dense regions, FCM assigns flexible memberships, and heuristic clustering adjusts clusters based on energy and base station proximity. The proposed method dynamically selects cluster heads based on energy levels and base station prox-imity. PSO optimizes selection by evaluating intra-cluster distances and residual energy, enabling dynamic reselection for improved efficiency, lower transmission costs, and extended network lifespan. Simulations show high energy savings in FCM (1.7213J for 50 nodes, 5 clusters), while K-Means depletes energy the fastest (1.5394J for 150 nodes) and DBSCAN consumes the most energy (1.0258J), rendering it unsuitable for longevity-focused applications. FCM ensures the longest network lifespan (904 iterations), compared to K-Means (483 iterations) and Heuristic (443 iterations). In terms of latency, K-Means experiences the highest delays (1.5s), while FCM and heuristic clustering maintain lower delays around 1.0s or less. The hybrid FCM-PSO approach reduces energy consumption by 12–15% and extends network lifespan by 20%.