Clustering with hierarchical routing (GMMCHR): a new gaussian mixture model for wireless sensor networks

Abstract

Military surveillance, industrial applications, and real-time environmental monitoring all depend on wireless sensor networks (WSNs). However, due to insufficient power sources for sensor nodes, energy efficiency (EE) and network lifetime (NL) extension are significant challenges. To vanquish these constraints, this investigation suggests a new GMMCHR (Gaussian Mixture Model Clustering with Hierarchical Routing) protocol that combines energy-aware routing with probabilistic clustering. The approach segregates network into NC (Near Clusters) and FC (Far Clusters) based on node distance from the BS. CHs are selected using a fitness function incorporating residual energy and spatial proximity, with FCs formed via Enhanced Gaussian Mixture Models (EGMM) and routing managed through a hierarchical structure. Simulations conducted in MATLAB R2021a under two scenarios—100 nodes in a 100×100 m² region and 200 nodes in a 200×200 m² region—demonstrate significant improvements over the benchmark EEHCHR protocol. In the 100-node scenario, GMMCHR delays the FND (First Node Dead) to 66 rounds, HND (Half Node Dead) to 911 rounds, and LND (Last Node Dead) to 1601 rounds, compared to EEHCHR’s 45, 735, and 1359, respectively. In the 200-node setup, GMMCHR achieves FND at 48 rounds, HND at 904, and LND at 1231, outperforming EEHCHR’s 31, 731, and 1024 rounds. Additionally, GMMCHR maintains over 70% coverage beyond 1200 rounds in Scenario 1 and delivers over 17,000 packets to the base station, significantly higher than EEHCHR. Moreover, the combination of soft clustering in GMM with the hierarchical routing would allow dynamic flexibility, superior load balancing, and improved scalability. Overall, GMMCHR provides an effective and capable method of enhancing the lifetime of the WSN in both small-scale and large-scale systems.