Multi-level clustering and Prediction based energy efficient routing protocol to eliminate Hotspot problem in Wireless Sensor Networks

Conserving energy of sensor nodes and ensuring balanced workloads among them are fundamental concerns in Wireless Sensor Network (WSN) design. Clustering strategies offer a promising avenue to minimize node energy consumption, thereby prolonging network lifespan. Nevertheless, numerous multi-hop routing protocols using clustering technique face the challenge of nodes nearer to the Base Station (BS) depleting their energy faster due to forwarding data from the entire network leading to premature node failure and network partitioning known as 'hotspot problem'. The paper introduces an Energy-Efficient Mega-Cluster-Based Routing (EEMCR) protocol specially designed for expansive coverage area. The primary principle behind designing this protocol is to eliminate the hotspot problem and restrict the transmission range of nodes to the threshold distance defined by the radio energy model, thereby enhancing the overall network lifespan. The protocol adopts a centralized approach employing fixed clustering wherein the BS partitions the network into square-shaped clusters. The cluster size is determined by the threshold transmission range of the sensor radio energy model, guaranteeing that all network communication stays within this threshold distance. Four such clusters form a mega-cluster with a Mega-Cluster-Head (MCH) elected among the four Cluster Heads (CHs). The MCH role is evenly distributed among nodes of all four clusters in subsequent rounds for uniform distribution of its overhead. Implementing data aggregation at two levels (CH level as well as MCH level) leads to reduced data traffic and energy consumption throughout the network. Moreover, data collection by two data mules based on odd-even round number ensures balanced data traffic and energy distribution across the network. Analysis indicates that the proposed protocol effectively mitigates the hot-spot problem and reduces data transmission overhead of sensor nodes. In simulation, the proposed protocol on an average improves network life by 34.5%, 23.5%, 14.5% and 5.5% as compared to existing protocols FCEEC, DBSCAN, LPGCR and FBECS respectively for deployment of nodes between 600 to 1200. Also, approximately 46%, 32%, 21% and 14% of lesser sensor nodes are dead for proposed protocol in respective rounds as compared to existing protocols FCEEC, DBSCAN, LPGCR and FBECS respectively. Comparative evaluations demonstrate improved network lifetime when compared to equivalent recent routing protocols.