Quality-of-Service-Aware and Energy-Efficient Cluster-Based Routing Protocol for Reliable Data Transfer in the Internet of Underwater Things

The Internet of Underwater Things (IoUT) supports marine applications such as environmental monitoring, offshore exploration, disaster prevention, and underwater infrastructure inspection. However, limited acoustic bandwidth, long propagation delays, unreliable communication links, and constrained node energy make reliable and energy-efficient routing challenging. We propose a Quality-of-Service (QoS)-Aware and Energy-Efficient Cluster-Based Routing (QEE-CBR) protocol that integrates cylindrical three-dimensional network partitioning, multi-criteria cluster head selection, hierarchical forwarding, and in-network data aggregation. The proposed Multi-Criteria Decision-Making (MCDM) mechanism selects cluster heads using residual energy, node degree, node proximity, and centroid distance, while regional cluster heads coordinate upward data forwarding toward a single surface sink. We implemented QEE-CBR in Network Simulator 3 (NS-3). We evaluated it against four benchmark protocols: Balanced Energy-Efficient Clustering (BEEC), Multi-Level Clustering Energy-Efficient Routing (MLCEE), Reliable QoS-Aware Routing (RQAR), and Energy-Aware Void Avoidance Routing Protocol (EAVARP). The performance metrics used included packet delivery ratio, throughput, end-to-end delay, network lifetime, and energy consumption for varying network densities. We conducted additional experiments to investigate the contributions of individual protocol components through ablation analysis, assess the sensitivity of routing performance to selected parameters, and compare the proposed clustered hierarchical architecture with our previous QoS-Aware and Energy-Optimised Cooperative Opportunistic Routing Protocol (QE-ORP). At 500 nodes, QEE-CBR achieves a packet delivery ratio of 99%, a throughput of 425 kbps, an end-to-end delay of 0.15 s, and a first-node-death time of 1130 s, along with a total energy consumption of 730 J. The architectural comparison highlights the distinction between QE-ORP’s flat cooperative opportunistic forwarding and QEE-CBR’s spatially organised hierarchical routing. The results obtained provide a comparative simulation-based assessment of the proposed protocol, its cluster head selection mechanism, and the effects of its principal design choices under the configurations used in our evaluation.