Abstract:
Internet of Underwater of Things (IoUT) networks offer real-time monitoring and data collection capabilities in marine environments for many application. However, the unique challenges of underwater acoustic communication and energy constraints requires a development of intelligent, adaptive and energy-efficient protocols. The thesis addresses these energy challenges particularly in IoUT heterogeneous networks by proposes an energy-aware clustering protocol to optimizing the cluster heads (CH) management and inter-cluster route decision based on Stable election protocol (SEP) computations. Two approaches are proposed, the first is hybrid Artificial Bee Colony (ABC) and Q-learning (QL) based energy-aware Stable Election Protocol (ABCQL-EASE) and its improved version as an energy-aware and Quality of Service (QoS) driven by Underwater Vehicles (UVs) relaying integrated Stable Election Protocol (EQURI-SE).
The ABCQL-EASE presents intelligence approach that combines the ABC algorithms with the adaptive decision-making by QL and multi-criteria decision-making (MCDM) approach. This integration enables the dynamic CHs management based on underwater nodes (UNs) status. The protocol continuously learns from network observations, balancing exploration and exploitation to avoid premature convergence and extend network lifetime. Building upon this foundation, EQURI-SE further enhances energy efficiency by incorporating UVs as mobile relays between CHs. To improve the selection of UVs, a comprehensive framework proposed which integrates Neural Network (NN) and Genetic Algorithm (GA) with MCDM to evaluate UV candidates based on energy, depth, distance and signal quality. A dynamic cooldown mechanism ensures fair usage of UV relaying and prevents overuse of any single UV, thereby extending relay availability and system lifetime.
Results show that, ABCQL-EASE reduces the energy consumption and dead UNs approximate between 30% and 61% compared to the baseline and benchmark studies. The EQURI-SE achieves a significant reduction in energy consumption compared to conventional protocols, substantially contributing to extending network lifetime while maintaining high throughput and achieving superior network efficiency. Compared to ABCQL-EASE, the EQURI-SE reduces energy consumption by up to 48%. Both ABCQL-EASE and EQURI-SE provides higher network efficiency compared to baseline given between 92% and 95%.
