An Energy‐Aware Dual‐Intelligence Protocol for Dynamic and Self‐Optimizing Underwater Wireless Sensor Networks

Underwater wireless sensor networks (UWSN) have been proven essential in providing solutions for various purposes ranging from aquatic surveillance to marine exploration and defense. However, such networks suffer from the drawback of being hindered by several aspects including lengthy propagation delay times, scarcity of energy supplies, changing network structures, and unstable acoustic mediums—issues that heuristic and static methods cannot resolve. Driven by these challenges, the Optimized Clustering and Energy‐Aware Network Protocol called OCEANet has been proposed in this research. This protocol is based on a double intelligence approach for a self‐adapting dynamic underwater network environment. Its unique feature lies in combining the evolutionary biology‐based clustering technique with software‐defined radio configuration through artificial intelligence techniques. The OCEANet system comprises two operational phases that complement each other. The clustering and evolutionary bio‐beaconswarm routing algorithm selects the appropriate cluster heads considering their residual energy, location, and link quality and thus helps create the optimal multi‐hop route toward the sink node. The Self‐Updating Reconfigurable Framework uses the Deep Q‐Network reinforcement learning technique along with Graph Convolutional Network–based transfer learning. The proposed OCEANet outperforms the other models with a packet delivery ratio of 94% and throughput of 189.18 kbps when there are 100 nodes. Also, while lowering the energy usage and delays, and by increasing the network life cycle, OCEANet establishes it as a reliable, scalable solution for next‐generation underwater sensor networks.