Energy‐Efficient Multi‐Level Adaptive Clustering for Enhanced Performance in Underwater Sensor Networks

ABSTRACT Underwater wireless sensor networks (UWSNs) are vital for real‐time data collection and monitoring in underwater environments, with applications in environmental conservation, disaster management, marine biodiversity monitoring, and offshore energy production. However, UWSN deployment faces significant challenges such as high energy consumption, limited communication range, signal attenuation, and environmental factors like underwater currents and temperature variations. Existing solutions suffer from scalability and adaptability in deeper and more complex underwater environments. This paper presents an Energy‐Efficient Multi‐Level Adaptive Clustering (EEMLAC) scheme designed to optimize energy consumption and enhance the network lifetime of UWSNs. The proposed approach dynamically adjusts clustering structures based on depth and water pressure, optimizing signal management and minimizing energy loss. Nodes are categorized into three adaptive levels: the first level, at shallower depths, divides the outer and middle sections into three subsections at 120° angles; the second level, at mid‐depths, increases the division to four subsections at 90° angles to mitigate higher signal attenuation; and the third level, at the greatest depths, further refines clustering with six subsections at 60° angles to efficiently manage data flow and energy usage. Normal nodes are positioned in the middle section for direct communication with the cluster head, whereas advanced nodes in the outer section utilize helper nodes to relay data, reducing transmission energy consumption. By adapting to environmental factors such as water pressure and attenuation, the proposed scheme effectively addresses communication challenges in deep‐water UWSNs. Simulation results demonstrate that EEMLAC achieves superior performance, retaining approximately 0.6 J residual energy after 1000 rounds, reducing energy consumption to 2.3 × 10 −3 J per node, extending network lifetime beyond 1800 rounds, achieving a packet delivery ratio (PDR) of 0.586, and improving throughput to 3.7 × 10 6 bits/s, outperforming benchmark schemes such as EBREC, EAMC, and EGRCs.