Void Avoidance in Underwater Acoustic Sensor Networks with Deformable Bi-Level Routing Attention-Based Multiagent Adaptive Deep Deterministic Policy Gradient Routing Protocol

Underwater acoustic sensor networks (UASNs) have gained increasing importance in various military and civilian applications due to their ability to facilitate communication in challenging underwater environments. However, routing in UASNs faces unique challenges stemming from the harsh characteristics of underwater acoustic communication caused by water currents and node mobility. This work suggests a comprehensive routing framework that uniquely integrates void avoidance with intelligent, deep reinforcement learning–based routing techniques.The contribution from this research is the performance of void avoidance and routing for the effective transmission. Specifically, a novel HH-VBF (hop-by-hop vector-based forwarding) method is employed to detect and eliminate the void nodes, thereby enhancing reliability and reducing packet loss. Furthermore, the proposed deformable bi-level routing attention-based multi-agent adaptive deep deterministic policy gradient (DBRAMADDPG) algorithm, optimized using the improved nuclear reaction optimization algorithm (ImNROA), dynamically learns optimal routing paths to reduce end-to-end delay and energy consumption by avoiding the void area.This unified approach can effectively mitigate void-induced disruptions while balancing energy efficiency, extending the network lifetime, and improving the quality of service. Extensive simulations demonstrated that our model outperformed existing methods, reducing delay by up to 7%, while taking the number of nodes as 10, the delays of the IR2P (intelligent reinforcement learning routing protocol with integrated power-control), HMERP (hybrid message encoding-based routing protocol), ROAR (resource-optimized adaptive routing), VAHR (void-aware hybrid routing), and the proposed ImNROA-DBRAMADDPG were 0.8948%, 0.8639%, 0.84533%, 0.80878%, and 0.78594%, respectively, ensuring more stable data delivery in complex underwater scenarios.These advancements contribute to the state of the art by offering robust routing strategies capable of dynamically adapting to underwater environmental challenges and void formations. The findings provide potential implications for designing resilient and energy-efficient UASNs, paving the way for improved real-time underwater monitoring and secure acoustic communication systems. These findings pave the way for more reliable, sustainable, and intelligent underwater sensor networks, with significant implications for real-time marine exploration, defense, and environmental monitoring.