A Q-Learning-Based Distributed Energy-Efficient Routing Protocol in UASNs.

Area of Science:

• Computer Science
• Network Engineering
• Robotics

Background:

• Underwater acoustic sensor networks (UASNs) face significant challenges in energy efficiency and routing due to the harsh underwater environment.
• Existing routing protocols often struggle with high communication overhead and limited adaptability to dynamic channel conditions.

Purpose of the Study:

• To propose a novel Q-Learning-Based Distributed Energy-Efficient Routing (QDER) protocol for UASNs.
• To enhance the energy efficiency and network lifetime of UASNs through intelligent, distributed routing.
• To improve the robustness of routing strategies in noisy underwater acoustic environments.

Main Methods:

• Formulating the routing problem as a Markov Decision Process (MDP).
• Employing a distributed Q-learning approach where each node acts as an agent.
• Designing a reward function that incorporates node residual energy, depth, and link quality.

Main Results:

• The QDER protocol demonstrated superior performance compared to Depth-Based Routing (DBR) and Deep Q-Network-Based Intelligent Routing (DQIR).
• Significant improvements in network lifetime and energy efficiency were observed, particularly when considering channel attenuation and noise.
• The protocol showed good robustness and adaptability across various signal-to-noise ratio (SNR) conditions.

Conclusions:

• The QDER protocol offers an effective solution for energy-efficient and robust routing in UASNs.
• Distributed Q-learning, incorporating link quality, is a promising approach for optimizing UASN performance.
• The proposed method addresses key limitations of existing protocols in challenging underwater environments.