GTR: GAN-Based Trusted Routing Algorithm for Underwater Wireless Sensor Networks.

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Underwater wireless sensor networks face security challenges due to open transmission environments susceptible to malicious nodes.
• Distinguishing malicious nodes is difficult, especially in dynamic underwater conditions, impacting data integrity and network reliability.

Purpose of the Study:

• To propose a novel Generative Adversarial Network (GAN)-based trusted routing algorithm (GTR) for enhanced security and efficiency in underwater wireless sensor networks.
• To improve the detection of malicious nodes, including unknown intrusions, and optimize data forwarding routes in challenging underwater environments.

Main Methods:

• GTR defines trust feature attributes and evaluation matrices for network nodes.
• A trust evaluation model is constructed using a Generative Adversarial Network (GAN) for malicious node detection.
• The algorithm integrates trust evaluation with Q-Learning-based adaptive routing for optimal data forwarding.

Main Results:

• GTR demonstrated superior malicious node detection rates and reduced error detection rates compared to baseline algorithms.
• Significant improvements were observed in packet delivery rates, energy efficiency, and network throughput.
• The algorithm effectively handles unlabeled and imbalanced training data for robust malicious node identification.

Conclusions:

• The proposed GAN-based trusted routing algorithm (GTR) significantly enhances the security, reliability, and efficiency of underwater wireless sensor networks.
• GTR's adaptive routing and robust malicious node detection capabilities make it well-suited for dynamic underwater environments.
• The algorithm offers a promising solution for securing critical data transmission in underwater applications.