Satellite Network Security Routing Technology Based on Deep Learning and Trust Management.

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Conventional trust models in satellite networks struggle with accurate malicious node detection and adapting to novel cyber threats.
• Existing routing algorithms lack robustness against sophisticated attacks, compromising network integrity and performance.

Purpose of the Study:

• To introduce a novel secure satellite network routing technology leveraging deep learning and distributed trust management.
• To enhance the accuracy and adaptability of malicious node detection and network security in satellite communications.

Main Methods:

• Utilized D-S evidence theory for preprocessing satellite node communication behavior to mitigate interference for Variational Autoencoder (VAE) module training.
• Implemented a distributed trust management system with anomaly detection modules on each satellite node.
• Integrated processed trust vectors into a VAE module, followed by safety factor assessment and incorporation into an ant colony algorithm for path discovery.

Main Results:

• The proposed deep learning-based secure routing algorithm effectively mitigates the impact of malicious nodes on satellite network data transmission.
• Simulation results demonstrate significant improvements in average end-to-end delay, packet loss rate, and throughput compared to traditional models.

Conclusions:

• The developed secure routing technology offers a robust solution for enhancing satellite network security and performance.
• Deep learning and distributed trust management provide a promising approach for addressing the limitations of conventional trust models in dynamic network environments.