LEO Satellite-Ground Link Authentication via ML-Driven Channel Fingerprinting

Overview

Low Earth Orbit (LEO) satellite constellations (Starlink, OneWeb, Amazon Kuiper) are rapidly expanding to provide global connectivity, but their open broadcast nature makes them vulnerable to spoofing, replay attacks, and unauthorized access. Traditional cryptographic authentication adds latency and computational overhead that conflicts with the real-time requirements of satellite communication. This project proposes a novel physical-layer authentication (PLA) framework that exploits unique propagation characteristics of satellite-ground links, including Doppler shifts, time-varying delay profiles, and atmospheric effects, to authenticate ground terminals without modifying existing protocols.

The research will develop machine learning models that learn the “fingerprint” of legitimate satellite-ground channels, enabling real-time detection of rogue ground stations or man-in-the-middle attacks. Unlike terrestrial RF fingerprinting, satellite channels present unique challenges: rapid orbital dynamics, ionospheric scintillation, and multi-path fading from terrain reflections. This project will pioneer techniques to extract stable authentication features from these inherently dynamic channels.

Requirements

• Strong programming skills in Python and signal processing libraries (NumPy, SciPy).
• Background in wireless communications and channel modeling.
• Familiarity with machine learning frameworks (PyTorch/TensorFlow).
• Interest in satellite communications and space security.
• Experience with SDR (GNU Radio, USRP) is a plus.

Expected Outcomes

• A novel physical-layer authentication framework for LEO satellite communications.
• Open-source implementation of the ML-based authentication system.
• Comprehensive security analysis against various attack vectors.
• Potential for publication in top security venues (S&P, CCS, NDSS, USENIX).
• Foundation for future work on secure 6G non-terrestrial networks.

References

1. Y. Liu et al., “Physical Layer Security for Next Generation Wireless Networks: Theories, Technologies, and Challenges,” IEEE Communications Surveys & Tutorials, 2016.
2. K. Sankhe et al., “ORACLE: Optimized Radio Classification through Convolutional Neural Networks,” IEEE INFOCOM, 2019.
3. G. Oligeri et al., “PAST-AI: Physical-Layer Authentication of Satellite Transmitters via Deep Learning,” IEEE TIFS, 2022.
4. I. Alla et al., “Robust Device Authentication in Multi-Node Networks: ML-Assisted Hybrid PLA Exploiting Hardware Impairments,” ACSAC, 2024.