Title: SaTC: Core: Small: Building Resilience into LEO Satellite Networks by Exploiting Network Layer Characteristics, funded by NSF under grant CNS-2308761, 10/01/2023 – 09/30/2026

Project Summary:

Being a critical infrastructure of our society, the security of low Earth orbit (LEO) satellite communication (SATCOM) networks has a high stake. Because these systems typically implement their own proprietary security protocols that are not open to the public, it has long been believed that they are secure. However, the recent cyber attack on the Viasat satellite network, which happened on the first day of Russia’s invasion of Ukraine and caused widespread disruption of high-speed satellite services throughout Ukraine and other European countries, has revealed the astonishing fact that these SATCOM networks are not really secure, and the scale and sophistication of cyber attacks on these systems are far above our previous thoughts. As noted in a recent joint cybersecurity advisory by CISA and FBI, attacks to satellite networks could be “state sponsored”, which has the potential to mobilize almost unlimited computing and intelligence resources for the attack. Such a resource-unconstrained attack may be able to crack an existing crypto system in much shorter time, and hence seriously challenges the validity of existing computational-security-based security measures, which are developed based on the common-sense assumption that attackers are resource-constrained – a serious under-assumption when it comes to high-stake critical infrastructure targets such as a satellite network. Realizing the threats from this new attack model, the overarching goal of this project is to tap into the intrinsic security resources possessed by the LEO SATCOM at its network layer, such as rich and dense satellite connectivity and high satellite mobility, so as to develop a suite of physical-constrained, rather than computation/math-constrained, security methods that build LEO satellite network's resilience to resource-unconstrained attacks.

Project Goals:

This project has the following three goals:

1. Collect empirical real-world LEO satellite orbit data from publically accessible websites such as N2YO.com. Perform statistical analysis on these data to create models that characterize the spatial distribution, mobility, and connectivity of realistic LEO satellite constellations such as StarLink and OneWeb.

2. Exploit the global spatial distribution of LEO satellites and their dense connectivity and high mobility to develop physical-constrained security methods to protect real-world LEO satellite networks against resource-unconstrained cyber attacks.

3. Develop machine-learning and cryptographic mechanisms to detect compromised LEO satellites within the constellation when there are any.

Project Personnel

PI: Tao Shu, Ph.D.

Graduate Students

· Xueyang Hu

· Hairuo Xu

· Guan Huang

· Minarul Islam

Project Activities and Results

1. Creating real-world LEO satellite orbit dataset by developing a N2YO crawler

In this research the PI and his team proposed a novel web crawler, called N2YO-crawler, to collect real-time satellite orbit tracking data from N2YO.com, which provides live real-time trajectory tracking data for over 28,751 objects in outer space, including satellites and space debris. The data for LEO satellites, such as the 5,814 StartLink satellites and the 628 OneWeb satellites, are updated every second on this website. N2YO.com is a highly reputable provider of real-time satellite tracking services. The real-time data provided by N2YO.com are diligently collected by the US Space Surveillance Network (SSN), which operates under the rigorous oversight of the US Air Force Space Command (AFSPC). However, it should be noted that while N2YO.com is accessible in the public domain, the dataset collected by SSN is not open to the public, which hinders the empirical data-driven LEO satellite orbit research in our research community.

In contrast to existing open-source web crawlers, which often struggle to keep up with the rapid content updates of high-frequency websites and may compromise data integrity when handling large volumes of information, our N2YO crawler is adept at orchestrating massive information volumes without compromising data integrity and demonstrates an extraordinary capability to capture dynamic data with one-second resolution.

2. Developing a novel Transformer-based LEO satellite orbit prediction algorithm and training it over the real-world LEO satellite orbit dataset to obtain better prediction accuracy

In recent years, the proliferation of LEO (Low-Earth Orbit) satellites and the accumulation of space debris have made near-Earth space more and more crowded, and hence significantly increased the risk of collisions in this space. As a result, precise orbit prediction becomes essential for LEO satellites to avoid collision, maintain the right constellation, and retain stable communication between LEO satellites.

Conventionally, satellite orbit prediction approaches can be divided into two distinct categories: physics-based approaches and machine learning (ML)-based approaches. On one hand, physics-based methods are based on the classical laws of motion and principles of gravity, enabling the prediction of satellite trajectories by solving differential equations of motion, either analytically or numerically. However, these methods cannot consistently yield high-precision orbit predictions for LEO satellites due to the difficulty in taking into account a large number of trajectory-perturbing random factors, e.g., environmental influences such as atmospheric drag and forces of solar radiation, as well as particular satellite attributes, including mass, shape, and information regarding maneuvers. Consequently, the prediction errors can be substantial (e.g., in the order of hundreds of kilometers), potentially rendering them less effective for practical applications. On the other hand, ML-based methods have been used to complement those physics-based models by taking into account those aforementioned random factors. However, due to the absence of publically-available real-world LEO satellite orbit dataset, most existing ML-based methods are trained on simulated/synthetic orbit data, which essentially assumes a stationary orbit process and cannot reflect the non-stationary dynamic orbit changes in real-world LEO satellite constellations that are caused by satellite flight status adjustment (e.g., for the purpose of collision avoidance).

In this research, we propose a novel multi-range (global-local) self- attention transformer-based ML model, the GloLoSAT, and train the model over real-world LEO satellite orbit data crawled from N2YO.com to give more precise orbit prediction in case of a series of orbit adjustments. To overcome the limitations of existing transformer-based LSTF models that often focus solely on either global or local context, and thus lacking a comprehensive view of the input sequence, we introduce an innovative Global-Local Probsparse Self-Attention (GloLoSAT) mechanism. In particular, the global Probsparse self-attention module in this mechanism analyzes the entire input sequence to capture long-range dependencies, while local Probsparse self-attention modules provide a detailed analysis of each evenly segmented portion of the sequence to capture finer-grained details, collectively enriching the model with a multi-faceted perspective. Theoretical analysis demonstrates that our Global-Local Probsparse Self-Attention mechanism can achieve (𝐿 log(𝐿)) computational complexity with respect to the input sequence length. Extensive experiments conducted on real satellite orbit tracking datasets demonstrate the efficacy of GloLoSAT in achieving consistent performance improvements across various prediction scenarios compared to the counterparts.

3. Spoofing detection for LiDAR

In recent years the LiDAR (light detection and ranging) sensor has been adopted extensively by various IoT smart sensing applications, such as autonomous vehicles, UAVs, remote sensing, and LEO satellites. Unfortunately, LiDARs are susceptible to malicious spoofing attacks, which can lead to falsified LiDAR sensing outcomes. Most current work focuses on protecting LiDAR against spoofing attacks by using perception model-level defense methods, whose effectiveness unfortunately depends on the correctness of the LiDAR’s raw sensing outcome. A spoofer thus can elude from these methods as long as it fabricates points in the LiDAR sensing point cloud that maintain the right contextual relationship with other (legitimate) points in the point cloud. In this research, we propose to use the LiDAR signal’s Doppler frequency shift to verify the sender of the LiDAR signal and detect potential spoofing attacks. In particular, we first thoroughly analyze the working principle of LiDAR and conduct real-world experiments to deeply understand and reveal the vulnerability of LiDAR sensors. We then prove that the Doppler frequency shifts of legitimate and spoofing signals present different characteristics, which can be used to fundamentally authenticate the LiDAR sensing outcome. We then consider three different attack models, including static attacker, moving attacker, and moving attacker with control of both velocity and signal frequency. For each of the models, we first show how the spoofing attack is performed and then present our countermeasures. We also propose a statistical spoofing detection framework to jointly consider the impact of short-term uncertainty in vehicle velocity, which can provide more accurate spoofing detection results in realistic environments. Extensive numerical results are provided to verify the effectiveness of our proposed spoofing detection methods in a wide range of mobility and environment settings.

Broader Impacts

LEO satellite networks are critical infrastructures that have a huge economic and social impact. Because the stake is so high, the adversary's motivation to compromise these systems is also strong, as evidenced by the excessive attacks, some state-sponsored, received by these systems. The proposed research contributes to protecting the growing social and economic benefit of this critical infrastructure while ensuring its security. The proposed research will develop physical-constrained, rather than math-constrained, security measures under the new assumption of resource-unconstrained attacks, hence making contribution to the scientific foundation of cyber security. This project also carries out a comprehensive education plan to broaden its impacts, including research integration with curriculum development, recruitment and training of student researchers, and dissemination and outreach to the community, especially to minority and under-represented groups.