Edd Salkield, Sebastian Köhler, Simon Birnbach, Richard Baker (University of Oxford). Martin Strohmeier (armasuisse S+T), Ivan Martinovic (University of Oxford) Presenter: Edd Salkield

Data from Earth Observation satellites has become crucial in private enterprises, research applications, and in coordinating national responses to events such as forest fires. These purposes are supported by data derived from a variety of satellites, some of which do not secure the wireless downlink channel effectively. This opens the door for modern adversaries to conduct spoofing attacks by overshadowing the signal with commercially available radio equipment.

In this paper, we assess the vulnerability of current Earth Observation systems to spoofing attacks conducted at the physical layer. The effect of these attacks is amplified since the data is received at dedicated ground stations and distributed to hundreds of downstream systems, which are themselves not designed with security in mind. Specifically, we take NASA’s live forest fire detection system as a case study, and demonstrate that the attacker can achieve arbitrary manipulation of fires in the derived dataset to trigger false emergency responses or mislead crisis analysis. We also assess the attack surface presented by ground station software which implicitly trusts data from the RF port. Against the NASA system we uncover several new vulnerabilities that can be exploited to stealthily deny service.

We conclude with a discussion of physical-layer counter-measures to detect and defend against spoofing, which can be implemented in existing deployments at the ground station.

View More Papers

The “Beatrix” Resurrections: Robust Backdoor Detection via Gram Matrices

Wanlun Ma (Swinburne University of Technology), Derui Wang (CSIRO’s Data61), Ruoxi Sun (The University of Adelaide & CSIRO's Data61), Minhui Xue (CSIRO's Data61), Sheng Wen (Swinburne University of Technology), Yang Xiang (Digital Research & Innovation Capability Platform, Swinburne University of Technology)

Read More

BEAGLE: Forensics of Deep Learning Backdoor Attack for Better...

Siyuan Cheng (Purdue University), Guanhong Tao (Purdue University), Yingqi Liu (Purdue University), Shengwei An (Purdue University), Xiangzhe Xu (Purdue University), Shiwei Feng (Purdue University), Guangyu Shen (Purdue University), Kaiyuan Zhang (Purdue University), Qiuling Xu (Purdue University), Shiqing Ma (Rutgers University), Xiangyu Zhang (Purdue University)

Read More