GPS underpins many safety- and mission-critical systems, providing positioning, navigation, and timing (PNT) services that are attractive targets for physical-layer attacks. A new arXiv paper, GhostLock, proposes a learning-assisted method for blind GPS signal nullification that does not require victim-side information.
The approach appears to use machine learning to generate nulling signals that suppress GPS reception at a target location, even when the attacker has no knowledge of the victim's receiver configuration. This makes it a potentially powerful tool for jamming or protecting GPS-dependent systems.
The abstract is truncated, so details on the learning architecture and experimental results are not yet available. However, the stated goal is clear: a practical, blind technique for disabling GPS signals without needing to cooperate with or probe the victim side.