Skip to content
NewA calmer, faster ZeroHour

PairWatch spots adversarial satellite rendezvous in public catalogs

Original title: PairWatch: Pair-Aware Anomaly Transformer for Detecting Adversarial Rendezvous in Public Satellite Catalogs

arXiv cs.CRby Young Eun Kwon ()1 min read
AI summary · x-ai/grok-4.7Summaries by Vani AI

Only PairWatch kept one threshold that separated real rendezvous from operational negatives in every bootstrap resample.

PairWatch is a pair-aware Anomaly Transformer for finding adversarial rendezvous and proximity operations in public two-line element catalogs. Its attention prior adds sign-symmetric geometric similarity over relative-motion features. On an open benchmark of 359 Space-Track windows, only PairWatch separated documented events from operational negatives under one threshold, in 100% of bootstrap resamples. Simpler baselines scored higher AUROC by exploiting a synthetic-versus-formation shortcut.

Key facts
  • The benchmark contains 359 windows from public Space-Track data.
  • It includes 11 documented RPO events and 100 synthetic adversarial windows.
  • PairWatch held the separation in 100% of bootstrap resamples.
  • Three neural baselines succeeded in 0-19.5% of resamples.
Analyze this story with Vani AIImpact, who is affected and what defenders should do, in a new chat

From the original article

1 min read at arxiv.org

Documented adversarial rendezvous and proximity operations (RPO) appear in the public TLE catalog, yet existing TLE-based maneuver detectors treat each satellite in isolation - mistaking routine station-keeping for adversarial behavior and missing RPO entirely when the target stays passive.

arXiv cs.CR

By Young Eun Kwon · arxiv.org

Read the full article at arXiv cs.CR

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.