Adversarial hull patches and physical disguise evaluated for USV detection
Original title: When appearance and motion disagree: evaluating class--motion consistency against disguised uncrewed surface vessels
Physical disguise bypasses camera and radar checks while adversarial patches suppress detection in up to 29% of attempts.
The paper evaluates adversarial hull patches and physical disguises for defense systems distinguishing benign craft from uncrewed surface vessels using shore cameras and computer vision. Adversarial hull patches suppress detection in up to 29% of attempts but never cause misclassification as benign. Physical disguise can bypass both a camera check and a radar presence check. A hybrid replay protocol from automatic identification system tracks combines recorded benign motion with direct or off-axis attack run-ins. Closing-rate checks flag 94-95% at 15-20 knots for recreational-labeled vessels attacking via direct run-in within 4 nautical miles.
- Adversarial patches suppress detection in up to 29% of attempts
- Physical disguise bypasses both camera and radar checks
- Closing-rate check flags 94-95% at 15-20 knots
- False-alarm rate 4.9% overall and 12.3% on recreational traffic
From the original article
1 min read at arxiv.orgDefense systems for maritime assets must distinguish approaching benign craft from uncrewed surface vessels (USVs) with sufficient warning to act. Shore cameras using computer vision can separate hostile vessels from benign traffic, but these systems may be attacked using an adversarial hull patch or physical disguise.
By Jon Reifschneider · arxiv.org
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.