The success of machine learning for real-world robotic systems has created a new form of intellectual property: the trained policy. This raises a critical need for novel methods that verify ownership and detect unauthorized, possibly unsafe misuse. While watermarking is established in other domains, physical policies present a unique challenge: remote detection. Existing methods assume access to the robot's internal state, but auditors are often limited to external observations (e.g., video footage). This Physical Observation Gap means the watermark must be detected from signals that are noisy, asynchronous, and filtered by unknown system dynamics.
The policy auditor, who aims to identify the policy used on the robot, can only access glimpses of the policy behavior through remote sensing, such as a camera feed; these glimpses are passed through the detection function to identify the policy. In our experiments, we consider the following glimpses modalities.
Results of our CoNoCo method watermarking a policy trained for navigation task on the real-world RoboMaster platform, using only Remote Motion Capture Glimpses for detection.
Results of our CoNoCo method watermarking a policy trained for Mujoco HalfCheetah control, using Ground-Truth Actions, Noisy Onboard Sensors, and Remote Camera Feed Glimpses modalities for detection.
@misc{amir2025remotelydetectablerobotpolicy,
title={Remotely Detectable Robot Policy Watermarking},
author={Michael Amir and Manon Flageat and Amanda Prorok},
year={2025},
eprint={2512.15379},
archivePrefix={arXiv},
primaryClass={cs.RO},
url={https://arxiv.org/abs/2512.15379},
}