cs.RONov 5, 2026

Dynamic Watermarking and Action-Conditioned Latent Dynamics for Secure Vision-Based Controllers

Authors: Aditya Parameshwaran, Navid Aftabi, Abhishek Hanchate, Satish T. S. Bukkapatnam, Dan Li, Yue Wang

Organizations: Department of Mechanical Engineering, Clemson University, Clemson, SC 29634, USA · Department of Industrial and Systems Engineering, University of Washington, Seattle, WA 98195 USA · Wm Michael Barnes ’64 Department of Industrial and Systems Engineering, Texas A&M University, College Station, TX 77843 USA

Abstract

Replay attacks are difficult to detect in vision-based control systems because previously recorded observations can remain plausible while the physical system follows a different trajectory. This paper develops an adaptive dynamic-watermarking framework for detecting replay attacks against vision-based controllers. A private Gaussian watermark is added to the nominal control input, and a control input-conditioned latent dynamics network (LDN) predicts the corresponding evolution in a compact latent space learned from image observations. Attack evidence is obtained from the discrepancy between the predicted latent response and the latent representation of the received image observation. A reinforcement learning policy adjusts the watermark covariance using the latent state and detector confidence, thereby balancing attack detectability, watermark energy, and control distortion. A Lagrangian penalty on task-specific safety costs is included during training to limit the adverse effect of the excitation. The framework is evaluated on a vision-guided UR5 manipulation task in which the robot reaches a target while avoiding a moving human arm under nominal and replay-attacked observations.

Explore similar work

CardsList
  1. A Watermark for Vision-Language-Action and World Action Models

    Jun 22, 2026Yule Liu, Shuai Liu, Jiaheng Wei +1Language Model FingerprintingVision-Language-Action Models

  2. Sensitivity Shaping for Latent Modeling

    Jun 12, 2026Hongzhan Yu, Chenghao Li, Ruipeng Zhang +2Robotic ControlLatent Dynamics Modeling