ROBOT AUTHENTICATION USING ENCRYPTED VISUAL TAGS: A SECURE HUMAN-ROBOT INTERACTION APPROACH

Abstract

The rapid growth of robotic applications in sensitive human-centric environments demands secure and verifiable human-robot interaction (HRI). This paper presents a unified authentication framework using encrypted visual tags enhanced with post-quantum cryptographic techniques. The proposed method uses session-specific visual markers (QR or ArUco tags) that encode encrypted metadata, signed with the Dilithium algorithm and encrypted with ChaCha20-Poly1305. A Kyber-based key exchange protocol enables secure symmetric encryption between the user and robot in real time. The robot captures the visual tag via its onboard camera, verifies its authenticity, and performs controlled actions based on secure user identity. We implement and test the framework in simulated and physical robotic settings using Python, OpenCV, and ROS2. The system achieves an authentication success rate of 98.7%, 100% replay attack detection, and operates with an average latency of 30.9 ms, demonstrating high security, speed, and robustness against post-quantum threats.