| Autors: Landsberg-Stoilova, G. O., Ilieva, R. Y. Title: Simulation-Based Digital Twins for Robotic Cybersecurity Testing with AI-Driven Adaptive Defence Keywords: AI-Augmented Digital Twins, Intelligent Threat Mitigation Abstract: Robotic and mechatronic systems are increasingly connected via open communication protocols and cloud services, exposing safety-critical operations to cyber threats such as spoofing, denial-of-service, and data poisoning. Testing security measures directly on physical robots is costly, disruptive, and often unsafe. This paper proposes a simulation-based digital twin (DT) framework that faithfully replicates a robotic system’s kinematics, dynamics, and network behaviour in a virtual environment, enabling risk-free cybersecurity experimentation. Within this DT, an AI–driven adaptive defence module combines traffic-based anomaly detection with reinforcement learning agents that learn to respond to evolving attacks by reconfiguring control or communication policies. The study highlights digital twins as a practical, low-risk platform for developing and validating next-generation cybersecurity strategies for robotic systems. References Issue
Copyright Institute of Robotics, Bulgarian Academy of Sciences |
Вид: статия в списание, публикация в реферирано издание
