Researchers have developed a new framework to detect cyber-physical attacks by monitoring temporal inconsistencies in digital twins. This method trains a digital twin predictor on normal system behavior and then analyzes discrepancies between predicted and observed states. These inconsistencies are transformed into temporal features that capture the evolution of prediction residuals, which are then used by an unsupervised density model and a sequential change detection mechanism to identify deviations indicative of attacks. The framework has been evaluated on industrial control system datasets, demonstrating high detection reliability and low false alarm rates, even when the digital twin view is degraded. AI
IMPACT This research could enhance the security of industrial control systems by leveraging digital twins to detect sophisticated cyber-physical attacks.
RANK_REASON Research paper published on arXiv detailing a new detection framework for cyber-physical attacks. [lever_c_demoted from research: ic=1 ai=1.0]
- Båtadalshaugen
- Cyber-Physical Systems
- digital twin
- industrial control system
- Institute for Human-Centered Artificial Intelligence
- Konstantinos Evangelos Kampourakis
- Swat Industrial Control System Testbed
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
