Virginia Tech® home

Hybrid Knowledge and Data Driven Synthesis of Runtime Monitors for Cyber-Physical Systems

Paper Details


Advances in sensing and computing technology have led to the proliferation of Cyber-Physical Systems (CPS) in safety-critical domains. 

However, increasing device complexity, shrinking technology sizes, and shorter time to market have challenged CPS reliability, safety, and security. 

Researchers developed a hybrid knowledge and data-driven approach to designing run-time context-aware safety monitors that can detect and mitigate hazards, closing the gap between design-time hazard analysis and modeling and run-time verification. 

Two project simulations used an autonomous driving system and closed-loop artificial pancreas systems for treatment of diabetes. The results showed that a safety monitor developed with the proposed approaches increases the average prediction accuracy by up to 4.7 times over several baseline monitors while reducing both false-positive and false-negative rates in most scenarios.

By combining formal specification of domain knowledge with learning from closed-loop CPS data, the proposed method enables the early detection and mitigation of safety violations in CPS controllers with improved accuracy and transparency.