2026/07/01 by B. Ayvaz, E. Yumuk, D. Copot +2
Engineering · Medicine · #Anesthesia and Sedative Agents #Healthcare Technology and Patient Monitoring #Non-Invasive Vital Sign Monitoring
paper · doi:10.1016/j.ejcon.2026.101578
openalex publication_date 2026/07/01 · openalex created_date 2026/07/24 · openalex updated_date 2026/07/28
Anesthesia digital twin (DT) frameworks rarely capture sensor malfunctions or signal quality loss, limiting their capability to reflect reality and assess safety. This paper presents two main contributions addressing this limitation. First, a Hidden Markov Model (HMM)-based non-parametric statistical model of the BIS monitor’s Signal Quality Index (SQI) is developed using data from 5861 patients, enabling realistic representation of SQI behavior and sensor failures in anesthesia DTs. Second, a Control Barrier Function (CBF)-based Safety Filter (SF) is proposed to ensure safety in proportional-integral-derivative (PID) controlled closed-loop anesthesia under degraded SQI conditions. SQI model validation demonstrated close similarity between real and synthetic SQI data, with multiple numerical analyses indicating negligible differences in the statistical behavior. For the safety framework, comparative results showed that the Safety Filter improved safety by increasing the percentage of time within the adequate BIS range by 8.57%, decreasing the root-mean-squared violation error by 72.69%, while causing a 10.1% reduction in performance indices, highlighting the expected safety–performance trade-off under signal quality loss events. The proposed framework enhances the safety of existing anesthesia control systems while improving the realism of anesthesia digital twins through the integration of an HMM-based SQI model, enabling reliable operation under realistic signal degradation and sensor fault conditions.