Cortisol dysregulation associated with chronic stress and mood disorders is now mainly treated orally, according to an open-loop control logic without feedback on the patient's physiological state. The present work proposes and simulates a closed-cycle control system for the regulation of plasma cortisol, based on an implantable pharmacological release device of the MEMS type. The dynamics of the plant are described by a two-partment pharmacokinetic model in the form of a state, including the transport delay introduced by the actuator. Starting from open chain analysis (poles, zeros, frequency response), three control schemes are designed and compared: a PID regulator with filtered derivative, a Smith Predictor for dead time compensation and a predictive-anticipatory method (NAPE) that integrates wearable signals (HRV, EDA) to pre-compensate acute events. The simulation results show a marked improvement in the performance of the Smith Predictor compared to the PID (ISE −31%, set-up time −69%) and a safety-oriented NAPE strategy, with early rejection of disturbances and elimination of the undershoot. The work constitutes a proof of concept in simulation, a prerequisite for a future experimental validation.
La disregolazione del cortisolo associata allo stress cronico e ai disturbi dell’umore è oggi trattata prevalentemente per via orale, secondo una logica di controllo ad anello aperto priva di retroazione sullo stato fisiologico del paziente. Il presente lavoro propone e simula un sistema di controllo a ciclo chiuso per la regolazione del cortisolo plasmatico, basato su un dispositivo di rilascio farmacologico impiantabile di tipo MEMS. La dinamica dell’impianto è descritta mediante un modello farmacocinetico a due compartimenti in forma di stato, comprensivo del ritardo di trasporto introdotto dall’attuatore. A partire dall’analisi in catena aperta (poli, zeri, risposta in frequenza), vengono progettati e confrontati tre schemi di controllo: un regolatore PID con derivata filtrata, un Smith Predictor per la compensazione del tempo morto e un metodo predittivo-anticipatorio (NAPE) che integra segnali indossabili (HRV, EDA) per pre-compensare gli eventi acuti. I risultati in simulazione mostrano un netto miglioramento delle prestazioni dello Smith Predictor rispetto al PID (ISE −31%, tempo di assestamento −69%) e una strategia NAPE orientata alla sicurezza, con reiezione anticipata dei disturbi ed eliminazione dell’undershoot. Il lavoro costituisce una proof of concept in simulazione, presupposto per una futura validazione sperimentale.
Tecnologie MEMS per la salute mentale: analisi e simulazione di un controllo closed-loop del cortisolo.
STOCCO, FEDERICA
2025/2026
Abstract
Cortisol dysregulation associated with chronic stress and mood disorders is now mainly treated orally, according to an open-loop control logic without feedback on the patient's physiological state. The present work proposes and simulates a closed-cycle control system for the regulation of plasma cortisol, based on an implantable pharmacological release device of the MEMS type. The dynamics of the plant are described by a two-partment pharmacokinetic model in the form of a state, including the transport delay introduced by the actuator. Starting from open chain analysis (poles, zeros, frequency response), three control schemes are designed and compared: a PID regulator with filtered derivative, a Smith Predictor for dead time compensation and a predictive-anticipatory method (NAPE) that integrates wearable signals (HRV, EDA) to pre-compensate acute events. The simulation results show a marked improvement in the performance of the Smith Predictor compared to the PID (ISE −31%, set-up time −69%) and a safety-oriented NAPE strategy, with early rejection of disturbances and elimination of the undershoot. The work constitutes a proof of concept in simulation, a prerequisite for a future experimental validation.| File | Dimensione | Formato | |
|---|---|---|---|
|
STOCCO FEDERICA 2110220 PPT.pdf
accesso aperto
Dimensione
1.2 MB
Formato
Adobe PDF
|
1.2 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/115071