Background Two widely used artificial pancreas (AP) control algorithms are the model predictive control (MPC) and the proportional integral derivative (PID) algorithms. Numerous studies across different settings have used both algorithms with positive results, but there has never been a randomized clinical trial directly comparing the effectiveness of each. This study aimed to compare individual-personalized MPC and PID controls under nonideal but comparable clinical conditions. Methods After a pilot safety and feasibility study (n= 10), closed-loop control (CLC) was conducted and evaluated in a randomized, crossover trial that included 20 additional adults with type 1 diabetes. Both the MPC and PID algorithms were compared during supervised 27.5 hour CLC sessions. The algorithms were tested by evaluating control performance following a 65 g dinner, 50 g breakfast, and unannounced...
Closing the Loop
S Del Favero;F Boscari;R Visentin;R Calore;S Galasso;A Galderisi;V Vallone;F Di Palma;E Losiouk;A Avogaro;D Chernavvsky;L Magni;C Cobelli;D Bruttomesso;D Bruttomesso;R Visentin;R Calore;F Di Palma;P Magni;F Boscari;S Galasso;A Avogaro;S Del Favero;C Cobelli;L Magni;S Del Favero;R Visentin;M Monaro;F Di Palma;F Boscari;S Galasso;P Magni;A Avogaro;D Bruttomesso;C Cobelli;L Magni;
2017
Abstract
Background Two widely used artificial pancreas (AP) control algorithms are the model predictive control (MPC) and the proportional integral derivative (PID) algorithms. Numerous studies across different settings have used both algorithms with positive results, but there has never been a randomized clinical trial directly comparing the effectiveness of each. This study aimed to compare individual-personalized MPC and PID controls under nonideal but comparable clinical conditions. Methods After a pilot safety and feasibility study (n= 10), closed-loop control (CLC) was conducted and evaluated in a randomized, crossover trial that included 20 additional adults with type 1 diabetes. Both the MPC and PID algorithms were compared during supervised 27.5 hour CLC sessions. The algorithms were tested by evaluating control performance following a 65 g dinner, 50 g breakfast, and unannounced...Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
