Online Monitoring of the Artificial Pancreas

LUIS AVILA; ERNESTO C. MARTINEZ

Lasi

Conferencia; EHB 2013 IEEE E-Healt and Bioengineering Conference; 2013

IEEE

The integration of continuous glucose monitoring and subcutaneous insulin delivery technologies with novel control strategies, has brought closer the development of the Artificial Pancreas (AP). However, due to its condition of a safetycritical system makes online control performance monitoring of paramount importance for safe, yet optimal operation. Just a few works have reflected the importance and feasibility of implementing online performance monitoring in the face of unplanned events and abnormal glycemic variability. Here, control performance monitoring is implemented based on a Bayesian surprise index to detect functional degradation and guarantee adequate control of glycemic levels. A glucoregulatory system facing the effect of faults and abnormal glycemic variability is evaluated against an optimal specified behavior under uncertainty. Timely detection of performance degradation highlights the sensitivity of the proposed index to fast detect deviations from a specified system behavior.