Automatic Glycemia Regulation: Avoiding Hypoglycemia and Hyperglycemia

Driving Blood Glucose Concentration (BGC) from hyperglycemia to euglycemia as quickly as possible while preventing hypoglycemia has been a big challenge for Type-1 Diabetes (T1D) for decades. This study has contributed to handle this problem. In this study, a control algorithm is established that ensures the full avoidance of both hypoglycemia and hyperglycemia. The solution consists of a positive control law that finds the necessary hyperglycemia correction bolus in real-time to safely drive glycemia to the normal range. The action of this controller depends mainly on the amount of the pre-estimated meal. Where the estimated meal forms part of the control law. The other part is a state feedback control. The positive controller steered BGC to ensure avoiding both hyper. and hypoglycemia. This control is examined in Magdalene T1D simulator patients during a fasting-night scenario as well as a postprandial scenario. The control algorithm was tested in silico on 5 different T1DM patients in postprandial of three unknown's meal scenarios. The in silico results show that glycemia remained on average for the five T1DM patients. The percentage of glycemia in a normal range about (100 %) percent over the simulation time. No hypoglycemia where registered and time spent in hyperglycemia over the 24-hour simulation for the 5 patients was very limited. The controller showed high performance avoiding both hypoglycemia and hyperglycemia.