GLUCOSE FORECASTING USING A PHYSIOLOGICAL MODEL AND STATE ESTIMATION

Accurate glucose forecasting algorithms have been proven to be an effective solution for reducing the risk of hypo- and hyperglycaemia events when combined with glucose alarms and/or low-insulin suspension systems. • Effective glucose forecasting algorithm can be applied to control insulin delivery in automatic systems. • Daily routine information (e.g. meal intake, insulin injection, physical exercises) of patient can be included into forecasting algorithms to improve prediction accuracy. • In this work, we introduce a novel model-based glucose prediction algorithm which uses deconvolution of the continuous glucose monitoring (CGM) signal to estimate some of the model states in order to improve prediction accuracy.