Research On Glycemic Dynamic Control For Type 1 Diabetes Based On GPC Method

Nowadays,most of type 1 diabetes(T1D)control their glycemic dynamic by injecting insulin several times everyday,where the adjustment of insulin is the key part.The mis-injection of insulin can lead to serious hyperglycemia and hypoglycemia which have big influence on patients’ life.For the purpose of realizing automatic control of glycemic dynamic for T1D,glycemic dynamic control method for T1D based on generalized predictive control(GPC)algorithm is researched.GPC algorithm which contains the advantages of predictive control and adaptive control can overcome the difficults caused by time delays in the system and has high robustness.The predictive model in GPC algorithm is researched and in order to identify the parameters in multiple inputs models where the variation of model parameters are different,a modified multiple forgetting schedules recursive least square(MMFS-RLS)system parameters identification method is proposed.For improving identification accuracy and robustness of the algorithm,a forgetting matric is designed and the predictive error is used to update the forgetting factor.Aiming at providing reasonable,reliable and automatic control of glycemic dynamic for T1D,a glycemic control method based on GPC algorithm is proposed.The proposed algorithm contains two parts:meal insulin compensation strategy and a feedback controller based on GPC algorithm.The meal insulin compensation strategy works according to the meal detection information of patients and offers compensation of insulin for meals.Well,the feedback controller is used as the main controller to manage the glycemic dynamic of T1D.The proposed glycemic dynamic control method is tested through the UVa/Padova simulator,and compared with the GPC controller which is once used in glycemic dynamic control,to verify the validity of the proposed method.The proposed method lay a firm fundamental for further study in the control of glycemic dynamic for type 1 diabetes,having great theoretical significance and application value.