Generalized predictive control closed-loop insulin infusion system based on adaptive reference curve strategy

Abstract

The invention discloses a generalized predictive control closed-loop insulin infusion system based on an adaptive reference curve strategy. The system comprises a blood glucose detection module, a controller, a program module and an insulin pump, the blood glucose detection module is used for acquiring the blood glucose of a user, and an output end of the blood glucose detection module is connected with an input end of the controller through signals. The insulin pump is used for infusing insulin into the user, and an input end of the insulin pump is connected with an output end of the controller through signals. The program module is executed by the controller, acquires the blood glucose of the user through the blood glucose detection module and adjusts the infusion rate of the insulin pump according to a CARIMA (controlled autoregressive integral moving average) model, a minimum variance control algorithm and the adaptive reference curve strategy. According to the generalized predictive control closed-loop insulin infusion system based on the adaptive reference curve strategy, adaptive reference curves can be generated according to blood glucose fluctuation of different patients,so that the infusion rate of the insulin pump can be flexibly controlled.

Description

technical field [0001] The invention relates to the field of insulin pump infusion, and more specifically, to a generalized predictive control closed-loop insulin infusion system based on an adaptive reference curve strategy. Background technique [0002] The existing control algorithms of insulin infusion devices include proportional calculus control algorithms, fuzzy control algorithms, model control algorithms and so on. [0003] Proportional Integral Derivative (PID) control is a relatively mature and widely used industrial control algorithm. The design of the insulin pump based on the PID algorithm has been realized by the Steil team of Medtronic MiniMed. Its main design idea is: the three components of the PID algorithm, the proportional, differential and integral components, respectively simulate the physiological transmission process of insulin secreted by human β cells. Among them, the linear proportional component corresponds to the insulin secretion when the meas...

AI Technical Summary

Problems solved by technology

However, its disadvantage is that the selection and construction of the model is often based on the physiological model of the patient's glucose metabolism, and the parameters need to be set with reference to clinical medical experience.

Although the model itself often has the ability of rolling optimization, once the model construction deviates, it is often difficult to achieve the expected control effect

Images