Event Detection and Control of Blood Glucose Levels Using Deep Neural Network

Diabetes Mellitus is the most prevalent chronic illness worldwide that occurs because of the changes in the blood glucose regulation of the human body. Diabetes is also leading to many short-term and long-term complications if it is not managed properly. Therefore, monitoring and control of glucose level plays a vital part in diabetic research. This paper deals with the classification of glucose levels of the patients based on their clinical data and design of an event-triggered controller for type-1 diabetic patients. The main aim of this paper is to provide tight control of glucose level in the patient to avoid the complications of hypoglycaemia and hyperglycemia. Moreover, designing of control scheme for post-operative diabetic patients is a challenging task due to large fluctuations in glucose levels. In this paper, the classification of glucose levels is implemented using supervised and deep neural network model. After classification, the glucose-insulin regulatory system is modelled for hypoglycaemia and hyperglycaemia patients using Bergman Minimal Model (BMM). Based on the event detected the controller will take necessary control action using PI Controller. Further, Performance analysis is carried out quantitively and qualitatively and the result shows the feasibility of the proposed control scheme.