On Board Neuro Fuzzy Inverse Optimal Control for Type 1 Diabetes Mellitus Treatment: In-Silico Testing

Abstract

Type 1 Diabetes Mellitus (T1DM) is one of the most adverse diseases in the modern era; its treatment is mainly based on exogenous insulin injections. The scientific community has formulated strategies to improve insulin supply using state-of-the-art technology. Therefore, this article develops a multi-age glycemic control scheme, which can be implemented in an Artificial Pancreas (AP) device to enhance diabetics treatment. The procedure is based on the implementation of a neuro-fuzzy inverse optimal control (NF-IOC) algorithm on the Texas Instrument LAUNCHXL-F28069M development board; this controller communicates with the Uva/Padova simulator for diabetics' patients of different ages under predefined meal protocols running on a Personal Computer (PC). The novelty lies in the proposed NF-IOC capability to regulate glucose within safe levels for virtual populations of 10 adults, 10 adolescents and, 10 children.

Keywords

• biomedical electronics
• biomedical monitoring
• fuzzy neural networks
• neurocontrollers
• virtual prototyping