Adaptive controllers for intelligent monitoring

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Medicine
• Diabetes Technology

Background:

• Managing Insulin Dependent Diabetes Mellitus (IDDM) requires precise insulin dosing and regular metabolic control monitoring.
• Current patient self-management often relies on periodic physician evaluations and pre-defined dosage tables.
• There is a need for advanced systems to support out-patients in real-time insulin adjustment.

Purpose of the Study:

• To develop and describe a novel two-level system architecture for assisting out-patients with IDDM.
• To create a system that integrates medical knowledge and patient data for personalized insulin therapy.
• To explore the application of adaptive control in diabetes management for improved metabolic control.

Main Methods:

• A two-level system architecture was designed, comprising a High Level Module (HLM) and a Low Level Module (LLM).
• The HLM assesses insulin protocols using medical knowledge and clinical data.
• The LLM, featuring a Fuzzy Set Controller and an ARX model, adaptively suggests insulin dosages based on blood glucose measurements.

Main Results:

• The system architecture facilitates periodic evaluation of insulin protocol adequacy using patient data.
• The LLM's adaptive controller can be modified by the HLM for tailored patient management.
• Preliminary assessment involved analyzing a dataset of 60 patients.

Conclusions:

• The proposed system offers a framework for intelligent, adaptive insulin therapy management for out-patients with IDDM.
• The telemedicine context allows for remote monitoring and control, enhancing patient accessibility.
• Further implementation and validation are ongoing to refine the system's efficacy in diabetes care.