Automated Glucose Control for Undeclared Meal Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated blood glucose regulation systems, or artificial pancreases, are ineffective in managing meals not declared by the user, leading to prolonged periods of hyperglycemia due to their reliance on user input, which can result in inadequate insulin administration during undeclared meals.
Innovation Solution
An automated blood glucose regulation system that uses machine learning to detect undeclared meals through statistical data tables generated from user history, estimating meal times and sizes, and activates a specific meal management module to administer insulin based on these estimates, with optional user confirmation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the system relies on user input for meal declaration, then the device complexity is reduced, but the reliability of blood glucose management deteriorates due to undeclared meals
Solution Approach 1:
The system automatically detects and manages undeclared meals through machine learning algorithms that analyze user data history, blood glucose patterns, and sensor data. The meal management module autonomously identifies meal events, estimates meal sizes, and calculates appropriate insulin doses without requiring explicit user declaration, thereby improving reliability while maintaining simplicity.
Solution Approach 2:
The patent replaces the manual user declaration mechanism with an automated detection system using machine learning and pattern recognition. Instead of relying on users to manually input meal information, the system uses algorithms to infer meal events from physiological data and historical patterns, substituting mechanical user action with intelligent automated detection.
2Measurement precision
If the system activates meal management for detected events, then the blood glucose control precision is improved, but the productivity of the system decreases due to additional processing steps
Solution Approach 1:
The system performs preliminary machine learning training during a data collection phase to build personalized meal detection models. This preliminary action enables the system to quickly and accurately detect undeclared meals in real-time without extensive processing during critical glucose management moments, improving both precision and efficiency.
Solution Approach 2:
The system dynamically adjusts detection thresholds and processing parameters based on learned user patterns and current physiological states. By optimizing parameters such as probability thresholds for meal detection and insulin dosing calculations, the system achieves high precision while minimizing unnecessary processing steps and maintaining efficiency.
3Reliability
If the system uses machine learning to detect undeclared meals, then the reliability of meal detection is improved, but the device complexity increases due to additional algorithms and data processing
Solution Approach 1:
The machine learning model serves multiple functions: detecting meal events, estimating meal sizes, identifying meal timing, and predicting glucose responses. This multi-functional approach improves detection reliability while avoiding the need for separate specialized algorithms for each task, thereby limiting the increase in overall system complexity.
Solution Approach 2:
The system uses simplified representations and lookup tables that capture essential meal detection patterns without implementing complex full-scale machine learning models in the implanted device. By using pre-computed probability tables and simplified algorithms that replicate the essence of complex ML behavior, the system achieves high reliability with reduced computational complexity.
Data Source
Figure 1~2
Figure 3~4
AI summary
The present invention relates to an automated system for controlling the blood glucose level, comprising: - a blood glucose sensor (101); - an insulin injection device (103); and - a processing and control unit (105), wherein the processing and control unit is configured to implement a method for managing undeclared meals.