Analyte Sensor Error Correction via Training Mode
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional implantable sensors and pumps face challenges in accurately measuring blood analyte levels and delivering medicaments due to unmodeled system variables, such as user-specific and context-dependent factors, which cannot be pre-programmed or adapted effectively, leading to errors in both measurement and delivery.
Innovation Solution
An apparatus and method that includes a data processing system for communicating with both analyte sensors and pumps, utilizing an error correction operational model to correct sensor and pump errors, generated through a training mode that adapts to user-specific conditions, enabling improved accuracy in analyte measurement and medicament delivery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional implantable sensors and pumps are used with pre-programmed parameters, then device complexity is reduced and ease of operation is improved, but measurement precision and delivery accuracy deteriorate due to unmodeled system variables
Solution Approach 1:
The system performs preliminary actions by implementing a training mode before normal operation, where the operational model is generated and adapted to the specific user's physiological characteristics. This preliminary adaptation phase allows the system to learn user-specific parameters and context-dependent variables, which then improve measurement precision and delivery accuracy during subsequent normal operation without requiring complex real-time adjustments.
Solution Approach 2:
The system employs feedback mechanisms by continuously using sensor data to refine and update the operational model during the training phase. The model learns from the feedback provided by actual sensor measurements and pump delivery outcomes, adapting to unmodeled system variables and improving accuracy over time while maintaining manageable device complexity through iterative optimization.
2Reliability
If an error correction operational model with training mode is implemented, then measurement precision and delivery accuracy are improved, but device complexity and ease of operation worsen
Solution Approach 1:
The system implements self-service by enabling the operational model to automatically adapt and refine itself during the training mode without requiring manual intervention or complex user configuration. The model autonomously learns from sensor data and pump delivery outcomes, improving system reliability while maintaining ease of operation, as users simply need to wear the device and follow normal procedures without managing complex parameters.
3Measurement precision
If user-specific and context-dependent variables are accounted for, then measurement precision and delivery accuracy are improved, but device complexity increases due to additional modeling requirements
Solution Approach 1:
The system handles user-specific and context-dependent variables through preliminary action by dedicating a training mode phase to capture and model these variables before normal operation begins. During this training phase, the operational model learns and adapts to the specific user's physiological characteristics and context-dependent patterns, thereby improving measurement precision and delivery accuracy without requiring complex real-time modeling during actual use.
Data Source
AI summary
Apparatus and methods for error modeling and correction in one or both of (i) a partially or fully implanted or non-implanted medicant delivery mechanism (such as a pump), and (ii) implanted physiologic parameter sensor. In one exemplary embodiment, the apparatus and methods employ a training mode of operation, whereby the apparatus conducts “machine learning” to model one or more errors (e.g., unmodeled variable system errors) associated with the medicant dose calculation process, and (ii) generation of a medicant delivery operational model (based at least in part on data collected/received in the training mode), which is applied to correct or compensate for the errors during normal operation of the sensor and pump system. This enhances accuracy of medicant delivery, such as over the lifetime of an implanted pump at a single implantation site, or during multiple relocations of a transcutaneously implanted pump), and enables “personalization” of the pump to each user.