Sensor-Unspecific Calibration Using Unscented Kalman Filter

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current continuous glucose monitoring systems require frequent external calibration using finger sticks, which are inconvenient, prone to errors, and costly, and lack reliable self-calibration and diagnostics, leading to instability and inaccurate readings.

Innovation Solution

A method using an unscented Kalman filter to calculate an estimated calibration factor based on electrode current and blood glucose values, allowing for real-time calibration and fusion of signals from multiple electrodes to provide a stable and accurate glucose measurement without the need for frequent reference values.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If external calibration using finger sticks is performed frequently, then measurement precision is improved, but ease of operation deteriorates and loss of time increases

Engineering Contradiction:
Improveglucose measurement accuracyVSAvoidcalibration convenience
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The sensor system performs self-calibration using an unscented Kalman filter that automatically processes sensor signals and determines calibration factors without requiring user intervention. The system uses its own measured signals (ISIG, counter voltage, impedance) to autonomously calculate calibration parameters, eliminating the need for manual finger-stick calibrations while maintaining measurement accuracy.

Inventive Principle:
Principle #25Self-service

2Measurement precision

If external calibration using finger sticks is performed frequently, then measurement precision is improved, but loss of time increases

Engineering Contradiction:
Improveglucose measurement accuracyVSAvoidcalibration time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The calibration process operates continuously in the background using the unscented Kalman filter, which constantly updates calibration factors based on incoming sensor data. This continuous automated calibration eliminates the discrete time loss associated with manual finger-stick procedures, as the system maintains accurate calibration without requiring the user to pause and perform separate calibration actions.

Inventive Principle:
Principle #20Continuity of useful action

Solution Approach 2:

The system autonomously performs calibration calculations using its own sensor signals and an embedded unscented Kalman filter algorithm, eliminating the time required for manual blood sampling and meter-based calibration procedures.

Inventive Principle:
Principle #25Self-service

3Measurement precision

If sensor design-specific calibration methods are used, then measurement precision is improved, but device complexity increases

Engineering Contradiction:
Improvesensor calibration accuracyVSAvoidcalibration method complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The unscented Kalman filter implementation provides a universal calibration approach that works across different sensor designs and configurations. The same filter architecture and mathematical framework can be applied to various electrochemical sensor types, eliminating the need for design-specific calibration algorithms and reducing overall system complexity while maintaining precision.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3397161B1Sensor-unspecific calibration methods
Publication Date: 2020.06.24 MEDTRONIC MINIMED INC
  • EP3397161B1 patent drawingFigure 1
  • EP3397161B1 patent drawingFigure 2A
  • EP3397161B1 patent drawingFigure 2B

AI summary

An unscented Kalman filter is used to enable sensor calibration independently of the actual design of the subject sensors. By utilizing an unscented Kalman filter, an underlying calibration methodology is developed that is sensor-unspecific, such that a single calibration methodology and related systems may be used to calibrate various sensors, without the need to re-calculate a calibration factor for each specific sensor, and without the need to design a separate filtering mechanism to compensate for noise. In this way, various calibration inputs can be allowed to change over time without the need to change the codebase on which the calibration methodology otherwise operates. In multi-electrode systems, the methodology may incorporate a fusion algorithm to provide a single, fused sensor glucose value. The fusion algorithm may incorporate, and/or work in conjunction with, Electrochemical Impedance Spectroscopy (EIS) procedures.