Advancements in Fault Tolerance and Redundancy in CGM Devices

Improving accuracy and fault detection of glucose sensors using secondary physiological measurements. The method involves cross-referencing glucose level readings from a glucose sensor with secondary physiological measurements like lactate, ketone, or heart rate. If both metrics indicate a low glucose condition, it confirms true hypoglycemia. If the secondary metric doesn't support low glucose, it indicates a false low. This allows more aggressive lag correction and fault detection. It also enables discerning true vs false low glucose conditions to improve overall low-end accuracy.

Closed-loop insulin infusion systems using orthogonally redundant glucose sensors for improved accuracy and reliability. The system has two glucose sensors, one optical and one electrochemical, to provide orthogonal redundancy. An algorithm combines the sensor data to improve accuracy and reliability. If one sensor fails, the other can provide glucose values. The sensors have features like distributed electrodes and membrane barriers to reduce drift and fouling. The system uses on-demand calibration rather than frequent fingersticks.

Redundant glucose sensor system for diabetes management that provides continuous monitoring when one sensor is a warmup, stabilization, or end of life. Multiple glucose sensors attach to a patient simultaneously and communicate with each other. If a sensor is in a non-operational state, another sensor provides glucose data. This ensures continuous monitoring and reduces the risk of missed readings. The sensors can be separate devices or integrated into insulin pumps. The redundant sensors can also be used in a dual pump configuration for improved insulin delivery reliability.

Detection of sensor faults in biomedical devices like glucose monitors to improve accuracy and prevent misdiagnosis. The method involves calculating analyte metrics like rate of change and sum of levels from sensor data. When these metrics exceed predetermined thresholds, it indicates a suspected sensor fault. This allows early detection of issues like moisture ingress that can affect sensor performance.

Mitigating single-point failures in medical devices like glucose monitors by having them periodically check their components and report the results to other devices. If a component fails, an alarm is triggered. This allows redundant monitoring and detection of device failures.

Fault discrimination and responsive processing in continuous glucose monitoring systems that take into account clinical context to improve accuracy and user experience. The method involves detecting faults in the sensor signal and then discriminating the type of fault based on the signal and clinical context like age, activity level, drugs, etc. Appropriate responsive processing is then performed based on the fault type and clinical context. This allows targeted actions like filtering, alerts, or recalibration based on the nature of the fault and patient context.

Analyzing the reliability of a glucose sensor signal to determine if the sensor is still accurately measuring blood glucose levels over time. The reliability is assessed by detecting changes in sensor responsiveness based on metrics like dispersion of rate change in measurements and mean values over time intervals. Decreases in sensitivity indicate sensor degradation. This allows proactive replacement of sensors before they become unreliable.

Detecting faults in glucose monitoring systems without requiring user blood samples. The method involves analyzing glucose sensor data over time to identify faults like the end of sensor life or faulty calibration. It compares sensor readings taken at different times to determine if the sensor is providing data that is physiologically implausible or outside normal ranges. The method uses metrics like median glucose and variability to identify faults based on sensor data trends.

Real-time detection of declining sensitivity in continuous glucose monitoring (CGM) sensors to accurately identify when CGM sensors are losing sensitivity. It involves analyzing CGM current signals and comparing them to blood glucose levels to estimate the sensitivity ratio. If the estimated sensitivity ratio falls below a threshold, it indicates declining sensor sensitivity. A statistical analysis using a single blood glucose measurement confirms the sensitivity decline. This two-step process reduces false positives. The decline is verified by comparing the estimated sensitivity to a threshold based on the blood glucose.

Reducing false hypoglycemic alarms from continuous glucose monitoring (CGM) systems in diabetic patients. The technique involves delaying the hypoglycemic alarm response based on the user's glucose range and CGM signal artifact characteristics. When the user's glucose is mostly normal and the CGM signals have brief dropouts, the hypoglycemic threshold is increased to avoid false alarms due to artifacts. This allows longer delay before sounding an alarm, but still alerts in time for safe response to true hypoglycemic events.

Monitoring the reliability of a glucose sensor to determine when it is no longer accurate enough to trust for closed-loop glucose control in a medical device like an insulin pump. The reliability metric is based on trends in the sensor data, such as reduced sensitivity, anomalies, drift, or noise. When the metric indicates the sensor is unreliable, it triggers a transition to manual or open-loop operation instead of relying on the sensor for glucose control.

Closed loop insulin delivery system that uses redundant glucose sensors to improve reliability and fault detection. The system has two glucose sensors at different sites. It corroborates the sensor readings by predicting one sensor's value using the other sensor's reading and then comparing the predicted and actual values. If the sum of errors exceeds a threshold, it indicates a sensor failure. The system then suspends closed-loop operation and initiates sensor replacement. If one sensor has less error, it uses that sensor's value for insulin delivery. This allows continued partial closed-loop operation with sensor redundancy until replacement.

Automated insulin delivery system for diabetes management that integrates continuous glucose monitoring and closed-loop insulin delivery. The system uses a glucose sensor, insulin pump, and controller to automate insulin delivery based on real-time glucose levels. It adjusts basal and bolus insulin rates, suspends delivery during bolus, and resumes basal. It also has safety features to mitigate calibration errors, prevent hypoglycemia, and suggest carbohydrate intake. The system aims to improve diabetes management, especially during sleep, by automating insulin delivery from glucose monitoring.