System and method of screening biosensors

1. A method for testing a biosensor, the method comprising:

connecting to a biosensor to a biosensor screening test apparatus;

conducting a biosensor screening test using the biosensor screening apparatus on the biosensor connected to the biosensor screening test apparatus;

collecting data outputs from the biosensor after completing the biosensor screening test;

processing the collected data outputs, wherein the processing the collected data outputs includes:

determining whether the biosensor screening test is complete;

determining data attributes from the data outputs of the biosensor screening test; and

scaling the determined data attributes; and

determining a screening test score.

2. The method of claim 1 , wherein the biosensor screening test includes:

powering on the biosensor;

programming of test firmware;

initializing or rebooting the biosensor with the test firmware;

enabling a test mode operation of the test firmware;

applying a first average load current to the biosensor for a predetermined duration;

applying a test load current to the biosensor for a predetermined duration;

applying a second average load current to the biosensor for a predetermined duration;

disabling the test mode operation of the test firmware; and

restoring regular firmware for normal operation of the biosensor.

3. The method of claim 1 , wherein the determined data attributes include: an average baseline voltage, a rate of change of voltage during at least one of a stimulus and stimuli of the biosensor screening test, minimum voltage during the biosensor screening test, maximal change in voltage during the biosensor screening test, rate of change of voltage during at least one of a post-stimulus and post-stimuli of the biosensor screening test, and average recovery voltage.

4. The method of claim 1 , wherein the determining the screening test score includes applying a prediction model to the determined data attributes to determine the screening test score as a numerical or categorical value.

5. The method of claim 4 , wherein the applying the prediction model includes training the prediction model using an updated database of determined data attributes, and the training the prediction model includes:

obtaining and updating a database of determined data attributes from the biosensor screening test;

scaling the determined data attributes for scaled data attributes;

training the prediction model using at least one of:

an unsupervised model to learn to predict two or more groups of biosensors including anomalous non-anomalous biosensor groups using the scaled data attributes of a training database;

a supervised classifier model to learn to predict two or more classes of biosensors using the scaled data attributes of the training database; and

a semi-supervised classifier model to learn to predict two or more classes of biosensors using the scaled data attributes of the training database.

6. The method of claim 5 , wherein the obtaining and updating the database of determined data attributes for training of the prediction model includes:

conducting biosensor screening tests;

deploying the biosensors in field use and obtaining field observations from field use;

identifying data attributes of each biosensor with a group label according to the field observations; and

creating a set of two or more unique groups biosensor data attributes per the group label.

7. The method of claim 1 , wherein the screening testing score includes a predetermined set of numerical/categorical values.

8. The method of claim 1 , further comprising outputting an intermediate biosensor screening test score if the biosensor screening test is determined to be not complete.

9. The method of claim 5 , wherein the training the prediction model using the updated database of determined data attributes includes at least one of:

fitting a multivariate gaussian model to determine an anomalous biosensor group with failure risk from a normal biosensor group;

determining a clustering model that identifies a predetermined number of biosensor groups and respective centroids of biosensor cluster groups, and training a classifier model to predict the identified cluster group labels; and

training a classifier model to predict biosensor group labels identified according to the field observations or group labels.

10. A computer-readable medium storing executable instructions that, upon execution, cause a digital computing processor to test a biosensor by performing functions comprising:

conducting a biosensor screening test using the biosensor screening apparatus on the biosensor connected to the biosensor screening test apparatus;

collecting data outputs from the biosensor after completing the biosensor screening test;

processing the collected data outputs;

determining a screening test score; and

outputting an intermediate biosensor screening test score, wherein the intermediate biosensor screening test score includes a predetermined set of numerical/categorical values if the biosensor screening test is determined to be not complete.

11. The computer readable medium of claim 10 , wherein the biosensor screening test includes:

powering on the biosensor;

programming of test firmware;

initializing or rebooting the biosensor with the test firmware;

enabling a test mode operation of the test firmware;

applying a first average load current to the biosensor for a predetermined duration;

applying a test load current to the biosensor for a predetermined duration;

applying a second average load current to the biosensor for a predetermined duration;

disabling the test mode operation of the test firmware; and

restoring regular firmware for normal operation of the biosensor.

12. The computer-readable medium of claim 10 , wherein the processing the collected data outputs includes:

determining whether the biosensor screening test is complete;

determining data attributes from the data outputs of the biosensor screening test; and

scaling the determined data attributes.

13. The computer-readable medium of claim 12 , wherein the determined data attributes includes: an average baseline voltage, a rate of change of voltage during at least one of a stimulus and stimuli of the biosensor screening test, minimum voltage during the biosensor screening test, maximal change in voltage during the biosensor screening test, rate of change of voltage during at least one of a post-stimulus and post-stimuli of the biosensor screening test, and average recovery voltage.

14. The computer-readable medium of claim 12 , wherein the determining the screening test score includes applying a prediction model to the determined data attributes to determine the screening test score as a numerical or categorical value.

15. The computer-readable medium of claim 14 , wherein the applying the prediction model includes training the prediction mode using an updated database, and the training the prediction model includes:

obtaining and updating a database of determined data attributes from the biosensor screening tests;

scaling the determined data attributes for scaled data attributes;

training the prediction model using at least one of:

an unsupervised model to learn to predict two or more groups of biosensors including anomalous non-anomalous biosensor groups using the scaled data attributes of a training database;

a supervised classifier model to learn to predict two or more classes of biosensors using the scaled data attributes of the training database; and

a semi-supervised classifier model to learn to predict two or more classes of biosensors using the scaled data attributes of the training database.

16. The computer readable medium of claim 15 , wherein the obtaining and updating the database of determined data attributes for training of prediction model includes:

conducting biosensor screening tests;

deploying the biosensors in field use and obtaining field observations from field use;

identifying data attributes of each biosensor with a group label according to the field observations; and

creating a set of two or more unique groups biosensor data attributes per the group label.

17. The computer-readable medium of claim 15 , wherein the training the prediction model using the updated database of determined data attributes includes at least one of:

fitting a multivariate gaussian model to determine an anomalous biosensor group with failure risk from a normal biosensor group;

determining a clustering model that identifies predetermined number of biosensor groups and respective centroids of biosensor cluster groups, and training a classifier model to predict the identified cluster group labels; and

training a classifier model to predict biosensor group labels identified according to the field observations or group labels.

18. A biosensor screening test apparatus comprising:

a connector for connecting to a biosensor to the biosensor screening test apparatus;

a memory storing executable instructions;

a processor, coupled to the memory and executing the instructions stored on the memory, wherein upon execution, the stored instructions cause the processor to performing functions comprising:

conducting a biosensor screening test using the biosensor screening apparatus on the biosensor connected to the biosensor screening test apparatus;

collecting data attributes from the biosensor after completing the biosensor test;

scaling the collected data attributes; and

obtaining a screening test score;

wherein the biosensor screening test includes:

powering on the biosensor;

programming of test firmware;

initializing or rebooting the biosensor with the test firmware;

enabling a test mode operation of the test firmware;

applying a first average load current to the biosensor for a predetermined duration;

applying a test load current to the biosensor for a predetermined duration;

applying a second average load current to the biosensor for a predetermined duration;

disabling the test mode operation of the test firmware; and

restoring regular firmware for normal operation of the biosensor.