MANUFACTURING CONTROLS FOR SENSOR CALIBRATION USING FABRICATION MEASUREMENTS
Techniques disclosed herein relate to determining a calibrated measurement value indicative of a physiological condition of a patient using sensor calibration data determined based on fabrication measurements. In some embodiments, the techniques involve obtaining one or more electrical signals from a sensing element of a sensing arrangement, where the one or more electrical signals are influenced by a physiological condition in a body of a patient; obtaining calibration data associated with the sensing element, where the calibration data is based on fabrication process measurement data for the sensing element and a calibration model for a certain physiological condition; and determining, using the one or more electrical signals and the calibration data associated with the sensing element, a calibrated output value indicative of the physiological condition.
1 . A processor-implemented method, comprising:
obtaining one or more fabrication process measurements from a plurality of process control monitor (PCM) regions on a substrate having a plurality of sensing elements fabricated thereon;
obtaining one or more reference measurement outputs from the plurality of sensing elements in response to one or more known reference inputs;
assigning to the plurality of sensing elements at least one corresponding estimated fabrication process measurement;
maintaining associations between the respective estimated fabrication process measurements assigned to the plurality of sensing elements and the one or more reference measurement outputs obtained from the plurality of sensing elements; and
determining a predictive model based on the maintained associations that correlates combinations of fabrication process measurements to one or more calibration measurement parameters.
2 . The processor-implemented method of claim 1 , wherein the one or more fabrication process measurements for a given PCM region reflect physical characteristics of the given PCM region.
3 . The processor-implemented method of claim 2 , wherein the one or more fabrication process measurements reflecting physical characteristics for the given PCM region comprise at least one of: layer thickness measurements, material composition measurements, or physical dimension measurements.
4 . The processor-implemented method of claim 1 , wherein at least one fabrication process measurement is obtained directly from at least one of the plurality of sensing elements.
5 . The processor-implemented method of claim 1 , wherein the one or more calibration measurement parameters comprise at least one calibration factor, the at least one calibration factor being a ratio of a measured reference output to a design value for a respective sensing element.
6 . The processor-implemented method of claim 5 , wherein the at least one calibration factor comprises at least one of: an electrical current calibration factor or an electrochemical impedance spectroscopy (EIS) calibration factor.
7 . The processor-implemented method of claim 6 , wherein the electrical current calibration factor is determined by dividing a measured reference electrical current output by a design current value.
8 . The processor-implemented method of claim 6 , wherein the EIS calibration factor is determined by dividing a measured reference EIS value by a design EIS value.
9 . The processor-implemented method of claim 1 , wherein determining the predictive model comprises using machine learning to identify a subset of fabrication process measurement parameters that are correlated to the one or more calibration measurement parameters.
10 . The processor-implemented method of claim 9 , wherein the predictive model assigns different weights to different fabrication process measurement parameters based on a degree of correlation to a respective calibration measurement parameter.
11 . The processor-implemented method of claim 9 , wherein the machine learning comprises at least one of: neural networks, linear regression, genetic programming, support vector machines, Bayesian networks, or probabilistic machine learning models.
12 . The processor-implemented method of claim 1 , wherein the predictive model is trained to output a calibration factor for a given sensing element in response to input of one or more estimated fabrication process measurements assigned to the given sensing element.
13 . The processor-implemented method of claim 12 , further comprising applying the calibration factor to scale measurement outputs from the given sensing element.
14 . The processor-implemented method of claim 1 , wherein the at least one estimated fabrication process measurement assigned to a given sensing element is derived based on an estimate of physical characteristics corresponding to the given sensing element.
15 . The processor-implemented method of claim 14 , wherein the physical characteristics corresponding to the given sensing element are estimated based on the one or more fabrication process measurements obtained from one or more neighboring PCM regions.
16 . The processor-implemented method of claim 15 , wherein the estimate of the physical characteristics is derived based on interpolation of the one or more fabrication process measurements obtained from the one or more PCM neighboring regions.
17 . The processor-implemented method of claim 16 , wherein the interpolation accounts for spatial relationships between a location of the given sensing element on the substrate and respective locations of the one or more PCM regions that neighbor the given sensing element on the substrate.
18 . The processor-implemented method of claim 1 , wherein the plurality of sensing elements are interstitial glucose sensing elements.
19 . A system comprising:
one or more processors; and
one or more processor-readable media storing instructions which, when executed by the one or more processors, cause performance of:
obtaining one or more fabrication process measurements from a plurality of process control monitor (PCM) regions on a substrate having a plurality of sensing elements fabricated thereon;
obtaining one or more reference measurement outputs from the plurality of sensing elements in response to one or more known reference inputs;
assigning to the plurality of sensing elements at least one corresponding estimated fabrication process measurement;
maintaining associations between the respective estimated fabrication process measurements assigned to the plurality of sensing elements and the one or more reference measurement outputs obtained from the plurality of sensing elements; and
determining a predictive model based on the maintained associations that correlates combinations of fabrication process measurements to one or more calibration measurement parameters.
20 . A non-transitory computer-readable medium storing instructions which, when executed by one or more processors, cause performance of:
obtaining one or more fabrication process measurements from a plurality of process control monitor (PCM) regions on a substrate having a plurality of sensing elements fabricated thereon;
obtaining one or more reference measurement outputs from the plurality of sensing elements in response to one or more known reference inputs;
assigning to the plurality of sensing elements at least one corresponding estimated fabrication process measurement;
maintaining associations between the respective estimated fabrication process measurements assigned to the plurality of sensing elements and the one or more reference measurement outputs obtained from the plurality of sensing elements; and
determining a predictive model based on the maintained associations that correlates combinations of fabrication process measurements to one or more calibration measurement parameters.