Dual analyte sensor signal spike removal

Abstract

Described is a system for correcting for artifacts in raw analyte signals by operating, by an analyte sensor, in a first mode and in a second mode, the analyte sensor generating a first raw sensor signal indicative of a first analyte concentration of a host when operated in the first mode and a second raw sensor signal indicative of a second analyte when operated in the second mode; transitioning the analyte sensor from operating in the second mode to operating in the first mode; accessing a first value based at least on the first raw sensor signal taken during a first time period after the transitioning; generating a corrected first value of the raw sensor signal using corrective signal model configured to correct artifacts in the first raw sensor signal resulting from the transitioning; and generating an estimated first analyte concentration based at least on the corrected first value.

Description

Docket No. 4855.146WO1 / / 0958-PCT01DUAL ANALYTE SENSOR SIGNAL SPIKE REMOVALCLAIM OF PRIORITY

[0001] This application claims the benefit of priority to U.S. Provisional Application Serial No. 63 / 739,312, filed on Dec 27, 2024, which is incorporated herein by reference in its entirety.BACKGROUND

[0002] Diabetes is a metabolic condition relating to the production or use of insulin by the body. Insulin is a hormone that allows the body to use glucose for energy, or store glucose as fat.

[0003] When a person eats a meal that contains carbohydrates, the food is processed by the digestive system, which produces glucose in the person’s blood. Blood glucose can be used for energy or stored as fat. The body normally maintains blood glucose levels in a range that provides sufficient energy to support bodily functions and avoids problems that can arise when glucose levels are too high, or too low. Regulation of blood glucose levels depends on the production and use of insulin, which regulates the movement of blood glucose into cells.

[0004] When the body does not produce enough insulin, or when the body is unable to effectively use insulin that is present, blood sugar levels can elevate beyond normal ranges. The state of having a higher than normal blood sugar level is called "hyperglycemia." Chronic hyperglycemia can lead to a number of health problems, such as cardiovascular disease, cataract and other eye problems, nerve damage (neuropathy), and kidney damage. Hyperglycemia can also lead to acute problems, such as diabetic ketoacidosis - a state in which the body becomes excessively acidic due to the presence of blood glucose and ketones, which are produced when the body cannot use glucose. The state of having lower than normal blood glucose levels is called “hypoglycemia.” Severe hypoglycemia can lead to acute crises that can result in seizures or death.Docket No. 4855.146WO1 / / 0958-PCT01

[0005] A diabetes patient can receive insulin to manage blood glucose levels. Insulin can be received, for example, through a manual injection with a needle. Wearable insulin pumps are also available. Diet and exercise also affect blood glucose levels. A glucose sensor can provide an estimated glucose concentration level, which can be used as guidance by a patient or caregiver.

[0006] Diabetes conditions are sometimes referred to as “Type 1” and “Type 2.” A Type 1 diabetes patient is typically able to use insulin when it is present, but the body is unable to produce sufficient amounts of insulin, because of a problem with the insulin-producing beta cells of the pancreas. A Type 2 diabetes patient may produce some insulin, but the patient has become “insulin resistant” due to a reduced sensitivity to insulin. The result is that even though insulin is present in the body, the insulin is not sufficiently used by the patient’s body to effectively regulate blood sugar levels.

[0007] Blood sugar concentration levels may be monitored with an analyte sensor, such as a continuous glucose monitor. A continuous glucose monitor is used by a host (e.g., patient) to provide information, such as an estimated blood glucose value or a trend of estimated blood glucose levels.SUMMARY

[0008] This present application discloses, among other things, systems, devices, and methods related to analyte sensor, including, for example, the removal of spikes from sensor signals in analyte sensors.

[0009] Example l is a sensor system for dual analyte sensing, comprising: an analyte sensor configured to operate in a first mode and in a second mode, the analyte sensor generating a first raw sensor signal indicative of a first analyte concentration of a host when operated in the first mode and a second raw sensor signal indicative of a second analyte when operated in the second mode; and sensor electronics configured to perform operations comprising: transitioning the analyte sensor from operating in the second mode to operating in the first mode; accessing a first value based at least on the first raw sensor signal taken during a first time period after the transitioning; generating a corrected first value of the first raw sensor signal usingDocket No. 4855.146WO1 / / 0958-PCT01corrective signal model configured to correct one or more artifacts in the first raw sensor signal resulting from the transitioning; and generating an estimated first analyte concentration based at least on the corrected first value.

[0010] In Example 2, the subject matter of Example 1 optionally includes wherein the first analyte concentration includes a glucose concentration, wherein the transitioning of the analyte sensor from the second mode to the first mode includes transitioning the analyte sensor from generating the second raw sensor signal indicative of a second analyte concentration to generating the first raw sensor signal indicative of the glucose concentration.

[0011] In Example 3, the subject matter of any one or more of Examples 1- 2 optionally includes, where the second analyte concentration includes an oxygen concentration, wherein the transitioning of the analyte sensor from the second mode to the first mode includes transitioning the analyte sensor from generating the second raw sensor signal indicative of the oxygen concentration to generating the first raw sensor signal indicative of the glucose concentration.

[0012] In Example 4. the subject matter of any one or more of Examples 1- 3 optionally includes, where the second analyte concentration includes an oxygen concentration, wherein the transitioning of the analyte sensor from the second mode to the first mode includes transitioning the analyte sensor from generating the second raw sensor signal indicative of the oxygen concentration to generating the first raw sensor signal indicative of the first analyte concentration.

[0013] In Example 5, the subject matter of any one or more of Examples 1- 4 optionally includes, prior to transitioning the analyte sensor from operating in the second mode to operating in the first mode, causing the analyte sensor to operate in the first mode at least in part by providing a first bias condition to the analyte sensor.

[0014] In Example 6. the subject matter of Example 5 optionally includes, wherein transitioning the analyte sensor from operating in the second mode to operating in the first mode comprises applying a second bias condition toDocket No. 4855.146WO1 / / 0958-PCT01the analyte sensor, the second bias condition being different than the first bias condition.

[0015] In Example 7. the subject matter of Example 6 optionally includes, wherein the first bias condition is an opposite polarity to the second bias condition.

[0016] In Example 8, the subject matter of any one or more of Examples 1-7 optionally includes, wherein the one or more artifacts include a spike in a magnitude of the first raw sensor signal, wherein the corrected first value removes at least the spike of the one or more artifacts.

[0017] In Example 9, the subject matter of Example 8 optionally includes, where the one or more artifacts include a decay in a magnitude of the first raw sensor signal subsequent to the spike, wherein the corrected first value removes at least a delay portion of the one or more artifacts.

[0018] In Example 10, the subject matter of Example 9 optionally includes, wherein the one or more artifacts further include a spike in the magnitude of the first raw sensor signal, wherein the corrected first value further removes at least the spike portion of the one or more artifacts.

[0019] In Example 11, the subject matter of any one or more of Examples 1-10 optionally includes, wherein generating the estimated first analyte concentration includes subtracting the corrected first value from the first value.

[0020] In Example 12, the subject matter of any one or more of Examples 1-11 optionally includes, wherein the transitioning of the analyte sensor from operating in the second mode to operating in the first mode is in response to a predefined time period elapsing.

[0021] In Example 13, the subject matter of any one or more of Examples 1-12 optionally includes, wherein the transitioning of the analyte sensor from operating in the second mode to operating in the first mode is in response to a determination that a second value of the second raw sensor signal meets or exceeds a certain threshold indicative of an abnormality trigger.Docket No. 4855.146WO1 / / 0958-PCT01

[0022] In Example 14, the subject matter of any one or more of Examples 1-13 optionally includes, wherein the operations further comprise: identifying a current activity of a host; and identifying the corrective signal model among a plurality of corrective signal models based on the identified current activity; wherein generating the corrected first value of the first raw sensor signal is based on the identified current activity.

[0023] In Example 15, the subject matter of any one or more of Examples 1-14 optionally includes, wherein the corrective signal model includes a machine learning model, wherein generating the corrected first value of the first raw sensor signal using the corrective signal model comprises inputting the first value into the machine learning model, the machine learning model trained to generate a corrected first value of the first raw sensor signal.

[0024] In Example 16, the subject matter of any one or more of Examples 1-15 optionally includes, wherein the operations further comprise: accessing a second value of the first raw sensor signal; determining that the one or more artifacts resulting from the transitioning of the analyte sensor from operating in the second mode to operating in the first mode do not appear in the second value; and generating an estimated second analyte concentration based on an analyte concentration module; wherein generating the estimated first analyte concentration includes applying the corrected first value of the first raw sensor signal to the analyte concentration module.

[0025] In Example 17, the subject matter of any one or more of Examples 1-16 optionally includes, wherein the operations further comprise: retrieving historical dual analyte sensor data; applying the historical dual analyte sensor data to a retrospective artifact removal model; and receiving, from the retrospective artifact removal model, the corrective signal model configured to generate corrected values from raw sensor signals.

[0026] In Example 18, the subject matter of Example 17 optionally includes, wherein the retrospective artifact removal model includes a first machine learning model, the first machine learning model being used by the corrective signal model.Docket No. 4855.146WO1 / / 0958-PCT01

[0027] In Example 19, the subject matter of any one or more of Examples 17-18 optionally includes, wherein the corrective signal model includes a first machine learning model generated by the retrospective artifact removal model.

[0028] In Example 20, the subject matter of Example 19 optionally includes, wherein the retrospective artifact removal model includes a second machine learning model, the second machine learning model trained to generate the first machine learning model.

[0029] In Example 21, the subject matter of any one or more of Examples 19-20 optionally includes, wherein the corrective signal model is configured to continuously retain itself based on new values of the first raw sensor signal.

[0030] In Example 22, the subject matter of any one or more of Examples 17-21 optionally includes, wherein a sliding window of the historical dual analyte sensor data is applied to the retrospective artifact removal model to generate the corrective signal model.

[0031] In Example 23, the subject matter of any one or more of Examples 17-22 optionally includes, wherein the historical dual analyte sensor data includes historical dual analyte sensor data of the host.

[0032] In Example 24, the subject matter of any one or more of Examples 17-23 optionally includes, wherein the historical dual analyte sensor data includes historical dual analyte sensor data of other users other than the host.

[0033] In Example 25, the subject matter of Example 24 optionally includes, wherein the other users have one or more matching physiological characteristics of the host.

[0034] In Example 26, the subject matter of any one or more of Examples 17-25 optionally includes, wherein the corrective signal model is configured to generate corrected values from raw sensor signals in real time.

[0035] In Example 27, the subject matter of Example 26 optionally includes, wherein the retrospective artifact removal model is performed on an external server.Docket No. 4855.146WO1 / / 0958-PCT01

[0036] In Example 28, the subject matter of any one or more of Examples 17-27 optionally includes, wherein the operations further comprise: receiving historical single analyte sensor data; and applying the historical single analyte sensor data to the retrospective artifact removal model; wherein receiving, from the retrospective artifact removal model, the corrective signal model is further based on the historical single analyte sensor data.

[0037] In Example 29, the subject matter of Example 28 optionally includes, wherein the retrospective artifact removal generates the corrective signal model by subtracting values between the historical single analyte sensor data and the historical dual analyte sensor data.

[0038] In Example 30, the subject matter of Example 29 optionally includes, wherein prior to subtracting the values between the historical single analyte sensor data and the historical dual analyte sensor data, offsetting at least one of the historical single analyte sensor data and the historical dual analyte sensor data such that baseline values for the historical single analyte sensor data and the historical dual analyte sensor data align.

[0039] In Example 31, the subject matter of Example 30 optionally includes, wherein prior to using the corrective signal model, applying an opposite offset to the subtracted value based on the offset applied to the at least one of the historical single analyte sensor data and the historical dual analyte sensor data.

[0040] In Example 32, the subject matter of any one or more of Examples 30-31 optionally includes, wherein the offset is a magnitude offset of the at least one of the historical single analyte sensor data and the historical dual analyte sensor data.

[0041] In Example 33, the subject matter of any one or more of Examples 30-32 optionally includes, wherein the offset is a temporal offset of the at least one of the historical single analyte sensor data and the historical dual analyte sensor data.

[0042] In Example 34, the subject matter of any one or more of Examples 1-33 optionally includes, wherein a sliding window of the first raw sensorDocket No. 4855.146WO1 / / 0958-PCT01signal is applied to the corrective signal model to generate corrected values of the first raw sensor signal.

[0043] In Example 35, the subject matter of any one or more of Examples 1-34 optionally includes, wherein the operations further comprise: accessing a second value of the first raw sensor signal after the first time period; determining that the first raw sensor signal does not include the one or more artifacts resulting from the transitioning; and generating an estimated second analyte concentration based on the second value of the first raw sensor signal.

[0044] Example 36 is a method comprising: operating, by an analyte sensor, in a first mode and in a second mode, the analyte sensor generating a first raw sensor signal indicative of a first analyte concentration of a host when operated in the first mode and a second raw sensor signal indicative of a second analyte when operated in the second mode; transitioning the analyte sensor from operating in the second mode to operating in the first mode; accessing a first value based at least on the first raw sensor signal taken during a first time period after the transitioning; generating a corrected first value of the first raw sensor signal using corrective signal model configured to correct artifacts in the first raw sensor signal resulting from the transitioning; and generating an estimated first analyte concentration based at least on the corrected first value.

[0045] Example 37 is a non-transitory computer-readable storage medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising: operating, by an analyte sensor, in a first mode and in a second mode, the analyte sensor generating a first raw sensor signal indicative of a first analyte concentration of a host when operated in the first mode and a second raw sensor signal indicative of a second analyte when operated in the second mode; transitioning the analyte sensor from operating in the second mode to operating in the first mode; accessing a first value based at least on the first raw sensor signal taken during a first time period after the transitioning; generating a corrected first value of the first raw sensor signal using corrective signal model configured to correct artifacts in the first raw sensorDocket No. 4855.146WO1 / / 0958-PCT01signal resulting from the transitioning; and generating an estimated first analyte concentration based at least on the corrected first value.

[0046] Example 38 is at least one machine-readable medium including instructions that, when executed by processing circuitry, cause the processing circuitry to perform operations to implement any of Examples 1-37.

[0047] Example 39 is an apparatus comprising means to implement any of Examples 1-37.

[0048] Example 40 is a system to implement any of Examples 1-37.

[0049] Example 41 is a method to implement any of Examples 1-37.

[0050] This summary is intended to provide an overview of subject matter of the present patent application. It is not intended to provide an exclusive or exhaustive explanation of the disclosure. The detailed description is included to provide further information about the present patent application. Other aspects of the disclosure will be apparent to persons skilled in the art upon reading and understanding the following detailed description and viewing the drawings that form a part thereof, each of which are not to be taken in a limiting sense.BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0051] In the drawings, which are not necessarily drawn to scale, like numerals may describe similar components in different views. Like numerals having different letter suffixes may represent different instances of similar components. The drawings illustrate generally, by way of example, but not by way of limitation, various embodiments described in the present document.

[0052] FIG. 1 illustrates an aspect of the subject matter in accordance with one example.

[0053] FIG. 2 is a schematic illustration of an example analyte sensor system, which may for example, be the system shown in FIG. 1.Docket No. 4855.146WO1 / / 0958-PCT01

[0054] FIG. 3 is a diagram showing one example of a medical device system including the analyte sensor system of FIG. 1.

[0055] FIG. 4 is a side view of an example analyte sensor that may be implanted into a host.

[0056] FIG. 5 is a side view of another example analyte sensor in an arrangement including a mounting unit and an electronics unit.

[0057] FIG. 6 is an enlarged view of a distal portion of an analyte sensor, according to some examples.

[0058] FIG. 7 is a cross-sectional view through the sensor of FIG. 6 on plane 2-2 illustrating a membrane system, according to some examples.

[0059] FIG. 8 is a schematic illustration of a circuit that represents the behavior of an example analyte sensor, such as the analyte sensor shown in FIGS. 6-7, according to some examples.

[0060] FIG. 9 illustrates an example method for dynamically removing artifacts generated from transitioning between two modes, according to some examples.

[0061] FIG. 10 illustrates an architectural diagram for the removal of artifacts induced from switching modes in an analyte sensor, according to some examples.

[0062] FIG. 11 illustrates measurements taken from a dual analyte sensor and a single analyte sensor, according to some examples.

[0063] FIG. 12 illustrates baseline aligning of the dual sensor signal and the single analyte sensor signal, according to some examples.

[0064] FIG. 13 illustrates the generation of a corrective signal that can be applied to the raw glucose signal to generate a corrected glucose signal without the artifacts, according to some examples.

[0065] FIG. 14 illustrates the corrected glucose signal after the application of the corrective signal generated in FIG. 13, according to some examples.

[0066] FIG. 15 illustrates retrospective artifact removal modeling, according to some examples.

[0067] FIG. 16 illustrates prospective artifact removal modeling and realtime retraining, according to some examples.Docket No. 4855.146WO1 / / 0958-PCT01

[0068] FIG. 17 illustrates a machine-learning pipeline, according to some examples.

[0069] FIG. 18 illustrates training and use of a machine-learning program, according to some examples.

[0070] FIG. 19 is a block diagram illustrating a computing device hardware architecture, within which a set or sequence of instructions can be executed to cause a machine to perform examples of any one of the methodologies discussed herein.DETAILED DESCRIPTION

[0071] Various examples described herein are directed to analyte sensor systems and methods for using analyte sensor systems. An analyte sensor system includes an analyte sensor that is positioned in contact with a bodily fluid of a host to measure a concentration of an analyte, such as glucose, in the bodily fluid. In some examples, the analyte sensor is inserted into the host to contact the bodily fluid in vivo. In some examples, the analyte sensor is inserted subcutaneously to contact interstitial fluid below the host’s skin.

[0072] When the analyte sensor is exposed to analyte in the host’s bodily fluid, an electrochemical reaction between the analyte sensor and the analyte causes the analyte sensor to generate a raw sensor signal that is indicative of the analyte concentration in the bodily fluid. For example, the analyte sensor may include two or more electrodes. An analyte sensor system sensor electronics applies a bias condition or voltage to the electrodes. The bias condition may be, for example, a potential difference applied between a working electrode of the analyte sensor and a reference electrode of the analyte sensor. The bias condition promotes the electrochemical reaction between the analyte and the analyte sensor, resulting in a current between the working electrode and at least one other analyte sensor electrode. The raw sensor signal may be the current and / or may be based on the current.

[0073] The sensor electronics use the raw sensor signal to determine an estimated analyte concentration. In some examples, the sensor electronics are also programmed to output result data, which may include the estimated analyte concentration or other data. In some examples, the sensorDocket No. 4855.146WO1 / / 0958-PCT01electronics communicate the result data to one or more other external devices.

[0074] When measuring multiple analyte signals, some dual analyte sensors may switch between modes — the analyte sensors may measure an analyte signal in one mode and switch to another mode to measure a different analyte signal. Some dual analyte sensors suffer from disruptions when the analyte sensor switches from measuring one analyte in one mode to another analyte in another mode. These disruptions may manifest as spikes, noise, or other artifacts in the sensor data, which can compromise the accuracy and reliability of the measurements. Some systems often lack the sophisticated means to effectively identify, isolate, and correct these artifacts in real-time, leading to potential errors in analyte concentration estimation.

[0075] Many existing systems do not incorporate real-time data processing capabilities, are slow, and / or require a large amount of computing resources. As a result, they are unable to adjust on-the-fly to sudden changes in the sensor output caused by switching from measuring one analyte in one mode to another analyte in another mode. Without these real-time corrections, the data collected immediately after a mode change is often inaccurate and misleading, which can delay critical decisions in medical and environmental applications.

[0076] Existing sensors typically do not leverage historical data to improve the accuracy and responsiveness of the sensor dynamically. Existing sensors generally lack mechanisms to learn from past measurements, which means the sensors cannot anticipate or correct for recurring patterns of artifacts, missing an opportunity to enhance sensor performance over time.

[0077] In many existing systems, there is a significant underutilization of the data collected. These systems often do not have the capability to integrate new data with historical data to continuously refine their models. As a result, the potential for adaptive improvements in sensor accuracy and the ability to tailor responses based on specific user conditions or environments is not realized.

[0078] Various examples described herein may address these and other challenges. For example, an analyte sensor system may be implementedDocket No. 4855.146WO1 / / 0958-PCT01using a prospective model and / or a retrospective model to use both current and historical data to refine the correction processes. This may lead to increased accuracy and reliability of the analyte sensor, for example, during and / or after sensor mode switching (e.g., switching from measuring one analyte to another analyte), and may result in adaptive improvements in sensor accuracy that accounts for specific user conditions, specific environments, recurring patterns of artifacts, and / or the like. Further, the analyte sensor system consistent with the current subject matter effectively identifies, isolates, and corrects spikes, noise, or other artifacts in real-time, leading to a reduction or elimination in potential errors in analyte concentration estimation. The analyte sensor system consistent with the current subject matter may do so with improved speed and / or a reduction in required computing resources, leading to more accurate analyte concentration estimations that can be used in, for example, critical decision making environments.

[0079] In some examples of the current subject matter, by implementing a prospective model, the analyte sensor may continuously correct artifacts in real-time as they occur. This model dynamically adjusts to the artifacts introduced by mode switching. For example, the prospective model may make changes to the raw sensor signal generated by the analyte sensor in real time as the raw sensor signal is generated. This may ensure that the data being analyzed and used for analyte concentration estimation is clean and accurate during an analyte sensor session. This real-time correction may enable timely and reliable analyte concentration data for medical decisionmaking.

[0080] Complementing the real-time capabilities, the retrospective model according to some examples of the current subject matter may analyze historical data accumulated from both the current session and previous sessions. The retrospective model may identify long-term trends and recurring patterns in the artifacts, and may refine the analyte sensor system's understanding and response to these issues over time. By learning from a broad dataset, the retrospective model enhances the system’s ability to predict and mitigate similar artifacts in the future, continuously improving the accuracy and reliability of the sensor.Docket No. 4855.146WO1 / / 0958-PCT01

[0081] The analyte sensor system according to examples of the current subject matter may employ the prospective model, the retrospective model, and / or both the prospective model and the retrospective model. In some examples, the dual approach of using both the prospective and retrospective models may create a robust system that may deal with artifacts as they occur and may also use historical insights to prevent future occurrences. For example, the analyte sensor system may adapt its artifact correction strategies based on comprehensive feedback and learning, which is particularly effective for handling complex and dynamic environmental conditions where sensor responses may vary significantly. For example, the analyte sensor system may retrain the prospective model in real-time using new data as the data is received, alone, or together with insights from the retrospective analysis. Doing so may ensure that the models are up-to-date and optimized for current conditions. This continuous learning cycle may reduce the lag time in model responsiveness, allowing the system according to examples of the current subject matter to quickly adapt to changes in sensor behavior or the external environment. This improves not only the immediate output of the sensor system, but may also enriches the dataset used for ongoing training and refinement of the models. Such comprehensive data utilization leads to more nuanced and accurate sensor calibrations and enhances the predictive capabilities of the system according to examples of the current subject matter.

[0082] When the effects in this disclosure are considered in aggregate, one or more of the methodologies described herein may improve analyte sensor systems, providing additional functionality' (such as, but not limited to, the functionality mentioned above), making the analyte sensor systems easier, faster, and / or more intuitive to operate, and / or obviating a need for certain efforts or computing resources that otherwise would be involved in a dual analyte sensor process. Computing resources used by one or more machines, databases, or networks may thus be more efficiently utilized or even reduced.

[0083] FIG. 1 is a diagram showing one example of an environment 100 including an analyte sensor system 102. The analyte sensor system 102 is coupled to a host 101, which may be a human patient. In some examples, theDocket No. 4855.146WO1 / / 0958-PCT01host is subject to a temporary or permanent diabetes condition or other health condition that makes analyte monitoring useful.

[0084] The analyte sensor system 102 includes an analyte sensor 104. In some examples, the analyte sensor 104 is or includes a glucose sensor configured to measure a glucose concentration in the host 101. The analyte sensor 104 can be exposed to analyte at the host 101 in any suitable way. In some examples, the analyte sensor 104 is fully implantable under the skin of the host 101. In other examples, the analyte sensor 104 is wearable on the body of the host 101 (e.g.. on the body but not under the skin). Also, in some examples, the analyte sensor 104 is a transcutaneous device (e.g., with a sensor residing at least partially under or in the skin of a host). It should be understood that the devices and methods described herein can be applied to any device capable of detecting a concentration of an analyte, such as glucose, and providing an output signal that represents the concentration of the analyte.

[0085] In the example of FIG. 1, the analyte sensor system 102 also includes sensor electronics 106. In some examples, the sensor electronics 106 and analyte sensor 104 are provided in a single integrated enclosure (See FIG. 5). In other examples, the analyte sensor 104 and sensor electronics 106 are provided as separate components or modules (See FIG. 6). For example, the analyte sensor system 102 may include a disposable (e.g., single-use) sensor mounting unit that may include the analyte sensor 104, a component for attaching the sensor analyte sensor 104 to a host (e.g., an adhesive pad), and / or a mounting structure configured to receive a sensor electronics unit including some or all of the sensor electronics 106 shown in FIGS. 2 and 3. The sensor electronics unit 106 may be reusable.

[0086] The analyte sensor 104 may use any known method, including invasive, minimally-invasive, or non-invasive sensing techniques (e.g., optically excited fluorescence, microneedle, transdermal monitoring of glucose), to provide a raw sensor signal indicative of the concentration of the analyte in the host 101. The raw sensor signal may be converted into calibrated and / or filtered analyte concentration data used to provide a useful value of the analyte concentration (e.g., estimated blood glucoseDocket No. 4855.146WO1 / / 0958-PCT01concentration level) to a user, such as the host or a caretaker (e.g., a parent, a relative, a guardian, a teacher, a doctor, a nurse, or any other individual that has an interest in the wellbeing of the host 101).

[0087] In some examples, the analyte sensor 104 is or includes a continuous glucose sensor. A continuous glucose sensor can be or include a subcutaneous, transdermal (e.g., transcutaneous), and / or intravascular device. In some examples, such a sensor or device may recurrently (e.g., periodically, or intermittently) analyze sensor data. The glucose sensor may¬ use any method of glucose measurement, including enzymatic, chemical, physical, electrochemical, spectrophotometric, polarimetric, calorimetric, iontophoretic, radiometric, immunochemical, and the like. In various examples, the analyte sensor system 102 may be or include a continuous glucose monitor sensor available from DexCom™, (e g., the DexCom G5™ sensor, Dexcom G6™ sensor, the DexCom G7™ sensor, or any variation thereof), from Abbott™ (e.g., the Libre™ sensor), or from Medtronic™ (e.g., the Enlite™ sensor).

[0088] In some examples, analyte sensor 104 includes an implantable glucose sensor, such as described with reference to U.S. Patent 6,001,067 and U.S. Patent Publication No. US-2005-0027463-A1, which are incorporated by reference. In some examples, analyte sensor 104 includes a transcutaneous glucose sensor, such as described with reference to U.S. Patent Publication No. US-1806-0020187-A1, which is incorporated by reference. In some examples, analyte sensor 104 may be configured to be implanted in a host vessel or extracorporeally, such as is described in U.S. Patent Publication No. US-2007-0027385-A1, co-pending U.S. Patent Publication No. US-1808-0119703-Al filed October 4, 1806. U.S. Patent Publication No. US- 1808-0108942-Al filed on March 26, 2007, and U.S. Patent Application No. US-2007-0197890-A1 filed on February714, 2007, all of which are incorporated by reference. In some examples, the continuous glucose sensor may include a transcutaneous sensor such as described in U.S. Patent 6,565,509 to Say et al., which is incorporated by reference. In some examples, analyte sensor 104 may include a continuous glucose sensor that includes a subcutaneous sensor such as described with reference to U.S. Patent 6,579,690 to Bonnecaze et al. or U.S. Patent 6,484,046 to Say et al.,Docket No. 4855.146WO1 / / 0958-PCT01which are incorporated by reference. In some examples, the continuous glucose sensor may include a refillable subcutaneous sensor such as described with reference to U.S. Patent 6,512,939 to Colvin et al., which is incorporated by reference. The continuous glucose sensor may include an intravascular sensor such as described with reference to U.S. Patent 6,477,395 to Schulman et al., which is incorporated by reference. The continuous glucose sensor may include an intravascular sensor such as described with reference to U.S. Patent 6,424,847 to Mastrototaro et al., which is incorporated by reference.

[0089] The environment 100 may also include various other external devices including, for example, a medical device 108. The medical device 108 may be or include a drug delivery device such as an insulin pump or an insulin pen. In some examples, the medical device 108 includes one or more sensors, such as another analyte sensor, a heart rate sensor, a respiration sensor, a motion sensor (e.g., accelerometer), posture sensor (e.g., 3-axis accelerometer), acoustic sensor (e.g., to capture ambient sound or sounds inside the body). The medical device 108 may be wearable, e.g., on a watch, glasses, contact lens, patch, wristband, ankle band, or another wearable item, or may be incorporated into a handheld device (e.g., a smartphone). In some examples, the medical device 108 includes a multi-sensor patch that may, for example, detect one or more of an analyte levels (e.g., glucose, lactate, insulin, or other substance), heart rate, respiration (e.g., using impedance), activity (e.g.. using an accelerometer), posture (e.g., using an accelerometer), galvanic skin response, tissue fluid levels (e.g., using impedance or pressure).

[0090] In some examples, the analyte sensor system 102 and the medical device 108 communicate with one another. Communication between the analyte sensor system 102 and medical device 108 may occur over any suitable wired connection and / or via a wireless communication signal 110. For example, the analyte sensor system 102 (e.g., the sensor electronics 106 thereof) may be configured to establish a communication connection with the medical device 108 using a suitable short-range communications medium such as, for example, a radio frequency medium (e.g., Bluetooth, Medical Implant Communication System (MICS), Wi-Fi, near field communicationDocket No. 4855.146WO1 / / 0958-PCT01(NFC), radio frequency identification (RFID), Zigbee, Z-Wave or other communication protocols), an optical medium (e.g.. infrared), a sonic medium (e.g., ultrasonic), a cellular protocol-based medium (e.g.. Code Division Multiple Access (CDMA) or Global System for Mobiles (GSM)), and / or the like.

[0091] In some examples, the environment 100 also includes other external devices such as, for example, a wearable sensor 130. The wearable sensor 130 can include a sensor circuit (e.g., a sensor circuit configured to detect a glucose concentration or other analyte concentration) and a communication circuit, which may, for example, be an NFC circuit. In some examples, information from the wearable sensor 130 may be retrieved from the wearable sensor 130 using a user computing device 132, such as a smart phone, that is configured to communicate with the wearable sensor 130 via the wearable sensor’s communication circuit, for example, when the user device 132 is positioned near the wearable sensor 130. For example, swiping the user device 132 over the sensor 130 may retrieve sensor data from the wearable sensor 130 using NFC or other suitable wireless communication.

[0092] The use of NFC communication may reduce power consumption by the wearable sensor 130, which may reduce the size of a power source (e.g., battery7or capacitor) in the wearable sensor 130 or extend the usable life of the power source. In some examples, the wearable sensor 130 may be wearable on an upper arm as shown. In some examples, a wearable sensor 130 may additionally or alternatively be on the upper torso of the patient (e.g., over the heart or over a lung), which may, for example, facilitate detecting heart rate, respiration, or posture. A wearable sensor 136 may also be on the lower body (e.g., on a leg) or other part of the body (e.g., on the abdomen).

[0093] In some examples, an array or network of sensors may be associated with the patient. For example, one or more of the analyte sensor system 102, and / or external devices, such as the medical device 108, wearable device 120 such as a watch, an additional wearable sensor 130 and / or the like, may communicate with one another via a short-range communication medium (e.g., Bluetooth, MICS, NFC, or any of the other options described above.).Docket No. 4855.146WO1 / / 0958-PCT01The additional wearable sensor 130 may be any of the examples described above with respect to medical device 108. The analyte sensor system 102. medical device 108, and additional sensor 130 on the host 101 are provided for illustration and description and are not necessarily drawn to scale.

[0094] The environment 100 may also include one or more other external devices such as a hand-held smart device (e.g., smart phone) 112, tablet 114, smart pen 116 (e.g., insulin delivery pen with processing and communication capability ), computer 118, a wearable device 120 such as a watch, or peripheral medical device 122 (which may be a proprietary device such as a proprietary user device available from DexCom™), any of which may communicate with the analyte sensor system 102 via a short-range communication medium, such as indicated by wireless communication signal 110, and may also communicate over a network 124 with a server system (e.g., remote data center) or with a remote terminal 128 to facilitate communication with a remote user (not shown) such as a technical support staff member or a clinician.

[0095] The wearable device 120 may include an activity sensor, a heart rate monitor (e.g., light-based sensor or electrode-based sensor), a respiration sensor (e.g., acoustic- or electrode-based), a location sensor (e.g., GPS), or other sensors.

[0096] In some examples, the environment 100 includes a server system 126. The server system 126 can include one or more computing devices, such as one or more server computing devices. In some examples, the server system 126 is used to collect analyte data from the analyte sensor system 102 and / or analyte or other data from the plurality of other devices, and to perform analytics on collected data, generate, or apply universal or individualized models for glucose levels, and communicate such analytics, models, or information based thereon back to one or more of the devices in the environment 100. In some examples, the server system 126 gathers interhost and / or intra-host break-in data to generate one or more break-in characteristics, as described herein.

[0097] The environment 100 may also include a wireless access point (WAP) 138 used to communicatively couple one or more of analyte sensorDocket No. 4855.146WO1 / / 0958-PCT01system 102, network 124, server system 126, medical device 108 or any of the peripheral devices described above. For example, WAP 138 may provide Wi-Fi and / or cellular connectivity within environment 100. Other communication protocols, such as NFC or Bluetooth, may also be used among devices of the environment 100.

[0098] FIG. 2 is a schematic illustration of an example analyte sensor system 200, which may for example, be the system 102 shown in FIG. 1. The analyte sensor system may include an analyte sensor 202. The analyte sensor 202 may be configured to measure glucose or another suitable analyte. The analyte sensor system 200 may also comprise one or more temperature sensors 204, a processor 210, and a memory 206. The processor 210 may receive a signal indicative of an analyte concentration level from the analyte sensor 202 and receive a temperature signal indicative of a temperature parameter (e.g. absolute or relative temperature, or a temperature gradient) from the temperature sensor 204. The signal indicative of the analyte concentration may be a raw sensor signal or a processed sensor signal. The sensor system 200 may also include one or more additional sensors 208, which may include, for example, a heart rate sensor, activity sensor (e.g. accelerometer), or a pressure gauge (e.g. to measure compression of the sensor against a host).

[0099] The processor 210 may determine a temperature-compensated analyte concentration level based on the temperature sensor signal and optionally also based on one or more signals from additional sensor(s) 208. The processor 210 may determine a specific temperature-compensated sensitivity value (e.g., analyte sensor sensitivity value based on the temperature), or may determine a compensated estimated glucose value. The signal from the temperature sensor 204 may be used as an approximation of a temperature at an analyte sensor, or the signal from the temperature sensor 204 may be processed (e.g., using methods described in detail below) to determine an estimated analyte temperature sensor based on the signal from the temperature sensor 204.

[0100] In some examples, the processor 210 may retrieve instructions or information from a memory 206 to determine temperature-compensatedDocket No. 4855.146WO1 / / 0958-PCT01analyte concentration level. For example, the processor may access a look-up table, or apply an algorithm based on the signal indicative of analyte concentration and temperature sensor signal or apply the signal indicative of analyte concentration and temperature signal to a model (e.g., use a state model or neural network).

[0101] In some examples, the processor may retrieve executable instructions from the memory 206 (or a separate memory that may be operatively coupled to or integrated into the processor.) In some examples, the processor may include, or be part of, an application-specific integrated circuit (ASIC) that may be configured to determine a temperature-compensated glucose concentration level. In various examples, any one or more of the methods described herein may be executed by the processor 210 or temperature-compensated glucose sensor, either alone, or in combination with other processors or devices.

[0102] FIG. 3 is a diagram showing one example of a medical device system 300 including the analyte sensor system 102 of FIG. 1. In the example of FIG. 3. the analyte sensor system 102 includes sensor electronics 106 and an example sensor mounting unit 390, although in some examples, it will be appreciated that the analyte sensor 104 and sensor electronics 106 may be included in a common enclosure. While a specific example of division of components between the sensor mounting unit 390 and sensor electronics 106 is shown, it is understood that some examples may include additional components in the sensor mounting unit 390 or in the sensor electronics 106, and that some of the components (e.g., a battery' or supercapacitor) that are shown in the sensor electronics 106 may be alternatively or additionally (e.g.. redundantly) provided in the sensor mounting unit 390.

[0103] In the example shown in FIG. 3, the sensor mounting unit 390 includes the analyte sensor 104 and a battery 392. In some examples, the sensor mounting unit 390 may be replaceable, and the sensor electronics 106 may include a debouncing circuit (e.g., gate with hysteresis or delay) to avoid, for example, recurrent execution of a power-up or power down process when a battery is repeatedly connected and disconnected or avoidDocket No. 4855.146WO1 / / 0958-PCT01processing of noise signal associated with removal or replacement of a battery.

[0104] The sensor electronics 106 may include electronics components that are configured to process sensor information, such as raw sensor signals, and generate corresponding analyte concentration values. The sensor electronics 106 may, for example, include electronic circuitry associated with measuring, processing, storing, or communicating continuous analyte sensor data, including prospective algorithms associated with processing and calibration of the raw sensor signal. The sensor electronics 106 may include hardware, firmware, and / or software that enables measurement of levels of the analyte via a glucose sensor. Electronic components may be affixed to a printed circuit board (PCB), or the like, and can take a variety7of forms. For example, the electronic components may take the form of an integrated circuit (IC), such as an Application-Specific Integrated Circuit (ASIC), a microcontroller, and / or a processor.

[0105] In the example of FIG. 3, the sensor electronics 106 include a measurement circuit 302 (e.g., potentiostat) coupled to the analyte sensor 104 and configured to recurrently obtain analyte sensor readings using the analyte sensor 104. For example, the measurement circuit 302 may continuously or recurrently measure a raw sensor signal indicating a current flow at the analyte sensor 104 between a working electrode and a reference electrode. The sensor electronics 106 may include agate circuit 394, which may be used to gate the connection between the measurement circuit 302 and the analyte sensor 104. For example, the analyte sensor 104 may accumulate charge over an accumulation period. After the accumulation period, the gate circuit 394 is opened so that the measurement circuit 302 can measure the accumulated charge. Gating the analyte sensor 104 may improve the performance of the sensor system 102 by creating a larger signal to noise or interference ratio (e.g., because charge accumulates from an analyte reaction, but sources of interference, such as the presence of acetaminophen near a glucose sensor, do not accumulate, or accumulate less than the charge from the analyte reaction).Docket No. 4855.146WO1 / / 0958-PCT01

[0106] The sensor electronics 106 may also include a processor 304. The processor 304 is configured to retrieve instructions 306 from memory 308 and execute the instructions 306 to control various operations in the analyte sensor system 102. For example, the processor 304 may be programmed to control application of bias potentials to the analyte sensor 104 via a potentiostat at the measurement circuit 302, interpret raw sensor signals from the analyte sensor 104, and / or compensate for environmental factors.

[0107] The processor 304 may also save information in data storage memory 310 or retrieve information from data storage memory 310. In various examples, data storage memory' 310 may be integrated with memory 308, or may be a separate memory circuit, such as a non-volatile memory circuit (e.g., flash RAM). Examples of systems and methods for processing sensor analyte data are described in more detail herein and in U.S. Patent Nos. 7,310.544 and 6,931.327.

[0108] The sensor electronics 106 may also include one or more sensors, such as the sensor 312, which may be coupled to the processor 304. The sensor 312 may be a temperature sensor, accelerometer, or another suitable sensor. The sensor electronics 106 may also include a power source such as a capacitor or battery 314, which may be integrated into the sensor electronics 106, or may be removable, or part of a separate electronics unit. The battery 314 (or other power storage component, e.g., capacitor) may optionally be rechargeable via a wired or wireless (e.g., inductive or ultrasound) recharging system 316. The recharging system 316 may harvest energy or may receive energy from an external source or on-board source. In various examples, the recharge circuit may include a triboelectric charging circuit, a piezoelectric charging circuit, an RF charging circuit, a light charging circuit, an ultrasonic charging circuit, a heat charging circuit, a heat harvesting circuit, or a circuit that harvests energy from the communication circuit. In some examples, the recharging circuit may recharge the rechargeable battery using power supplied from a replaceable battery (e.g., a battery supplied with a base component).

[0109] The sensor electronics 106 may also include one or more supercapacitors in the sensor electronics unit (as shown), or in the sensorDocket No. 4855.146WO1 / / 0958-PCT01mounting unit 390. For example, the supercapacitor may allow energy to be drawn from the batten’ 314 in a highly consistent manner to extend the life of the battery 314. The battery 314 may recharge the supercapacitor after the supercapacitor delivers energy to the communication circuit or to the processor 304, so that the supercapacitor is prepared for delivery of energy during a subsequent high-load period. In some examples, the supercapacitor may be configured in parallel with the battery 314. A device may be configured to preferentially draw energy from the supercapacitor, as opposed to the battery 314. In some examples, a supercapacitor may be configured to receive energy from a rechargeable battery for short-term storage and transfer energy to the rechargeable battery for long-term storage.

[0110] The supercapacitor may extend an operational life of the battery' 314 by reducing the strain on the battery' 314 during the high-load period. In some examples, a supercapacitor removes at least 10% of the strain off the battery during high-load events. In some examples, a supercapacitor removes at least 30% of the strain off the battery during high-load events. In some examples, a supercapacitor removes at least 30% of the strain off the battery' during high-load events. In some examples, a supercapacitor removes at least 50% of the strain off the battery during high-load events.

[0111] The sensor electronics 106 may also include a wireless communication circuit 318, which may for example include a wireless transceiver operatively coupled to an antenna. The wireless communication circuit 318 may be operatively coupled to the processor 304 and may be configured to wirelessly communicate with one or more peripheral devices or other medical devices, such as an insulin pump or smart insulin pen.

[0112] In the example of FIG. 3, the medical device system 300 also includes optional external devices including, for example, a peripheral device 350. The peripheral device 350 may be any suitable user computing device such as. for example, a wearable device (e.g., activity monitor), such as awearable device 120. In other examples, the peripheral device 350 may be a hand-held smart device (e.g., smartphone or other device such as a proprietary handheld device available from Dexcom), a tablet 114, a smart pen 116, or special-purpose computer 118 shown in FIG. 1.Docket No. 4855.146WO1 / / 0958-PCT01

[0113] The peripheral device 350 may include a UI 352, a memory circuit 354, a processor 356, a wireless communication circuit 358, a sensor 360, or any combination thereof. The peripheral device 350 may not necessarily include all the components shown in FIG. 3. The peripheral device 350 may also include a power source, such as a battery7.

[0114] The UI 352 may, for example, be provided using any suitable input / output device or devices of the peripheral device 350 such as, for example, a touch-screen interface, a microphone (e.g., to receive voice commands), or a speaker, a vibration circuit, or any combination thereof. The UI 352 may receive information from the host or another user (e.g., instructions, glucose values). The UI 352 may also deliver information to the host or other user, for example, by displaying UI elements at the UI 352. For example, UI elements can indicate glucose or other analyte concentration values, glucose or other analyte trends, glucose, or other analyte alerts, etc. Trends can be indicated by UI elements such as arrows, graphs, charts, etc.

[0115] The processor 356 may be configured to present information to a user, or receive input from a user, via the UI 352. The processor 356 may also be configured to store and retrieve information, such as communication information (e.g., pairing information or data center access information), user information, sensor data or trends, or other information in the memory circuit 354. The wireless communication circuit 358 may include a transceiver and antenna configured to communicate via a wireless protocol, such as any of the wireless protocols described herein. The sensor 360 may, for example, include an accelerometer, a temperature sensor, a location sensor, biometric sensor, or blood glucose sensor, blood pressure sensor, heart rate sensor, respiration sensor, or another physiologic sensor.

[0116] The peripheral device 350 may be configured to receive and display sensor information that may be transmitted by sensor electronics 106 (e.g., in a customized data package that is transmitted to the display devices based on their respective preferences). Sensor information (e.g., blood glucose concentration level) or an alert or notification (e.g., “high glucose level”, “low glucose level” or “fall rate alert” may be communicated via the UI 352 (e.g., via visual display, sound, or vibration). In some examples, theDocket No. 4855.146WO1 / / 0958-PCT01peripheral device 350 may be configured to display or otherwise communicate the sensor information as it is communicated from the sensor electronics 106 (e.g., in a data package that is transmitted to respective display devices). For example, the peripheral device 350 may transmit data that has been processed (e.g., an estimated analyte concentration level that may be determined by processing raw sensor data), so that a device that receives the data may not be required to further process the data to determine usable information (such as the estimated analyte concentration level). In other examples, the peripheral device 350 may process or interpret the received information (e.g., to declare an alert based on glucose values or a glucose trend). In various examples, the peripheral device 350 may receive information directly from sensor electronics 106, or over a network (e.g., via a cellular or Wi-Fi network that receives information from the sensor electronics 106 or from a device that is communicatively coupled to the sensor electronics 106).

[0117] In the example of FIG. 3, the medical device system 300 includes an optional medical device 370. For example, the medical device 370 may be an external device used in addition to or instead of the peripheral device 350. The medical device 370 may be or include any suitable type of medical or other computing device including, for example, the medical device 108, peripheral medical device 122, wearable device 120, wearable sensor 130, or wearable sensor 136 shown in FIG. 1. The medical device 370 may include a UI 372. a memory circuit 374, a processor 376, a wireless communication circuit 378, a sensor 380, a therapy circuit 382, or any combination thereof.

[0118] Similar to the UI 352, the UI 372 may be provided using any suitable input / output device or devices of the medical device 370 such as, for example, a touch-screen interface, a microphone, or a speaker, a vibration circuit, or any combination thereof. The UI 372 may receive information from the host or another user (e.g., glucose values, alert preferences, calibration coding). The UI 372 may also deliver information to the host or other user, for example, by displaying UI elements at the UI 352. For example, UI elements can indicate glucose or other analyte concentration values, glucose or other analyte trends, glucose, or other analyte alerts, etc. Trends can be indicated by UI elements such as arrows, graphs, charts, etc.Docket No. 4855.146WO1 / / 0958-PCT01

[0119] The processor 376 may be configured to present information to a user, or receive input from a user, via the UI 372. The processor 376 may also be configured to store and retrieve information, such as communication information (e.g., pairing information or data center access information), user information, sensor data or trends, or other information in the memory7circuit 374. The wireless communication circuit 378 may include a transceiver and antenna configured communicate via a wireless protocol, such as any?of the wireless protocols described herein.

[0120] The sensor 380 may, for example, include an accelerometer, a temperature sensor, a location sensor, biometric sensor, or blood glucose sensor, blood pressure sensor, heart rate sensor, respiration sensor, or another physiologic sensor. The medical device 370 may include two or more sensors (or memories or other components), even though only one sensor 380 is shown in the example in FIG. 3. In various examples, the medical device 370 may be a smart handheld glucose sensor (e.g., blood glucose meter), drug pump (e.g., insulin pump), or other physiologic sensor device, therapy device, or combination thereof.

[0121] In examples where medical device 370 is or includes an insulin pump, the pump and analyte sensor system 102 may be in two-way communication (e.g., so the pump can request a change to an analyte transmission protocol, e g., request a data point or request data on a more frequent schedule), or the pump and analyte sensor system 102 may communicate using one-way communication (e.g., the pump may receive analyte concentration level information from the analyte sensor system). In one-way communication, a glucose value may be incorporated in an advertisement message, which may be encrypted with a previously shared key. In a two-way communication, a pump may request a value, which the analyte sensor system 102 may share, or obtain and share, in response to the request from the pump, and any or all of these communications may be encrypted using one or more previously shared keys. An insulin pump may receive and track analyte (e.g., glucose) values transmitted from analyte sensor system 102 using one-way communication to the pump for one or more of a variety of reasons. For example, an insulin pump may suspend orDocket No. 4855.146WO1 / / 0958-PCT01activate insulin administration based on a glucose value being below or above a threshold value.

[0122] In some examples, the medical device system 300 includes two or more peripheral devices and / or medical devices that each receive information directly or indirectly from the analyte sensor system 102.Because different display devices provide many different user interfaces, the content of the data packages (e.g., amount, format, and / or type of data to be displayed, alarms, and the like) may be customized (e.g., programmed differently by the manufacturer and / or by an end user) for each device. For example, referring now to the example of FIG. 1, a plurality of different peripheral devices may be in direct wireless communication with sensor electronics 106 (e.g., such as an on-skin sensor electronics 106 that are physically connected to the continuous analyte sensor 104) during a sensor session to enable a plurality of different types and / or levels of display and / or functionality associated with the displayable sensor information, or, to save battery power in the sensor system 102, one or more specified devices may communicate with the analyte sensor system 102 and relay (i.e., share) information to other devices directly or through a server system (e g., a network-connected data center) 126.

[0123] FIG. 4 is a side view of an example analyte sensor 434 that may be implanted into a host. An enclosure 402 may be adhered to the host's skin using an adhesive pad 408. The adhesive pad 408 may be formed from an extensible material, which may be removably attached to the skin using an adhesive. Sensor electronics may be positioned within the enclosure 402. The sensor 434 may extend from the enclosure 402 and under the skin of a host, as shown.

[0124] FIG. 5 is a side view of another example analyte sensor 534 in an arrangement including a mounting unit 514 and an electronics unit 518. The mounting unit 514 may be adhered to the host's skin using an adhesive pad 508, which may be like the adhesive pad 408 described herein. The electronics unit 518 comprises an enclosure 502 that may have sensor electronics positioned thereon. In some examples, the electronics unit 518 and mounting unit 514 are arranged in a manner like the sensor electronicsDocket No. 4855.146WO1 / / 0958-PCT01106 and sensor mounting unit 490 shown in FIGS. 1 and 4. For example, the sensor 534 may extend from the enclosure 502 via the mounting unit 514.

[0125] FIG. 6 is an enlarged view of a distal portion of an analyte sensor 634. The analyte sensor 634 illustrates one example arrangement that may be used to implement the analyte sensors described herein, such as, for example, the analyte sensors 104, 434, 534. The analyte sensor 634 may be adapted for insertion under the host's skin and may be mechanically coupled to an enclosure, such as the enclosures 502, and / or to a mounting unit 514, such as the mounting unit 514. The analyte sensor 634 may be electrically coupled to sensor electronics, which may be positioned within the enclosure 402, 502.

[0126] The example analyte sensor 634 shown in FIG. 6 includes an elongated conductive body 641. The elongated conductive body 641 can include a core with various layers positioned thereon. A first layer 638 that at least partially surrounds the core and includes a working electrode, for example located in window 639). In some examples, the core and the first layer 638 are made of a single material (such as. for example, platinum). In some examples, the elongated conductive body 641 is a composite of two conductive materials, or a composite of at least one conductive material and at least one non-conductive material. A membrane system 632 is located over the working electrode and may cover other layers and / or electrodes of the sensor 634, as described herein.

[0127] The first layer 638 may be formed of a conductive material. The working electrode (at window 639) is an exposed portion of the surface of the first layer 638. Accordingly, the first layer 638 is formed of a material configured to provide a suitable electroactive surface for the working electrode. Examples of suitable materials include, but are not limited to, platinum, platinum-iridium, gold, palladium, iridium, graphite, carbon, a conductive polymer, an alloy, and / or the like.

[0128] A second layer 640 surrounds at least a portion of the first layer 638, thereby defining boundaries of the working electrode. In some examples, the second layer 640 serves as an insulator and is formed of an insulating material, such as polyimide, polyurethane, parylene, or any otherDocket No. 4855.146WO1 / / 0958-PCT01suitable insulating materials or materials. In some examples, the second layer 640 is configured such that the working electrode (of the layer 638) is exposed via the window 639.

[0129] In some examples, the sensor 634 further includes a third layer 643 comprising a conductive material. The third layer 643 may comprise a reference electrode. In some examples, the third layer 643, including the reference electrode, is formed of a silver-containing material that is applied onto the second layer 640 (e.g., an insulator). The silver-containing material may include various materials and be in various forms such as, for example. Ag / AgCl-polymer pasts, paints, polymer-based conducting mixtures, inks, etc.

[0130] The analyte sensor 634 may include two (or more) electrodes, e.g., a working electrode at the layer 638 and exposed at window 639 and at least one additional electrode, such as a reference electrode of the layer 643. In the example arrangement of FIGS. 6-7, the reference electrode also functions as a counter electrode, although other arrangements can include a separate counter electrode. While the analyte sensor 634 may be used with a mounting unit in some examples, in other examples, the analyte sensor 634 may be used with other types of sensor systems. For example, the analyte sensor 634 may be part of a system that includes a battery and sensor in a single package, and may optionally include, for example, anear-field communication (NFC) circuit.

[0131] FIG. 7 is a cross-sectional view through the sensor 634 of FIG. 6 on plane 2-2 illustrating a membrane system 632. The membrane system 632 may include a number of domains (e.g., layers). In an example, the membrane system 632 may include an enzyme domain 642, a diffusion resistance domain 644, and a bioprotective domain 646 located around the working electrode. In some examples, a unitary diffusion resistance domain and bioprotective domain may be included in the membrane system 632 (e g., wherein the functionality of both the diffusion resistance domain and bioprotective domain are incorporated into one domain).

[0132] The membrane system 632, in some examples, also includes an electrode layer 647. The electrode layer 647 may be arranged to provide anDocket No. 4855.146WO1 / / 0958-PCT01environment between the surfaces of the working electrode and the reference electrode that facilitates the electrochemical reaction between the electrodes. For example, the electrode layer 647 may include a coating that maintains a layer of water at the electrochemically reactive surfaces of the sensor 634.

[0133] In some examples, the sensor 634 may be configured for short-term implantation (e.g., from about 1 to 30 days). However, it is understood that the membrane system 632 can be modified for use in other devices, for example, by including only one or more of the domains, or additional domains. For example, a membrane system 632 may include a plurality of resistance layers, or a plurality of enzyme layers. In some examples, the resistance domain 644 may include a plurality of resistance layers, or the enzyme domain 642 may include a plurality of enzyme layers.

[0134] The diffusion resistance domain 644 may include a semipermeable membrane that controls the flux of oxygen and glucose to the underlying enzyme domain 642. As a result, the upper limit of linearity of glucose measurement is extended to a much higher value than that which is achieved without the diffusion resistance domain 644.

[0135] In some examples, the membrane system 632 may include a bioprotective domain 646. also referred to as a domain or biointerface domain, comprising a base polymer. However, the membrane system 632 of some examples can also include a plurality of domains or layers including, for example, an electrode domain, an interference domain, or a cell disruptive domain, such as described in more detail elsewhere herein and in U.S. Patent Nos. 7,494,465, 8,682,608. and 9,044.199, which are incorporated herein by reference in their entirety'.

[0136] It is to be understood that sensing membranes modified for other sensors, for example, may include fewer or additional layers. For example, in some examples, the membrane system 632 may comprise one electrode layer, one enzyme layer, and two bioprotective layers, but in other examples, the membrane system 632 may comprise one electrode layer, two enzyme layers, and one bioprotective layer. In some examples, the bioprotective layer may be configured to function as the diffusion resistance domain 644Docket No. 4855.146WO1 / / 0958-PCT01and control the flux of the analyte (e.g., glucose) to the underlying membrane layers.

[0137] In some examples, one or more domains of the sensing membranes may be formed from materials such as silicone, polytetrafluoroethylene, polyethylene-co-tetrafluoroethylene, polyolefin, polyester, polycarbonate, biostable polytetrafluoroethylene, homopolymers, copolymers, terpolymers of polyurethanes, polypropylene (PP), polyvinylchloride (PVC), polyvinylidene fluoride (PVDF), polybutylene terephthalate (PBT), polymethylmethacrylate (PMMA), polyether ether ketone (PEEK), polyurethanes, cellulosic polymers, poly(ethylene oxide), polypropylene oxide) and copolymers and blends thereof, polysulfones and block copolymers thereof including, for example, di-block, tri-block, alternating, random and graft copolymers.

[0138] In some examples, the sensing membrane can be deposited on the electroactive surfaces of the electrode material using known thin or thick film techniques (for example, spraying, electro-depositing, dipping, or the like). The sensing membrane located over the working electrode does not have to have the same structure as the sensing membrane located over the reference electrode; for example, the enzyme domain 642 deposited over the working electrode does not necessarily need to be deposited over the reference or counter electrodes.

[0139] Although the examples illustrated in FIGS. 6-7 involve circumferentially extending membrane systems, the membranes described herein may be applied to any planar or non-planar surface, for example, the substrate-based sensor structure of U.S. Pat. No. 6,565,509 to Say et al., which is incorporated by reference.

[0140] In an example in which the analyte sensor 634 is a glucose sensor, glucose analyte can be detected utilizing glucose oxidase. Glucose oxidase reacts with glucose to produce hydrogen peroxide (H2O2). The hydrogen peroxide reacts with the surface of the working electrode, producing two protons (2H+), two electrons (2e ) and one molecule of oxygen (O2). This produces an electronic current that may be detected by the sensor electronics 106. The amount of current is a function of the glucose concentration level.Docket No. 4855.146WO1 / / 0958-PCT01A calibration curve may be used to provide an estimated glucose concentration level based on a measured current. The amount of current is also a function of the diffusivity of glucose through the sensor membrane. The glucose diffusivity may change over time, which may cause the sensor glucose sensitivity to change over time, or “drift.’'

[0141] FIG. 8 is a schematic illustration of a circuit 800 that represents the behavior of an example analyte sensor, such as the analyte sensor 634 shown in FIGS. 6-7. As described above, the interaction of hydrogen peroxide (generated from the interaction between glucose analyte and glucose oxidase) and working electrode (WE) 804 produces a voltage differential between the working electrode (WE) 804 and reference electrode (RE) 806 which drives a current. The current may make up all or part of a raw sensor signal that is measured by sensor electronics, such as the sensor electronics 106 of FIGS. 1-2, and used to estimate an analyte concentration (e.g., glucose concentration).

[0142] The circuit 800 also includes a double-layer capacitance (Cdl) 808, which occurs at an interface between the working electrode (WE) 804 and the adjacent membrane (not shown in FIG. 8, see, e.g., FIGS. 6-7 above). The double-layer capacitance (Cdl) may occur at an interface between the working electrode 804 and the adjacent membrane due to the presence of two layers of ions with opposing polarity, as may occur during application of an applied voltage between the working electrode 804 and reference electrode. The equivalent circuit 800 may also include a polarization resistance (Rpol) 810, which may be relatively large, and may be modeled, for example, as a static value (e.g., 100 mega-Ohms), or as a variable quantity that varies as a function of glucose concentration level.

[0143] An estimated analyte concentration may be determined from a raw sensor signal based upon a measured current (or charge flow) through the analyte sensor membrane 812 when a bias potential is applied to the sensor circuit 800. For example, sensor electronics or another suitable computing device can use the raw7sensor signal and a sensitivity of the sensor, which correlates a detected current flow to a glucose concentration level, to generate the estimated analyte concentration.Docket No. 4855.146WO1 / / 0958-PCT01

[0144] With reference to the equivalent circuit 800, when a voltage is applied across the working and reference electrodes 804 and 806, a current may be considered to flow (forward or backward depending on polarity) through the internal electronics of transmitter (represented by R Tx internal) 811; through the reference electrode (RE) 806 and working electrode (WE) 804, which may be designed to have a relatively low resistance; and through the sensor membrane 812 (Rmembr, which is relatively small). Depending on the state of the circuit, current may also flow through, or into, the relatively large polarization resistance 810 (which is indicated as a fixed resistance but may also be a variable resistance that varies with the body's glucose level, where a higher glucose level provides a smaller polarization resistance), or into the double-layer capacitance 808 (i.e., to charge the double-layer membrane capacitor formed at the working electrode 804), or both.

[0145] The impedance (or conductance) of the membrane (Rmembr) 812 is related to electrolyte mobility in the membrane, which is in turn related to glucose diffusivity in the membrane. As the impedance goes down (i.e., conductance goes up, as electrolyte mobility in the membrane 812 goes up), the glucose sensitivity goes up (i.e., a higher glucose sensitivity means that a particular glucose concentration will produce a larger signal in the form of more current or charge flow). Impedance, glucose diffusivity7, and glucose sensitivity are further described in U.S. Patent Publication No.US2012 / 0262298, which is incorporated by reference in its entirety.

[0146] FIG. 9 illustrates an example method 900 for dynamically removing artifacts generated from transitioning between two modes, according to some examples. Although the example method 900 depicts a particular sequence of operations, the sequence may be altered without departing from the scope of the present disclosure. For example, some of the operations depicted may be performed in parallel or in a different sequence that does not materially affect the function of the method 900. In other examples, different components of an example device or system that implements the method 900 may perform functions at substantially the same time or in a specific sequence. For example, in the systems illustrated in Figures 1-8, the functionality described in method 900 can be performed by one or moreDocket No. 4855.146WO1 / / 0958-PCT01components of Figures 1-8 to transition the analyte sensor from one mode to another and process the sensor signals accordingly. Specifically, as shown in Figure 1 and 2, the analyte sensor 104 or the analyte sensor 202 along with the sensor electronics are configured to operate in different modes corresponding to the different analytes being monitored. This transition can include the mode transition described in block 904 of method 900, where the analyte sensor switches from measuring one analyte to another, thereby generating distinct raw sensor signals as illustrated in Figure 9.

[0147] Furthermore, the sensor electronics 106 of Figure 3 and / or the analyte sensor 434 of Figure 4 can process these signals to calculate corrected analyte concentrations, applying corrective models as described in block 908 of method 900.

[0148] FIG. 9 is described as being performed by certain systems or applying certain processes, such as a particular machine learning model, but the processes described herein can be performed by one or more other or the same machine learning models, or a combination thereof. In accordance with the configurations and functionalities detailed in FIG. 9, the machine learning models and pipelines depicted in Figures 6-8 exemplify how various or combined machine learning strategies can be implemented to process and analyze sensor data.

[0149] Figures 6-8 demonstrate the application of different machine learning models, including both supervised and unsupervised learning techniques, which can be utilized individually or in combination to enhance the accuracy of artifact correction and analyte concentration estimation as described in FIG. 9. The descriptions for Figures 6-8 showcase the flexibility and adaptability of the system to employ various machine learning approaches to achieve optimal results in real-time sensor data processing and analysis.

[0150] At block 902, the analyte sensor system (e.g., sensor system 102) operates in a first mode and in a second mode, the analyte sensor system generating a first raw sensor signal indicative of a first analyte concentration of a host when operated in the first mode and a second raw sensor signal indicative of a second analyte when operated in the second mode.Docket No. 4855.146WO1 / / 0958-PCT01

[0151] Although examples described herein describe a sensor (e.g., analyte sensor 202), senor sensor electronics (e.g., sensor electronics 106), or a sensor system (e.g., sensor system 102) performing certain steps (e.g., the sensor performing a particular feature), it is appreciated that other parts of the system can perform such steps (e g., the sensor electronics).

[0152] The analyte sensor system 102 can be capable of operating in two separate modes. Each mode is configured to measure a different type of analyte within the host's body. This dual-mode functionality allows the sensor to toggle between different types of biological or chemical measurements, depending on the mode it is currently operating in.

[0153] The first mode is configured for measuring a first analyte in the host's body. The first analyte can include glucose. In some examples, the first analyte includes any one of glucose, lactate, ketones, aptamer, potassium, oxygen, etc. When the sensor 104 operates in the first mode, the analyte sensor 104 generates a first raw sensor signal.

[0154] The first raw sensor signal can be directly indicative of the concentration of the first analyte (e g., glucose). The signal can be an electrical current, voltage, or other measurable response that varies proportionally with the concentration of the first analyte (e.g., glucose) in the interstitial fluid or blood.

[0155] The analyte sensor 104 can apply specific electrode configurations, use of particular reagents or enzymes that react with glucose, or specific voltage settings that optimize the detection of glucose-related changes.

[0156] The analyte sensor system 102 can operate in the first mode for a first period of time. The analyte sensor system 102 can transition from the first mode to a second mode in which a second analyte signal is measured for a second period of time. The second period of time can be less than the first period of time. For example, the analyte sensor 104 may measure a glucose signal in the first mode, switch to the second mode to measure an oxygen signal, and switch back to the first mode to again measure the glucose signal. The analyte sensor 104 may operate in the second mode to measure the oxygen signal for a short period of time, less than the period of time in which the analyte sensor 104 measures the glucose signal in the first mode.Docket No. 4855.146WO1 / / 0958-PCT01Thus, the analyte sensor 104 may switch from operating in the first mode, to the second mode, and back to the first mode between consecutive glucose measurements. This helps to ensure that glucose measurements determined based on the glucose signal measured in the first mode are not missed. In other examples, the first period of time and the second period of time are the same, or the second period of time may be longer than the first period of time. Additionally, while examples are described herein in which the analyte sensor system 102 only measures a signal from one analyte in a first mode or a second mode at one time, examples are contemplated in which the analyte sensor system 102 measures multiple analyte signals at the same time. For example, the analyte sensor system 102 can measure multiple analyte signals at one time in a first mode, switch to a second mode, and switch back to the first mode.

[0157] The capability to switch sensor modes on a scheduled basis is crucial for systems designed to monitor multiple analytes over time, particularly in medical or environmental contexts where different measurements may be needed at different times. Scheduled switching allows for automated transitions between modes according to a predefined timetable, optimizing the sensor's utility by gathering diverse data without manual intervention. This can be implemented in various ways depending on the desired frequency and precision of data collection.

[0158] In some examples, the sensor electronics 106 performs time-based switching of modes, where the sensor mode is changed at regular intervals such as every' 30 minutes, 1 hour, 1 day, or 1 week, etc. This approach is particularly useful in continuous monitoring systems where conditions are expected to vary predictably over time. For instance, a biomedical sensor may switch between glucose and lactate (or other analyte) monitoring throughout the day to manage diabetes while also assessing oxygen. This type of scheduling ensures that data is collected consistently over time, providing comprehensive monitoring without requiring user input at each switch.

[0159] The frequency of switching can be tailored based on the biological or environmental rhythms relevant to the analytes being monitored. ForDocket No. 4855.146WO1 / / 0958-PCT01example, hourly measurements might be used in a setting where patient conditions can change rapidly, while daily measurements might suffice for tracking slower metabolic changes in a home health monitoring scenario.

[0160] Switching can be event-triggered, where changes in the monitored environment or the condition of the subject prompt a change in the sensor's mode. For instance, if a sensor detects an abnormality in glucose levels, the sensor 104 can automatically switch modes to monitor other critical parameters like ketones, lactate, etc. to provide deeper insights into the patient's metabolic state. As another example, an error condition (e.g.. progressive sensor decline, end of life, or other error) may be detected in the glucose sensing. In response, the mode can be switched to monitoring oxygen to confirm whether the error condition is actually occurring. This type of switching is adaptive and can provide targeted data following specific triggers, which enhances the responsiveness of the monitoring system to emerging conditions.

[0161] In some examples, the analyte sensor system 102 can switch modes via an algorithm or model where data from the sensor itself or external sources influence the switching schedule. These algorithms can analyze ongoing data and determine the optimal times for switching based on predictive models or detected trends. For instance, a sensor system 102 may learn from past data that certain environmental conditions predict changes in analyte concentrations and adjust its mode switching schedule to anticipate these changes.

[0162] The switching between these two modes can be controlled by the sensor system 102, which might adjust parameters such as bias condition. This bias condition can include an electrical setting such as voltage or current that dictates the sensor’s chemical reactivity7and physical properties. By adjusting the bias condition, the sensor's electrochemical environment is altered, enabling the sensor to become selectively sensitive to different analytes or to optimize its performance for different measurement tasks.

[0163] The transition from one mode to another involves applying a specific bias condition that either enhances or suppresses certain electrochemical reactions at the sensor's electrodes, thus switching theDocket No. 4855.146WO1 / / 0958-PCT01sensor’s focus between different analytes without needing physical adjustments or replacement of the sensor itself. This method allows for rapid and flexible switching between modes, making it highly effective for multianalyte monitoring in real-time applications.

[0164] In some examples, switching is based on a bias voltage including changing the electric potential applied to the sensor electrode(s) to optimize the electrochemical environment for detecting different analytes, a chemical environment by altering the chemical composition at the sensor site (e.g., injecting or removing certain chemicals or enzymes that are selective to either glucose or oxygen), or mechanical or physical changes to the sensor’s exposure area or orientation.

[0165] In some examples, a bias voltage can be used to determine the sensor's response to each analyte. By flipping the bias from a first voltage (e.g., between -50 to -10, -10 to -5, -5 to -1, -1 to -.5, -.5 to -.2, -.2 to 0 V) to a second voltage (e.g., 0 to 0.3, 0.3 to 0.6, 0.6 to 1, 1 to 5, 5 to 10, 10 to 50 V), the sensor 104 switches operational modes between measuring a first analyte (e.g.. glucose) in the first mode to a second analyte (e.g.. oxygen) in the second mode.

[0166] In some examples, the bias being is being applied between the working electrode and the reference electrode. In some examples, the bias is applied such that the positive bias occurs when the working electrode is at a higher potential than the reference electrode, and a negative bias is being applied when the working electrode is at a lower potential than the reference electrode.

[0167] At a bias of a first voltage, the sensor environment is optimized for the reduction reactions necessary to measure oxygen levels, primarily involving the reduction of oxygen to water or hydroxide ions. Conversely, when the bias is increased to a second voltage, the sensor's electrochemical conditions favor the oxidation of glucose to gluconolactone, thereby enabling glucose concentration measurement.

[0168] In the example of FIG. 10, the analyte system controller 1020 controls the mode of the analyte sensor. The analyte system 102 controllerDocket No. 4855.146WO1 / / 0958-PCT01can configure the analyte sensor to operate at different biases, such as between the first voltage to the second voltage.

[0169] This transition between bias voltages, however, may not be instantaneous and involves a settling time during which residual signals from the previous measurement can persist and interfere with the new measurement. If these residual signals are not properly removed or accounted for, they can lead to inaccuracies in the sensor outputs.

[0170] The sensor system 102 includes versatile dual-mode capability, where the sensor is capable of transitioning between two distinct modes to measure different analytes.

[0171] FIG. 10 illustrates an architectural diagram for the removal of artifacts induced from switching modes in an analyte sensor system 102, according to some examples. The architectural diagram of FIG. 10 shows an analyte system 1010 that includes an analyte sensor 1016 and sensor electronics 1018. The analyte system 1010 can be used to implement one or more aspects of the method of FIG. 9. The sensor electronics 1018 includes an analyte system controller 1020, a prospective artifact removal model 1012, and an analyte estimator 1014.

[0172] The analyte system controller 1020, prospective artifact removal model 1012, and analyte estimator 1014 described in the current example functionally align with the components and systems outlined in earlier figures, such as those depicted in Figures 1-4. In these earlier examples, the analyte system controller similarly orchestrates the operational dynamics of the sensors, managing the transition between different measurement modes and ensuring the timely collection of data. Likewise, the role of the prospective artifact removal model in these figures involves dynamically correcting the raw sensor signals as they are generated.

[0173] The analyte system 1010 can be used by a host 1008. As shown in FIG. 10 and described in more detail herein, a retrospective artifact removal model 1006 can be used with and / or employed by the sensor electronics 1018 or another system in communication with the sensor electronics 1018. As shown, the retrospective artifact removal model 1006 may receiveDocket No. 4855.146WO1 / / 0958-PCT01historical dual-analyte data 1002 and / or historical analyte data 1004 as inputs.

[0174] In some cases, the retrospective model includes a model that analyzes historical data collected from dual and single analyte sensors over extended periods. This model incorporates a structured framework that can include statistical analysis tools and / or machine learning algorithms capable of identifying and learning from patterns and anomalies in past sensor data. The retrospective model can collect historical data, filter and normalize this data, and apply algorithms such as neural networks or decision trees to generate insights that continuously improve the artifact correction process.

[0175] In some cases, the prospective model operates in real-time, applying immediate corrections to the sensor data as it is generated. This model is structured to quickly identify and adjust for artifacts introduced by mode switching and other immediate sensor disturbances. The model can capture live data streams, use algorithms to identify discrepancies indicative of artifacts, and apply adjustments based on the model’s current understanding of artifact behaviors.

[0176] Referring to FIG. 10, the analyte sensor 1016 of the analyte system 1010 monitors analytes, such as glucose and oxygen (though as noted, the analyte sensor 1016 can measure other analytes). The analyte sensor 1016 is attached to the host 1008 and includes one or more sensors to generate such measurements of different analytes.

[0177] The dual-mode functionality of the analyte sensor 1016 allows the analyte sensor 1016 to switch between modes to measure different analytes as required. The analyte sensor 1016 records (e.g., continuously) data from these interactions with the host's body, which is then processed either locally on the device or transmitted to an external device such as a smartphone or a server.

[0178] Data previously collected from the host and / or from many other hosts using analyte sensors can be stored in the external device to generate a database of historical dual-analyte data. In some examples, the host and / or other hosts are also wearing a single analyte sensor where measurements from such historical analyte data 1004 are also stored.Docket No. 4855.146WO1 / / 0958-PCT01

[0179] Referring back to FIG. 9, at block 904, the sensor electronics 106 transitions the analyte sensor 104 from operating in the second mode to operating in the first mode. The second mode switches the focus to another analyte, such as oxygen levels in interstitial fluid or in the blood of the host. In the second mode, the sensor generates a second raw sensor signal that is indicative of the second analyte's concentration (e.g., oxygen concentration).

[0180] For oxygen, the sensor 104 can measure the reduction or oxidation currents that occur when oxygen interacts with the sensor's electrochemical configuration. Like the first mode, the second mode may require different electrode materials, catalysts, or electronic settings tailored to capture oxygen-specific interactions accurately.

[0181] The analyte sensor 104 can be transitioned from the second mode to the first mode in the same and / or similar manner as the analyte sensor 104 was transition from the first mode to the second mode. For example, the applied bias voltage can be flipped from a second voltage (e.g., 0 to 0.3, 0.3 to 0.6, 0.6 to 1. 1 to 5, 5 to 10, 10 to 50 V) to the first voltage (e.g., between -50 to -10, -10 to -5. -5 to -1, -1 to -.5, -.5 to -.2. -.2 to 0 V) to switch operational modes between the second mode and the first mode.

[0182] In some examples, the sensor 104 transitions from the second mode, where the sensor 104 measures an analyte (e.g., oxygen) other than glucose, back to a first mode dedicated to measuring glucose concentration in the host's body.

[0183] This transition can involve a shift in the sensor's operational focus and output — from generating a raw sensor signal that indicates the concentration of a second analyte, back to producing a signal that specifically indicates glucose levels.

[0184] In some examples, the sensor system 102 extends this dual-mode functionality to include oxygen saturation as the second analyte measured by the analyte sensor in its second mode. Here, the transition involves the sensor switching its operational output from a raw sensor signal that indicates oxygen saturation levels back to mode that measures another analyte, such as glucose concentration.Docket No. 4855.146WO1 / / 0958-PCT01

[0185] The examples described herein describe the measurement of a certain analyte in certain modes, such as switching between glucose and oxygen measurements. However, it is appreciated that the modes can measure other types of analytes, such as lactate, ketones, aptamer, potassium, etc.

[0186] In some examples, the analyte sensor system 102 can include a mode to measure Carbon Dioxide (CO2) levels that can be vital in settings such as anesthesia monitoring, emergency medicine, and patient monitoring in critical care, lactate used to evaluate the severity of critical illnesses, such as sepsis and trauma, or to monitor athletic performance, hemoglobin that measures the amount of hemoglobin in the blood, important for diagnosing anemia and other conditions, pH Levels to assess acid-base balance in the body, crucial in critical care settings, and / or the like.

[0187] In some examples, the analyte sensor system 102 can include a mode to measure electrolytes including sodium, potassium, calcium, and chloride levels for managing a variety of metabolic functions and conditions, urea and creatinine for assessing kidney function and renal health, cholesterol and lipids used to evaluate cardiovascular health and the risk of coronary heart disease, or other biomarkers.

[0188] The examples described herein describe two modes between which the analyte sensor system 102 can transition to and from. However, it is appreciated that the analyte sensor system 102 can include more modes, such as a third mode to measure hemoglobin.

[0189] At block 906, the sensor electronics 106 access a first value based at least on the first raw sensor signal taken during a first time period after the transitioning. The sensor electronics 106 may access a first value of the raw sensor signal taken during a first time period. The sensor system 102 acquires the raw sensor signal from a glucose sensor configured to generate the raw sensor signal indicative of an analyte concentration (such as a glucose) of a host. The first value may be based at least on the first raw sensor signal. For example, the first value may correspond to an estimated analyte concentration of the host.Docket No. 4855.146WO1 / / 0958-PCT01

[0190] The sensor electronics 106 can be configured to generate the raw sensor signal that is indicative of an analyte concentration of a host. The analyte sensor 104 and / or the sensor electronics 106 may perform post processing on the raw sensor signal to generate the first value.

[0191] The glucose sensor measures the raw glucose signal indicative of the glucose concentration in the interstitial fluid. The raw glucose signal is the direct, unprocessed output from the sensor, which may not yet have been subjected to any filtering or calibration. In some examples, the glucose sensor and / or the sensor electronics performs one or more processing to convert the raw sensor signal to an estimated analyte concentration to generate a first value of the raw sensor signal.

[0192] The raw sensor signal includes an electrical signal that corresponds to the glucose concentration in the interstitial fluid. The raw sensor signal is represented in units of electrical current or voltage, which are proportional to the glucose levels.

[0193] This raw sensor signal is transmitted from the sensor to the sensor electronics 106. The sensor electronics accesses the raw sensor signal data and logs this data along with a timestamp to indicate the specific time when the measurement was taken. The received data is stored at a memory' of the sensor electronics and / or sent to a connected device (such as a smartphone app) for further processing and analysis. In some examples, the sensor electronics and / or the sensor performs further processing on the raw glucose signals to generate an estimated glucose concentration for the host.

[0194] For example, upon insertion of the sensor, the sensor electronics may begin generating the raw sensor signal. The raw sensor signal is transmitted to the sensor electronics, and the sensor electronics logs the raw sensor signal along with the timestamp in memory.

[0195] At block 908, the sensor electronics 106 generate a corrected first value of the raw sensor signal using corrective signal model configured to correct artifacts in the first raw sensor signal resulting from the transitioning. The corrective signal model may include a prospective model (e.g., the prospective artifact removal model 1012) and / or the retrospective model (e.g., the retroactive artifact removal model 1006).Docket No. 4855.146WO1 / / 0958-PCT01

[0196] When the analyte sensor 104 measures analytes such as glucose and oxygen and flips between different bias voltages, the analyte sensor can generate transient artifacts, such as disturbances or "spikes," in the output signal. These spikes are abrupt deviations in the measured current or voltage that can occur immediately after the change in the bias condition. These spikes may be referred to herein as “re-break-in spikes.”

[0197] The initial spike includes an immediate reaction to the change in bias voltage. For instance, when the voltage is switched from -0.2 V (optimized for oxygen reduction as measured in the second mode) to 0.6 V (optimized for glucose oxidation as measured in the first mode), the electrical environment of the sensor's surface changes drastically. This sudden change can cause a sharp, temporary' overshoot in the recorded signal.

[0198] This initial spike can be characterized by a rapid increase in signal intensity that does not immediately reflect the true concentration of the new target analyte (glucose, in this scenario). The magnitude and duration of this spike can vary depending on the sensor design, the materials used, and the surrounding electrochemical environment.

[0199] After the initial spike, the signal can undergo a decay phase where the signal gradually returns to a steady state that accurately reflects the concentration of the new analyte. This return can be slower than the spike and may exhibit a tailing effect where the signal decreases exponentially back to the baseline level.

[0200] The rate of decay and how quickly the signal stabilizes can depend on various factors, including the sensor’s response time, the diffusivity of the analytes, and the efficiency of the electrochemical reactions at the new bias voltage.

[0201] The presence of spikes can significantly compromise the accuracy of the sensor unless properly managed. During the transition period marked by the initial spike and the subsequent return to baseline, the sensor outputs can be misleading, reflecting electrical noise and transitional artifacts rather than true analyte levels.Docket No. 4855.146WO1 / / 0958-PCT01

[0202] To mitigate these effects, the sensor electronics 106 can apply one or more algorithms, correction signals, and / or machine learning models that specifically address these spikes. For instance, the sensor electronics can filter out the spike from the actual sensor reading, using techniques such as digital smoothing filters, spike clipping, or adaptive thresholding.

[0203] As shown in FIG. 10, the sensor electronics 106 can apply a prospective artifact removal model 1012 that continuously gathers data from the analyte sensor and applies real-time corrections based on patterns and insights, such as those derived from previous retrospective artifact removal modeling.

[0204] The sensor 104 collects raw signals that may include various artifacts, including spikes from changes in measurement conditions or sensor bias. These artifacts are analyzed using a model that has been developed and refined, such as through retrospective analysis via the retrospective artifact removal model 1006 (as further described herein), where historical data is used to identify patterns of artifact characteristics for their removal.

[0205] In order to accurately model and mitigate the effects of artifacts, the sensor electronics 106 first identify and characterize the typical patterns of these artifacts, which occur when the sensor switches modes, causing an immediate spike, followed by a slower return to the baseline or correct signal measurement.

[0206] The retrospective artifact removal model 1006 can gather extensive data, such as historical dual-analyte data 1002 and historical analyte data 1004, on how the sensor behaves during and after the transition between different measurement modes. This data collection involves repeatedly inducing mode changes to provoke the artifacts and record their manifestations.

[0207] By analyzing this data, the model 1006 can define amplitude and duration of the initial spike, as well as the exponential decay rate as the signal returns to its stable state. These characteristics might include the height of the spike relative to the baseline, the time it takes for the spike to peak, and the duration and shape of the return to baseline.Docket No. 4855.146WO1 / / 0958-PCT01

[0208] Once these patterns are established, such as by the retrospective artifact removal model 1006, the prospective artifact removal model then uses this information to predict and correct similar artifacts in real-time sensor data. When the sensor switches modes and a spike is detected, the model generates a corrective adjustment based on the previously identified patterns.

[0209] For example, if it is known that spikes typically overshoot the baseline by a certain percentage and return to stability within a specific timeframe, the model 1006 can proactively adjust the sensor output when the mode change occurs, reducing the spike's impact and hastening the return to accurate readings.

[0210] Although examples herein describe spike or decay characteristics of artifacts, it is appreciated that features described herein can apply to other characteristics of the artifact. Characteristics of artifacts can include transient spikes which can include abrupt increases or decreases in sensor output that occur immediately after a change in mode. The artifacts can be caused by the sensor's initial reaction to new electrical or chemical conditions, such as a change in bias voltage or sensor environment.

[0211] Characteristics of artifacts can include settling overshoots, where after a mode change, the sensor output may temporarily exceed the stable measurement level before settling back. This overshoot can be seen as a spike that goes beyond the normal range before returning to a typical response level.

[0212] Characteristics of artifacts can include a ring down, which is an artifact that appears as oscillations in the sensor output following a mode switch. The signal may bounce above and below the equilibrium state several times due to the system’s attempt to stabilize under new conditions, similar to the damping oscillations seen in mechanical systems.

[0213] Characteristics of artifacts can include a baseline shift in the baseline level of the sensor output after a change in mode, which may not return to the original baseline. This can be caused by a permanent alteration in the sensor’s chemical or physical properties due to the new operational mode.Docket No. 4855.146WO1 / / 0958-PCT01

[0214] Characteristics of artifacts can include signal attenuation which includes a reduction in the amplitude of the sensor's output signal immediately following a switch in modes, often due to the sensor or its environment not yet being fully adapted to the new measurement conditions.

[0215] Characteristics of artifacts can include mode interaction artifacts where remnants of the previous mode’s conditions affect measurements in the new mode. For example, residual chemicals from a prior measurement phase may temporarily interact with elements of the new phase, distorting readings.

[0216] Characteristics of artifacts can include thermal equilibration effects that involve significant shifts in temperature or power application may lead to temporary thermal artifacts as the sensor comes to thermal equilibrium with its surroundings.

[0217] Characteristics of artifacts can include electrochemical memory when a sensor retains some memory of the previous analyte or condition due to incomplete clearing or washing of the sensor surface, which can lead to erroneous readings post-transition.

[0218] Characteristics of artifacts can include sensor reconditioning delay that includes a time required for a sensor to recondition itself to a new analyte or a new set of conditions after a mode change, during which the data might be unreliable.

[0219] Characteristics of artifacts can include noise increase where changes in operational modes can sometimes lead to an increase in electrical or mechanical noise, especially if the new mode operates under different frequencies, voltages, or pressures.

[0220] Incorporating activity-specific artifact models into an analyte sensor system 102 significantly enhances its accuracy and responsiveness by tailoring the handling of artifacts according to the physiological state or activity of the host . Different activities such as resting (steady state), eating, or exercising can impact the physiological parameters being monitored, leading to variations in artifact characteristics like spikes and decay rates.

[0221] First, the sensor electronics 106 need to reliably detect the host's current activity, which can significantly influence the readings of variousDocket No. 4855.146WO1 / / 0958-PCT01analytes. This detection can be achieved using auxiliary sensors or by¬ analyzing changes in the primary sensor signals, such as glucose levels where sharp increases in glucose levels might indicate that the host has recently eaten, prompting the system to switch to a post-meal artifact model that accounts for rapid changes in glucose. The sensor electronics can use temperature sensor readings where a rise in skin temperature detected bytemperature sensors could indicate a hot environment or activities like taking a shower, which might affect the sensor's physical properties and hence its readings. The sensor electronics can use heart rate monitors where an increase in heart rate, as detected by wearable devices, often correlates with physical activity. A higher heart rate could trigger the system to adopt an exercise-specific artifact model.

[0222] Once an activity- is detected, the sensor electronics 106 can apply a specifically tailored artifact model that adjusts for the expected changes in sensor response due to that activity. For example, the sensor electronics can apply a steady state model where when the host is at rest, artifact disturbances may be minimal and predictable, a post-meal model where after eating, glucose sensors, in particular, may experience rapid fluctuations, and an exercise model where physical activity can cause physiological changes that affect sensor readings, such as increased perspiration, changes in blood flow, or rapid fluctuations in analytes like lactate.

[0223] The system can dynamically select and apply the appropriate artifact correction model based on the detected activity. These models are trained by both historical data collected during similar activities and real-time data, allowing the system to leam and improve its predictions over time.

[0224] This approach not only- compensates for immediate artifacts but also adapts to changes in how activities affect the sensor's behavior over longterm use, such as changes in the host’s fitness level or metabolic responses.

[0225] The historical databases can include various types of historical data that inform the modeling process. For example, the historical databases can include historical data of the host. The historical database can include data of previously collected measurements from the same host .Docket No. 4855.146WO1 / / 0958-PCT01

[0226] By analyzing this personal historical data, the sensor electronics 106 can identify individual-specific patterns and trends in sensor artifacts that occur due to personal physiological responses, habits, or environmental factors. For example, if a particular host consistently shows a specific spike pattern after meals, the sensor electronics can learn to anticipate and correct this specific artifact more effectively. The use of individual historical data helps in customizing the artifact model to the host’s unique biological responses, enhancing the personalization of the monitoring system.

[0227] In some examples, the historical database can include historical data of other hosts. General historical data from a broader host base provides a comprehensive overview of common artifact patterns observed across different individuals. This data is invaluable for developing robust baseline models that are effective across the general population. By understanding how sensors react under various conditions for a wide range of hosts, the sensor electronics 106 can establish generalized artifact correction protocols that are widely applicable, ensuring the system performs well for new hosts or under unencountered conditions.

[0228] In some examples, the historical database can include historical data of hosts with similar physiology. Data segmented according to specific physiological profiles, such as age, gender, medical condition, or lifestyle, allows the sensor electronics 106 to tailor artifact correction models to specific subgroups.

[0229] For instance, diabetic patients might experience different sensor artifacts compared to non-diabetic individuals due to variations in glucose fluctuation patterns. Similarly, athletes might show different physiological responses during exercise compared to non-athletes. By segmenting the historical data in this way, the system can apply more precisely targeted models that are optimized for hosts with similar phy siological characteristics.

[0230] In some examples, the sensor electronics 106 can use machine learning (ML) models modeling such complex phenomena in sensor data, including identifying and correcting artifacts such as spikes and decay phases in analyte sensors. By leveraging ML models, the sensor electronicsDocket No. 4855.146WO1 / / 0958-PCT01can learn from data to predict and mitigate undesired signal behaviors, enhancing the accuracy and reliability of sensor outputs. For example, the prospective model and / or the retrospective model may be a trained ML model.

[0231] In some examples, the sensor electronics 106 apply regression analysis which can be used to predict a continuous outcome. The regression analysis can be applied to predict the magnitude and duration of spikes based on various input features such as previous spike characteristics, sensor mode, and external conditions. For instance, linear regression could be used to establish a predictive model, whereas the sensor electronics may apply polynomial regression or ridge / lasso regression to handle multicollinearity and prevent overfitting.

[0232] In some examples, the sensor electronics 106 apply neural networks, such as deep learning models, for more complex datasets where the relationships between input features and target outcomes are non-linear and involve high-dimensional interactions. These networks could learn to predict how the spike will evolve over time after a mode switch and determine the optimal adjustments needed to correct it in real-time.

[0233] Once trained and validated by the retrospective artifact removal model, the prospective artifact removal model can be implemented within the analyte sensor to operate in real-time, predicting and correcting spikes as they occur. Moreover, to maintain accuracy as new data are collected or when the sensor is used under new conditions, the sensor electronics 106 can apply adaptive learning enabling the ML model to update its parameters continually as new data comes in, ensuring the model stays relevant and accurate over time.

[0234] The retrospective model provides a framework for understanding typical artifact behaviors, which is then applied prospectively to new sensor data as it is collected. The sensor electronics 106 generate a corrective signal that can be used to adjust the incoming data in real-time, effectively minimizing the impact of these artifacts before the data is used for further analysis or decision-making. By leveraging past insights, the prospective model enhances the accuracy and reliability of real-time analyteDocket No. 4855.146WO1 / / 0958-PCT01measurement, ensuring that the data reflects true analyte levels rather than transient disturbances or sensor biases.

[0235] Referring back to FIG. 9, at block 910, the sensor electronics 106 generate an estimated first analyte concentration, such as by using the analyte estimator 1014, based on the corrected first value of the first raw sensor signal. The sensor electronics translate the sensor's raw and now corrected electrical or chemical outputs into meaningful biological or chemical information that can be used for monitoring, diagnostics, or therapeutic management. The analyte estimator 1014 may include at least one data processor and / or a memory for storing operations performed by the at least one data processor.

[0236] Prior to block 910, the first raw sensor signal, which is an initial electrical or chemical reading from the sensor when detecting the first analyte (such as glucose), has undergone correction procedures to remove any artifacts that may have been introduced during the sensing process. These artifacts can include spikes or abnormal patterns resulting from mode changes, sensor drift, or external interference (as further described herein).

[0237] Once the first raw sensor signal is corrected, the sensor electronics 106 proceed to estimate the concentration of the first analyte based at least on this refined data, such as by converting the corrected signal into an actual or estimated concentration value using a conversion function (which may be based on calibration information developed and / or obtained prior to sensor insertion). The conversion function correlates specific sensor signal values (such as voltage, current, or optical properties) with known concentrations of the analyte, allowing the system to interpolate or extrapolate the concentration of the analyte in the sample.

[0238] The corrected signal value can be mapped against the calibration information to find the corresponding analyte concentration. In some examples, especially those using ML techniques, the models may predict analyte concentration directly from the corrected signals. These models can account for nonlinearities and complex relationships between the sensor output and the analyte concentration, which may be influenced byDocket No. 4855.146WO1 / / 0958-PCT01environmental factors, the presence of other substances, or changes in the sensor's sensitivity over time.

[0239] After the detecting of mode change-induced artifacts and / or estimating the analyte concentration after such correction of the signal, the sensor electronics 106 can perform one or more actions. The system can send notifications or alerts to the host, informing them that the sensor is experiencing such mode change-induced artifacts or that the glucose reading is a corrected signal. These alerts could include visual, auditory', or haptic notifications to ensure that the host is promptly aware of the issue.

[0240] The sensor electronics 106 can end the sensor session and / or may not enable the host to start a new session until a new sensor is inserted. The host may be required to insert a new sensor to begin a new session.

[0241] The system may prompt the host to recalibrate the sensor, such as by performing a manual calibration with a blood glucose meter to restore accuracy and adjust for any detected mode change-induced artifacts. If the mode change-induced artifacts are significant or persistent, the system could recommend or remind the host to replace the glucose sensor, ensuring that the host continues to receive accurate glucose readings.

[0242] The system can log the data associated with the mode change-induced artifact, including glucose and oxygen signals, which can be later analyzed to provide insights into the cause of the mode change-induced artifacts and to improve future sensor designs.

[0243] The system may initiate diagnostic checks to determine if the mode change-induced artifacts are due to external factors, such as incorrect sensor placement or environmental conditions, which can help isolate the problem and provide solutions. The system may provide enhanced reporting features to track the performance trends of the sensor over time, helping hosts and healthcare providers to monitor and assess the sensor's long-term reliability’ .

[0244] The system may be integrated with healthcare provider platforms, automatically sharing information about mode change-induced artifacts and allowing healthcare professionals to review the data and make informed recommendations. In some examples, the healthcare provider platforms canDocket No. 4855.146WO1 / / 0958-PCT01automatically schedule an appointment or order a replacement as a result of the mode change-induced artifacts.

[0245] The system may perform self-diagnostics and apply software updates to optimize sensor performance and address potential issues, ensuring that the monitoring system remains up-to-date and effective. The system may offer preventive measures, such as guidance on optimal sensor placement, avoiding certain activities that may accelerate mode change-induced artifacts, or an indication of an unreliable glucose reading.

[0246] The system can adjust its monitoring algorithms to account for the detected mode change-induced artifacts by modifying the way glucose readings are interpreted based on the current performance of the sensor. For instance, if performance degradation causes the sensor to produce consistently lower or erratic glucose readings, the system can apply correction factors or adjustments to normalize the data. The sensor system 102 can use historical calibration data or real-time corrections based on observed deviations in the glucose and / or oxygen signals. The algorithms can be fine-tuned to filter out noise and minimize the impact of degradation, thereby enhancing the reliability of glucose measurements.

[0247] Additionally, the system can adapt its algorithms by incorporating adaptive filtering techniques or dynamic models that adjust to changing sensor conditions. For example, the system can use predictive modeling to estimate and correct for anticipated errors based on the current mode change-induced artifact pattern. By continuously updating these algorithms based on ongoing performance metrics, the system can provide more accurate glucose readings and maintain effective diabetes management, even as the sensor’s performance evolves over time.

[0248] In some examples, one advantage of removing artifacts from the sensor signal before estimating the analyte concentration is that some examples of the system described herein may allow for the consistent use of the same estimation process or model, regardless of the presence or ty pe of artifact. This consistency in the analytical process may offer one or more several key benefits, such as enhanced accuracy, reliability, and simplicity in system design.Docket No. 4855.146WO1 / / 0958-PCT01

[0249] By implementing artifact correction prior to the concentration estimation, the model or process used to convert sensor readings into analyte concentration estimates in some examples can operate under the assumption that the input data are clean and free of distortions. This approach may negate the need to develop and maintain multiple models or algorithms tailored to different artifact conditions. Instead, a single, robust estimation model can be employed that focuses solely on the relationship between the sensor signal and the analyte concentration, without having to account for variable noise factors like spikes, drifts, or other transient changes.

[0250] In some cases, removing artifacts prior to concentration estimation may simplify the overall system design. This simplification may be evident in the training phase of machine learning models, where the focus can be on understanding the true dynamics of analyte changes rather than compensating for anomalies in the data. Moreover, the examples herein may reduce the complexity of the algorithm needed for the estimation, as it may not need to be as adaptive or flexible to changing artifact patterns, which might vary over time or across different sensor batches or conditions.

[0251] In some cases, the accuracy of the analyte concentration estimates is enhanced when the artifacts may be reduced or in some examples removed. Artifacts can introduce substantial errors in concentration estimation, especially if their characteristics overlap with the true analyte signal changes. For example, a spike caused by a sensor mode change could be misinterpreted as a sudden rise in analyte concentration, leading to incorrect medical decisions if not corrected. By removing these artifacts in some cases, the system may ensure that such misinterpretations are minimized, leading to more reliable and precise analyte measurements.

[0252] Another benefit to some examples described herein is the uniform application of the estimation process across various host conditions and environments. Whether in a clinical setting, at home, or in mobile health monitoring, the same estimation model can be applied confidently, knowing that the preliminary artifact correction step has standardized the input data. This uniformity' can be important for regulatory approval and host trust, as itDocket No. 4855.146WO1 / / 0958-PCT01guarantees that the system's performance remains stable and predictable, irrespective of external conditions.

[0253] Systems and methods described herein include training a machine learning network, such as training to identify artifact characteristics. The machine learning network can be trained to identify artifact characteristics due to mode changes in analyte sensors and to generate corrective signals. The machine learning algorithm can be trained using historical information that include historical dual analyte sensor measurements and a single analyte sensor measurement (as further described herein).

[0254] Training of models, such as artificial intelligence models, is necessarily rooted in computer technology, and improves modeling technology by using training data to train such models and thereafter applying the models to new inputs to make inferences on the new inputs. Here, the new inputs can be real-time measurements of a host by a dual analyte sensor. The trained machine learning model can identify and remove artifacts resulting from the change in modes of the dual analyte sensor.

[0255] Such training involves complex processing that typically requires a lot of processor computing and extended periods of time with large training data sets, which are typically performed by massive serversystems. Training of models can require logistic regression and / or forward / backward propagating of training data that can include input data and expected output values that are used to adjust parameters of the models. Such training is the framework of machine learning algorithms that enable the models to be applied to new and unseen data (such as new analyte data) and make predictions that the model was trained for based on the weights or scores that were adjusted during training. Such training of the machine learning models described herein reduces false positives and increases the performance.

[0256] FIG. 11 illustrates measurements (e.g., the glucose measurement) taken from a dual analyte sensor system and a single analyte sensor system, according to some examples. The dual analyte sensor system is capable of measuring two different ty pes of analytes (such as glucose and oxygen),Docket No. 4855.146WO1 / / 0958-PCT01whereas the single analyte sensor system is dedicated to measuring only one analyte, for example, glucose. FIG. 11 illustrates the an analyte measurement (e.g., glucose measurement) taken from both the dual analyte sensor system and the single analyte sensor system.

[0257] The dual analyte sensor system's ability to switch between measuring two different analytes (e.g., glucose and oxygen) introduces complexities in its output. Each time the sensor switches modes — to measure glucose after measuring oxygen or vice versa — the sensor adjusts to the new analyte's chemical and physical properties. This adjustment process isn't instantaneous and can result in transient disturbances in the dual analyte sensor signal 1102, manifested as spikes and subsequent decay phases.

[0258] These are abrupt, temporary increases or decreases in the sensor reading that occur right after the sensor switches modes. For example, if the sensor switches from oxygen to glucose measurement, the change in electrical bias or chemical reagents needed for the new measurement can cause an initial overreaction or underreaction by the sensor, leading to a spike.

[0259] After the spike, the sensor signal does not immediately stabilize but rather exhibits a decay phase where it gradually returns to a level that accurately reflects the true concentration of the new analyte. This decay can be exponential and represents the sensor's return to equilibrium after the disruption caused by the mode switch.

[0260] The single analyte sensor system (e.g., dedicated solely to measuring glucose) does not need to switch modes and hence maintains a consistent measurement without mode-switching artifacts. Consequently, the single analyte sensor signal 1104 does not show the spikes and decays seen in the dual sensor's output. Instead, this sensor provides a stable and continuous indication of the glucose levels, which can be considered as closer to the "true" glucose readings in the absence of mode-switching disturbances.

[0261] There could be an offset 1106 between the two signals, where the magnitudes of the signals are offset, or there is a temporal offset such asDocket No. 4855.146WO1 / / 0958-PCT01where the dual sensor's readings may lag behind or lead the single sensor's readings, which can be attributed to several factors.

[0262] Each sensor type may have different calibration settings or inherent sensitivities, affecting how quickly and accurately they respond to changes in glucose concentration. The dual sensor’s need to stabilize after each mode switch may delay its response to changes in glucose levels compared to the single sensor, which is continuously attuned to detecting glucose without interruption. The artifacts introduced by mode switching in the dual sensor (spikes and decay) can temporarily distort the readings, making them appear higher or lower than the actual glucose levels until the sensor stabilizes.

[0263] The offset in magnitude between readings from a dual analyte sensor system and a single analyte sensor system can be caused by calibration differences where each sensor type may be calibrated using different standards or protocols, which can lead to variations in how7readings are scaled and interpreted. The offset can be caused by sensor sensitivity and specificity where the sensitivity of each sensor to glucose might differ due to the materials used in the sensor, the design of the electrode, or the presence of additional components that interact with glucose and the dual sensor, for example, may have reduced sensitivity due to the need to accommodate components for detecting another analyte (like oxygen), which could affect its ability to measure glucose concentrations as precisely as the single analyte sensor system.

[0264] The offset can be caused by physical and chemical interferences where for example, in the case of a dual analyte sensor system, the need to switch between different measurement modes may introduce additional chemical or physical interferences that are not present in the single analyte sensor system. These interferences can temporarily or permanently alter the sensor's response characteristics, leading to a measurable offset in the output.

[0265] The offset can be caused by wear and tear, where over time, sensors can degrade differently based on their construction and usage. For instance, a dual sensor that frequently switches modes may experience more rapidDocket No. 4855.146WO1 / / 0958-PCT01wear and tear than a single sensor dedicated to one type of measurement. This degradation can affect the accuracy and magnitude of the readings.

[0266] In some cases, an impact of frequent mode switching in dual sensors is on the battery budget. Switching potentials between two modes requires additional energy, which can lead to a faster depletion of the battery. This increased energy consumption not only can necessitate more frequent battery replacements or charges but also can limit the practical usability of the device in scenarios where long battery life is crucial, such as in continuous health monitoring or remote environmental sensing.

[0267] The offset can also be caused by environmental factors where variations in environmental conditions such as temperature, humidity', and the presence of other chemical substances can differently affect sensors in different ways depending on their design and protective measures. If one sensor is more susceptible to environmental changes than the other, this could result in a consistent difference in their readings.

[0268] FIG. 12 illustrates baseline aligning of the dual sensor signal and the single analyte sensor signal, according to some examples. FIG. 12 illustrates the two signals — the dual analyte sensor signal 1202 and the single analyte sensor 1204 — with the offset removed. This adjustment ensures that the comparison between the two becomes more direct and meaningful, allowing for a more accurate analysis of their performance and any remaining discrepancies due solely to sensor behavior rather than calibration or setup errors.

[0269] To remove the offset where the dual sensor signal was consistently higher than the single analyte sensor signal, the sensor electronics 106 can employ one or more processes, such as a simple linear correction factor by either adding or subtracting a constant value or scaling the output by a factor determined through calibration tests to match the single sensor's levels, regression analysis to find the best fit line or curve that aligns the dual sensor data with the single sensor data, algorithmic correction to adjust the readings dynamically, and / or the like.

[0270] Removing the offset allows for an "apples-to-apples" comparison between the two sensors, focusing on how well each sensor tracks actualDocket No. 4855.146WO1 / / 0958-PCT01analyte concentrations under identical conditions without the confounding factor of systematic calibration errors.

[0271] FIG. 13 illustrates the generation of a corrective signal that can be applied to the raw glucose signal to generate a corrected glucose signal without the artifacts, according to some examples. With the offset removed, any remaining differences in the readings between the dual and single sensors would likely be due to the effects of having to switch between different measurement modes in the dual sensor. Identifying these differences can help generate a corrective signal to remove such artifacts from the raw glucose measurements.

[0272] This corrective signal 1302 can be derived by subtracting the dual sensor readings from the single sensor readings after making necessary’ adjustments for any offsets. This approach directly utilizes the differential between the dual sensor's complex, possibly artifact-laden readings and the more stable, focused readings of the single analyte sensor.

[0273] As described, the sensor electronics 106 may subtract the adjusted dual sensor readings from the single sensor readings. The result is a corrective signal that represents the error or noise introduced by the dual sensor’s mode-switching.

[0274] FIG. 14 illustrates the corrected glucose signal 1402 after the application of the corrective signal generated in FIG. 13, according to some examples. This corrective signal can then be applied to the raw data from the dual sensor to bring its readings in line with what the single sensor indicates.

[0275] In some examples, the sensor electronics 106 may use regression analysis to model the relationship between the dual sensor readings and the single sensor readings. This model can predict what the readings from the dual sensor should be based on the single sensor's data. The residuals from this regression (the differences between the predicted values from the dual sensor and the actual readings) can then serve as the corrective signal.

[0276] In some examples, the sensor electronics 106 may apply machine learning models to learn patterns or relationships between the outputs of the dual and single sensors. Models such as neural networks or support vectorDocket No. 4855.146WO1 / / 0958-PCT01machines can be trained on a dataset where inputs are the dual sensor readings, and the target outputs are the single sensor readings. The model’s predictions can then be used to determine the corrective signal by assessing the deviation of the dual sensor's readings from the model's predictions.

[0277] In some examples, the sensor electronics 106 may apply filtering, such as digital filters (e.g., a low-pass filter) to both sets of sensor data to smooth out high-frequency noise (which may represent transient errors or spikes) and then calculate the corrective signal from the smoothed data.

[0278] In some examples, the sensor electronics 106 take multiple readings from both sensors over time, where averaging can be employed. By averaging several consecutive measurements, random noise can be reduced, and a more stable baseline and / or artifact characteristic identification can be established for both sensors. The corrective signal can then be computed from these averaged readings.

[0279] In some examples, the readings can be weighted such that more recent measurements are weighed heavier than the older measurements. For example, in prospective modeling where multiple readings are averaged over time, more recent measurements can be weighted to have a bigger effect on the average than older measurements. Furthermore, in retrospective modeling, older measurements can be weighted lower than more recent measurements to generate a model that is better fit to more recent data.

[0280] In some examples, the readings can be weighted such that more recent measurements are weighed heavier than the older measurements. For example, in prospective modeling where multiple readings are averaged over time, more recent measurements can be weighted to have a bigger effect on the average than older measurements. Additionally, the weighting of sensor readings can also be adjusted based on factors such as the sensor configuration, including specific designs and specifications of the sensor membrane, as well as the manufacturing batch. This allows for a more nuanced approach to data analysis, particularly in a prospective training model, where sensors that share the same configuration, chemistry, and originate from the same manufacturing batch can be given greater weight.Docket No. 4855.146WO1 / / 0958-PCT01

[0281] When utilizing a corrective signal to adjust the readings from a dual sensor, the sensor electronics 106 aligns not only with the true analyte concentration but also with the calibration standards of the dual sensor system. The sensor electronics modifies the corrective signal to match the calibration of the dual sensor by adding back the offset to the corrective signal. This offset might be a constant value across all readings or a variable adjustment based on the magnitude of the readings, the specific analyte concentration, or other contextual factors identified during the calibration process.

[0282] In some examples, the sensor electronics 106 applies a sliding window in either or both prospective and retrospective models. In prospective modeling, a sliding window technique can be used to continuously update the corrective signal based on the most recent data points.

[0283] The sensor electronics 106 can define a window size (the number of recent data points to include) and slide this window across the data stream as new data arrives. For each new data point, the oldest point in the window is discarded, and the newest point is added, maintaining a constant window size.

[0284] As the sensor generates data, the model uses the values within the current window to calculate the corrective signal. This may involve averaging the data points, applying a digital filter, or machine learning model to predict and correct for any artifacts identified within this window.

[0285] In retrospective modeling, the sliding window is used to analyze historical data and refine the model that generates the corrective signal for prospective use. The sensor electronics 106 can process segments of historical data to identify common patterns or anomalies and use this insight to enhance the model's accuracy and reliability.

[0286] By moving the window across a comprehensive dataset, the retrospective model can learn from diverse scenarios and conditions reflected in the historical data. This learning can include identifying the typical duration and amplitude of spikes associated with mode changes, theDocket No. 4855.146WO1 / / 0958-PCT01normal decay patterns following such spikes, and how these patterns vary under different operational conditions.

[0287] The sensor electronics 106 can also use sliding windows to validate the model's effectiveness and to test different window sizes or modeling techniques. By comparing the model's output against known true values within each window, the accuracy of the model can be assessed and optimized before being applied in real-time scenarios.

[0288] FIG. 15 illustrates retrospective artifact removal modeling, according to some examples. FIG. 15 depicting the retrospective artifact removal modeling highlights an advanced approach to refining sensor data accuracy by leveraging historical data from both a single analyte sensor 1516 and a dual analyte sensor 1522 of the analyte system 1510 taking measurements from a host 1508.

[0289] The single analyte sensor system is dedicated to monitoring one specific substance (like glucose), providing a focused and often more stable set of data. In contrast, the dual analyte sensor system can switch between monitoring two different substances (like glucose and oxygen), which introduces complexity due to the need to change operational modes. Both sensors are controlled by an analyte system controller 1520, which is responsible for configuration of the sensors such as initiating measurements at designated times and switching the operational modes of the dual analyte sensor as needed.

[0290] The data collected from both sensors is transmitted to an analyte estimator 1514 of the sensor electronics 1518. This sensor electronics can process the raw data to generate a corrective signal that is used to adjust the data from the dual analyte sensor, correcting any artifacts introduced by the mode switching, such as spikes or unusual decay patterns in the signal.

[0291] The measurements from both sensors are also sent to their respective historical databases — historical dual analyte database 1502 and historical single analyte database 1504. These databases store data over time, providing a rich source of information for post-processing and analysis.Docket No. 4855.146WO1 / / 0958-PCT01

[0292] The retrospective artifact removal modeling 1506 retrieves such historical data to analyze the historical data, and identify patterns, trends, and anomalies that are characteristic of the sensors under various conditions.

[0293] As such, the retrospective model is not limited to using only the data available at the moment or the immediate past (like the last three spikes as in prospective models). Instead, the models can analyze data from entire sessions, and even from past and future sessions, providing a more comprehensive view of the sensor's performance over time.

[0294] Moreover, the retrospective model can leverage data from other sessions beyond the current one and even data from other hosts who have similar physiological profiles or environmental conditions. This capability allows the model to learn from a broader array of scenarios, enhancing its ability to predict and correct artifacts more accurately.

[0295] Moreover, given that retrospective models are not constrained by the need for real-time processing, the retrospective model can utilize more sophisticated algorithms and more substantial computational resources. This setup enables deeper and more complex data analysis, leading to more refined corrections.

[0296] FIG. 16 illustrates prospective artifact removal modeling and realtime retraining, according to some examples. A dual analyte sensor 1616 of the analyte system 1610 is attached to a host 1608. This sensor is capable of monitoring two different analytes, which could include parameters such as glucose and oxygen levels, depending on its design and the needs of the host.

[0297] The operation of this sensor is managed by an analyte system controller 1620 of the sensor electronics 1618.

[0298] The raw measurements from the dual analyte sensor, which may contain inaccuracies due to mode switching or other operational artifacts, are sent to a prospective artifact removal model 1612 to identify and correct errors in real-time. Once the corrected signal is generated, the analyte estimator 1614 uses the cleaned data to estimate the concentration of the analytes more accurately.

[0299] The sensor electronics can continuously retrain the prospective artifact removal model using the training module 1622, such as in real-time.Docket No. 4855.146WO1 / / 0958-PCT01This dynamic learning process involves using the measured signal, the corrected signal, and / or the estimated analyte concentrations as new training data.

[0300] As the model receives this data, it adjusts its parameters and improves its artifact detection and correction algorithms. This ongoing training process allows the model to adapt to changes over time, such as to account for sensor aging, variations in the host’s condition, or environmental factors that might affect sensor readings.

[0301] In some examples, the sensor electronics 106 access a second value of the first raw sensor signal, which is captured after a transition between measurement modes and after a specified first time period. The sensor electronics captures the sensor data once it has had sufficient time to stabilize from any transient effects caused by the mode change, ensuring the data reflects the true state of the first analyte without being influenced by initial disturbances.

[0302] This second value is used to generate an estimated first analyte concentration without applying the previously discussed corrective model. By bypassing the corrective model in this instance, the estimation is based purely on the raw, stabilized sensor data. Advantageously, the same concentration estimation model can be used regardless of whether artifacts are present in the raw sensor signal.

[0303] In some examples, the polarity of the first bias condition can be opposite to that of the second bias condition refers to the electrical charge orientation used to operate the sensor in its different measurement modes. For example, if the first bias condition for measuring one analyte, such as glucose, involves applying a positive voltage, the second bias condition for another analyte, like oxygen, can involve applying a negative voltage, or vice versa.

[0304] By switching the electrical polarity7, the sensor can effectively alter the electrochemical reactions at the electrode surface, facilitating the reduction or oxidation processes necessary for the accurate detection of each specific analyte. This capability7allows the sensor to switch between measuring substances that require fundamentally different electrochemicalDocket No. 4855.146WO1 / / 0958-PCT01conditions for their detection, thereby expanding the versatility and utility of the sensor in various applications.

[0305] In some examples, the sensor electronics 106 measure and correct signals from an analyte sensor by applying a first bias condition to the sensor, setting the electrochemical environment needed for the specific detection of one analyte. Subsequently, this condition is switched to a second bias condition, adjusting the sensor's settings to either measure a different analyte or prepare for another measurement cycle under changed conditions.

[0306] After this transition, the sensor electronics 106 capture a first value of the raw sensor signal within a designated time period following the change in bias conditions. This timing ensures that the signal reflects the new electrochemical setting but is still early enough to capture transient responses. This captured raw signal value is processed through a corrective model, which adjusts for any predictable artifacts or disturbances introduced by the change in bias conditions, ensuring the output is a more accurate representation of the true analyte concentration.

[0307] FIG. 18 is a flowchart depicting a machine-learning pipeline 1800, according to some examples. The machine-learning pipelines 1800 may be used to generate a trained model, for example the trained machine-learning program 1802 of FIG. 18, described herein to perform operations associated with searches and query responses.

[0308] Machine learning may involve using computer algorithms to automatically learn patterns and relationships in data, potentially without the need for explicit programming to do so after the algorithm is trained.Examples of machine learning algorithms may include models based on supervised learning, models based on unsupervised learning, and models based on reinforcement learning models. Supervised learning involves training a model using labeled data to predict an output for new, unseen inputs. Examples of supervised learning algorithms include linear regression, decision trees, and neural networks.

[0309] Unsupervised learning involves training a model on unlabeled data to find hidden patterns and relationships in the data. Examples ofDocket No. 4855.146WO1 / / 0958-PCT01unsupervised learning algorithms include clustering, principal component analysis, and generative models like autoencoders.

[0310] Reinforcement learning involves training a model to make decisions in a dynamic environment by receiving feedback in the form of rewards or penalties. Examples of reinforcement learning algorithms include Q-leaming and policy gradient methods.

[0311] Examples of specific machine learning algorithms that may be deployed, according to some examples, include logistic regression, which is a type of supervised learning algorithm used for binary classification tasks. Logistic regression models the probability of a binary response variable based on one or more predictor variables. Another example ty pe of machine learning algorithm is Naive Bayes, which is another supervised learning algorithm used for classification tasks. Naive Bayes is based on Bayes' theorem and assumes that the predictor variables are independent of each other. Random Forest is another type of supervised learning algorithm used for classification, regression, and other tasks. Random Forest builds a collection of decision trees and combines their outputs to make predictions. Further examples include neural networks which consist of interconnected layers of nodes (or neurons) that process information and make predictions based on the input data. Matrix factorization is another type of machine learning algorithm used for recommender systems and other tasks. Matrix factorization decomposes a matrix into two or more matrices to uncover hidden patterns or relationships in the data. Support Vector Machines (SVM) are a ty pe of supervised learning algorithm used for classification, regression, and other tasks. SVM finds a hyperplane that separates the different classes in the data. Other types of machine learning algorithms include decision trees, k-nearest neighbors, clustering algorithms, and deep learning algorithms such as convolutional neural networks (CNN), recurrent neural networks (RNN), and transformer models. The choice of algorithm depends on the nature of the data, the complexity of the problem, and the performance requirements of the application.

[0312] The performance of machine learning models is typically evaluated on a separate test set of data that was not used during training to ensure thatDocket No. 4855.146WO1 / / 0958-PCT01the model can generalize to new, unseen data. Evaluating the model on a separate test set helps to mitigate the risk of overfitting, a common issue in machine learning where a model learns to perform exceptionally well on the training data but fails to maintain that performance on data it hasn't encountered before. By using a test set, the system obtains a more reliable estimate of the model's real-world performance and its potential effectiveness when deployed in practical applications.

[0313] Although several specific examples of machine learning algorithms are discussed herein, the principles discussed herein can be applied to other machine learning algorithms as well. Deep learning algorithms such as convolutional neural networks, recurrent neural networks, and transformers, as well as more traditional machine learning algorithms like decision trees, random forests, and gradient boosting may be used in various machine learning applications.

[0314] Two example ty pes of problems in machine learning are classification problems and regression problems. Classification problems, also referred to as categorization problems, aim at classifying items into one of several category values (for example, is this object an apple or an orange?). Regression algorithms aim at quantifying some items (for example, by providing a value that is a real number).

[0315] Generating a trained machine-learning program 1802 may include multiple types of phases that form part of the machine-learning pipeline 1800, including for example the following phases 1700 illustrated in FIG.17.• Data collection and preprocessing 1702 may include acquiring and cleaning data to ensure that it is suitable for use in the machine learning model. Data can be gathered from user data and labeled using a machine learning algorithm trained to label data. Data can be generated by applying a machine learning algorithm to identify or generate similar data. This may also include removing duplicates, handling missing values, and converting data into a suitable format.• Feature engineering 1704 may include selecting and transforming the training data 1804 to create features that are useful for predicting theDocket No. 4855.146WO1 / / 0958-PCT01target variable. Feature engineering may include (1) receiving features 1806 (e.g., as structured or labeled data in supervised learning) and / or (2) identifying features 1806 (e.g., unstructured or unlabeled data for unsupervised learning) in training data 1804.• Model selection and training 1706 may include specifying a particular problem or desired response from input data, selecting an appropriate machine learning algorithm, and training it on the preprocessed data. This may further involve splitting the data into training and testing sets, using cross-validation to evaluate the model, and tuning hyperparameters to improve performance. Model selection can be based on factors such as the type of data, problem complexity, computational resources, or desired performance.• Model evaluation 1708 may include evaluating the performance of a trained model (e.g., the trained machine-learning program 1802) on a separate testing dataset. This can help determine if the model is overfitting or underfitting and if it is suitable for deployment.• Prediction 1710: This involves using a trained model (e.g., trained machine-learning program 1802) to generate predictions on new, unseen data.• Validation, refinement or retraining 1712 may include updating a model based on feedback generated from the prediction phase, such as new data or user feedback.• Deployment 1714 may include integrating the trained model (e.g., the trained machine-learning program 1802) into a larger system or application, such as a web service, mobile app, medical device, or loT device. This can involve setting up APIs, building a user interface, and ensuring that the model is scalable and can handle large volumes of data.

[0316] FIG. 18 illustrates two example phases, namely a training phase 1808 (part of the model selection and trainings 1706) and a prediction phase 1810 (part of prediction 1710). Prior to the training phase 1808, feature engineering 1704 is used to identify features 1806. This may include identifying informative, discriminating, and independent features for the effective operation of the trained machine-learning program 1802 in pattern recognition, classification, and regression. In some examples, the trainingDocket No. 4855.146WO1 / / 0958-PCT01data 1804 includes labeled data, which is known data for pre-identified features 1806 and one or more outcomes.

[0317] Each of the features 1806 may be a variable or attribute, such as individual measurable property of a process, article, system, or phenomenon represented by a data set (e.g., the training data 1804). Features 1806 may also be of different types, such as numeric features, strings, vectors, matrices, encodings, and graphs, and may include one or more of content 1812, concepts 1814, attributes 1816, historical data 1818 and / or user data 1820, merely for example. Concept features can include abstract relationships or patterns in data, such as determining a topic of a document or discussion in a chat window between users. Content features include determining a context based on input information, such as determining a context of a user based on user interactions or surrounding environmental factors.

[0318] In training phases 1808, the machine-learning pipeline 1800 uses the training data 1804 to find correlations among the features 1806 that affect a predicted outcome or prediction / inference data 1822.

[0319] With the training data 1804 and the identified features 1806, the trained machine-learning program 1802 is trained during the training phase 1808 during machine-learning program training 1824. The machine-learning program training 1824 appraises values of the features 1806 as they correlate to the training data 1804. The result of the training is the trained machinelearning program 1802 (e.g., a trained or learned model).

[0320] Further, the training phase 1808 may involve machine learning, in which the training data 1804 is structured (e.g., labeled during preprocessing operations), and the trained machine-learning program 1802 implements a relatively simple neural network 1826 capable of performing, for example, classification and clustering operations. In other examples, the training phase 1808 may involve deep learning, in which the training data 1804 is unstructured, and the trained machine-learning program 1802 implements a deep neural network 1826 that is able to perform both feature extraction and classification / clustering operations.Docket No. 4855.146WO1 / / 0958-PCT01

[0321] A neural network 1826 may, in some examples, be generated during the training phase 1808, and implemented within the trained machinelearning program 1802. The neural network 1826 includes a hierarchical (e.g., layered) organization of neurons, with each layer including multiple neurons or nodes. Neurons in the input layer receive the input data, while neurons in the output layer produce the final output of the network. Between the input and output layers, there may be one or more hidden layers, each including multiple neurons.

[0322] Each neuron in the neural network 1826 operationally computes a small function, such as an activation function that takes as input the weighted sum of the outputs of the neurons in the previous layer, as well as a bias term. The output of this function is then passed as input to the neurons in the next layer. If the output of the activation function exceeds a certain threshold, an output is communicated from that neuron (e.g., transmitting neuron) to a connected neuron (e.g., receiving neuron) in successive layers. The connections between neurons have associated weights, which define the influence of the input from a transmitting neuron to a receiving neuron. During the training phase, these weights are adjusted by the learning algorithm to optimize the performance of the network. Different types of neural networks may use different activation functions and learning algorithms, which can affect their performance on different tasks. Overall, the layered organization of neurons and the use of activation functions and weights enable neural networks to model complex relationships between inputs and outputs, and to generalize to new inputs that were not seen during training.

[0323] In some examples, the neural network 1826 may also be one of a number of different types of neural networks or a combination thereof, such as a single-layer feed-forward network, a Multilayer Perceptron (MLP), an Artificial Neural Network (ANN), a Recurrent Neural Network (RNN), a Long Short-Term Memory’ Network (LSTM), a Bidirectional Neural Network, a symmetrically connected neural network, a Deep Belief Network (DBN), a Convolutional Neural Network (CNN), a Generative Adversarial Network (GAN), an Autoencoder Neural Network (AE), a Restricted Boltzmann Machine (RBM), a Hopfield Network, a Self-Organizing MapDocket No. 4855.146WO1 / / 0958-PCT01(SOM), a Radial Basis Function Network (RBFN), a Spiking Neural Network (SNN), a Liquid State Machine (LSM), an Echo State Network (ESN), a Neural Turing Machine (NTM), or a Transformer Network, merely for example.

[0324] In addition to the training phase 1808, a validation phase may be performed evaluated on a separate dataset known as the validation dataset. The validation dataset is used to tune the hyperparameters of a model, such as the learning rate and the regularization parameter. The hyperparameters are adjusted to improve the performance of the model on the validation dataset.

[0325] The neural network 1826 is iteratively trained by adjusting model parameters to minimize a specific loss function or maximize a certain objective. The system can continue to train the neural network 1826 by adjusting parameters based on the output of the validation, refinement, or retraining block 1712, and rerun the prediction 1710 on new or already run training data. The system can employ optimization techniques for these adjustments such as gradient descent algorithms, momentum algorithms, Nesterov Accelerated Gradient (NAG) algorithm, and / or the like. The system can continue to iteratively train the neural network 1826 even after deployment 1714 of the neural network 1826. The neural network 1826 can be continuously trained as new data emerges, such as based on user data or system-generated training data.

[0326] Once a model is fully trained and validated, in a testing phase, the model may be tested on a new dataset that the model has not seen before. The testing dataset is used to evaluate the performance of the model and to ensure that the model has not overfit the training data.

[0327] In prediction phase 1810, the trained machine-learning program 1802 uses the features 1806 for analyzing query data 1828 to generate inferences, outcomes, or predictions, as examples of a prediction / inference data 1822. For example, during prediction phase 1810, the trained machinelearning program 1802 is used to generate an output. Query data 1828 is provided as an input to the trained machine-learning program 1802, and theDocket No. 4855.146WO1 / / 0958-PCT01trained machine-learning program 1802 generates the prediction / inference data 1822 as output, responsive to receipt of the query data 1828.

[0328] In some examples the trained machine-learning program 1802 may be a generative Al model. Generative Al is a term that may refer to any ty pe of artificial intelligence that can create new data from training data 1804. For example, generative Al can produce text, images, video, audio, code or synthetic data that are similar to the original data but not identical.

[0329] Some of the example techniques that may be used in generative Al include:• Convolutional Neural Networks (CNNs): CNNs are commonly used for image recognition and computer vision tasks. They are designed to extract features from images by using filters or kernels that scan the input image and highlight important patterns.• Recurrent Neural Networks (RNNs): RNNs are designed for processing sequential data, such as speech, text, and time series data. They have feedback loops that allow them to capture temporal dependencies and remember past inputs.• Generative adversarial networks (GANs): These are models that consist of two neural networks: a generator and a discriminator. The generator tries to create realistic data that can fool the discriminator, while the discriminator tries to distinguish between real and fake data. The two networks compete with each other and improve over time.• Variational autoencoders (VAEs): These are models that encode input data into a latent space (a compressed representation) and then decode it back into output data. The latent space can be manipulated to generate new variations of the output data. They may use self-attention mechanisms to process input data, allowing them to handle long sequences of text and capture complex dependencies.• Transformer models: These are models that use attention mechanisms to leam the relationships between different parts of input data (such as words or pixels) and generate output data based on these relationships. Transformer models can handle sequential data such as text or speech as well as non-sequential data such as images or code.Docket No. 4855.146WO1 / / 0958-PCT01

[0330] In generative Al examples, the prediction / inference data 1822 that is output include trend assessment and predictions, translations, summaries, image or video recognition and categorization, natural language processing, face recognition, user sentiment assessments, advertisement targeting and optimization, voice recognition, or media data generation, recommendation, and personalization.

[0331] FIG. 19 is a block diagram illustrating a computing device hardware architecture 1200, within which a set or sequence of instructions can be executed to cause a machine to perform examples of any one of the methodologies discussed herein. The hardware architecture 1200 can describe various computing devices including, for example, the sensor electronics 106, the peripheral medical device 122, the smart device 112, the tablet 114, etc.

[0332] The architecture 1800 may operate as a standalone device or may be connected (e.g., networked) to other machines. In a networked deployment, the architecture 1800 may operate in the capacity of either a server or a client machine in server-client network environments, or it may act as a peer machine in peer-to-peer (or distributed) network environments. The architecture 1800 can be implemented in a personal computer (PC), a tablet PC, a hybrid tablet, a set-top box (STB), a personal digital assistant (PDA), a mobile telephone, a web appliance, a network router, a network switch, a network bridge, or any machine capable of executing instructions (sequential or otherwise) that specify operations to be taken by that machine.

[0333] The example architecture 1800 includes a processor unit 1902 comprising at least one processor (e.g., a central processing unit (CPU), a graphics processing unit (GPU), or both, processor cores, compute nodes). The architecture 1800 may further comprise a main memory 1904 and a static memory 1906, which communicate with each other via a link 1908 (e.g., bus). The architecture 1800 can further include a video display unit 1910, an input device 1912 (e.g., a keyboard), and a UI navigation device 1914 (e.g., a mouse). In some examples, the video display unit 1910, input device 1912, and UI navigation device 1914 are incorporated into a touchscreen display. The architecture 1800 may additionally include aDocket No. 4855.146WO1 / / 0958-PCT01storage device 1916 (e.g., a drive unit), a signal generation device 1918 (e.g., a speaker), a network interface device 1920, and one or more sensors (not shown), such as a Global Positioning System (GPS) sensor, compass, accelerometer, or another sensor.

[0334] In some examples, the processor unit 1902 or another suitable hardware component may support a hardware interrupt. In response to a hardware interrupt, the processor unit 1902 may pause its processing and execute an ISR, for example, as described herein.

[0335] The storage device 1916 includes a machine-readable medium 1922 on which is stored one or more sets of data structures and instructions 1924 (e.g., software) embodying or used by any one or more of the methodologies or functions described herein. The instructions 1924 can also reside, completely or at least partially, within the main memory 1904, within the static memory 1906, and / or within the processor unit 1902 during execution thereof by the architecture 1800, with the main memory 1904, the static memory 1906, and the processor unit 1902 also constituting machine-readable media.EXECUTABLE INSTRUCTIONS AND MACHINE-STORAGE MEDIUM

[0336] The various memories (i.e., 1904, 1906, and / or memory of the processor unit(s) 1902) and / or storage device 1916 may store one or more sets of instructions and data structures (e.g., instructions) 1924 embodying or used by any one or more of the methodologies or functions described herein. These instructions, when executed by processor unit(s) 1902 cause various operations to implement the disclosed examples.

[0337] As used herein, the terms "‘machine-storage medium,” “devicestorage medium,” “computer-storage medium” (referred to collectively as “machine-storage medium 1922”) mean the same thing and may be used interchangeably in this disclosure. The terms refer to a single or multiple storage devices and / or media (e.g., a centralized or distributed database, and / or associated caches and servers) that store executable instructions and / or data, as well as cloud-based storage systems or storage networks that include multiple storage apparatus or devices. The terms shall accordingly beDocket No. 4855.146WO1 / / 0958-PCT01taken to include, but not be limited to, solid-state memories, and optical and magnetic media, including memory internal or external to processors.Specific examples of machine-storage media, computer-storage media, and / or device-storage media 1922 include non-volatile memory, including by way of example semiconductor memory devices, e.g., erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), FPGA, and flash memory devices; magnetic disks such as internal hard disks and removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The terms machine-storage media, computer-storage media, and device-storage media 1922 specifically exclude carrier waves, modulated data signals, and other such media, at least some of which are covered under the term '‘signal medium” discussed below.SIGNAL MEDIUM

[0338] The term “signal medium” or “transmission medium” shall be taken to include any form of modulated data signal, carrier wave, and so forth. The term “modulated data signal” means a signal that has one or more of its characteristics set or changed in such a matter as to encode information in the signal.COMPUTER-READABLE MEDIUM

[0339] The terms “machine-readable medium,” “computer-readable medium” and “device-readable medium” mean the same thing and may be used interchangeably in this disclosure. The terms are defined to include both machine-storage media and signal media. Thus, the terms include both storage devices / media and carrier waves / modulated data signals.

[0340] The instructions 1824 can further be transmitted or received over a communications network 1826 using a transmission medium via the network interface device 1820 using any one of a number of well-known transfer protocols (e.g., HTTP). Examples of communication networks include a LAN, a WAN, the Internet, mobile telephone networks, plain old telephone service (POTS) networks, and wireless data networks (e.g., Wi-Fi, 3G, 4G LTE / LTE-A, 5G or WiMAX networks). The term “transmission medium”Docket No. 4855.146WO1 / / 0958-PCT01shall be taken to include any intangible medium that is capable of storing, encoding, or carrying instructions for execution by the machine, and includes digital or analog communications signals or other intangible media to facilitate communication of such software.

[0341] Throughout this specification, plural instances may implement components, operations, or structures described as a single instance.Although individual operations of one or more methods are illustrated and described as separate operations, one or more of the individual operations may be performed concurrently, and nothing requires that the operations be performed in the order illustrated. Structures and functionality presented as separate components in example configurations may be implemented as a combined structure or component. Similarly, structures and functionality presented as a single component may be implemented as separate components. These and other variations, modifications, additions, and improvements fall within the scope of the subject matter herein.

[0342] Various components are described in the present disclosure as being configured in a particular way. A component may be configured in any suitable manner. For example, a component that is or that includes a computing device may be configured with suitable software instructions that program the computing device. A component may also be configured by¬ virtue of its hardware arrangement or in any other suitable manner.

[0343] The above description is intended to be illustrative, and not restrictive. For example, the above-described examples (or one or more aspects thereof) can be used in combination with others. Other examples can be used, such as by one of ordinary skill in the art upon reviewing the above description. The Abstract is to allow the reader to quickly ascertain the nature of the technical disclosure, for example, to comply with 37 C.F.R. § 1.72(b) in the United States of America. It is submitted with the understanding that it will not be used to interpret or limit the scope or meaning of the claims.

[0344] Also, in the above Detailed Description, various features can be grouped together to streamline the disclosure. However, the claims cannot set forth every feature disclosed herein, as examples can feature a subset ofDocket No. 4855.146WO1 / / 0958-PCT01said features. Further, examples can include fewer features than those disclosed in a particular example. Thus, the following claims are hereby incorporated into the Detailed Description, with each claim standing on its own as a separate example. The scope of the examples disclosed herein is to be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled.

[0345] Each of these non-limiting examples in any portion of the above description may stand on its own or may be combined in various permutations or combinations with one or more of the other examples.

[0346] The above detailed description includes references to the accompanying drawings, which form a part of the detailed description. The drawings show, by way of illustration, specific embodiments in which the subject matter can be practiced. These embodiments are also referred to herein as “examples.” Such examples can include elements in addition to those shown or described. However, the present inventors also contemplate examples in which only those elements shown or described are provided. Moreover, the present inventors also contemplate examples using any combination or permutation of those elements shown or described (or one or more aspects thereof), either with respect to a particular example (or one or more aspects thereof), or with respect to other examples (or one or more aspects thereof) shown or described herein.

[0347] In the event of inconsistent usages between this document and any documents so incorporated by reference, the usage in this document controls.

[0348] In this document, the terms “a” or “an” are used, as is common in patent documents, to include one or more than one, independent of any other instances or usages of “at least one” or “one or more.” In this document, the term “or” is used to refer to a nonexclusive or. such that “A or B” includes “A but not B,” “B but not A,” and “A and B,” unless otherwise indicated. In this document, the terms “including” and “in which” are used as the plain-English equivalents of the respective terms “comprising” and “wherein.” Also, in the following claims, the terms “including” and “comprising” are open-ended, that is. a system, device, article, composition, formulation, or process that includes elements in addition to those listedDocket No. 4855.146WO1 / / 0958-PCT01after such a term in a claim are still deemed to fall within the scope of that claim. Moreover, in the following claims, the terms “first,” “second.” “third,” etc., are used merely as labels, and are not intended to impose numerical requirements on their objects.

[0349] Geometric terms, such as “parallel”, “perpendicular”, “round”, or “square” are not intended to require absolute mathematical precision, unless the context indicates otherwise. Instead, such geometric terms allow for variations due to manufacturing or equivalent functions. For example, if an element is described as “round” or “generally round”, a component that is not precisely circular (e.g., one that is slightly oblong or is a many-sided polygon) is still encompassed by this description.

[0350] Method examples described herein can be machine or computer-implemented at least in part. Some examples can include a computer-readable medium or machine-readable medium encoded with instructions operable to configure an electronic device to perform methods as described in the above examples. An implementation of such methods can include code, such as microcode, assembly language code, a higher-level language code, or the like. Such code can include computer readable instructions for performing various methods. The code may form portions of computer program products. Further, in an example, the code can be tangibly stored on one or more volatile, non-transitory, or non-volatile tangible computer-readable media, such as during execution or at other times. Examples of these tangible computer-readable media can include, but are not limited to, hard disks, removable magnetic disks, removable optical disks (e.g., compact disks and digital video disks), magnetic cassettes, memory cards or sticks, random access memories (RAMs), read only memories (ROMs), and the like.

[0351] The above description is intended to be illustrative, and not restrictive. For example, the above-described examples (or one or more aspects thereof) may be used in combination with each other. Other embodiments can be used, such as by one of ordinary skill in the art upon reviewing the above description. The Abstract is provided to allow the reader to quickly ascertain the nature of the technical disclosure. It isDocket No. 4855.146WO1 / / 0958-PCT01submited with the understanding that it will not be used to interpret or limit the scope or meaning of the claims. Also, in the above Detailed Description, various features may be grouped together to streamline the disclosure. This should not be interpreted as intending that an unclaimed disclosed feature is essential to any claim. Rather, inventive subject matter may he in less than all features of a particular disclosed embodiment. Thus, the following claims are hereby incorporated into the Detailed Description as examples or embodiments, with each claim standing on its own as a separate embodiment, and it is contemplated that such embodiments can be combined with each other in various combinations or permutations. The scope of the subject matter should be determined with reference to the claims, along with the full scope of equivalents to which such claims are entitled.

[0352] Unless the context clearly requires otherwise, throughout the description and the claims, the words “comprise,” “comprising,” and the like are to be construed in an inclusive sense, as opposed to an exclusive or exhaustive sense, i.e., in the sense of “including, but not limited to.” As used herein, the terms “connected,” “coupled,” or any variant thereof means any connection or coupling, either direct or indirect, between two or more elements; the coupling or connection between the elements can be physical, logical, or a combination thereof. Additionally, the words “herein,” “above,” “below,” and words of similar import, when used in this application, refer to this application as a whole and not to any particular portions of this application. Where the context permits, words using the singular or plural number may also include the plural or singular number respectively. The word “or” in reference to a list of two or more items, covers all of the following interpretations of the word: any one of the items in the list, all of the items in the list, and any combination of the items in the list. Likewise, the term “and / or” in reference to a list of two or more items, covers all of the following interpretations of the word: any one of the items in the list, all of the items in the list, and any combination of the items in the list.

[0353] Although some examples, e.g., those depicted in the drawings, include a particular sequence of operations, the sequence may be altered without departing from the scope of the present disclosure. For example, some of the operations depicted may be performed in parallel or in aDocket No. 4855.146WO1 / / 0958-PCT01different sequence that does not materially affect the functions as described in the examples. In other examples, different components of an example device or system that implements an example method may perform functions at substantially the same time or in a specific sequence.

[0354] The various features, steps, and processes described herein may be used independently of one another, or may be combined in various ways. All possible combinations and subcombinations are intended to fall within the scope of this disclosure. In addition, certain method or process blocks may be omitted in some implementations.

Claims

Docket No. 4855.146WO1 / / 0958-PCT01What is claimed is:

1. A sensor system, comprising:an analyte sensor configured to operate in a first mode and in a second mode, the analyte sensor generating a first raw sensor signal indicative of a first analyte concentration of a host when operated in the first mode and a second raw sensor signal indicative of a second analyte when operated in the second mode; andsensor electronics configured to perform operations comprising: transitioning the analyte sensor from operating in the second mode to operating in the first mode;accessing a first value based at least on the first raw sensor signal taken during a first time period after the transitioning;generating a corrected first value of the first raw sensor signal using corrective signal model configured to correct one or more artifacts in the first raw sensor signal resulting from the transitioning; andgenerating an estimated first analyte concentration based at least on the corrected first value.

2. The sensor system of claim 1, wherein the first analyte concentration includes a glucose concentration, wherein the transitioning of the analyte sensor from the second mode to the first mode includes transitioning the analyte sensor from generating the second raw sensor signal indicative of a second analyte concentration to generating the first raw sensor signal indicative of the glucose concentration.

3. The sensor system of claim 2, where the second analyte concentration includes an oxygen concentration, wherein the transitioning of the analyte sensor from the second mode to the first mode includes transitioning the analyte sensor from generating the second raw sensor signal indicative of the oxygen concentration to generating the first raw sensor signal indicative of the glucose concentration.Docket No. 4855.146WO1 / / 0958-PCT014. The sensor system of any of claims 1-3, where the second analyte concentration includes an oxygen concentration, wherein the transitioning of the analyte sensor from the second mode to the first mode includes transitioning the analyte sensor from generating the second raw sensor signal indicative of the oxygen concentration to generating the first raw sensor signal indicative of the first analyte concentration.

5. The sensor system of any of claims 1-4, prior to transitioning the analyte sensor from operating in the second mode to operating in the first mode, causing the analyte sensor to operate in the first mode at least in part by providing a first bias condition to the analyte sensor.

6. The sensor system of claim 5, wherein transitioning the analyte sensor from operating in the second mode to operating in the first mode comprises applying a second bias condition to the analyte sensor, the second bias condition being different than the first bias condition.

7. The sensor system of claim 6, wherein the first bias condition is an opposite polarity to the second bias condition.

8. The sensor system of any of claims 1-7, wherein the one or more artifacts include a spike in a magnitude of the first raw sensor signal, wherein the corrected first value removes at least the spike of the one or more artifacts.

9. The sensor system of claim 8, where the one or more artifacts include a decay in a magnitude of the first raw sensor signal subsequent to the spike, wherein the corrected first value removes at least a delay portion of the one or more artifacts.

10. The sensor system of claim 9, wherein the one or more artifacts further include a spike in the magnitude of the first raw sensor signal, wherein the corrected first value further removes at least the spike portion of the one or more artifacts.

11. The sensor system of any of claims 1-10, wherein generating the estimated first analyte concentration includes subtracting the corrected first value from the first value.Docket No. 4855.146WO1 / / 0958-PCT0112. The sensor system of any of claims 1-11, wherein the transitioning of the analyte sensor from operating in the second mode to operating in the first mode is in response to a predefined time period elapsing.

13. The sensor system of any of claims 1-12, wherein the transitioning of the analyte sensor from operating in the second mode to operating in the first mode is in response to a determination that a second value of the second raw sensor signal meets or exceeds a certain threshold indicative of an abnormality trigger.

14. The sensor system of any of claims 1-13, wherein the operations further comprise:identifying a current activity of a host; andidentifying the corrective signal model among a plurality of corrective signal models based on the identified current activity;wherein generating the corrected first value of the first raw sensor signal is based on the identified current activity.

15. The sensor system of any of claims 1-14, wherein the corrective signal model includes a machine learning model, wherein generating the corrected first value of the first raw sensor signal using the corrective signal model comprises inputting the first value into the machine learning model, the machine learning model trained to generate a corrected first value of the first raw sensor signal.

16. The sensor system of any of claims 1-15, wherein the operations further comprise:accessing a second value of the first raw sensor signal; determining that the one or more artifacts resulting from the transitioning of the analyte sensor from operating in the second mode to operating in the first mode do not appear in the second value; and generating an estimated second analyte concentration based on an analyte concentration module;wherein generating the estimated first analyte concentration includes applying the corrected first value of the first raw sensor signal to the analyte concentration module.Docket No. 4855.146WO1 / / 0958-PCT0117. The sensor system of any of claims 1-16, wherein the operations further comprise:retrieving historical dual analyte sensor data;applying the historical dual analyte sensor data to a retrospective artifact removal model; andreceiving, from the retrospective artifact removal model, the corrective signal model configured to generate corrected values from raw sensor signals.

18. The sensor system of claim 17, wherein the retrospective artifact removal model includes a first machine learning model, the first machine learning model being used by the corrective signal model.

19. The sensor system of claim 17, wherein the corrective signal model includes a first machine learning model generated by the retrospective artifact removal model.

20. The sensor system of claim 19, wherein the retrospective artifact removal model includes a second machine learning model, the second machine learning model trained to generate the first machine learning model.

21. The sensor system of claim 19, wherein the corrective signal model is configured to continuously retain itself based on new values of the first raw sensor signal.

22. The sensor system of claim 17, wherein a sliding window of the historical dual analyte sensor data is applied to the retrospective artifact removal model to generate the corrective signal model.

23. The sensor system of claim 17, wherein the historical dual analyte sensor data includes historical dual analyte sensor data of the host.

24. The sensor system of claim 17, wherein the historical dual analyte sensor data includes historical dual analyte sensor data of other users other than the host.Docket No. 4855.146WO1 / / 0958-PCT0125. The sensor system of claim 24, wherein the other users have one or more matching physiological characteristics of the host.

26. The sensor system of claim 17, wherein the corrective signal model is configured to generate corrected values from raw sensor signals in real time.

27. The sensor system of claim 26, wherein the retrospective artifact removal model is performed on an external server.

28. The sensor system of claim 17, wherein the operations further comprise:receiving historical single analyte sensor data; andapplying the historical single analyte sensor data to the retrospective artifact removal model;wherein receiving, from the retrospective artifact removal model, the corrective signal model is further based on the historical single analyte sensor data.

29. The sensor system of claim 28, wherein the retrospective artifact removal generates the corrective signal model by subtracting values between the historical single analyte sensor data and the historical dual analyte sensor data.

30. The sensor system of claim 29, wherein prior to subtracting the values between the historical single analyte sensor data and the historical dual analyte sensor data, offsetting at least one of the historical single analyte sensor data and the historical dual analyte sensor data such that baseline values for the historical single analyte sensor data and the historical dual analyte sensor data align.

31. The sensor system of claim 30, wherein prior to using the corrective signal model, applying an opposite offset to the subtracted value based on the offset applied to the at least one of the historical single analyte sensor data and the historical dual analyte sensor data.

32. The sensor system of claim 30, wherein the offset is a magnitude offset of the at least one of the historical single analyte sensor data and the historical dual analyte sensor data.Docket No. 4855.146WO1 / / 0958-PCT0133. The sensor system of claim 30, wherein the offset is a temporal offset of the at least one of the historical single analyte sensor data and the historical dual analyte sensor data.

34. The sensor system of any of claims 1-33, wherein a sliding window of the first raw sensor signal is applied to the corrective signal model to generate corrected values of the first raw sensor signal.

35. The sensor system of any of claims 1-34, wherein the operations further comprise:accessing a second value of the first raw sensor signal after the first time period;determining that the first raw sensor signal does not include the one or more artifacts resulting from the transitioning; andgenerating an estimated second analyte concentration based on the second value of the first raw sensor signal.

36. A method comprising:operating, by an analyte sensor, in a first mode and in a second mode, the analyte sensor generating a first raw sensor signal indicative of a first analyte concentration of a host when operated in the first mode and a second raw sensor signal indicative of a second analyte when operated in the second mode;transitioning the analyte sensor from operating in the second mode to operating in the first mode;accessing a first value based at least on the first raw sensor signal taken during a first time period after the transitioning;generating a corrected first value of the first raw sensor signal using corrective signal model configured to correct artifacts in the first raw sensor signal resulting from the transitioning; andgenerating an estimated first analyte concentration based at least on the corrected first value.

37. A non-transitory computer-readable storage medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising:Docket No. 4855.146WO1 / / 0958-PCT01operating, by an analyte sensor, in a first mode and in a second mode, the analyte sensor generating a first raw sensor signal indicative of a first analyte concentration of a host when operated in the first mode and a second raw sensor signal indicative of a second analyte when operated in the second mode;transitioning the analyte sensor from operating in the second mode to operating in the first mode;accessing a first value based at least on the first raw sensor signal taken during a first time period after the transitioning;generating a corrected first value of the first raw sensor signal using corrective signal model configured to correct artifacts in the first raw sensor signal resulting from the transitioning; andgenerating an estimated first analyte concentration based at least on the corrected first value.

Images