Aggregation of zoned sensor glucose models

Abstract

A processor-implemented method includes: receiving sensor measurement data from a glucose sensor; selecting, based on the sensor measurement data, a first regional sensor glucose (SG) model from a plurality of first regional sensor glucose (SG) models for respective ones of a plurality of first regions of an input parameter space associated with the sensor measurement data, and selecting, from a plurality of second regional SG models, a second regional SG model for a corresponding region of a plurality of second regions of the input parameter space; estimating a first SG value and a second SG value using the first regional SG model and the second regional SG model, respectively; and determining a predicted SG value based on a combination of the first SG value and the second SG value. The input parameter space is partitioned into the plurality of first regions and the plurality of second regions using different partitioning schemes.

Description

Technical Field

[0001] The present disclosure generally relates to glucose level measurement and / or management. Background Art

[0002] In a healthy person, the pancreas produces insulin in response to elevated glucose levels and releases it into the bloodstream. More specifically, the beta cells (β-cells) in the pancreas produce insulin as needed and secrete it into the bloodstream. If the β-cells become incapable or die (a condition called type 1 diabetes), insulin may need to be delivered to the diabetic's body using, for example, a syringe, pen, or infusion pump to manage blood sugar levels in order to maintain health or life.

[0003] Some blood glucose level management systems may be closed-loop systems that may include automatic or semi-automatic control to deliver insulin to a patient's pump. Some blood glucose level management systems may require manual administration of insulin based on a dose determined by an insulin calculator. The time and amount of insulin delivered to the patient may be controlled, and these times and amounts are determined based on, for example, the amount of carbohydrates ingested by the patient and / or the real-time measurement of glucose levels by a glucose sensor. A glucose sensor may detect and / or quantify the amount of glucose (e.g., interstitial glucose or blood glucose) in the patient's body, which enables the patient and medical staff to monitor the physiological conditions in the patient's body. In many cases (e.g., diabetic patients), continuous monitoring of glucose levels may be beneficial. Such information may help monitor and / or adjust a treatment regimen that typically includes administering insulin to the patient. For the purpose of continuous glucose monitoring (CGM) or determining an operating command for an infusion pump, continuous sampling and monitoring of the user's blood glucose level, which may damage battery life, is not adopted, but intermittently sensed glucose data samples may be generally utilized. Summary of the invention

[0004] The present disclosure as a whole relates to glucose level measurement and / or management. More specifically, the technology disclosed herein relates to estimating blood glucose levels using the aggregation of measurement data from a subcutaneous sensor and a partitioned sensor glucose model. The technology disclosed herein can be practiced in a variety of ways, such as using a processor-implemented method, a system including one or more processors and one or more processor-readable media, and / or one or more (non-transient) processor-readable media.

[0005] According to certain embodiments, the technology may involve: receiving sensor measurement data measured by a glucose sensor, the sensor measurement data including at least one signal indicating a glucose level and at least one signal indicating a sensor health condition; selecting a first regional sensor glucose (SG) model from a plurality of first regional SG models for a corresponding region in a plurality of first regions of an input parameter space associated with the sensor measurement data based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of first regions based on a first partitioning scheme; estimating a first SG value using the first regional SG model and the sensor measurement data; selecting a second regional SG model from a plurality of second regional SG models for a corresponding region in a plurality of second regions of the input parameter space based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of second regions based on a second partitioning scheme different from the first partitioning scheme; estimating a second SG value using the second regional SG model and the sensor measurement data; and determining a predicted SG value based on a combination of the first SG value and the second SG value.

[0006] This summary is neither intended to identify the main or essential features of the claimed subject matter nor is it intended to be used alone to determine the scope of the claimed subject matter. The subject matter should be understood by reference to appropriate portions of the entire specification of the disclosure, any or all of the drawings, and each claim. The foregoing and other features and examples will be described in more detail in the following description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Exemplary embodiments are described in detail below with reference to the accompanying drawings.

[0008] Figure 1 An example of blood glucose level management according to certain embodiments is illustrated.

[0009] Figure 2 is a block diagram of an example of a glucose control system according to certain embodiments.

[0010] Figure 3 An example of a continuous glucose monitoring (CGM) system according to certain embodiments is illustrated.

[0011] Figure 4A An example of a glucose sensor according to certain embodiments is illustrated.

[0012] Figure 4B Another example of a glucose sensor according to certain embodiments is illustrated.

[0013] Figure 4C An exemplary cross-section of a working electrode of a glucose sensor according to certain embodiments is illustrated.

[0014] Figure 5 A schematic diagram of a glucose sensor according to certain embodiments is shown.

[0015] Figure 6 Included is a flow chart illustrating an example of a method of determining a sensor glucose value using measurement data from a glucose sensor according to certain embodiments.

[0016] Figure 7 An example of a sensor signature generator according to certain embodiments is illustrated.

[0017] Figure 8 Examples of using one or more models to determine sensor glucose values ​​in accordance with certain embodiments are illustrated.

[0018] Fig. 9A An example of a zonal model including a set of regional models for modeling the relationship between sensor glucose (SG) values ​​and sensor measurement data is shown according to certain embodiments.

[0019] Fig. 9B Another example of a zonal model including a set of regional models for modeling the relationship between SG values ​​and sensor measurement data is shown in accordance with certain embodiments.

[0020] Fig. 10A An example of a zonal model including a set of regional models for modeling the relationship between SG values ​​and sensor measurement data is shown according to certain embodiments.

[0021] Fig. 10B Another example of a zonal model including a set of regional models for modeling the relationship between SG values ​​and sensor measurement data is shown in accordance with certain embodiments.

[0022] Fig.11 An example of estimating SG values ​​using aggregation of multiple regional model sets according to certain embodiments is illustrated.

[0023] Fig. 12A An example of an aggregate model including two sets of regional SG models for estimating SG values ​​based on sensor measurement data generated by a glucose sensor is illustrated according to certain embodiments.

[0024] Fig. 12B Examples of performance improvements over baseline techniques using the techniques disclosed herein according to certain embodiments are illustrated.

[0025] Fig.13 Included is a flow chart illustrating an example of a process for estimating SG values ​​using aggregation of partitioned SG models according to certain embodiments.

[0026] Fig.14A and Fig. 14B Depicted are examples of integrated glucose sensor devices that can implement some of the examples disclosed herein, according to certain embodiments.

[0027] Fig.15 Depicted are examples of delivery devices that may implement some of the examples disclosed herein.

[0028] Fig.16 is a simplified block diagram of an example of an electronic device that may implement some of the examples disclosed herein.

[0029] The accompanying drawings depict embodiments of the present disclosure for purposes of illustration only. Those skilled in the art will readily recognize from the following description that alternative embodiments of the illustrated structures and methods may be employed without departing from the principles or claimed benefits of the present invention.

[0030] In the drawings, similar components and / or features may have the same reference numeral. In addition, various components of the same type may be distinguished by following the reference numeral with a dash and a second reference numeral that distinguishes the similar components. If only the first reference numeral is used in the specification, the description applies to any of the similar components having the same first reference numeral, without regard to the second reference numeral. DETAILED DESCRIPTION

[0031] The present disclosure as a whole relates to glucose level measurement and / or management. More specifically, the technology disclosed herein relates to estimating blood glucose levels using the aggregation of measurement data from a subcutaneous glucose sensor and a partitioned sensor glucose model. The technology disclosed herein can be practiced in a variety of ways, such as using a processor-implemented method, a system including one or more processors and one or more processor-readable media, and / or one or more (non-transient) processor-readable media.

[0032] Diabetes is a disease of the patient's glucose regulation system, wherein due to insufficient insulin production and / or insulin resistance, the naturally produced insulin in the body may not be enough to control the glucose level in the patient's bloodstream. Therefore, diabetic patients may need to receive insulin from a pump or another delivery device (such as an injection or infusion device (e.g., a syringe)) to control the glucose level in the patient's bloodstream. In order to control glucose levels, the diabetic's therapy routine may generally include a basal insulin dose and a bolus insulin dose. Basal insulin (also referred to as background insulin) may include a continuous or constant release of a small amount of insulin to keep blood sugar levels at a consistent level during a long period of time. Bolus insulin can be taken specifically before, during or after mealtimes or at other times when blood sugar levels may rise rapidly.

[0033] The insulin dose to be delivered can be determined based on, for example, the carbohydrate count of the meal and / or the patient's glucose level measured using a blood glucose monitor (such as a fingerstick blood glucose measurement device or a continuous glucose monitoring (CGM) sensor). The CGM sensor can use the measurement results (e.g., current, voltage level, resistance, etc.) from the sensor inserted into the subcutaneous layer of the user to estimate the user's blood glucose level based on the glucose level in the interstitial fluid (ISF). In some specific implementations, when the sensor is inserted and exposed to glucose and oxygen diffused from the interstitial fluid, the glucose sensor can generate a signal through a series of electrochemical reactions. For example, the glucose sensor may include a glucose limiting membrane (GLM) that limits the amount of glucose and oxygen delivered to the glucose oxidase (GOx) layer of the working electrode of the glucose sensor to ensure that the reaction is glucose-limited. The GOx layer or another active enzyme layer on the working electrode of the glucose sensor can decompose glucose and oxygen into gluconic acid and hydrogen peroxide. The generated hydrogen peroxide molecules can interact with the working electrode (e.g., under the condition of applying a voltage level) to decompose hydrogen peroxide into two hydrogen ions, oxygen and two electrons at the surface of the working electrode. Due to the potential difference, the charge can be forced to move between the electrodes, thereby generating a sensor current signal (Isig) that can be measured by the sensor electronics. Other signals can also be measured, such as the counter voltage (Vcntr, the voltage potential difference between the counter electrode and the working electrode), electrochemical impedance spectroscopy (EIS) at different frequencies, etc. The signals measured using the sensor (including Isig, Vcntr, and EIS) can be processed (e.g., filtered, calibrated, or otherwise transformed) to generate some other signals or parameters, such as a filtered Isig signal, real and imaginary impedance at various frequencies, derivative signals, transformed signals, etc. The raw signals and / or processed signals measured using the sensor can be used in one or more sensor glucose (SG) models to determine a SG value, which can be an estimate of the user's blood glucose (BG) level.

[0034] The one or more SG models may include one or more trained machine learning models, equations, functions, etc. that may map input parameters (e.g., raw and / or processed sensor measurement data) to SG (or BG) values. For example, the SG model may be represented as:

[0035]

[0036] Wherein the years can be the operating time of the glucose sensor after insertion. In a simplified example, the SG value can be calculated according to the following formula:

[0037] ,

[0038] Wherein α, β, γ and δ can be coefficients associated with age, Isig, Vcntr and EIS, and Δ is an offset value. In some embodiments, at least some of these coefficients may include vectors or tensors, and at least some of age, Isig, Vcntr and EIS may include vectors or tensors.

[0039] In order to achieve the desired accuracy and reliability and reduce the effects of noise and other stray signals, CGM sensors may need to be calibrated periodically or occasionally to improve the accuracy and / or reliability of the glucose sensor. For example, it may be necessary to calibrate the sensor data using a known good blood glucose value, which is typically obtained via a so-called "fingerstick measurement" using a blood glucose meter that measures blood glucose levels in the capillaries. Performing such calibration measurements can increase the patient's burden and perceived complexity, and may be inconvenient, uncomfortable, or otherwise unpopular with the patient. In addition, due to the time required for glucose to diffuse from the capillaries to the interstitial space, ISF glucose measurements may lag behind blood glucose measurements, which may require certain signal processing (e.g., filtering) or other techniques to compensate for the physiological lag. Additionally, various factors may cause instantaneous changes in the sensor output, which may affect the accuracy of calibration and measurement. Deterioration of sensor performance over time and production variations between sensors may further exacerbate these challenges.

[0040] Sensor properties associated with a glucose sensor, such as age (or time in use or time worn), battery level, calibration accuracy, sensing environment (e.g., oxygen level, temperature, humidity, and pH, etc.), physiological dynamics, production variability, and other properties, may affect enzymatic glucose level measurements in complex ways that may be difficult to model. As a result, the accuracy of many CGM sensors may be suboptimal and may exhibit large differences in measurement results. In addition to sensor-to-sensor production variability, sensor performance degradation over time, etc., the lower-than-expected performance of CGM sensors may also be at least partially attributed to the complexity of modeling the in vivo behavior of CGM sensors. In many cases, a single SG model may not accurately estimate blood glucose levels, or a single SG model may be complex and difficult to train because various sensor properties may affect blood glucose level measurements in complex ways that may be difficult to characterize.

[0041] According to certain embodiments, different SG models may be used for different input parameter ranges (e.g., different subspaces of the input parameter space of the SG model), such as different ranges of age, Vcntr, Isig, EIS or other sensor measurement data. Each SG model may be optimized for a corresponding region of the input parameter space (referred to herein as a regional SG model or regional model), and may therefore be able to more accurately characterize the relationship between the glucose level and the sensor measurement data in the corresponding region. A plurality of regional SG model sets may be generated and used to determine an SG value approximating the blood glucose level of the user, wherein different regional SG model sets may be generated for different partitions of the input parameter space (e.g., different combinations of input parameter ranges). The input parameters may include, for example, Isig, Vcntr, EIS, age, sensor sensitivity, etc. The input parameter space may be divided in different ways according to different partitioning schemes (e.g., based on different combinations of input parameters and / or by dividing the range of the same combination of input parameters in different ways). A regional SG model set may be generated for a region of the input parameter space divided according to a corresponding partitioning scheme. Based on the region that the current input parameter falls into, a regional SG model may be selected from each set of regional SG models of a corresponding input parameter space partitioning scheme. Multiple regional SG models selected from multiple sets of regional SG models for different partitioning schemes may be used to estimate SG values ​​individually, and then the outputs of the selected multiple regional SG models may be combined (e.g., by averaging or weighted averaging) to determine a predicted SG value.

[0042] Aggregation using multiple regional SG model sets disclosed herein can more accurately model sensor variations (e.g., sensitivity variations) and can significantly improve the performance (e.g., accuracy and consistency) of CGM sensors. Each regional model can be trained individually at a faster speed using a smaller training data set, and the response of the glucose sensor to different concentrations of analytes, such as sensitivity and nonlinearity of sensitivity, can be better characterized. Using certain parameters such as Isig as additional parameters for partitioning the input parameter space can also help improve the performance of the SG model, due to, for example, the improved granularity of partitioning the input parameter space and the correlation of additional parameters for modeling the response of the glucose sensor. Compared with using one complex SG model or one regional SG model set with a large number of models for a large number of regions partitioned using a large number of input parameters, aggregation using multiple regional SG model sets for different partitioning schemes (e.g., using different input parameters for different partitioning schemes) can further improve the performance (e.g., accuracy and consistency) of the CGM sensor, due to, for example, using additional parameters to improve the granularity of the partitions, and using additional partitioning schemes to include additional parameters or degrees of freedom to tune and improve the SG model without significantly increasing the overall complexity of the SG model and the difficulty of training the SG model.

[0043] Aspects of the present disclosure and each embodiment can be practiced with one or more types of insulin (e.g., rapid-acting insulin, intermediate-acting insulin and / or slow-acting insulin). Unless the context indicates, terms such as "dose", "insulin", "basal" and "bolus" may not represent a specific type of insulin. For example, rapid-acting insulin can be used for both basal dose and bolus dose. The term "basal" may refer to and include insulin that is intended to be delivered in an amount and frequency corresponding to a healthy body between meals and during sleep to release insulin. The term "bolus" may refer to and include insulin that is intended to be delivered in an amount and timing corresponding to a healthy body to release insulin to offset a high glucose level such as caused by dietary consumption. Meals may include food or beverage consumption of any type or amount, including breakfast, lunch, dinner, snacks and beverages, etc.

[0044] In the following description, for the purpose of explanation, specific details are set forth in order to provide a thorough understanding of the examples of the present disclosure. However, it is apparent that various examples can be practiced without these specific details. For example, devices, systems, structures, assemblies, methods and other components may be shown as components in the form of block diagrams, so as not to obscure the examples with unnecessary details. In other instances, well-known devices, processes, systems, structures, operations and techniques may be shown or may not be shown without the necessary details, so as to avoid obscuring the examples. The drawings and descriptions are not intended to be restrictive. The terms and expressions used in the present disclosure are used as descriptive terms rather than restrictive terms, and when using these terms and expressions, it is not intended to exclude any equivalents of the features shown and described or parts thereof. The word "example" is used herein to mean "serving as an example, example or illustration". Any embodiment or design described as "example" herein is not necessarily interpreted as being preferred or advantageous compared to other embodiments or designs.

[0045] Carbohydrates in food intake can be broken down into glucose (sugar) in the stomach and / or intestines and can be absorbed into the bloodstream at the small and / or large intestines. The bloodstream can transport glucose to the body's capillaries, where some of the glucose can diffuse into the interstitial fluid between cells (e.g., fat or muscle cells), which can use the glucose as energy. The human endocrine system (working with the nervous system) that directs and regulates the body's functions and activities may secrete chemicals that send messages to tissues and organs. For example, an endocrine gland or organ may release a hormone into the bloodstream for delivery to a target cell with a receptor. In one example, insulin may be produced by the beta cells of the pancreas and may, for example, increase glucose intake, enhance glucose utilization, stop hepatic glucose production, stimulate glycogen formation in the liver and skeletal muscle, promote protein synthesis, and increase fat storage.

[0046] Insulin acts as a key that unlocks cells and helps glucose move into cells where it can be used for energy. Without insulin, glucose may not be able to enter cells for use as energy and may accumulate in the interstitial fluid and bloodstream. A healthy pancreas continuously releases small amounts of regular human insulin 24 hours a day, including between meals and during sleep. This small amount of insulin matches the glucose released by the liver. A healthy pancreas may also secrete more insulin after eating to match the amount of food consumed.

[0047] A person's liver may take in excess glucose from the digestive process and convert the excess glucose into glycogen for storage so that the glucose can be used when needed. For example, when blood sugar levels are low, the alpha cells of the pancreas may secrete glucagon. Glucagon may cause the liver to release a stored form of glucose (glycogen) into the bloodstream to help increase glucose levels. The balance between insulin secreted by the beta cells and glucagon secreted by the alpha cells can help maintain normal blood sugar levels, such as within a range of about 80 mg / dL to 120 mg / dL before a meal.

[0048] Due to insufficient insulin production and / or insulin resistance, the insulin naturally produced in the body of a diabetic patient may not be sufficient to control the glucose level in the patient's bloodstream. Therefore, a diabetic patient may need to receive insulin from a pump or another delivery device (such as an injection or infusion device) to control the glucose level in the patient's bloodstream. In order to control glucose levels, a diabetic patient's therapy routine may generally include a basal insulin dose and a bolus insulin dose. The dose of insulin to be delivered can be determined based on, for example, a carbohydrate count of a meal and / or a patient's glucose level measured using a glucose monitor (such as a continuous glucose monitor (CGM)).

[0049] In some implementations (e.g., in a closed-loop system), the insulin delivery device may communicate with a sensor device (including but not limited to a CGM) or otherwise use the sensor device to perform various measurements for the patient. In one example, the sensor device may include subcutaneously implanted electrodes to simultaneously monitor the patient's response to a meal and insulin introduced by the insulin delivery device. The sensor device and the insulin delivery device may be in a communication network (wired or wireless) with one or more processors and / or patient devices (such as a patient's smartphone equipped with an application or other software) to create an overall system for monitoring the patient's disease state and for promoting its treatment.

[0050] Figure 1An example of a blood glucose level management system 100 according to certain embodiments is illustrated. The blood glucose level management system 100 can be used to monitor and adjust the blood glucose level of a patient 101. In the illustrated example, the blood glucose level management system 100 may include a delivery device 102, a monitoring device 104, a computing device 106, and an optional remote or cloud computing system 108. The delivery device 102, the monitoring device 104, and the computing device 106 may be embodied in various ways, including being disposed in one or more device housings. For example, in some embodiments, all of the devices 102 to 106 may be disposed in a single device housing. In some embodiments, each of the devices 102 to 106 may be disposed in a separate device housing. In some embodiments, two or more of the devices 102 to 106 may be disposed in the same device housing. In some embodiments, a single device 102, 104, or 106 may have two or more parts disposed in two or more housings. For example, the monitoring device 104 may include an on-body portion and a display and control portion that communicates with the on-body portion by wire or wireless means. The delivery device 102 may include an on-body site (eg, including a cannula) and portions including a reservoir, a pump, and a control unit. These and other embodiments and combinations thereof are contemplated to be within the scope of the present disclosure.

[0051] The blood glucose level management system 100 may include multiple communication links, such as communication links 112 to 118. The communication links 112 to 118 may each be a wired connection and / or a wireless connection. In an embodiment where the two devices are located in the same housing, the communication links may include, for example, wires, cables, and / or a communication bus located on a printed circuit board, etc. In an embodiment where the two devices are separated from each other in different device housings, the communication links may be wired connections and / or wireless connections. The wired connection may include, for example, an Ethernet connection, a universal serial bus (USB) connection, and / or another type of physical connection. The wireless connection may include, for example, a cellular connection, a Wi-Fi connection, a Bluetooth connection, etc. ® connection, a mesh network connection, and / or another type of connection using a wireless communication protocol. Some embodiments of communication links 112 to 118 may use a direct connection such as Bluetooth ® The communication links 112 to 118 may be connected and / or may use connections routed through one or more networks or network devices (not shown), such as Ethernet networks, Wi-Fi networks, cellular networks, satellite networks, intranets, extranets, the Internet, and / or other types of networks such as Internet backbone networks. Various combinations of wired connections and / or wireless connections may be used for the communication links 112 to 118.

[0052] The delivery device 102 may be configured to deliver a therapeutic substance to the patient 101. The therapeutic substance may include, for example, insulin, HIV drugs, drugs for treating pulmonary hypertension, iron chelation drugs, painkillers, anti-cancer treatments, medicines, vitamins, hormones, nutritional supplements, dyes, tracing media, saline media, hydration media, etc. The delivery device 102 may be fixed to the patient 101 (e.g., fixed to the body or clothing of the patient 101) or may be at least partially implanted in the body of the patient 101. In some embodiments, the delivery device 102 may include a reservoir, an actuator, a delivery mechanism, and a cannula (not shown). The reservoir may be configured to store a certain amount of therapeutic substance. In some embodiments, the reservoir may be refillable or replaceable. The actuator may be configured to drive the delivery mechanism. In some examples, the actuator may include a motor, such as an electric motor. The delivery mechanism may be configured to move the therapeutic substance from the reservoir through the cannula. In some examples, the delivery mechanism may include a pump and / or a plunger. The cannula may facilitate fluid connection between the reservoir and the body of the patient 101. The cannula and / or needle may facilitate delivery of a therapeutic substance to a tissue layer, vein, interstitial fluid, or body cavity of the patient 101. During operation, the actuator may drive the delivery mechanism in response to a signal (e.g., a command signal), thereby moving the therapeutic substance from the reservoir, through the cannula, and into the body of the patient 101.

[0053] The components of the delivery device 102 described above are provided as examples only. The delivery device 102 may include other components, such as, but not limited to, a power source, a communication transceiver, one or more processors or other computing resources, a memory device, and / or a user interface (e.g., buttons, keys, displays, etc.), among others. In some implementations, the delivery device 102 may host an App (e.g., an insulin calculator) that may calculate the desired amount of therapeutic substance to be delivered to the patient 101. Those skilled in the art will recognize various implementations of the delivery device 102 and components of such implementations. All such implementations and components are contemplated to be within the scope of the present disclosure.

[0054] The monitoring device 104 may be configured to detect a physiological condition (e.g., a glucose concentration level) of the patient 101, and may also be configured to detect other matters. The monitoring device 104 may be secured to the body of the patient 101 (e.g., secured to the skin of the patient 101 via an adhesive) and / or may be at least partially implanted in the body of the patient 101. Depending on the particular location or configuration, the monitoring device 104 may be in contact with biological matter (e.g., interstitial fluid and / or blood) of the patient 101.

[0055] The monitoring device 104 may include one or more sensors (not shown), such as, but not limited to, electrochemical sensors, electrical sensors, and / or optical sensors. As will be appreciated by those skilled in the art, an electrochemical sensor may be configured to respond to the interaction or binding of a biomarker with an electrode by generating an electrical signal based on, for example, the potential, conductance, current, and / or impedance of an electrical path through the electrode. The electrode may include a material selected to interact with a specific biomarker such as glucose. The potential, current, conductance, and / or impedance may be related to the concentration of a specific biomarker. In one example, the electrochemical sensor may include a glucose limiting membrane (GLM) that limits the amount of glucose and oxygen delivered to a glucose oxidase (GOx) layer of a working electrode of the sensor to ensure that the reaction is glucose-limited. The GOx layer or another active enzyme layer on the working electrode of the sensor may decompose glucose and oxygen into gluconic acid and hydrogen peroxide. When a voltage signal is supplied to the working electrode, the generated peroxide molecules may interact with the working electrode to decompose hydrogen peroxide into two hydrogen ions, oxygen, and two electrons at the surface of the working electrode. Charge may be forced to move between electrodes (e.g., between a working electrode and a counter electrode), thereby generating a sensor current signal (Isig) that may be measured by the sensor electronics. Other signals may also be measured, such as a counter voltage (Vcntr, the voltage potential difference between the counter electrode and the working electrode), electrochemical impedance spectroscopy (EIS) at different frequencies, and the like. The signals measured using the sensor (including Isig, Vcntr, and EIS) may be processed (e.g., filtered or transformed) to generate some other signals or parameters, such as a filtered Isig signal, real impedance and imaginary impedance at various frequencies, and the like. These signals and / or processed parameters may be used in one or more sensor glucose (SG) models (e.g., machine learning models or mathematical models) to determine a SG value, which may be an estimate of a patient's blood glucose (BG) level.

[0056] As will be appreciated by those skilled in the art, the electrical sensor may be configured to respond to the electrical biosignal by generating an electrical signal based on the amplitude, frequency, and / or phase of the electrical biosignal. The electrical biosignal may include a change in the current generated by the sum of the potential differences across the tissues of the patient 101 (such as the nervous system). In some embodiments, the electrical biosignal may include a portion of the potential changes (e.g., recorded as an electrocardiogram) generated by the heart of the patient 101 over time that may indicate the patient 101's 101 glucose level. The optical sensor may be configured to respond to the interaction or binding of the biomarker with the substrate, for example, by generating an electrical signal based on a change in the brightness of the substrate. In one example, the substrate may include a material selected to fluoresce in response to contact with a selected biomarker (such as glucose). The fluorescence may be proportional to the concentration of the selected biomarker.

[0057] In some embodiments, the monitoring device 104 may include other types of sensors that may be worn, carried, or coupled to the patient 101 to measure activities of the patient 101 that may affect the glucose level or glycemic response of the patient 101. For example, the sensor may include an accelerometer sensor configured to detect acceleration of the patient 101 or a portion of the patient 101 (such as a person's hand or foot), whose position changes may be associated with the activities of the patient 101. For example, the acceleration or movement (or lack of acceleration or movement) of the body part of the patient 101 may indicate exercise, sleep, or food / drink consumption activities of the patient 101, which may affect the glycemic response of the patient 101. In some embodiments, the sensor may measure heart rate and / or body temperature, which may indicate the amount of physical exercise experienced by the patient 101. In some embodiments, the sensor may include a global positioning system (GPS) receiver that may detect GPS signals to determine the location of the patient 101.

[0058] The above sensors are provided as examples only. Other sensors or other types of sensors for monitoring physiological conditions, activities and / or positions, etc. will be recognized by those skilled in the art and are conceivable within the scope of the present disclosure. For any sensor, the signal provided by the sensor may be referred to as a "sensor signal" herein. As used herein, the term "sensed data" may mean and include information represented by a sensor signal or by a preprocessed sensor signal. In some embodiments, the sensed data may include a glucose level in the patient 101, an acceleration of a portion of the patient 101, a heart rate of the patient 101, a temperature of the patient 101, and / or a geographic location (e.g., GPS location) of the patient 101, etc. The monitoring device 104 may communicate the sensed data to the delivery device 102 via the communication link 112 and / or communicate the sensed data to the computing device 106 via the communication link 114. The use of the sensed data by the delivery device 102 and / or the computing device 106 is described in more detail below.

[0059] In some embodiments, the monitoring device 104 may include components and / or circuits configured to pre-process sensor signals. Pre-processing may include, for example, amplification, filtering, attenuation, scaling, isolation, normalization, transformation, sampling and / or analog-to-digital conversion, etc. In some embodiments, the monitoring device 104 may host an App for processing sensor signals. In some embodiments, the monitoring device 104 may include a wired or wireless transceiver for sending sensor signals or receiving commands or instructions as described above. Those skilled in the art will recognize the various specific implementations of such pre-processing, including but not limited to specific implementations using processors, controllers, integrated circuits, application-specific integrated circuits (ASICs), hardware, firmware, programmable logic devices and / or machine executable instructions, etc. The types of pre-processing and their specific implementations are provided only as examples. It is conceivable that other types of pre-processing and specific implementations are within the scope of the present disclosure. In some embodiments, the monitoring device 104 may not perform pre-processing.

[0060] The computing device 106 may provide processing capabilities and may be implemented in a variety of ways. In some embodiments, the computing device 106 may be a consumer device such as a smartphone, a computerized wearable device (e.g., a smartwatch), a tablet computer, a laptop computer, or a desktop computer, etc., or may be a dedicated device (e.g., a portable control device) provided by, for example, a manufacturer of the delivery device 102. In some embodiments, the computing device 106 may be a processing circuit that may be integrated with another device such as the delivery device 102. In some embodiments, the computing device 106 may be secured to the patient 101 (e.g., secured to the body or clothing of the patient 101), may be at least partially implanted in the body of the patient 101, and / or may be held by the patient 101.

[0061] For each of the embodiments of the computing device 106, the computing device 106 may include various types of logic circuits, including but not limited to microprocessors, controllers, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), central processing units (CPUs), graphics processing units (GPUs), programmable logic devices, memories (e.g., random access memories, volatile memories, non-volatile memories, etc.), or other discrete or integrated logic circuits, as well as combinations of such components. The term "processing circuit" may generally refer to any of the foregoing logic circuits, alone or in combination with other logic circuits, or any other circuits for performing computations.

[0062] Some aspects of the delivery device 102, the monitoring device 104, and the computing device 106 have been described above. One or more of the devices 102 to 106 may include a user interface (not shown) that presents information to the patient 101 and / or receives information from the patient 101. The user interface may include a graphical user interface (GUI), a display device, a keyboard, a touch screen, a speaker, a microphone, a vibration motor, a button, a switch, and / or other types of user interfaces. Those skilled in the art will recognize the various types of user interfaces that can be used, and all such user interfaces are conceivable to be within the scope of the present disclosure. For example, in the case where the computing device 106 is a consumer device such as a smart phone, a tablet computer, a laptop computer, etc., the user interface will include a display device, a physical and / or virtual keyboard, and / or an audio speaker, etc. provided by such a consumer device. In some embodiments, the user interface may notify the patient 101 of the sensed data (e.g., glucose level) and / or insulin delivery data (e.g., historical, current, or future insulin delivery rate) and may present an alarm to the patient 101. In some embodiments, the user interface may receive input from the patient 101, which may include, for example, requested insulin delivery setting changes and / or meal instructions, etc. The above description and embodiments of user interfaces are provided merely as examples, and other types and other uses of user interfaces are contemplated as being within the scope of the present disclosure.

[0063] In one specific example, the communication between devices 102 to 106 and the collaboration between devices 102 to 106 may be used for insulin delivery. Figure 1As depicted and described above, devices 102 to 106 may communicate with each other via communication links 112 to 116. In some embodiments, computing device 106 may control the operation of delivery device 102 and / or monitoring device 104. For example, computing device 106 may generate one or more signals (e.g., command signals) that cause delivery device 102 to deliver insulin to patient 101, for example, at a basal dose and / or a bolus dose. In some embodiments, computing device 106 may receive data associated with insulin delivery (e.g., insulin delivery data) from delivery device 102 and / or sensed data (e.g., glucose levels) from monitoring device 104, and may perform calculations based on the insulin delivery data, sensed data, and / or other data to control delivery device 102. Insulin delivery data may include, but is not limited to, the type of insulin delivered, historical insulin delivery rate and / or amount, current insulin delivery rate and / or amount, insulin delivery time, and / or user input that affects insulin delivery. As will be appreciated by those skilled in the art, in the closed-loop mode of operation, computing device 106 may communicate dosage commands to delivery device 102 based on, for example, the difference between the current glucose level in the body of patient 101 (e.g., received from monitoring device 104) and the target glucose level (e.g., determined by computing device 106 or set on delivery device 102). The dosage commands may indicate the amount of insulin to be delivered and / or the rate (or time) of insulin delivery, and may adjust the current glucose level toward the target glucose level.

[0064] The remote / cloud computing system 108 can be any proprietary remote / cloud computing system or commercial cloud computing system including one or more server computing devices. When the computing resources of the client computing device (e.g., computing device 106) are insufficient, the remote / cloud computing system 108 can provide alternative or additional computing resources as needed. The computing device 106 and the remote / cloud computing system 108 can communicate with each other through a communication link 118, which can traverse one or more communication networks (not shown). The communication network may include, for example, an Ethernet network, a Wi-Fi network, a cellular network, a satellite network, an intranet, an extranet, the Internet, and / or an Internet backbone network. Those skilled in the art will recognize the specific implementation of the remote / cloud computing system 108 and how to interact with such a system through various types of networks. For example, the remote / cloud computing system 108 may include a processing circuit array and may execute machine-readable instructions. It is conceivable that such specific implementations, interfaces, and networks are within the scope of the present disclosure.

[0065] In some embodiments, the remote / cloud computing system 108 can make a treatment determination (e.g., an amount of insulin or an adjusted amount of insulin) and can communicate the treatment to the delivery device 102 via the computing device 106. In some embodiments, the computing device 106 can make a treatment determination and communicate the treatment determination to the delivery device 102. In some embodiments, the monitoring device 104 can make a treatment determination and communicate the treatment determination to the delivery device 102 directly or through an intermediary such as the computing device 106.

[0066] Figure 2 2 is a block diagram of an example of a glucose control system 200 according to certain embodiments. In the illustrated example, the glucose control system 200 may include a glucose sensor subsystem 210, a controller 220, an insulin delivery subsystem 230, a glucose delivery subsystem 240, and a glucagon delivery subsystem 250. The glucose sensor subsystem 210 may generate sensor glucose (SG) signals (e.g., SG levels) (these SG signals may be estimates of blood glucose levels in the body 260), and may provide the SG signals to the controller 220. The controller 220 may receive the SG signals and generate commands to the insulin delivery subsystem 230, and in some implementations, to the glucose delivery subsystem 240 and / or the glucagon delivery subsystem 250. The insulin delivery subsystem 230 may receive commands from the controller 220 and deliver insulin to the body 260 in accordance with the commands. In some embodiments, the glucose delivery subsystem 240 may receive commands from the controller 220 and provide glucose to the body 260 in accordance with the commands. In some embodiments, glucagon delivery subsystem 250 may receive commands from controller 220 and deliver glucagon into body 260 in accordance with the commands.

[0067] In some implementations, the glucose sensor subsystem 210 may include a glucose sensor, a sensor electronics configured to generate a SG signal, a sensor communication system configured to transmit the SG signal to the controller 220, and a housing for the sensor electronics and the sensor communication system. The glucose sensor may measure blood glucose levels, for example, directly from the bloodstream or indirectly via interstitial fluid using a subcutaneous sensor, as described in more detail below.

[0068] The controller 220 may include electrical components and software that generate commands for the insulin delivery subsystem 230, the glucose delivery subsystem 240, and / or the glucagon delivery subsystem 250. The controller 220 may include a controller communication system to receive sensor signals and provide commands to the insulin delivery subsystem 230, the glucose delivery subsystem 240, and / or the glucagon delivery subsystem 250. In some specific implementations, the controller 220 may implement a glucose calculator. In some specific implementations, the controller 220 may include a user interface and / or an operator interface (not shown), which includes a data input device and / or a data output device. Such a data output device may, for example, generate a signal to initiate an alarm and / or include a display or printer to display the status of the controller 220 and / or the patient's vital signs. Such a data input device may include a dial, button, pointing device, manual switch, alphanumeric key, touch-sensitive display, a combination thereof, etc. for receiving user input and / or operator input. For example, such a data input device can be used to schedule and / or initiate an insulin bolus injection for a meal. It should be understood, however, that these are merely examples of input and output devices that may be part of an operator and / or user interface and that claimed subject matter is not limited in these respects.

[0069] The insulin delivery subsystem 230 may include, for example, an infusion device and / or an infusion tube to infuse insulin into the body 260. Similarly, the glucose delivery subsystem 240 may include, for example, an infusion device and / or an infusion tube to infuse glucose into the body 260. Similarly, the glucagon delivery subsystem 250 may include, for example, an infusion device and / or an infusion tube to infuse glucagon into the body 260. In some embodiments, a shared delivery system and / or infusion tube may be used to infuse insulin, glucagon, and / or glucose into the body 260. In some embodiments, an intravenous system for providing fluids to a patient (e.g., in a hospital or other medical environment) may be used to infuse insulin, glucagon, and / or glucose. However, it should be understood that certain exemplary embodiments may include the insulin delivery subsystem 230 without the glucagon delivery subsystem 250 and / or without the glucose delivery subsystem 240. In some embodiments, each of the insulin delivery subsystem 230, the glucose delivery subsystem 240, and the glucagon delivery subsystem 250 may include an infusion electrical component that activates an infusion motor based on commands from the controller 220, an infusion communication system that receives commands from the controller 220, and a delivery subsystem housing.

[0070] In some embodiments, the controller 220 may be housed in a delivery subsystem housing, and the infusion communication system may include electrical traces or wires that carry commands from the controller 220 to the delivery subsystem. In some embodiments, the controller 220 may be housed in a sensor system housing, and the sensor communication system may include electrical traces or wires that carry sensor signals from sensor electrical components to controller electrical components. In some embodiments, the controller 220 may have its own housing or may be included in an auxiliary device. In some embodiments, the controller 220 may be co-located with the delivery subsystem and the sensor system in a single housing. In some embodiments, the sensors, controllers, and / or infusion communication systems may utilize cables; wires; fiber optic lines; RF, IR, or ultrasonic transmitters and receivers; combinations thereof; and / or the like in place of electrical traces, to name just a few examples.

[0071] In some embodiments, the glucose control system 200 may also include a meal intake monitoring subsystem 215. The meal monitoring subsystem 215 may be used to record the user's food intake, or may automatically detect the user's food intake. For example, in some specific implementations, the user may input an estimated amount of carbohydrates in a food item and / or a meal during a meal. In some specific implementations, the meal intake monitoring subsystem 215 may include sensors (e.g., cameras or accelerometers) that may automatically detect meal events, food items during a meal, and / or an estimated amount of carbohydrates in a meal. The estimated amount of carbohydrates in the meal may be transmitted to the controller 220, which may determine the appropriate amount of meal bolus and generate an order for the insulin delivery subsystem 230 to deliver the meal bolus. In some specific implementations, the meal intake monitoring subsystem 215 may not be used in the glucose control system 200, and the dosage of the meal bolus may be determined based on the measured glucose level.

[0072] Figure 3 310 is a perspective view of an example of a CGM system 300 according to certain embodiments. In the illustrated example, the CGM system 300 may include a subcutaneous sensor kit 302 that is provided for subcutaneously placing an active portion of a flexible sensor 310 or the like at a selected site within a user's body. The subcutaneous or transcutaneous portion of the sensor kit 302 includes a hollow slotted insertion needle 326 having a sharp tip 312 and a cannula 322. The sensor kit 302 may facilitate accurate placement of the flexible sensor 310 into the user's body. The glucose sensing portion 318 of the flexible sensor 310 is located inside the cannula 322. The glucose sensor portion 318 includes one or more glucose sensor electrodes 316 that may be exposed to a user's bodily fluid 390, for example, through a window 314 formed in the cannula 322. The sensor electrodes 316 may include, for example, a counter electrode, a reference electrode, and one or more working electrodes.

[0073] The proximal portion of the flexible sensor 310 may be mounted in a mounting base 305 suitable for placement on the skin of a user. In some embodiments, the mounting base 305 may be a pad having a lower surface coated with a suitable pressure-sensitive adhesive layer 334, wherein a release paper tape 336 is typically provided to cover and protect the adhesive layer 334 until the sensor kit 302 is ready for use. The mounting base 305 may include an upper layer 330 and a lower layer 331, wherein a connecting portion 328 of the flexible sensor 310 is sandwiched between the upper layer 330 and the lower layer 331. The connecting portion 328 may include a front section coupled to the glucose sensing portion 318 of the flexible sensor 310, which may be folded at an angle to extend downward through a hole 320 formed in the lower layer 331 of the mounting base 305. Optionally, the adhesive layer 334 (or another portion of the device that contacts the tissue in the body) may include an anti-inflammatory agent for reducing an inflammatory response and / or an antimicrobial agent for reducing the chance of infection. The insertion needle 326 may be adapted to be received by a snug fit through a needle port 324 formed in an upper layer 330 of the mounting base 305 and through a lower hole 320 in a lower layer 331 of the mounting base 305. In some embodiments, the insertion needle 326 may be withdrawn after insertion to leave the glucose sensing portion 318 (including the glucose sensor electrode 316) and / or the cannula 322 in place at a selected insertion site.

[0074] The flexible sensor 310 of the subcutaneous sensor kit 302 can be connected to the sensor electronics 304, which can be referred to as a transmitter. For example, the mounting base 305 can be designed so that the glucose sensing portion 318 can be coupled to the connecting portion 328, which terminates at a conductive contact pad, etc. The connecting portion 328 and the contact pad can be adapted to be electrically connected to the sensor electronics 304 to determine the user's glucose level in response to the signal from the sensor electrode 316. The connecting portion 328 can be electrically connected to the sensor electronics 304 through a connector block 338. The sensor electronics 304 can include, for example, a housing 350 supporting a printed circuit board 344, a battery 354, and an antenna 346. In some embodiments, the housing 350 can include an upper housing 356 and a lower housing 352, which are sealed by, for example, ultrasonic welding to form a waterproof (or water-resistant) seal, thereby allowing cleaning by immersion (or wiping) with water, detergent, alcohol, etc. In some embodiments, the upper housing 356 and the lower housing 352 can be formed of medical grade plastic. In alternative embodiments, the upper housing 356 and the lower housing 352 may be connected together and bonded together by other methods such as snap fit, sealing rings, RTV (silicone sealant), etc., or formed of other materials such as metals, composites, ceramics, etc. In other embodiments, the components may be encapsulated in an epoxy or other moldable material that is compatible with the electronics and reasonably moisture resistant. In some embodiments, the sensor electronics 304 may be connected to the flexible sensor 310 via the connector block 338 and the connector 340, and may be mounted to the mounting substrate 305. In some embodiments, the sensor electronics 304 may be connected to the flexible sensor 310 via the connector block 338, the connector 340, and the cable 342. In some embodiments, the bottom surface of the lower housing 352 may be coated with a suitable pressure sensitive adhesive layer 358, wherein a release paper tape 348 is typically provided to cover and protect the adhesive layer 358 until the sensor electronics 304 are ready for use.

[0075] The battery 354 may include a rechargeable battery or a non-rechargeable battery. In some embodiments, the battery 354 may include three silver oxide battery cells connected in series. In other examples, different battery chemistries may be used, such as lithium-based chemistries, alkaline batteries, nickel metal hydride, etc., and different numbers of batteries may be used. The battery 354 may provide power to the sensor assembly 302 via the sensor electronics 304 through, for example, the connector block 338, the connector 340, and / or the cable 342.

[0076] The sensor electronics 304 may include a processor, a transceiver unit, a signal processing / conditioning circuit, and a memory device on a printed circuit board 344. The processor may be configured to receive a current and / or voltage signal from a sensor electrode 316 of the flexible sensor 310. The sensor electrode 316 may generate a sensor signal indicating the concentration of an analyte (e.g., glucose) being measured. In one example, the sensor signal may include a current measured at a working electrode, which may include glucose oxidase (GOx) to catalyze a reaction with glucose. The sensor signal may also include a voltage signal measured at a counter electrode of the sensor electrode 316 or between the working electrode and the counter electrode (or reference electrode). In some examples, other electrical parameters such as capacitance, resistance, impedance, etc. may also be measured using the sensor electrode 316 inserted into the user's body.

[0077] Examples of impedance parameters measured using the sensor electrodes 316 may include electrochemical impedance spectroscopy (EIS) values. EIS can provide additional information in the form of sensor impedance and impedance-related parameters at different frequencies. In addition, for certain frequency ranges, the impedance and / or impedance-related data may be substantially independent of glucose. This glucose independence can enable the use of a variety of EIS-based markers or indicators to produce robust, highly reliable sensor glucose values ​​(e.g., through fusion methods), and also to assess the condition, health, age, and efficiency of individual electrodes and the overall sensor substantially independently of the glucose-related current signal. In one example, analysis of glucose-independent impedance data can provide information about the efficiency of the glucose sensor, i.e., the speed at which the glucose sensor is hydrated and ready for data acquisition. In addition, glucose-independent impedance data (e.g., using the value of 1kHz actual impedance) can provide information about potential occlusions that may exist on the surface of the sensor membrane, where the occlusion may temporarily prevent glucose from entering the sensor and thus cause a signal drop. In addition, for example, using the values ​​of phase angle and / or imaginary impedance at 1kHz and higher frequencies, glucose-independent impedance data can provide information about sensor sensitivity loss during long-term wear.

[0078] The processor of the sensor electronic device 304 may receive the sensor signal and calibrate the sensor signal using a calibration function, which may be determined, for example, using a reference value. In some embodiments, the parameters of the calibration function (e.g., calibration factor or sensor sensitivity) may be determined (e.g., using calibration data) or pre-determined and stored in a memory for use by the processor. For example, in some embodiments, the processor may implement or otherwise execute a calibration application module that calculates or otherwise determines calibration parameters based on measured values ​​and reference values ​​(e.g., using, for example, a blood glucose value measured by a finger-stick blood glucose meter). Based on the sensor signal and the determined calibration function, the processor may estimate the blood glucose level of the user. In some embodiments, the processor may store the glucose measurement in a memory. The sensor measurement result may be displayed on a display of the sensor electronic device 304 or may be transmitted to the receiver 306 for display.

[0079] The memory of the sensor electronics 304 can be any type of memory device and can be configured to store raw sensor signals from the flexible sensor 310 and / or glucose measurements (e.g., SG values) generated by the processor, calibration parameters for determining glucose measurements from the sensor signals, or other data used and / or generated by the processor. The memory can also store software and / or firmware that can be executed by the processor. In some embodiments, the memory can also include configuration information for configuring the operation of the sensor electronics 304.

[0080] The signal processing / conditioning circuitry of the sensor electronics 304 may include one or more amplifiers, filters, current-to-voltage converters, integrators, analog-to-digital converters, potentiostats, etc. The signal processing / conditioning circuitry may filter the signal, apply a constant reference voltage level, amplify the current or voltage signal, sample the analog current or voltage signal and convert the analog current or voltage signal to a digital data value, etc.

[0081] The transceiver unit of the sensor electronics 304 may include a wireless transceiver that may transmit the glucose measurements determined by the processor and stored in the memory to the receiver 306 using a suitable wireless communication protocol such as Bluetooth (e.g., Bluetooth Low Energy (BLE)), Wi-Fi, WiMax, etc. For example, the transceiver unit may transmit a radio signal indicating the glucose level at a regular time period (e.g., once every 5 minutes) to provide a real-time sensor glucose (SG) value. The transceiver unit may also receive data from the receiver 306, such as configuration data and calibration data (e.g., reference glucose values). In some embodiments, the transceiver unit of the sensor electronics 304 may transmit data to the receiver 306 via a cable or wire.

[0082] The receiver 306 may include a monitor 335 for monitoring physiological characteristic readings and / or generating warning signals, and a display 332 for displaying physiological characteristic readings and / or warning signals. For example, the SG value / graph may be displayed on the display 332 of the receiver 306 so that the user can monitor blood glucose levels and administer insulin using an insulin pump. In various embodiments, the receiver 306 may be implemented in a desktop computer, a pager, a television including communication capabilities, a notebook computer, a server, a network computer, a personal digital assistant (PDA), a smart phone, an infusion pump including a display, a glucose sensor including a display, a combination infusion pump / glucose sensor, etc.

[0083] Figure 4A An example of a glucose sensor 400 according to certain embodiments is illustrated. The glucose sensor 400 may be an example of the flexible sensor 310 described above. In the illustrated example, the glucose sensor 400 may include a sensing portion 410 including one or more electrodes, such as a first working electrode (WE1) 412, a reference electrode (RE) 414, a second working electrode (WE2) 416, and a counter electrode (CE) 418, wherein these electrodes may be electrically isolated by an isolation layer. The sensing portion 410 may be an in vivo portion (also referred to as a distal end) that may be inserted into the subcutaneous layer of a user using a hollow needle (such as the insertion needle 326). The glucose sensor 400 may include a bending region 420 that may be bent, for example, up to about 90° before or after insertion. The glucose sensor 400 may also include a contact pad region 430 that may be an out-of-body portion of the glucose sensor 400. As illustrated, the contact pad region 430 may include one or more contact pads electrically connected to one or more electrodes in the sensing portion 410 by conductive traces (not shown). The contact pads may include, for example, a contact pad 438 connected to the counter electrode 418, a contact pad 434 connected to the reference electrode 414, and two contact pads 432 and 436 connected to the first working electrode 412 and the second working electrode 416, respectively. The contact pads in the contact pad region 430 of the glucose sensor 400 may include, for example, a metal layer and may be connected to one or more contact pads on a printed circuit board of the sensor electronics. In the illustrated example, the contact pad region 430 of the glucose sensor 400 may also include a pair of guide holes 440 that may cooperate with a pair of guide pins on the printed circuit board to align the contact pads on the contact pad region 430 with the contact pads on the printed circuit board so that the voltage required for glucose sensing can be applied and a sensor current proportional to the interstitial glucose concentration can be measured by the sensor electronics on the printed circuit board.

[0084] Figure 4BAnother example of a glucose sensor 402 according to certain embodiments is illustrated. The glucose sensor 402 may be another example of the flexible sensor 310 described above. In the illustrated example, the glucose sensor 402 may include a sensing portion 450 including one or more electrodes, such as a reference electrode (RE) 452, one or more working electrodes (WE) 454, and one or more counter electrodes (CE) 456, wherein the electrodes may be electrically isolated by an isolation layer. The sensing portion 450 may be an in vivo portion that may be inserted into the subcutaneous layer of a user using a hollow needle, such as the insertion needle 326. The glucose sensor 402 may include a bending region 460 that may be bent, for example, up to about 90° before or after insertion. The glucose sensor 402 may also include a contact pad region 470 that may be an ex vivo portion of the glucose sensor 402. As illustrated, the contact pad region 470 may include one or more contact pads electrically connected to one or more electrodes in the sensing portion 450 by conductive traces (not shown). The contact pads may include, for example, a contact pad 472 connected to the reference electrode 452, a contact pad 474 connected to the working electrode 454, and a contact pad 476 connected to the counter electrode 456. The contact pads in the contact pad region 470 of the glucose sensor 402 may include, for example, a metal layer and may be connected to one or more contact pads on a printed circuit board of the sensor electronics. In the illustrated example, the contact pad region 470 of the glucose sensor 402 may also include a pair of guide holes 478 that may cooperate with a pair of guide pins on the printed circuit board to align the contact pads on the contact pad region 470 with the contact pads on the printed circuit board so that the voltage required for glucose sensing can be applied and a sensor current proportional to the interstitial glucose concentration can be measured by the sensor electronics on the printed circuit board.

[0085] Figure 4CA cross section of an example of a working electrode 404 of a glucose sensor according to certain embodiments is illustrated. As illustrated, the working electrode 404 may include a substrate or base layer 480 (e.g., comprising polyimide) that provides a base and rigidity for a flexible substrate. An overlying electroplated metal layer 482 (e.g., comprising chromium and / or gold) may be deposited on the base layer 480. The chromium layer may promote adhesion between the polyimide and the gold layer, which enables the sensor to transfer charge when a bias is applied. An electroplated layer 484 (e.g., comprising platinum) may be formed on the metal layer 482, such as in a groove formed in an insulating layer 486 (e.g., comprising polyimide). The insulating layer 486 may help improve sensor integrity by providing electrical insulation between the working electrode, the counter electrode, and the reference electrode, and may also serve as a support barrier for the electrode and the sensor's flexible chemistry. The electroplated layer 484 may be an active electrode for analyte (glucose) sensing, where glucose may be sensed at the working electrode by generating charge through oxidation of hydrogen peroxide. The working electrode is also responsible for maintaining the bias potential V applied by the sensor electronics. 设定 For example, a glucose oxidase (GOx) layer 488 is formed on the insulating layer 486 by depositing a glucose oxidase solution in a groove formed in the insulating layer 486 (e.g., via slot coating). A human serum albumin (HSA) layer 490 is formed on the glucose oxidase layer 488. An adhesion promoter layer 492 may be used on the HSA layer 490 to attach a glucose limiting membrane (GLM) layer 494 to the working electrode 404. Note that Figure 4C A simplified representation of one example of a working electrode is depicted, and actual embodiments may include any number of additional layers and / or alternative layers (eg, high density amine (HDA) layers, etc.). Thus, the subject matter described herein is not intended to be limited to Figure 4C The reference electrode or counter electrode may have a structure similar to Figure 4C The structure of the structure shown in , but some modifications may be made. For example, in some embodiments, the reference electrode may include an Ag layer instead of a platinum layer, and may not include a glucose oxidase layer.

[0086] The glucose oxidase in the GOx layer 488 may comprise the active enzyme of the sensor. In the presence of oxygen, GOx may decompose glucose into gluconic acid and hydrogen peroxide. The HSA layer 490 may decompose hydrogen peroxide into water. The adhesion promoter layer 492 may comprise, for example, 3-aminopropyltriethoxysilane (APTES), which is a compound that provides a better bonding surface between the HSA layer 490 and the GLM layer 494. The GLM layer 494 may limit the amount of glucose and oxygen delivered to the GOx layer 488 of the working electrode 404 to ensure that the reaction is glucose-limited. The GOx layer 488 or another active enzyme layer on the working electrode 404 may decompose glucose and oxygen into gluconic acid and hydrogen peroxide. The generated peroxide molecules may interact with the working electrode 404 (e.g., biased at about 0.2V to about 0.7V) to decompose hydrogen peroxide into two hydrogen ions, oxygen, and two electrons at the surface of the working electrode 404. The charge may be forced to move between the electrodes, thereby generating a sensor current signal (Isig) that may be measured by the sensor electronics. For example, when GOx enzyme is used as a catalyst in a glucose sensor, if oxygen is present near the enzyme and sensor electrodes, current can flow from the counter electrode to the working electrode. Other signals can also be measured, such as the counter voltage (Vcntr, e.g., the voltage potential difference between the counter electrode and the working electrode), electrochemical impedance spectroscopy (EIS) signals at different frequencies, etc.

[0087] In some embodiments, the measurement of the parameters of the electrodes of the above-mentioned electrochemical glucose sensor can be performed during and / or after the manufacturing process. For example, the imaginary impedance of the working electrode (and the counter electrode) can be measured after the platinum electroplating process of forming the electroplating layer 484. The GOx solution activity measurement can be performed during or after the GOx solution preparation process before deposition, and the thickness of the GOx layer 488 can be measured after the slot coating and selective patterning process on the working electrode 404. The HSA concentration can be measured during or after the solution preparation process before coating the HSA layer 490, and the thickness of the GLM layer 494 can be measured after forming the GLM layer 494. In some embodiments, these measurements can be performed using sacrificial or monitoring instances of glucose sensors manufactured on wafers. In some embodiments, additional manufacturing process measurements (such as surface roughness or other morphology measurements) can be performed on the electrodes during or after the manufacturing process (e.g., via interferometry).

[0088] Figure 5 Schematic diagram of an example of a glucose sensor 500 according to some embodiments is shown. The glucose sensor 500 may include a potentiostat 530 electrically connected to electrodes of the glucose sensor 500. As illustrated, the potentiostat 530 may include an operational amplifier (op amp) 532 (or other servo control device) including two inputs V 设定 and V测量 . V 测量 V may be the measured voltage between the reference electrode 512 and the working electrode 510 of the sensor flex circuit as described above. 设定 The output of the operational amplifier 532 may be connected to the counter electrode 514 to apply a voltage on the counter electrode 514. The Op amp 532 may utilize feedback through the sensor electrode to attempt to maintain a desired voltage between the reference electrode 512 and the working electrode 510 by adjusting the voltage at the counter electrode 514. The counter electrode 514 may balance the chemical reaction occurring at the working electrode 510. As described above, due to the electrochemical reaction between glucose and the sensor chemical substance of the working electrode 510, a current 520 may flow between the counter electrode 514 and the working electrode 510. This current may indicate the level of glucose in the interstitial fluid of the user, and may be measured by the ammeter 540 as a sensor current (Isig), or may be output to the processing circuit of the sensor electronics for measurement. The sensor electronics of the glucose sensor may also be used to measure other signals, such as a counter voltage (Vcntr, e.g., a voltage potential difference between the counter electrode and the working electrode), electrochemical impedance spectroscopy (EIS) values ​​at different frequencies, and the like. Based on these measured raw sensor signals, one or more models may be used to determine the user's blood glucose level.

[0089] Figure 6 Included is a flowchart 600 illustrating an example of a method for determining a sensor glucose value using input data from a glucose sensor according to certain embodiments. Figure 6In the illustrated example, the method may include sensor feature generation at box 610, optional blood glucose calibration at box 620, sensor glucose modeling at box 630, and optional conditional blanking and termination at box 640. In some embodiments, the sensor feature generation at box 610 may use the raw gap current signal, electrochemical impedance spectroscopy signal, and reverse voltage signal to extract input features used by a downstream SG model (e.g., a machine learning model). The optional blood glucose calibration at box 620 may include adjusting the input sensor features from box 610 based on calibration parameters determined using a reference blood glucose value. The sensor glucose modeling at box 630 may include applying machine learning techniques to the input features to convert the input features into sensor glucose values. The optional conditional blanking and termination at box 640 may include applying various tests and logic to determine, for example, when to stop displaying the sensor output signal or terminate the sensor to reduce the probability of displaying noise or erroneous information to a user or receiving device. In some embodiments, terminating the sensor may include stopping sending sensor data from the sensor device. In some embodiments, input data (e.g., Isig, Vcntr, EIS, and / or BG values ​​of gap measurements) may pass through blocks 610 to 640 and be transformed into SG values. Table 1 below shows examples of inputs and outputs for each of the four blocks of flowchart 600. In some embodiments, the process of flowchart 600 may be implemented using one or more machine learning models.

[0090] Table 1. Information Transmission in Flowchart 600

[0091] Figure 7 An example of a sensor feature generator 700 according to certain embodiments is illustrated. The sensor feature generator 700 may be used to perform Figure 6The operation of sensor signature generation at block 610 of the sensor signature generator 700 is described. In the illustrated example, the input signals 710 to the sensor signature generator 700 may include an Isig signal 712 (e.g., at intervals of about 1 minute), a Vcntr signal 714 (e.g., at intervals of about 5 minutes), and an EIS signal 716 (e.g., at intervals of about 15 minutes to 30 minutes). At block 720, the Isig signal 712 may be collected over, for example, 5 minutes or more, and the collected Isig signal 712 may be used to generate a filtered Isig signal 722 (e.g., within a 5 minute time window). For example, the Isig signal 712 for the most recent 8 minutes may be filtered using, for example, a seventh-order finite impulse response (FIR) filter to produce the filtered Isig signal 722. In some specific implementations, for example, if the value of the Isig signal 722 for the last 8 minutes is greater than twice the average value of the Isig signal 712 for 8 minutes or less than half the average value of the Isig signal 712 for 8 minutes, any calculated invalid filtered Isig signal 712 may be discarded. In some embodiments, artifact detection may be performed to identify large and small drops in the Isig signal 712 or the filtered Isig signal 722. The EIS signal 716 may include measurements made using signals of various frequencies (e.g., a frequency set between about 0.1 Hz and about 8 kHz). The EIS signal 716 may be verified to generate a verified EIS signal 730. In some embodiments, the verified EIS signal 730 may be converted to 5-minute sampling using a zero-order hold. In some embodiments, the frequency of the measurement signal may be changed according to the hardware design. For example, Vcntr may be measured once every 1 minute, while EIS may be measured more frequently, depending on factors such as battery life and memory size limitations.

[0092] The filtered Isig signal 722, the verified EIS signal 730, and the Vcntr signal 714 may be used to generate features 740, wherein the features for each time period (e.g., a 5 minute time period) may be included in the corresponding data packet. Derived features 750 may be added to the data packet based on the features 740 in the current data packet and the previous features 740 in the previous data packet. In some embodiments, the transformed features ( Figure 7). For example, principal component analysis (PCA) or Autoencoder may be used to generate transformed features. Final features 760 including features 740, derived features 750, and / or transformed features may include, for example, filtered Isig signal 722, previously filtered Isig signal, cumulative average of Isig signal, moving average of Isig signal, filtered Isig signal, Vcntr signal 714 of current data group, Vcntr signal of previous data group, sensor age, real and imaginary components of EIS signal at various frequencies, changes in real and imaginary components of EIS signal at various frequencies, etc. In some embodiments, sensor feature generator 700 may process Isig signal, Vcntr signal, and EIS signal based on different time intervals (e.g., shorter or longer than 5 minutes). Data packets including all or some of the features 740 , derived features 750 , and / or transformed features may be used for optional blood glucose calibration at box 620 , sensor glucose modeling at box 630 , and optional conditional blanking and termination at box 640 .

[0093] Due to production variations and aging of glucose sensors, some glucose sensors may not provide accurate readings based on the measured signals and derived signal characteristics. Therefore, glucose sensors may require occasional or periodic in vitro and / or in vivo calibration to improve the accuracy and / or reliability of the glucose sensors. Figure 6 As shown in the example of , the sensor signal can be calibrated at box 620 using a calibration function that is determined using a known good blood glucose value (such as a finger stick measurement from a blood glucose meter that measures blood glucose levels in capillaries). For example, a calibration ratio CR (or calibration factor) can be determined according to CR=BG / (Isig+offset). The calibration ratio and / or offset can then be used to calibrate an input signal such as an Isig signal. In some embodiments, a weighted calibration ratio can be determined based on a previous calibration ratio and the time of the previous calibration and applied to all subsequent data packets until the next calibration is performed.

[0094] At block 630, one or more SG models may be used to derive sensor glucose values ​​from the generated features. The SG model may include, for example, a function, an equation, a machine learning model, or other transformation models that may map the generated features, Isig signals, Vcntr signals, and / or EIS signals to sensor glucose (SG) values ​​indicating estimated blood glucose levels. For example, the features generated at blocks 610 and / or 620 may be input to one or more SG models as feature vectors, and the SG value may be determined as a dot product of the feature vector and a coefficient vector having corresponding coefficients (e.g., weights). For example, the SG model may be represented as:

[0095]

[0096] In a simplified example, the SG value can be calculated as follows:

[0097] ,

[0098] Where α, β, γ, and δ may be coefficients associated with age, Isig, Vcntr, and EIS, and Δ is an offset value. In some embodiments, at least one of Isig, Vcntr, or EIS may include a vector, and at least one of β, γ, or δ may be a coefficient vector. In some embodiments, at least one of α, β, γ, and / or δ may be a function of time (e.g., age of the sensor).

[0099] As described above, sensor properties associated with glucose sensors (such as age (or time in use or time worn), battery level, calibration accuracy, sensing environment (e.g., oxygen level, temperature, humidity, and pH, etc.), physiological dynamics, production variability, and other properties) may affect enzymatic glucose level measurements in complex ways that may be difficult to model. Therefore, the accuracy of many CGM sensors may not be optimal and may show large differences in measurement results. In addition to production variability from sensor to sensor, degradation of sensor performance over time, etc., the lower-than-expected performance of CGM sensors may also be at least partially attributed to the complexity of modeling the in vivo behavior of CGM sensors. In many cases, a single SG model (e.g., a single equation) may not accurately estimate blood glucose levels because various sensor properties may affect blood glucose level measurements in complex ways that may be difficult to characterize. In some embodiments, different SG models may be used for different input parameter ranges, such as different ranges of age, Vcntr, or other parameters of sensor measurement data. A model may be selected from one or more models based on the range in which the input parameters of the sensor measurement data fall.

[0100] In some embodiments, the one or more models may include a machine learning model. The machine learning model may include a neural network or other machine learning models. For example, the machine learning model may include a neural network that includes a large number of nodes (neurons) arranged in one or more layers. Each node of the neural network may be connected to one or more other nodes of the neural network, wherein these connections may have associated coefficients or weights. In some embodiments, each individual node may implement a multiplication and accumulation function that combines the weighted values ​​of all its inputs based on the associated weights. In some embodiments, each connection (or node) may have a threshold function so that only signals exceeding the threshold may be propagated to other nodes. The neural network may be trained or self-learned using training samples rather than being explicitly programmed. In some embodiments, the neural network may include multiple layers, each layer including multiple nodes, wherein the input signal may traverse multiple layers to generate the final output. In some embodiments, the neural network may be trained using techniques such as back propagation techniques, such as determining coefficients or weights associated with connections between nodes. The data used to train the machine learning model may include, for example, sensor data (e.g., Isig, Vcntr, EIS, etc.) and corresponding known SG or BG values.

[0101] In some embodiments, the machine learning model may receive a multidimensional feature vector as input. For example, the input to the machine learning model may include features such as EIS features, Isig trend features, Isig rate of change features, previous Isig values, previous Vcntr values, etc. Some examples of input features and models for determining SG values ​​can be found in, for example, U.S. Patent Application No. 16 / 773,422, entitled "Methods, Systems, and Devices for Continuous Glucose Monitoring", which is incorporated herein by reference in its entirety. The output of the machine learning model may depend on a combination of input features. In some embodiments, one or more models may include multiple (e.g., N) models (e.g., functions, equations, machine learning models, etc.). These models may use the same or different input features and / or may use different weights to weight the input features differently. In some embodiments, one or more models may be selected based on the input data and the SG value may be determined using the one or more models. For example, in some embodiments, the outputs of two or more models selected from multiple models may be combined to generate a final SG value.

[0102] Figure 8 Examples of using one or more models to determine sensor glucose values ​​according to certain embodiments are illustrated. Figure 8As shown, the input 810 to one or more models may include Isig, EIS, Vcntr or other input signals. In some embodiments, the input 810 may additionally include previously modeled (e.g., trend) information. In the illustrated example, the signal modeling and feature engineering engine 820 may include N models, such as models 822, 824, ..., 826, which may be trained or otherwise determined for various combinations of inputs or scenarios. As described above, these models may include machine learning models and may be trained to use different inputs or weight the inputs differently. The model controller 830 may be an average, weighted average, median, maximum, minimum, etc. of the outputs of a subset of models 822, 824, ..., 826. For example, the model controller 830 may select a model based on the input 810 or other information (such as the age of the sensor) (e.g., from models 822, 824, ..., 826) to determine the SG value. In some embodiments, the model controller 830 may average the outputs of two or more of the models 822, 824, ..., 826 to generate a final sensor glucose value 840. In some embodiments, the model controller 830 may take the average, weighted average, median, maximum, minimum, etc. of the outputs of a subset of models 822, 824, ..., 826. In some embodiments, the model controller 830 may rank or weight the models 822, 824, ..., 826, and may select a high-ranking model. In some embodiments, a confidence score associated with the output of each of the models 822, 824, ..., and 826 may be generated to indicate a confidence level (or probability) that the output is correct, and the model controller 830 may adjust the model output to support one or more models associated with the highest confidence level. The final SG value 840 may be displayed to the user or may be transmitted to an insulin pump or insulin calculator. In some embodiments, a confidence score associated with the SG value 840 may be generated to indicate a confidence level that the final SG value 840 is correct.

[0103] At box 640, conditional blanking and termination may be performed on the SG values ​​generated by the SG model. For example, if the SG value is determined to be an outlier or otherwise erroneous, some SG values ​​generated by one or more SG models may be blanked and not transmitted or displayed to the user. In some embodiments, the conditional blanking and termination at box 640 may be performed by one or more models (e.g., the machine learning model used at box 630 or other machine learning models). For example, the machine learning model may classify the SG value as an outlier / non-outlier, having a large negative deviation / large positive deviation / nominal accuracy, having poor accuracy / medium accuracy / good accuracy, etc. The SG value may be blanked based on the output from the machine learning model. For example, if the output of the machine learning model indicates that the SG value is not an outlier (e.g., within a specific threshold range), the SG value may not be blanked. Otherwise, the SG value may be blanked. Similarly, if the output of the machine learning model indicates that the SG value has medium, normal, or good accuracy, the SG value may not be blanked. Otherwise, the SG value may be blanked. In some embodiments, when a large number of abnormal values ​​are generated, for example, using a sensor measurement signal from a sensor, the sensor can be terminated (e.g., stop calculating or sending any sensor data from the sensor and / or stop displaying any sensor data from the sensor).

[0104] In some embodiments, in order to achieve the desired accuracy and reliability and reduce the impact of noise and other stray signals, the CGM sensor may need to be calibrated in vitro and / or in vivo periodically or occasionally to improve the accuracy and / or reliability of the glucose sensor. For example, it may be necessary to calibrate the sensor data using a known good blood glucose value, which is usually obtained via a so-called "fingerstick measurement" using a blood glucose meter that measures blood glucose levels in capillaries. Performing such calibration measurements can increase the patient's burden and perceived complexity, and may be inconvenient, uncomfortable or otherwise unpopular with the patient. In addition, due to the time required for glucose to diffuse from the capillaries to the interstitial space, ISF glucose measurements may lag behind blood glucose measurements, which may require certain signal processing (e.g., filtering) or other techniques to compensate for physiological lags. Additionally, various factors may cause instantaneous changes in the sensor output, which may affect the accuracy of calibration and measurement. Deterioration of sensor performance over time and production variations between sensors may further exacerbate these challenges. Occasional or periodic calibration can improve the performance of the CGM sensor, but may not adequately address all challenges associated with the accuracy of the CGM sensor.

[0105] According to certain embodiments disclosed herein, different SG models may be used for different input parameter ranges (e.g., different subspaces of the input parameter space), such as age, Vcntr, Isig, or different ranges of other sensor measurement data. Each SG model (referred to as a regional SG model or regional model) may be optimized for a corresponding region of the input parameter space using, for example, machine learning techniques, and thus may be able to more accurately characterize the relationship between the glucose level and the sensor measurement data in the corresponding region. For example, in some embodiments, the input parameter space may be partitioned based on the age of the sensor, Vcntr, sensor current Isig, another input parameter, or any combination thereof. In one example, the input parameter space may be partitioned based on two parameters, wherein one of the two parameters may include sensor current Isig. In some examples, the input parameter space may be partitioned based on three or more parameters, wherein one of the three or more parameters may include sensor current Isig. The input parameter space may be partitioned uniformly or unevenly into regions having regular or irregular shapes. The SG model for each region may be trained separately using a corresponding training data set, or may be trained together using the same training data set. In some embodiments, the SG models may use the same input parameters and have a similar structure, but may have different coefficients or weights for the same input parameters. In some embodiments, the SG models for different regions of the input parameter space may use different input parameters and / or may include different types of models or different model structures.

[0106] Fig. 9A An example of a partition model 900 including a set of regional models for modeling the relationship between SG values ​​and sensor measurement data according to certain embodiments is shown. In the illustrated example, the input parameter space can be partitioned based on the range of two parameters 910 and 920, wherein one of the two parameters may include the sensor current Isig, and the other of the two parameters may include, for example, sensor age, Vcntr, in vivo characteristics of the glucose sensor (e.g., sensitivity), another input parameter of the SG model, or an environmental property of the glucose sensor (e.g., humidity or temperature). Each region 930 of the input parameter space can be associated with a corresponding regional SG model. The regions 930 may have the same or different sizes. Each region 930 may have a regular shape (e.g., a rectangle) or an irregular shape, wherein the boundary between two adjacent regions may be a straight line or may be a curve. In one example, the boundary between the regions 930 may be determined using a machine learning technique such as a support vector machine (SVM).

[0107] Fig. 9BAnother example of a partition model 902 including a set of regional models for modeling the relationship between SG values ​​and sensor measurement data according to certain embodiments is shown. In the illustrated example, the input parameter space can be partitioned based on the ranges of three or more parameters 940, 950, and 960, wherein one of the three or more parameters may include the sensor current Isig, and another of the three or more parameters may include, for example, sensor age, Vcntr, in vivo characteristics of the glucose sensor (e.g., sensor sensitivity), another input parameter of the SG model, an environmental property of the glucose sensor (e.g., humidity or temperature), or a combination thereof. A corresponding regional SG model can be generated for each region 970 of the input parameter space. The regions 970 may have the same or different sizes. Each region 970 may have a regular shape (e.g., a prism, such as a rectangular prism) or an irregular shape, wherein the boundary between two adjacent regions may be a flat surface or may be a curved surface.

[0108] like Fig. 9A and Fig. 9B As shown in , the input parameter space can be partitioned into multiple regions, which together form a continuous input parameter space. In some embodiments, in order to partition the input parameter space, each region of the input parameter space in which the sensor may have similar characteristics or can be modeled using a similar model can be identified. In some embodiments, the stability of the sensor trend, the regularity of the sensor signal amplitude, or other characteristics of the sensor signal can be identified and used to partition the input parameter space. In some embodiments, machine learning techniques such as clustering, dimensionality reduction (e.g., principal component analysis (PCA)), SVM, etc. can be used to partition the input parameter space, including determining the parameters used for partitioning and the boundaries of the partitions. In some embodiments, if the model used in one or more regions generates incorrect SG values ​​(e.g., with a large mean absolute relative difference (MARD)), if the models in adjacent regions output significantly different SG values ​​for similar input parameters, etc., the partitions and / or models can be modified.

[0109] In some embodiments, different models may be used in different regions 930 or 970 of the input parameter space. The model may include, for example, a linear regression model, a neural network, another machine learning model, a function, an equation, or another model. In some embodiments, multiple candidate models for each region of the input parameter space may be ranked based on the complexity, accuracy, and / or other metrics of the model, and one or more models may be selected for the region based on the ranking. This process may be performed for each region. In some embodiments, a simpler model may be sufficient for some regions, while a more complex model may be required for other regions. Therefore, a combination of different types of models may be used. In some embodiments, some models may use the same input parameters and have similar structures, but may have different coefficients or weights for the same input parameters.

[0110] Fig. 10A An example of a scheme 1000 for partitioning an input parameter space into a plurality of regions 1010 associated with different corresponding regional SG models according to certain embodiments is shown. In the illustrated example, the input parameter space may also be partitioned based on the age range of the glucose sensor and the Vcntr of the sensor measurement data. Each region 1010 of the input parameter space may have the same or different sizes and may be associated with a corresponding regional SG model. Each region 1010 may have a regular shape (e.g., a rectangle) or an irregular shape, wherein the boundary between two adjacent regions may be a straight line or may be a curve, which may be determined, for example, by a support vector machine or other machine learning technique. Fig. 10A In the illustrative example shown in , a regional SG model can be used for input measurement data with Vcntr values ​​greater than a certain threshold (e.g., greater than about 0.7), and three different regional SG models can be used for input measurement data with Vcntr values ​​below the threshold and measured during different ages of the glucose sensor. It should be noted that Fig. 10B The specific partitioning scheme illustrated in is for illustrative purposes only and is not intended to limit the scope of the present disclosure to only the depicted examples.

[0111] Fig. 10B Another example of a scheme 1005 for partitioning an input parameter space into a plurality of regions 1020 associated with different corresponding regional SG models according to certain embodiments is shown. In the illustrated example, the input parameter space can be partitioned based on the age range of the glucose sensor and the sensor current Isig of the sensor measurement data. Each region 1020 of the input parameter space can have the same or different sizes and can be associated with a corresponding regional SG model. Each region 1020 can have a regular shape (e.g., a rectangle) or an irregular shape, wherein the boundary between two adjacent regions can be a straight line or can be a curve, which can be determined by, for example, a support vector machine or other machine learning techniques. Fig. 10B In the illustrative example shown in , when the age of the glucose sensor is greater than a certain threshold, a regional SG model can be used, and two different regional SG models can be used for input measurement data whose Isig values ​​are in two different ranges and measured before the age of the glucose sensor reaches the threshold. It should be noted that Fig. 10B The specific partitioning scheme illustrated in is for illustrative purposes only and is not intended to limit the scope of the present disclosure to only the depicted examples.

[0112] In some embodiments, various smoothing or blending techniques may be used in the regional models of adjacent regions to avoid sudden changes in the output SG values ​​at the boundaries between adjacent regions due to different models used for adjacent regions. For example, in some embodiments, the output SG value of an input parameter near the boundary between two regions may be a weighted average of the output SG values ​​of the two models from the two adjacent regions. The output SG value of an input parameter at a region adjacent to three or more adjacent regions (e.g., a corner) may be a weighted average of the output SG values ​​of three or more models from the three or more adjacent regions. In other embodiments, other techniques for smoothing or blending the output SG values ​​of input parameters at or near region boundaries may be used.

[0113] According to certain embodiments, a plurality of regional SG model sets may be generated and used to determine the SG value of the blood glucose level of the approximate user, wherein different regional SG model sets may be generated for different partitions of the input parameter space (e.g., different combinations of input parameter ranges). Input parameters may include, for example, Isig, Vcntr, EIS, age, sensor sensitivity, etc. The input parameter space may be divided in different ways according to different partitioning schemes (e.g., based on different combinations of input parameters and / or by dividing the range of the same combination of input parameters in different ways). In some embodiments, different partitioning schemes may use different numbers of input parameters. For example, one partitioning scheme may partition the input parameter space based on different ranges of two parameters, and another partitioning scheme may partition the input parameter space based on different ranges of three or more parameters. A regional SG model set of a zone of the input parameter space divided according to the corresponding partitioning scheme may be generated. Based on the area into which the current input parameter falls, a regional SG model may be selected from each regional SG model set for a corresponding input parameter space partitioning scheme. Thus, multiple regional SG models selected from a plurality of regional SG model sets for different zoning schemes may be used to estimate SG values ​​individually, and the outputs of the selected multiple regional SG models may be combined (e.g., by averaging or weighted averaging) to determine the predicted SG value. In some embodiments, the manner in which the outputs of the selected multiple regional SG models are combined (e.g., the weights of the different outputs) may be determined using, for example, machine learning techniques.

[0114] Fig.11An example of a process 1100 for estimating SG values ​​using an aggregation of multiple regional model sets according to certain embodiments is illustrated. In the illustrated example, the input parameter space of the SG model used to estimate the SG value can be partitioned in different ways according to K (e.g., ≥2) different partitioning schemes, wherein the K different partitioning schemes can use, for example, different combinations of input parameters, different ranges of the same input parameters, different numbers of input parameters, or combinations thereof, and the input parameter space can be partitioned into the same or different numbers of regions of the same or different sizes, shapes, and / or dimensions. A corresponding regional SG model set can be determined for each partitioning scheme. For example, the first partitioning scheme 1120 can use one or more parameters (e.g., selected from Isig, age, Vcntr, and other input parameters) to partition the input parameter space into M regions, and a regional SG model 1122 (e.g., model 1-1, model 1-2, ..., model 1-(M-1) or model 1-M) can be determined for each corresponding region in the M regions. The Kth partitioning scheme 1130 can use one or more parameters (e.g., selected from Isig, age, Vcntr and other input parameters) to partition the input parameter space into N regions, and a regional SG model 1132 (e.g., model K-1, model K-2, ..., model K-(N-1) or model KN) can be determined for each corresponding region in the N regions.

[0115] When receiving input data 1110 (e.g., sensor measurement data, its derivatives, and / or transformed sensor measurement data), the controller or processor may select a first regional SG model 1-i (1124) from Model 1-1, Model 1-2, ..., Model 1-(M-1), and Model 1-M based on the value of the input parameter and the first partitioning scheme, and estimate a first SG value (SG value 1) using the first regional SG model 1-i (1124) and the input data 1110. The controller or processor may also select a corresponding regional SG model from each corresponding regional SG model set for a corresponding partitioning scheme of other partitioning schemes based on the value of the input parameter and other partitioning schemes, and estimate a corresponding SG value using the selected regional SG model and the input data 1110. For example, the controller or processor may select a K-th regional SG model Kj (1134) from model K-1, model K-2, ..., model K-(N-1), and model Kn, and estimate a K-th SG value (SG value K) using the K-th regional SG model Kj (1134) and input data 1110. In some embodiments, each regional SG model may also output an estimated probability, confidence level, or accuracy of the estimated SG value.

[0116] The outputs of the K selected regional SG models may be combined by the combiner 1140 to generate a predicted SG value for the input data 1110. For example, the combiner 1140 may determine an average or weighted average of the K SG values ​​and use the average or weighted average as the predicted SG value for the input data 1110. In one example, the weights of the K SG values ​​may be determined based on an estimated probability or confidence level for each of the K SG values. In another example, the weights of the K SG values ​​may be predetermined during a training or learning phase, where the weights may be determined for each corresponding set of regional SG models associated with a corresponding partitioning scheme.

[0117] Aggregation using multiple regional SG model sets disclosed herein can more accurately model sensor variations (e.g., sensitivity variations) and can significantly improve the performance (e.g., accuracy and consistency) of CGM sensors. Each regional model can be trained individually at a faster speed using a smaller training data set, and the response of the glucose sensor to different concentrations of analytes, such as sensitivity and nonlinearity of sensitivity, can be better characterized. Using certain parameters such as Isig as additional parameters for partitioning the input parameter space can also help improve the performance of the SG model, due to, for example, the improved granularity of partitioning the input parameter space and the correlation of additional parameters for modeling the response of the glucose sensor. Compared with using one complex SG model or one regional SG model set with a large number of models for a large number of regions partitioned using a large number of input parameters, aggregation using multiple regional SG model sets for different partitioning schemes (e.g., using different input parameters for different partitioning schemes) can further improve the performance (e.g., accuracy and consistency) of the CGM sensor, due to, for example, using additional parameters to improve the granularity of the partitions, and using additional partitioning schemes to include additional parameters or degrees of freedom to tune and improve the SG model without significantly increasing the overall complexity of the SG model and the difficulty of training the SG model.

[0118] Fig. 12A An example of an aggregate model 1200 including a plurality of regional SG model sets for estimating SG values ​​based on sensor measurement data generated by a glucose sensor according to certain embodiments is illustrated. It should be noted that Fig. 12AThe specific partitioning schemes in the aggregation model illustrated in are for illustrative purposes only and are not intended to limit the scope of the present disclosure to the depicted examples. In the illustrated examples, the input parameter space can be partitioned according to at least two different partitioning schemes. In partitioning scheme 1210, the input parameter space can be partitioned into, for example, four regions based on the age of the glucose sensor and the sensor current Isig value of the received input measurement data, wherein each of the four regions can be modeled using a corresponding regional SG model (e.g., model 1212, 1214, 1216, or 1218). In partitioning scheme 1220, the input parameter space can be partitioned into, for example, four regions based on the Vcntr value of the received input measurement data and the sensor current Isig value, wherein each of the four regions can be modeled using a corresponding regional SG model (e.g., model 1222, 1224, 1226, or 1228).

[0119] When input data (e.g., sensor measurement data, its derivatives, and / or transformed sensor measurement data) is received, a regional SG model may be selected from models 1212, 1214, 1216, and 1218 based on the values ​​of the input parameters and the two partitioning schemes, and another regional SG model may be selected from models 1222, 1224, 1226, and 1228. The outputs of the two selected regional SG models may be used in combination to generate a predicted SG value for the input data. For example, an average or weighted average of two SG values ​​estimated using the two selected regional SG models may be used as the predicted SG value for the input data, as described above with respect to, for example, Fig.11 as described.

[0120] Fig. 12B An example of performance improvement over a baseline technique using some of the techniques disclosed herein according to certain embodiments is illustrated. In the illustrated example, the baseline technique can partition the input parameter space into four subspaces using Vcntr and the age of the sensor. Using Isig and the age of the sensor to partition the input parameter space can improve the accuracy of the CGM sensor compared to the baseline technique, for example, by increasing the proportion of SG values ​​estimated on day 1 that are within 20% of the reference glucose level by about 2%, increasing the proportion of SG values ​​estimated on day 1 that are within 15% of the reference glucose level by about 5.4%, reducing the MARD on day 1 by about 1.0%, increasing the proportion of all estimated SG values ​​within 20% of the reference glucose level by about 0.2%, increasing the proportion of all estimated SG values ​​within 15% of the reference glucose level by about 0.8%, reducing the MARD of hypoglycemic conditions (e.g., BG<70 mg / dL) by about 2.6%, and reducing the MARD of hyperglycemic conditions (e.g., BG>240 mg / dL) by about 0.4%.

[0121] Also like Fig. 12B As shown in , when using aggregation of multiple regional model sets as disclosed herein, the accuracy of the CGM sensor can be further improved. For example, compared to the baseline technique, using aggregation of multiple regional model sets can increase the proportion of SG values ​​estimated on day 1 that are within 20% of the reference glucose level by about 1.6%, increase the proportion of SG values ​​estimated on day 1 that are within 15% of the reference glucose level by about 5.5%, reduce the MARD on day 1 by about 1.3%, increase the proportion of all estimated SG values ​​within 20% of the reference glucose level by about 0.6%, increase the proportion of all estimated SG values ​​within 15% of the reference glucose level by about 1.9%, reduce the MARD of hypoglycemic conditions (e.g., BG<70 mg / dL) by about 3.6%, and reduce the MARD of hyperglycemic conditions (e.g., BG>240 mg / dL) by about 0.5%. It should be noted that Fig. 12B The examples and data illustrated in the accompanying drawings are for illustrative purposes only and are not intended to limit the scope of the present disclosure to the illustrated examples.

[0122] Fig.13 The flowchart 1300 is included to illustrate an example of a process for estimating SG values ​​using aggregation of partitioned SG models according to certain embodiments. The operations in the flowchart 1300 may be performed by, for example, the monitoring device 104, the glucose sensor subsystem 210, the sensor electronics 304, the following FIG. 14A to FIG. 14B The glucose sensor device 1400 described in the flowchart 1300 is performed. Although the flowchart 1300 may describe some operations as being performed in parallel or simultaneously, some of these operations may be performed in a sequential process. In addition, the order of the operations may be rearranged. The process may have additional steps not included in the figure. Some operations may be optional or may be omitted.

[0123] The operations in block 1310 may include receiving sensor measurement data measured by a glucose sensor. The sensor measurement data may include, for example, sensor current (Isig), reverse voltage (Vcntr), electrochemical impedance spectroscopy (EIS) data, age of the glucose sensor, or a combination thereof. In some embodiments, derived features or other features may be derived from, for example, the sensor current (Isig), reverse voltage (Vcntr), electrochemical impedance spectroscopy (EIS) data, age of the glucose sensor, or a combination thereof. Figures 6 to 8 The sensor measurement data described is derived, extracted, or transformed and can be used in combination with the sensor measurement data as input parameters to the SG model for estimating SG values. In some embodiments, the operations in block 1310 (or after block 1310) can include generating derived features and / or transformed features based on the sensor measurement data.

[0124] The operation in box 1320 may include selecting a first regional SG model from a plurality of first regional SG models for corresponding regions in a plurality of first regions of an input parameter space associated with the sensor measurement data based on the sensor measurement data. The input parameter space may be partitioned into a plurality of first regions based on a first partitioning scheme. In some embodiments, the first partitioning scheme may partition the input parameter space based at least in part on a sensor current of a glucose sensor. In some embodiments, the first partitioning scheme may partition the input parameter space based at least in part on the sensor current and the age of the glucose sensor. The regions in the plurality of first regions of the input parameter space may be characterized by different sizes, different shapes (regular or irregular), different dimensions, or a combination thereof. In some embodiments, the input parameters of the plurality of first regional SG models may include derived features and / or transformed features. Each of the plurality of first regional SG models may include one or more machine learning models, equations, functions, or a combination thereof.

[0125] The operations in block 1330 may include using the first regional SG model and sensor measurement data (and in some embodiments, as described above with respect to Figures 6 to 8 In some embodiments, the first regional SG model may estimate not only the first SG value but also the probability or confidence level associated with the first SG value.

[0126] The operation in box 1322 may include selecting a second regional SG model from a plurality of second regional SG models for corresponding regions of a plurality of second regions of the input parameter space based on sensor measurement data, wherein the input parameter space is partitioned into a plurality of second regions based on a second partitioning scheme different from the first partitioning scheme. In some embodiments, the second partitioning scheme may partition the input parameter space based at least in part on the sensor current of the glucose sensor. In some embodiments, the second partitioning scheme may partition the input parameter space based at least in part on the sensor current and the age of the glucose sensor. The regions in the plurality of second regions of the input parameter space may be characterized by different sizes, different shapes (regular or irregular), different dimensions, or a combination thereof. In some embodiments, the input parameters of the plurality of second regional SG models may include derived features and / or transformed features. Each of the plurality of second regional SG models may include one or more machine learning models, equations, functions, or a combination thereof.

[0127] In some embodiments, the first partitioning scheme and the second partitioning scheme may partition the input parameter space using different combinations of input parameters. In some embodiments, the first partitioning scheme and the second partitioning scheme may partition the input parameter space using different numbers of input parameters. In some embodiments, the first partitioning scheme and the second partitioning scheme may partition the input parameter space using different ranges of the same set of input parameters. The number of regions in the plurality of first regions of the input parameter space may be the same as or different from the number of regions in the plurality of second regions of the input parameter space.

[0128] The operations in block 1332 may include using the second regional SG model and sensor measurement data (and in some embodiments, as described above with respect to Figures 6 to 8 In some embodiments, the second regional SG model may estimate not only the second SG value but also the probability or confidence level associated with the second SG value.

[0129] The operations in block 1340 may include determining a predicted SG value based on a combination of the first SG value and the second SG value. For example, determining the predicted SG value based on a combination of the first SG value and the second SG value may include determining an average or a weighted average of the first SG value and the second SG value. The weights associated with the first SG value and the second SG value for the weighted average may be predetermined (e.g., during a learning phase) or may be determined based on, for example, a first probability or confidence level and a second probability or confidence level.

[0130] In some embodiments, the input parameter space may be partitioned according to three or more different partitioning schemes, and three or more regional SG model sets may be created for the three or more different partitioning schemes. For the corresponding partitioning schemes, a regional SG model may be selected from each corresponding regional SG model set. The predicted SG value may then be determined based on the outputs of the selected three or more regional SG models. For example, in some embodiments, the operation of flowchart 1300 may also include selecting a third regional SG model from a plurality of third regional SG models for corresponding regions of a plurality of third regions of the input parameter space based on sensor measurement data, wherein the input parameter space may be partitioned into a plurality of third regions based on a third partitioning scheme different from the first partitioning scheme and the second partitioning scheme; and estimating a third SG value using the third regional SG model and the sensor measurement data, wherein determining the predicted SG value may include determining the predicted SG value based on a combination (e.g., an average or a weighted average) of the first SG value, the second SG value, and the third SG value.

[0131] Fig.14A and Fig. 14B An example of an integrated glucose sensor device 1400 according to certain embodiments is depicted, which can implement some of the examples disclosed herein. The glucose sensor device 1400 can be an example of a monitoring device 104, a glucose sensor subsystem 210, or a CGM system 300. In the illustrated example, the glucose sensor device 1400 may include a housing formed by a base 1410 and a top cover 1420, a sensor electronics enclosed in the housing, an electrochemical sensor 1440 partially located within the housing and connected to the sensor electronics, and an adhesive patch 1430 attached to the base 1410. The electrochemical sensor 1440 can be an example of the electrochemical sensor described above (e.g., glucose sensor 400 or 402), and may include a plurality of electrodes located at a distal end that can be inserted into the subcutaneous layer of a user. The proximal end of the electrochemical sensor 1440 can be bent and held against the printed circuit board (PCB) 1412 by an elastomer or rubber block 1414, so that the contacts on the proximal end of the electrochemical sensor 1440 can be physically and electrically connected to the contact pads on the PCB 1412. The housing can be hermetically sealed to protect the sensor electronics from moisture. The electrochemical sensor 1440 can be inserted into the subcutaneous layer of the user using an inserter. The inserter can include a needle and can encapsulate the entire glucose sensor device 1400 before insertion. When activated (e.g., pushed), the inserter can move the entire glucose sensor device 1400 toward the user's skin to attach the adhesive patch 1430 to the user's skin and insert the distal end of the electrochemical sensor 1440 into the subcutaneous layer of the user using a needle, wherein the needle can be automatically retracted, for example, after the distal end of the electrochemical sensor 1440 is inserted into the user's body.

[0132] The sensor electronics in the housing may include a battery 1416, an antenna 1418, and circuits on the PCB 1412. The circuits on the PCB 1412 may include, for example, a processor, a controller, a potentiostat, a filter, an amplifier, a wireless (e.g., Bluetooth) transceiver, one or more memory devices, a power conversion and management device, or a combination thereof. The predicted in vivo characteristics of the electrochemical sensor 1440 and the SG model described above may be stored in one or more memory devices. Before or after the distal end of the electrochemical sensor 1440 is inserted into the subcutaneous layer of the user, the sensor electronics may be activated to initialize the sensor for measuring sensor measurement data (e.g., Isig, Vcntr, EIS, etc.). The processor or controller may execute instruction codes stored on one or more memory devices to obtain sensor measurement data, process (e.g., filter, amplify, transform, etc.) the sensor measurement data, estimate SG values ​​using the SG model and the sensor measurement data, optionally clear certain estimated SG values, and send the estimated SG values ​​to a receiver (e.g., a smart phone or insulin pump) via the wireless transceiver and antenna 1418, as described above.

[0133] Fig.15 An example of a delivery device 1500 is depicted that can implement the above-described delivery devices (e.g., delivery device 102 or insulin delivery subsystem 230). In an insulin delivery device implemented using delivery device 1500, insulin delivery can be performed based on internal communications between a central computing module (e.g., a microprocessor of delivery device 1500) and an insulin delivery module (e.g., including a microcontroller, a motor, and a pump). For example, insulin delivery can be caused by the central computing module transmitting a delivery command in the form of an electrical signal that travels to the insulin delivery module via a communication structure. The central computing module can also be configured to communicate with a remote or cloud computing system (e.g., a microprocessor of delivery device 1500) to communicate with the insulin delivery module. Figure 1 A computing device (e.g., a remote / cloud computing system 108) Figure 1 According to the techniques described herein, delivery device 1500 can communicate various event data toward a remote or cloud computing system, which can communicate insulin delivery determinations toward delivery device 1500.

[0134] The delivery device 1500 can provide insulin through a small tube 1510, which is configured to be connected to the cannula fluid inserted subcutaneously. The delivery device 1500 can be configured to deliver two types of doses: a basic dose, which can be delivered in micro-doses periodically (e.g., every five minutes) throughout the day and night; and a bolus dose to cover the blood glucose increase caused by a meal and / or correct high blood glucose levels. In the depicted example, the delivery device 1500 includes a user interface with a button element 1520, which can be manipulated to administer an insulin bolus, change therapy settings, change user preferences, and select display features, etc. The delivery device 1500 may also include a display device 1530, which can be used to present various types of information or data to the user. According to various aspects of the present disclosure, the user of the delivery device 1500 can use the button element 1520 to enter certain event data (e.g., event type, event start time, event details, etc.), and can use the display device 1530 to confirm the user input. Fig.15 The delivery device 1500 is provided by way of example only, and other types of insulin delivery devices and other technologies than those described above are considered to be within the scope of the present disclosure.

[0135] Fig.16 1 is a block diagram of an example of an electronic device 1600 that can implement some of the examples disclosed herein. For example, the electronic device 1600 can be used to implement the delivery device 102, the monitoring device 104, the computing device 106 and the remote / cloud computing system 108, the glucose sensor subsystem 210, the controller 220, the insulin delivery subsystem 230, the glucose delivery subsystem 240 and the glucagon delivery subsystem 250, the CGM system 300, the glucose sensor device 1400, the delivery device 1500, or another device that can be used to calculate insulin dosage and / or predict future blood glucose levels. In the illustrated example, the computing device 1600 may include one or more processors 1610 and a memory 1620. The processor 1610 may be configured to execute instructions stored in the memory 1620 to perform one or more of the methods and other applications described herein. The processor 1610 may include, for example, one or more central processing units, microprocessors, microcontrollers, special-purpose processors (e.g., digital signal processors), ASICs, DSPs, FPGAs, or other processors suitable for implementation in portable electronic devices. The processor 1610 may be communicatively coupled to a number of components within the electronic device 1600. To achieve this communicatively coupling, the processor 1610 may communicate with the other illustrated components across a bus 1640. The bus 1640 may be any subsystem suitable for transmitting data within the electronic device 1600. The bus 1640 may include multiple computer buses and additional circuits for transmitting data.

[0136] The memory 1620 may include one or more transient and / or non-transient storage devices, such as, for example, static random access memory (SRAM), dynamic random access memory (DRAM), read access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), flash-EPROM, secure digital (SD) card, and any other memory chip or cartridge. Such storage devices may be configured to implement any suitable data storage, including but not limited to various file systems, data structures, computer-readable instructions, program modules, and the like. In some embodiments, the memory 1620 may be distributed into different hardware modules.

[0137] The memory 1620 may include an operating system 1625 loaded therein. The operating system 1625 may be operable to initiate the execution of instructions provided by the application modules 1622 to 1824 and / or manage other hardware modules 1670, as well as interface with the communication subsystem 1630 which may include one or more wired and / or wireless transceivers. The operating system 1625 may be adapted to perform other operations across the components of the electronic device 1600, including threading, virtualization, resource management, data storage control, and other similar functionality. In some embodiments, the memory 1620 may store a plurality of application modules 1622 to 1624, which may include any number of applications. Examples of applications may include insulin calculators, blood sugar level monitors or predictors, glucose level management applications, etc. The application modules 1622 to 1424 may include specific instructions to be executed by the processor 1610. In some embodiments, certain applications or portions of the application modules 1622 to 1424 may be executed by other hardware modules 1670.

[0138] The communication subsystem 1630 may include, for example, an infrared communication device, a wireless communication device and / or a chipset (such as a Bluetooth ®Device, IEEE 802.11 device, Wi-Fi device, WiMax device, cellular communication device, etc.) and / or similar communication interfaces. One or more antennas (not shown) may be used for wireless communication as part of the communication subsystem 1630 or as a separate component (e.g., a wireless charging receiver or a near field communication receiver) coupled to any portion of the electronic device 1600. In some embodiments, the communication subsystem 1630 may include circuits for wired communication technologies, such as Ethernet, coaxial communications, a universal serial bus (USB), etc. The communication subsystem 1630 may permit data to be exchanged with a network, other computer systems, and / or any other device. For example, the communication subsystem 1630 may be used to receive a therapeutic determination of the delivery of a therapeutic fluid (e.g., insulin) from a cloud computing system, such as via an intermediate computing device (e.g., a controller) communicatively coupled to the electronic device 1600, wherein the processor 1610 may transmit a command to the actuator controller based on the therapeutic determination to cause the delivery of an appropriate amount of therapeutic fluid (e.g., insulin) to the user. In another example, the communication subsystem 1630 may be used to communicate a measurement result (e.g., a sensor glucose level) to a computing device (e.g., a smart phone or a personal health monitoring device) and / or to a remote server via the computing device, or to receive data (e.g., calibration data, configuration data, etc.) from the computing device or the remote server via the computing device.

[0139] Sensor 1650 may include, for example, a camera, an infrared sensor, an accelerometer, a pressure sensor, a temperature sensor, a proximity sensor, a magnetometer, a gyroscope, an inertial sensor (e.g., an inertial measurement unit (IMU)), an ambient light sensor, a position sensor, a depth sensor, a gesture detector, or any other similar module operable to provide sensory output and / or receive sensory input. In one example, sensor 1650 may be used to implement a meal detector.

[0140] The input / output user interface 1660 may allow a user to transmit an action request to the electronic device 1600 to perform a specific action, and may provide information to the user (e.g., the status of the electronic device 1600, measurement results, alarms, etc.). The input / output user interface 1660 may include one or more input devices, such as a touch screen, a touch pad, a microphone, a button, a dial, a switch, a keyboard, a mouse, a game controller, or any other suitable device for receiving an action request and communicating the received action request to the processor 1610. In some embodiments, the input / output user interface 1660 may include one or more output devices, such as a display, a speaker, a light emitting device, a tactile device, etc., to provide feedback or alarms to the user.

[0141] In some embodiments, the electronic device 1600 may include multiple other hardware modules 1670. Each of the other hardware modules 1670 may be a physical module within the electronic device 1600. Although each of the other hardware modules 1670 may be permanently configured as a structure, some of the other hardware modules 1670 may be temporarily configured to perform a specific function or be temporarily activated. Examples of other hardware modules 1670 may include, for example, an audio output and / or input module (e.g., a microphone or a speaker), a near field communication (NFC) module, a rechargeable battery, a battery management system, a wired / wireless battery charging system, an actuator controller, etc. In some embodiments, one or more functions of the other hardware modules 1670 may be implemented in software.

[0142] In one example, the electronic device 1600 can be part of an insulin delivery device (e.g., a pump) that can deliver rapid-acting insulin through a small tube configured to be connected to a cannula fluid inserted subcutaneously. The electronic device 1600 can cause the delivery of two types of doses—basal doses, which can be delivered in micro-amounts periodically (e.g., every five minutes) throughout the day and night; and bolus doses, which cover the blood sugar increase caused by eating and / or correct high blood sugar levels. The insulin delivery device may include a user interface with button elements that can be manipulated to administer insulin boluses, change therapy settings, change user preferences, and select display features, etc. The insulin delivery device may also include a display device that can be used to present various types of information or data to the user. According to various aspects of the present disclosure, a user of the insulin delivery device can use button elements to enter certain event data (e.g., event type, event start time, event details, etc.), and can use a display device to confirm user input.

[0143] In various specific implementations, the above hardware and modules may be implemented on a single device or on multiple devices that can communicate with each other using wired or wireless connections. In alternative configurations, different and / or additional components may be included in the electronic device 1600. Similarly, the functionality of one or more components may be distributed among the components in a manner different from that described above.

[0144] The methods, systems and devices discussed above are examples. Various programs or components may be omitted, substituted or added to various embodiments as appropriate. For example, in an alternative configuration, the described method may be performed in a different order than described, and / or various stages may be added, omitted and / or combined. In addition, the features described with respect to certain embodiments may be combined in various other embodiments. Different aspects and elements of the embodiments may be combined in a similar manner. In addition, technology is constantly evolving, and therefore many elements are examples, and the scope of this disclosure is not limited to those specific examples.

[0145] Specific details are given in the description to provide a thorough understanding of the embodiments. However, embodiments may be practiced without these specific details. For example, well-known circuits, processes, systems, structures, and techniques have been shown without the necessary details to avoid blurring the embodiments. This description only provides example embodiments and is not intended to limit the scope, applicability, or configuration of the present invention. On the contrary, the foregoing description of the embodiments will provide enabling descriptions for realizing various embodiments for those skilled in the art. Various changes may be made to the functions and arrangements of the elements without departing from the spirit and scope of the present disclosure.

[0146] In addition, some embodiments are described as processes depicted as flow charts or block diagrams. Although each process can describe the operation as a sequential process, many operations can be performed in parallel or concurrently. In addition, the order of the operations can be rearranged. The process may have additional steps not included in the figure. In addition, the embodiments of these methods can be implemented by hardware, software, firmware, middleware, microcode, hardware description language or any combination thereof. When implemented in software, firmware, middleware or microcode, the program code or code segment used to perform the associated task can be stored in a computer-readable medium such as a storage medium. The processor can perform the associated task.

[0147] It will be apparent to those skilled in the art that substantial variations may be made depending on specific requirements. For example, customized or dedicated hardware may also be used, and / or specific elements may be implemented in hardware, software (including portable software, such as applets, etc.), or both. In addition, connections to other computing devices, such as network input / output devices, may be employed.

[0148] Each embodiment disclosed herein is an example of the present disclosure and can be embodied in various forms. For example, although some embodiments herein are described as separate embodiments, each of these embodiments herein can be combined with one or more of the other embodiments herein. The specific structural and functional details disclosed herein should not be construed as restrictive, but as the basis of the claims, and as a representative basis for teaching those skilled in the art to use the present disclosure differently with any appropriate detailed structure. In the entire description of the drawings, similar reference numerals may refer to similar elements.

[0149] Any of the techniques, operations, methods, programs, algorithms or codes described herein may be converted to or expressed in a programming language or computer program embodied on a computer, processor or machine-readable medium. As used herein, the terms "programming language" and "computer program" each include any language for specifying instructions for a computer or processor, and include (but are not limited to) the following languages ​​and their derivatives: assembler, Basic, Batch file, BCPL, C, C+, C++, Delphi, Fortran, Java, JavaScript, machine code, operating system command language, Pascal, Perl, PL1, Python, script processing language, visual Basic, meta-language self-specified programming and all first, second, third, fourth, fifth or higher generation computer languages. Databases and other data patterns and any other meta-languages ​​are also included. There is no distinction between interpretation, compiled languages ​​or languages ​​that use compilation and interpretation methods at the same time. There is no distinction between the compiled and source versions of the program. Therefore, the reference to a program that can exist in more than one state (such as source, compilation, target or link) of a programming language is a reference to any and all such states. References to procedures may cover the actual instructions and / or the intent of those instructions.

[0150] It should be understood that the foregoing description only illustrates the present disclosure. To the extent consistent, any or all aspects described in detail herein may be used in combination with any or all other aspects described in detail herein. Without departing from the present disclosure, a person skilled in the art may devise various alternatives and modifications. Therefore, the present disclosure is intended to cover all such alternatives, modifications and variations. The embodiments described with reference to the accompanying drawings are presented only to show certain examples of the present disclosure. Other elements, steps, methods and techniques that are not significantly different from those described in the above and / or attached claims are also intended to be included within the scope of the present disclosure. Although several embodiments of the present disclosure have been shown in the drawings, it is not intended to limit the present disclosure to this, because it is desired to make the present disclosure as broad as the scope allowed by the art and the present specification should be read in the same manner. Therefore, the above description should not be understood as limiting, but only as an illustration of a specific embodiment. A person skilled in the art can envision other modifications within the scope and essence of the claims attached hereto. The aspects and features of the present disclosure can be embodied in various forms. Therefore, specific structural and functional details disclosed herein are not to be interpreted as limiting, but rather as a basis for the claims and as a representative basis for teaching one skilled in the art to variously employ the present disclosure in any appropriately detailed structure.

[0151] With reference to the accompanying drawings, a component that may include a memory may include a non-transitory machine-readable medium. The term "machine-readable medium", "processor-readable medium" or "computer-readable medium" may refer to any storage medium that participates in providing data that causes a machine to operate in a particular manner. In the embodiments provided above, various machine-readable media may involve providing instructions / codes to a processing unit and / or other devices for execution. Additionally or alternatively, a machine-readable medium may be used to store and / or carry such instructions / codes. In many embodiments, a computer-readable medium is a physical and / or tangible storage medium. Such a medium may take a variety of forms, including but not limited to non-volatile media, volatile media, and transmission media. Common forms of computer-readable media include, for example, magnetic and / or optical media, such as compact disks (CDs) or digital versatile disks (DVDs), punch cards, paper tapes, any other physical media with hole patterns, RAM, programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), flash-EPROM, any other memory chip or cartridge, a carrier as described below, or any other medium from which a computer can read instructions and / or codes. A computer program product may include code and / or machine executable instructions which may represent a procedure, function, subprogram, program, routine, application program (App), subroutine, module, software package, class, or any combination of instructions, data structures, or program statements.

[0152] Those skilled in the art will appreciate that any of a variety of different technologies and techniques may be used to represent information and signals used to convey the messages described herein. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be referenced throughout the above description may be represented by voltage, current, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.

[0153] As used herein, the terms "and" and "or" may include multiple meanings, which are also expected to depend at least in part on the context in which the terms are used. Generally, if used in an associative list (such as A, B, or C), "or" is intended to mean A, B, and C (used herein in an inclusive sense) as well as A, B, or C (used herein in an exclusive sense). In addition, the term "one or more" used herein may be used to describe any feature, structure, or characteristic of the singular, or may be used to describe some combination of feature, structure, or characteristic. However, it should be noted that this is merely an illustrative example, and the subject matter claimed is not limited to this example. In addition, if used in an associative list (such as A, B, or C), the term "at least one" may be interpreted to mean A, B, C, or any combination of A, B, and / or C, such as AB, AC, BC, AA, ABC, AAB, AABBCCC, etc.

[0154] In addition, although certain embodiments have been described using a specific combination of hardware and software, it should be appreciated that other combinations of hardware and software are also possible. Certain embodiments may be implemented with hardware only, or with software only, or using a combination thereof. In one example, the software may be implemented with a computer program product that includes computer program codes or instructions that may be executed by one or more processors for performing any or all of the steps, operations, or processes described in this disclosure, wherein the computer program may be stored on a non-transient computer-readable medium. The various processes described herein may be implemented in any combination on the same processor or on different processors.

[0155] Where a device, system, component or module is described as being configured to perform certain operations or functions, such configuration may be achieved, for example, by designing electronic circuits to perform the operations, by programming programmable electronic circuits (such as microprocessors) to perform the operations (such as by executing computer instructions or code), or a processor or core programmed to execute code or instructions stored on a non-transitory memory medium, or any combination thereof. Processes may communicate using a variety of techniques, including but not limited to conventional techniques for inter-process communication, and different pairs of processes may use different techniques, or the same pair of processes may use different techniques at different times.

[0156] In view of this description, embodiments may include different combinations of features. Implementation examples are described in the following numbered clauses:

[0157] Clause 1. A processor-implemented method, the processor-implemented method comprising:

[0158] receiving sensor measurement data measured by a glucose sensor, the sensor measurement data comprising at least one signal indicative of a glucose level and at least one signal indicative of a health of the sensor;

[0159] selecting, based at least in part on the sensor measurement data, a first regional sensor glucose (SG) model from a plurality of first regional SG models for a corresponding region of a plurality of first regions of an input parameter space associated with the sensor measurement data, wherein the input parameter space is partitioned into the plurality of first regions based on a first partitioning scheme;

[0160] estimating a first SG value using the first regional SG model and the sensor measurement data;

[0161] selecting a second regional SG model from a plurality of second regional SG models for corresponding regions of a plurality of second regions of the input parameter space based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of second regions based on a second partitioning scheme different from the first partitioning scheme;

[0162] estimating a second SG value using the second regional SG model and the sensor measurement data; and

[0163] A predicted SG value is determined based on a combination of the first SG value and the second SG value.

[0164] Clause 2. A processor-implemented method according to clause 1, wherein at least one of the first partitioning scheme or the second partitioning scheme partitions the input parameter space based at least in part on sensor current (Isig), counter voltage (Vcntr), electrochemical impedance spectroscopy (EIS) data, age of the glucose sensor, temperature, age of a user of the glucose sensor, body mass index (BMI) of the user, or a combination thereof.

[0165] Clause 3. The processor-implemented method of clause 1 or 2, wherein the first partitioning scheme or the second partitioning scheme partitions the input parameter space based at least in part on a sensor current and an age of the glucose sensor.

[0166] Clause 4. A processor-implemented method as described in any one of clauses 1 to 3, wherein the first partitioning scheme and the second partitioning scheme use different combinations of input parameters to partition the input parameter space.

[0167] Clause 5. The processor-implemented method of any one of clauses 1 to 4, wherein the first partitioning scheme and the second partitioning scheme use different numbers of input parameters to partition the input parameter space.

[0168] Clause 6. A processor-implemented method as described in any one of Clauses 1 to 5, wherein the first partitioning scheme and the second partitioning scheme use different ranges of the same set of input parameters to partition the input parameter space.

[0169] Clause 7. A processor-implemented method according to any one of clauses 1 to 6, wherein the regions in the plurality of first regions of the input parameter space are characterized by different sizes, different shapes, different dimensions, or a combination thereof.

[0170] Clause 8. A processor-implemented method according to any one of clauses 1 to 7, wherein the number of regions in the plurality of first regions of the input parameter space is the same as or different from the number of regions in the plurality of second regions of the input parameter space.

[0171] Clause 9. A method implemented by a processor according to any one of clauses 1 to 8, the method further comprising generating at least one of a derived feature or a transformed feature based on sensor measurement data, wherein input parameters of the plurality of first regional SG models and the plurality of second regional SG models include at least one of the derived feature or the transformed feature.

[0172] Clause 10. A processor-implemented method according to any one of clauses 1 to 9, wherein determining the predicted SG value based on the combination of the first SG value and the second SG value comprises determining an average or weighted average of the first SG value and the second SG value.

[0173] Clause 11. The processor-implemented method of claim 10, further comprising:

[0174] determining a first probability or confidence level associated with the first SG value; and

[0175] determining a second probability or confidence level associated with the first SG value,

[0176] Wherein the weights associated with the first SG value and the second SG value for the weighted average are determined based on the first probability or confidence level and the second probability or confidence level.

[0177] Clause 12. A processor-implemented method according to any one of clauses 1 to 11, the processor-implemented method further comprising:

[0178] selecting a third regional SG model from a plurality of third regional SG models for corresponding regions of a plurality of third regions of the input parameter space based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of third regions based on a third partitioning scheme different from the first partitioning scheme and the second partitioning scheme; and

[0179] estimating a third SG value using the third regional SG model and the sensor measurement data,

[0180] Determining the predicted SG value includes determining the predicted SG value based on a combination of the first SG value, the second SG value, and the third SG value.

[0181] Clause 13. A processor-implemented method as defined in any one of clauses 1 to 12, wherein:

[0182] Each of the plurality of first regional SG models and the plurality of second regional SG models comprises one or more machine learning models, equations, functions, or a combination thereof; and

[0183] The first regional SG model and the second regional SG model in the plurality of first regional SG models and the plurality of second regional SG models use different input parameters.

[0184] Clause 14. A system, comprising:

[0185] one or more processors; and

[0186] One or more processor-readable media storing instructions that, when executed by the one or more processors, cause operations including:

[0187] receiving sensor measurement data measured by a glucose sensor, the sensor measurement data comprising at least one signal indicative of a glucose level and at least one signal indicative of a health of the sensor;

[0188] selecting, based at least in part on the sensor measurement data, a first regional sensor glucose (SG) model from a plurality of first regional SG models for corresponding regions of an input parameter space associated with the sensor measurement data, wherein the input parameter space is partitioned into the plurality of first regions based on a first partitioning scheme;

[0189] estimating a first SG value using the first regional SG model and the sensor measurement data;

[0190] selecting a second regional SG model from a plurality of second regional SG models for corresponding regions of the input parameter space based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of second regions based on a second partitioning scheme different from the first partitioning scheme;

[0191] estimating a second SG value using the second regional SG model and the sensor measurement data; and

[0192] A predicted SG value is determined based on a combination of the first SG value and the second SG value.

[0193] Clause 15. One or more non-transitory processor-readable storage media storing instructions that, when executed by one or more processors, cause performance of operations comprising:

[0194] receiving sensor measurement data measured by a glucose sensor, the sensor measurement data comprising at least one signal indicative of a glucose level and at least one signal indicative of a health of the sensor;

[0195] selecting, based at least in part on the sensor measurement data, a first regional sensor glucose (SG) model from a plurality of first regional SG models for corresponding regions of an input parameter space associated with the sensor measurement data, wherein the input parameter space is partitioned into the plurality of first regions based on a first partitioning scheme;

[0196] estimating a first SG value using the first regional SG model and the sensor measurement data;

[0197] selecting a second regional SG model from a plurality of second regional SG models for corresponding regions of the input parameter space based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of second regions based on a second partitioning scheme different from the first partitioning scheme;

[0198] estimating a second SG value using the second regional SG model and the sensor measurement data; and

[0199] A predicted SG value is determined based on a combination of the first SG value and the second SG value.

Claims

1. A processor-implemented method, the processor-implemented method comprising: receiving sensor measurement data measured by a glucose sensor, the sensor measurement data comprising at least one signal indicative of a glucose level and at least one signal indicative of a sensor health; selecting, based at least in part on the sensor measurement data, a first regional sensor glucose (SG) model from a plurality of first regional SG models for a corresponding region of a plurality of first regions of an input parameter space associated with the sensor measurement data, wherein the input parameter space is partitioned into the plurality of first regions based on a first partitioning scheme; estimating a first SG value using the first regional SG model and the sensor measurement data; selecting a second regional SG model from a plurality of second regional SG models for respective ones of a plurality of second regions of the input parameter space based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of second regions based on a second partitioning scheme different from the first partitioning scheme; estimating a second SG value using the second regional SG model and the sensor measurement data; as well as A predicted SG value is determined based on a combination of the first SG value and the second SG value.

2. The processor-implemented method of claim 1 , wherein at least one of the first partitioning scheme or the second partitioning scheme partitions the input parameter space based at least in part on sensor current (Isig), back voltage (Vcntr), electrochemical impedance spectroscopy (EIS) data, age of the glucose sensor, temperature, age of a user of the glucose sensor, body mass index (BMI) of the user, or a combination thereof. 3 . The processor-implemented method of claim 1 , wherein the first partitioning scheme or the second partitioning scheme partitions the input parameter space based at least in part on a sensor current and an age of the glucose sensor. 4 . The processor-implemented method of claim 1 , wherein the first partitioning scheme and the second partitioning scheme use different combinations of input parameters to partition the input parameter space. 5 . The processor-implemented method of claim 1 , wherein the first partitioning scheme and the second partitioning scheme use different numbers of input parameters to partition the input parameter space. 6 . The processor-implemented method of claim 1 , wherein the first partitioning scheme and the second partitioning scheme partition the input parameter space using different ranges of the same set of input parameters.

7. The processor-implemented method of claim 1, wherein regions of the plurality of first regions of the input parameter space are characterized by different sizes, different shapes, different dimensions, or a combination thereof.

8. The processor-implemented method of claim 1, wherein the number of regions in the plurality of first regions of the input parameter space is the same as or different from the number of regions in the plurality of second regions of the input parameter space.

9. The processor-implemented method of claim 1 , further comprising generating at least one of a derived feature or a transformed feature based on the sensor measurement data, wherein input parameters of the plurality of first regional SG models and the plurality of second regional SG models include at least one of the derived feature or the transformed feature.

10. The processor-implemented method of claim 1, wherein determining the predicted SG value based on the combination of the first SG value and the second SG value comprises determining an average or a weighted average of the first SG value and the second SG value.

11. The processor-implemented method of claim 10, further comprising: determining a first probability or confidence level associated with the first SG value; as well as determining a second probability or confidence level associated with the first SG value, Wherein weights associated with the first SG value and the second SG value for the weighted average are determined based on the first probability or confidence level and the second probability or confidence level.

12. The processor-implemented method of claim 1 , further comprising: selecting a third regional SG model from a plurality of third regional SG models for corresponding regions of a plurality of third regions of the input parameter space based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of third regions based on a third partitioning scheme different from the first partitioning scheme and the second partitioning scheme; as well as estimating a third SG value using the third regional SG model and the sensor measurement data, Wherein determining the predicted SG value comprises determining the predicted SG value based on a combination of the first SG value, the second SG value and the third SG value.

13. The processor-implemented method of claim 1 , wherein: Each of the plurality of first regional SG models and the plurality of second regional SG models comprises one or more machine learning models, equations, functions, or a combination thereof; and The first regional SG model and the second regional SG model among the plurality of first regional SG models and the plurality of second regional SG models use different input parameters.

14. A system, comprising: one or more processors; and One or more processor-readable media storing instructions that, when executed by the one or more processors, cause operations including: receiving sensor measurement data measured by a glucose sensor, the sensor measurement data comprising at least one signal indicative of a glucose level and at least one signal indicative of a sensor health; selecting, based at least in part on the sensor measurement data, a first regional sensor glucose (SG) model from a plurality of first regional SG models for corresponding regions of an input parameter space associated with the sensor measurement data, wherein the input parameter space is partitioned into the plurality of first regions based on a first partitioning scheme; estimating a first SG value using the first regional SG model and the sensor measurement data; selecting a second regional SG model from a plurality of second regional SG models for respective regions of the input parameter space based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of second regions based on a second partitioning scheme different from the first partitioning scheme; estimating a second SG value using the second regional SG model and the sensor measurement data; as well as A predicted SG value is determined based on a combination of the first SG value and the second SG value.

15. The system of claim 14, wherein at least one of the first partitioning scheme or the second partitioning scheme partitions the input parameter space based at least in part on sensor current (Isig), back voltage (Vcntr), electrochemical impedance spectroscopy (EIS) data, age of the glucose sensor, temperature, age of a user of the glucose sensor, body mass index (BMI) of the user, or a combination thereof.

16. The system of claim 14, wherein the first partitioning scheme or the second partitioning scheme partitions the input parameter space based at least in part on a sensor current and an age of the glucose sensor. 17 . The system of claim 14 , wherein determining the predicted SG value based on the combination of the first SG value and the second SG value comprises determining an average or a weighted average of the first SG value and the second SG value.

18. One or more non-transitory processor-readable storage media storing instructions that, when executed by one or more processors, cause operations including: receiving sensor measurement data measured by a glucose sensor, the sensor measurement data comprising at least one signal indicative of a glucose level and at least one signal indicative of a sensor health; selecting, based at least in part on the sensor measurement data, a first regional sensor glucose (SG) model from a plurality of first regional SG models for corresponding regions of an input parameter space associated with the sensor measurement data, wherein the input parameter space is partitioned into the plurality of first regions based on a first partitioning scheme; estimating a first SG value using the first regional SG model and the sensor measurement data; selecting a second regional SG model from a plurality of second regional SG models for respective regions of the input parameter space based at least in part on the sensor measurement data, wherein the input parameter space is partitioned into the plurality of second regions based on a second partitioning scheme different from the first partitioning scheme; estimating a second SG value using the second regional SG model and the sensor measurement data; as well as A predicted SG value is determined based on a combination of the first SG value and the second SG value.

19. One or more non-transitory processor-readable media according to claim 18, wherein at least one of the first partitioning scheme or the second partitioning scheme partitions the input parameter space based at least in part on sensor current (Isig), counter voltage (Vcntr), electrochemical impedance spectroscopy (EIS) data, age of the glucose sensor, temperature, age of a user of the glucose sensor, body mass index (BMI) of the user, or a combination thereof.

20. The one or more non-transitory processor-readable media of claim 18, wherein determining the predicted SG value based on the combination of the first SG value and the second SG value comprises determining an average or a weighted average of the first SG value and the second SG value.

Images