Methods, systems, and devices for improved skin temperature monitoring

Abstract

The present disclosure provides computer-implemented methods, systems, and devices for improved skin temperature monitoring. Accurate estimation of skin and ambient temperatures is generated based on determinations and comparisons of skin and internal device temperature sensor measurements contained on or within an example device. Estimation of skin and ambient temperature measurements helps monitor skin and core temperature changes, detect physiological events of a wearer of the example device, and determine when skin temperature changes are caused by external environmental or physiological causes.

Description

Technical Field

[0001] This disclosure generally relates to health monitoring and wearable devices for health monitoring. More specifically, this disclosure relates to methods and devices for improved skin temperature monitoring. Background Technology

[0002] Skin temperature data can be influenced by changes in ambient temperature. Therefore, devices providing skin temperature monitoring may introduce inaccuracies unrelated to sensor error. Methods, systems, and devices are needed to distinguish between physiologically induced skin temperature changes and those induced by external environmental factors. Summary of the Invention

[0003] Aspects and advantages of embodiments of the present invention will be set forth in part in the description which follows, or may be learned from the description or by practice of the embodiments.

[0004] One example aspect of this disclosure relates to a computer-implemented method for providing improved skin temperature monitoring. The method includes: a computing system including one or more computing devices determining the internal device temperature of a wearable device based on sensor data received from an internal device temperature sensor contained within a wearable device worn by a user. The method further includes: the computing system determining a first estimate of the user's skin temperature based on sensor data received from a skin temperature sensor contained on or within the wearable device. Next, the method involves: the computing system estimating an ambient air temperature based at least in part on the first estimate of the skin temperature and the internal device temperature. The method then includes: the computing system refining the first estimate of the skin temperature based at least in part on the estimated ambient air temperature to generate a second estimate of the skin temperature.

[0005] Another example aspect of this disclosure relates to a wearable device, comprising: a device housing configured for wear by a user; one or more processors contained within the device housing; one or more skin temperature sensors included on or within the device housing and configured to generate skin temperature sensor data; one or more internal device temperature sensors included within the device housing and configured to generate internal device temperature sensor data; and a non-transitory computer-readable storage device included within the device housing and storing instructions that, when executed by the one or more processors, cause the wearable device to perform operations. Specifically, the operations include: determining an internal device temperature within the device housing based at least in part on internal device temperature sensor data received from one or more internal device temperature sensors contained within the wearable device; determining a first estimate of the user's skin temperature based on sensor data received from skin temperature sensors included on or within the device housing; estimating an ambient air temperature based at least in part on the first estimate of the skin temperature and the internal device temperature; and refining the first estimate of the skin temperature based at least in part on the estimated ambient air temperature to generate a second estimate of the skin temperature.

[0006] Another example aspect of this disclosure relates to a computer-implemented method for providing improved skin temperature monitoring. The method includes: a computing system including one or more computing devices determining the internal device temperature of a wearable device based on sensor data received from an internal device temperature sensor contained within a wearable device worn by a user. Next, the method includes: the computing system determining a first estimate of the user's skin temperature based on sensor data received from a skin temperature sensor included on or within the wearable device. The method then involves: the computing system refining the first estimate of the skin temperature, at least partially based on the internal device temperature of the wearable device, to generate a second estimate of the skin temperature. The method then includes: the computing system determining one or more physiological events, at least partially based on the second estimate of the skin temperature. Attached Figure Description

[0007] The specification provides a detailed discussion of embodiments for those skilled in the art, with reference to the accompanying drawings, in which:

[0008] Figure 1 The illustration shows a front view of an example wearable device according to an example embodiment of the present disclosure.

[0009] Figure 2 The illustration shows a rear view of an example wearable device according to an example embodiment of the present disclosure.

[0010] Figure 3 An internal side view of an example wearable device according to an exemplary embodiment of the present disclosure is illustrated, showing an internal device temperature sensor contained within a housing.

[0011] Figure 4 An internal side view of an example wearable device according to an exemplary embodiment of the present disclosure is illustrated, showing a skin temperature sensor contained within a housing.

[0012] Figure 5 The illustration shows the flow of steps in an exemplary skin temperature denoising method according to an exemplary embodiment of the present disclosure.

[0013] Figure 6 An example physiological event detection system according to an example embodiment of the present disclosure is illustrated.

[0014] Figure 7 Another example of a physiological event detection system according to an exemplary embodiment of the present disclosure is illustrated.

[0015] Figure 8 Another example of a physiological event detection system according to an exemplary embodiment of the present disclosure is illustrated.

[0016] Figure 9 An example computing system according to an example embodiment of the present disclosure is illustrated.

[0017] Figure 10 A graph illustrating example temperature data from a wearable device according to an example embodiment of the present disclosure is shown.

[0018] Figure 11 The illustration shows a graph of example internal devices and skin temperature sensor data of a wearable device according to an example embodiment of the present disclosure.

[0019] Figure 12 A graph illustrating example internal device and skin temperature difference data according to an example embodiment of the present disclosure is shown.

[0020] Figure 13 A graph illustrating example temperature data from several wearable devices according to exemplary embodiments of the present disclosure is shown.

[0021] Figure 14 The illustration shows a graph of an example linear regression model for predicting ambient temperature according to an example embodiment of the present disclosure.

[0022] Figure 15 A graph illustrating example skin temperature changes related to the number of days of fever, according to an example embodiment of the present disclosure.

[0023] Figure 16 A graph illustrating example temperature changes including sleep data is shown according to an example embodiment of the present disclosure. Detailed Implementation

[0024] Overview

[0025] Generally, this disclosure relates to health monitoring and wearable devices for health monitoring, such as methods and apparatus for improved skin temperature monitoring. Specifically, exemplary aspects of this disclosure relate to computer-implemented methods for providing improved skin temperature monitoring. The methods of this disclosure allow for the observation of meaningful changes in a person's core temperature and the identification of physiological trends and events related to a person's health.

[0026] More specifically, human skin temperature is expected to change in response to variations in core temperature. Sometimes, a positive correlation may exist between skin temperature and core temperature, while at other times, this relationship may lead to the opposite. For example, sometimes the body raises its core temperature to induce fever, and does so by reducing perfusion to the extremities; therefore, when a person's core temperature rises, a decrease in skin temperature is expected. The development of these physiological evoked changes allows for the observation of trends by comparing temperature data streams that might not be possible by observing other data streams. On the other hand, changes in skin temperature can also be influenced by changes in ambient temperature, making it difficult to identify skin temperature changes associated with physiological variations. Therefore, methods are needed to distinguish between physiologically induced changes in skin temperature and those induced by the external environment.

[0027] In view of these challenges, this disclosure provides systems and methods for computer-implemented improved skin temperature monitoring. In some embodiments, the method of this disclosure involves a computing system determining the internal temperature of a wearable device based at least in part on sensor data from one or more sensors contained within a wearable device worn by a user. In examples, the wearable device may be a wristband, bracelet, watch, armband, ring worn around a user's finger, or other wearable products that may be equipped with sensors as described in this disclosure. As another example, one or more sensors measuring the internal temperature of the device may be contained within the housing of the wearable device but may not be in thermal contact with the user's skin. When one or more sensors measuring the internal temperature of the device are arranged according to this example, the sensors may be associated with the user's skin temperature, but may also be more influenced by the ambient air temperature, for example, compared to dedicated skin temperature sensors.

[0028] According to another aspect, the method disclosed herein relates to a computing system determining a first estimate of a user's skin temperature based on sensor data received from one or more skin temperature sensors that are in thermal contact with the user and are included in a wearable device. In another example, the wearable device may include a thermally conductive baseplate configured to contact the user's skin. In this example, one or more skin temperature sensors may be configured to measure the temperature of the baseplate. When one or more sensors measuring the first estimate of the user's skin temperature are arranged in this manner, for example, the sensors may be less affected by external air temperature compared to the aforementioned internal device temperature sensors and / or dedicated ambient air temperature sensors.

[0029] According to another aspect of this disclosure, the computing system can estimate the ambient air temperature based at least in part on a first estimate of skin temperature and the internal temperature of the wearable device. In some embodiments, the computing system estimates the ambient air temperature by determining the difference between the first estimate of skin temperature and the internal temperature of the wearable device. In examples, the estimated ambient air temperature may be the product of one or more smoothing or curve fitting processes on a dataset corresponding to the temperature differences described above. In some embodiments, the computing system can estimate the ambient air temperature by processing the first estimate of skin temperature and the internal temperature of the wearable device using a machine learning model. As an example, the machine learning model may include a linear regression model, a neural network (e.g., a recurrent neural network), or a clustering model. In some embodiments, the computing system may receive the ambient air temperature as a predicted output of the machine learning model.

[0030] According to another aspect of this disclosure, the computing system can generate an intermediate estimate of skin temperature based at least in part on a first estimate of skin temperature and the internal device temperature. This intermediate estimate of skin temperature can be further refined based at least in part on an ambient temperature estimate to help generate a second estimate of skin temperature. In an example, the intermediate estimate of skin temperature may be the product of one or more smoothing or curve-fitting processes on a dataset corresponding to the aforementioned temperature differences. In some embodiments, the computing system can generate an intermediate estimate of skin temperature by processing the first estimate of skin temperature and the internal temperature of the wearable device using a machine learning model. As an example, the machine learning model may include a linear regression model, a neural network (e.g., a recurrent neural network), or a clustering model. In some embodiments, the computing system may receive the intermediate estimate of skin temperature as a prediction output of the machine learning model.

[0031] In some implementations, the computing system may adjust the ambient air temperature estimate based at least in part on additional environmental sensor data acquired from additional environmental sensors included within the wearable device. In some implementations, the additional environmental sensors may include location sensors (e.g., GPS sensors), geosensors, weather sensors, motion sensors, altitude sensors, altimeter temperature sensors, ambient light sensors, heart rate sensors, and other physiological sensors. In one example, the computing system may observe whether the ambient light sensor included within the wearable device is covered or uncovered. In this example, the computing system may infer a warmer ambient temperature based on whether the ambient light sensor is covered, which can help determine skin temperature estimates based on skin temperature, skin temperature changes, and skin temperature change rates, and thus help determine physiological trends. In some implementations, the computing system may adjust the ambient air temperature estimate based at least in part on sleep data of the user collected by the wearable device. In some implementations, examples of sleep data may include data indicating whether the user is awake or asleep, the user's sleep coefficient, and the user's heart rate. If the user is asleep, the data may also include information indicating whether the user is moving (e.g., repositioning) and characteristics of the type of sleep the user is experiencing (e.g., restless sleep, light sleep, deep sleep, REM sleep). In the example, the computing system can infer that the user's repositioning during sleep is a function of the user's thermoregulatory cycle, thus helping to determine the user's core temperature based at least in part on a second estimate of the user's skin temperature.

[0032] According to another aspect of this disclosure, the method involves a computing system refining a first estimate of skin temperature, at least in part, based on an intermediate estimate of the estimated ambient air temperature or skin temperature, to generate a second estimate of skin temperature. In examples, the second estimate of skin temperature may be the product of one or more smoothing or curve-fitting processes corresponding to a dataset of estimated ambient air temperatures. In some embodiments, the computing system modifies a confidence value associated with the second estimate of skin temperature or the estimated ambient air temperature. As an example, the confidence value may increase as the temperature difference between the first estimate of skin temperature or the estimated ambient air temperature and the internal temperature of the wearable device decreases. In other words, a smaller temperature difference between the first estimate of skin temperature or the estimated ambient air temperature and the internal temperature of the wearable device can be equivalent to greater confidence in the second estimate of skin temperature. As another example, a smaller temperature difference also means that the ambient temperature is unlikely to cause a change in the user's skin temperature.

[0033] According to another aspect of this disclosure, the method relates to a computational system determining one or more physiological events of a user based at least in part on a second estimate of skin temperature. In some embodiments, physiological events may include fever, circadian rhythm, menstrual cycle, ovulation, heat stress, and thermal comfort. Detection of a physiological event may include detection or determination of the condition of the physiological event, such as detection of the onset of the physiological event, determination of the ongoing condition of the event, and / or prediction of the future condition of the event (e.g., a fever peak expected to occur in the next six hours). In an example, the computational system may detect fever by smoothing a dataset of second estimates of skin temperature collected after an increase in the user's skin temperature.

[0034] In some implementations, the computing system may estimate a user's core temperature based on a second estimate of skin temperature. In examples, the estimated user core temperature may be based on a second estimate of skin temperature collected at different frequencies and during different sleep stages of the user (e.g., awake or asleep). In some implementations, the computing system distinguishes between physiologically induced core temperature changes and those induced by external environmental factors. In some implementations, the computing system may monitor the rate of change of the second estimate of skin temperature to detect shifts in the second estimate. As an example, the computing system may determine physiological events based on detected shifts. In some examples, physiological events may include the onset of fever, circadian rhythms, menstrual cycles, ovulation, heat stress, and thermal comfort.

[0035] The methods and apparatus disclosed herein provide several technical effects and benefits, including determining higher confidence estimates of human true skin temperature, observing meaningful changes in human core temperature, and identifying physiological trends and events related to human health.

[0036] This disclosure further refines the data received from sensors within a wearable device by combining and analyzing data from multiple sensors in a manner that allows users to observe not only the raw data but also trends and events inferred from the data. In this way, the disclosure also eliminates the need for additional sensors within the wearable device, thereby saving device space and processor usage. The disclosure also allows for more accurate monitoring of skin temperature by implementing the device's sensors in a more efficient, predictable, and useful manner.

[0037] Exemplary embodiments of this disclosure will now be discussed in more detail with reference to the accompanying drawings.

[0038] Example methods, systems, and devices

[0039] Figure 1 A front view of an example wearable device 100 according to an exemplary embodiment of the present disclosure is illustrated. Although Figure 1Wearable device 100 is a watch, but the systems and methods of this disclosure can be applied to any different type of wearable device, such as wristbands, bracelets, watches, armbands, headbands, glasses, earplugs, rings placed around a user's finger, clothing items with embedded computers, and / or other wearable products that may be equipped with sensors as described in this disclosure. Wearable device 100 may be configured with a display 102, a device housing 104, and a strap 106.

[0040] In some implementations, wearable device 100 may be configured to transmit data to a user via display 102 and / or any other data output device such as a haptic device, a light-emitting diode, etc. The transmitted data may include data related to skin temperature, heart rate, sleep status (e.g., light sleep, deep sleep, and REM sleep), sleep coefficient (e.g., sleep score), and other physiological data of the user (e.g., blood oxygen level). Display 102 may also be configured to transmit data from additional environmental sensors included in or on wearable device 100. Example information transmitted by display 102 from these additional environmental sensors may include location, altitude, and weather associated with the user's position. Display 102 may also transmit data about the user's movement (e.g., whether the user is stationary, walking, and / or running).

[0041] In some implementations, the display 102 may be configured to display information about a user's physiological events. Examples of physiological events that may be displayed include the user's fever, circadian rhythm, menstrual cycle, ovulation, heat stress, and thermal comfort. The display 102 may also convey information related to the detection of a user's physiological events. Configured in this way, the display 102 may communicate to the user the detection or determination of the status of a physiological event, such as the detection of the onset of a physiological event, the determination of the ongoing status of the event, and / or a prediction of the status of future events (e.g., a high fever is expected to occur within the next six hours).

[0042] Display 102 or other data input devices (e.g., various touch sensors and / or physical buttons, switches, or toggle keys) may also be configured to receive user-inputted data. Examples of user-inputted data may include information about symptoms, sleep patterns, ovulation, menstruation, and other physiological information related to the user's health.

[0043] Device housing 104 may be configured to include one or more sensors. Example sensors included in device housing 104 may include a skin temperature sensor, an internal device temperature sensor, a location sensor (e.g., GPS), a motion sensor, an altitude sensor, a heart rate sensor, and other physiological sensors (e.g., a blood oxygen level sensor). Device housing 104 may be configured to include one or more processors.

[0044] The strap 106 can be configured to secure the wearable device 100 to a user's arm by, for example, connecting the end of the strap 106 to a buckle, loop or other similar fastening device, thereby allowing the wearable device 100 to be worn by the user.

[0045] Figure 2 The illustration shows a rear view of an example wearable device 100 according to an exemplary embodiment of the present disclosure. The wearable device 100 may be configured with a base plate 202. The wearable device 100 may also be configured with one or more sensors 204 that can be fixed to the base plate 202. Example sensors 204 may include skin temperature sensors, location sensors (e.g., GPS), motion sensors, altitude sensors, heart rate sensors, altimeter temperature sensors, and other physiological sensors (e.g., blood oxygen level sensors). The base plate 202 may be thermally conductive and may be configured to come into thermal contact with the user and / or the skin temperature sensors when the wearable device 100 is worn by the user. For example, when worn, the base plate may press against the user's skin. Similarly, the skin temperature sensors may be configured to generate a temperature reading of the base plate. Thus, the base plate 202 may be configured to measure the user's skin temperature when the base plate 202 is in thermal contact with the user (e.g., the temperature of the base plate is used as a representative of the user's skin temperature).

[0046] While some example implementations use a base plate to read the user's skin temperature, other methods can also be used, including temperature sensors configured to be in direct contact with the user's skin. In another example, an infrared sensor may be included in device 100 (e.g., within housing 104) and the user's skin temperature may be measured using infrared light. For example, infrared light may pass through holes in the base plate 202. In another example, one or more temperature sensors may be incorporated into a strip 106 and may be used to measure the user's skin temperature. For example, one or more temperature sensors may be woven into a fabric version of strip 106.

[0047] Figure 3The illustration shows an internal side view of an example wearable device 100 according to an exemplary embodiment of the present disclosure, illustrating an internal device temperature sensor 302 contained within a device housing 104. The internal device temperature sensor 302 can be configured to measure the temperature (e.g., internal device temperature) of the internal space of the device housing 104 and generate data relating to the temperature (e.g., internal device temperature) of the internal space of the device housing 104. The internal device temperature sensor 302 may also consist of multiple internal device temperature sensors. The internal device temperature sensor 302 may be contained within the device housing 104 of the wearable device 100 but is generally not in thermal contact with the user's skin. In one example, the internal device temperature sensor 302 may be attached to the inside of the device housing 104 (e.g., rather than the base plate 202). In another example, the internal device temperature sensor 302 may be attached to a printed circuit board (e.g., a "motherboard") of the device 100.

[0048] Figure 4 An internal side view of an example wearable device 100 according to an exemplary embodiment of the present disclosure is illustrated, showing a skin temperature sensor 402 contained within a device housing 104. In some embodiments, the skin temperature sensor 402 may be configured to physically contact a user. In this way, the skin temperature sensor 402 can measure the user's skin temperature and generate data related to the user's skin temperature. In other embodiments, the wearable device 100 may include a thermally conductive base plate 202 configured to have thermal contact with the skin temperature sensor 402 and the user's skin. In such an embodiment, the skin temperature sensor 402 may be configured to measure the temperature of the base plate 202 and generate data related to the temperature of the base plate 202, which also represents the temperature of the user's skin. In another embodiment, an infrared sensor may be included in the device 100 (e.g., within the housing 104) and the temperature of the user's skin may be measured using infrared light. For example, infrared light may pass through holes in the base plate 202. The skin temperature sensor 402 may also be composed of multiple skin temperature sensors (e.g., multiple different sensors employing various different methods described herein). Other arrangements are also possible.

[0049] Figure 5The illustration shows the flow of steps in an exemplary skin temperature denoising method 500 according to an exemplary embodiment of the present disclosure. In step 502, the skin temperature denoising method 500 can determine the internal device temperature of a wearable device 100 worn by a user. The determination of the internal temperature of the wearable device 100 in step 502 can be based at least in part on sensor data from one or more internal device temperature sensors 302 contained on or within the wearable device 100. In another example, determining the internal temperature of the wearable device 100 in step 502 can be based at least in part on data from temperature sensors contained within other components of the device 100, such as temperature sensors present in the accelerometer and charging circuitry, which generate temperature data as part of monitoring and calibrating their own output. In the next step 504, the skin temperature denoising method 500 can determine a first estimate of the user's skin temperature. Determining the first estimate of the user's skin temperature in step 504 can be based at least in part on sensor data received from one or more skin temperature sensors 402 that are in physical and / or thermal contact with the user and are contained within the wearable device 100. In another example, the first estimate of the user's skin temperature determined in step 504 may be based at least in part on sensor data from one or more skin temperature sensors 402 that are in thermal contact with the base plate 202. In this example, the base plate 202 may also be configured to be in thermal contact with the user's skin when the user wears the wearable device 100.

[0050] In step 506, the skin temperature denoising method 500 may estimate the ambient air temperature at least in part based on a first estimate of the skin temperature from step 504 and the internal device temperature of the wearable device 100 from step 502. The estimation of the ambient air temperature in step 506 may be based on determining the difference between the first estimate of the skin temperature from step 504 and the internal device temperature of the wearable device 100 from step 502. In another example, the estimation of the ambient air temperature at step 506 may be the product of one or more smoothing or curve fitting processes on a dataset corresponding to the temperature difference between the first estimate of the skin temperature from step 504 and the internal device temperature of the wearable device 100 at step 502.

[0051] In step 506, the skin temperature denoising method 500 may also generate an intermediate estimate of the skin temperature based at least in part on the first estimate of the skin temperature from step 504 and the internal device temperature of the wearable device 100 from step 502. The intermediate estimate of the skin temperature from step 506 may be further refined based at least in part on the ambient temperature estimate to help generate a second estimate of the skin temperature in step 508 of the skin temperature denoising method 500. The intermediate estimate of the skin temperature at step 506 may be the product of one or more smoothing or curve fitting processes on a dataset corresponding to the temperature difference between the first estimate of the skin temperature from step 504 and the internal device temperature of the wearable device 100 at step 502.

[0052] In yet another example, an estimate of the ambient air temperature or an intermediate estimate of the skin temperature at step 506 can be determined by processing a first estimate of the skin temperature from step 504 and the internal temperature of the wearable device from step 502 using a machine learning model. In such an implementation, the estimate of the ambient air temperature or the intermediate estimate of the skin temperature at step 506 can be a predictive output of the machine learning model. For example, the machine learning model can be trained over time using a supervised learning method on a set of training data that includes pairs of input data and baseline fact labels. For example, for each pair, the input data may include example skin temperature and internal device temperature readings, while the baseline fact labels may include baseline fact ambient temperature. For example, the training dataset can be collected over time from cases where the wearable device operates in an external environment with known ambient temperatures. In some implementations, the training dataset can be generated using sensors present in a test chamber that enables control of various temperatures, including temperatures representative of the user's skin and the surrounding external environment. By training on such a training dataset, the machine learning model can be configured to generate a predicted intermediate estimate of the ambient temperature or skin temperature based on the skin temperature and the internal device temperature. Machine learning models can be any type of model, including various neural networks (e.g., feedforward neural networks, recurrent neural networks, transformer networks, etc.), linear models, support vector machines, clustering models, etc.

[0053] In another example, the estimate of the ambient air temperature at step 506 may be adjusted at least in part based on additional environmental sensor data acquired from additional environmental sensors contained on or within the wearable device 100. In this example, the estimate of the ambient air temperature at step 506 may be adjusted at least in part based on sleep data input by the user and collected by the wearable device 100 and / or other environmental data as described herein.

[0054] In step 508, the skin temperature denoising method 500 generates a second estimate of the user's skin temperature by refining the first estimate of the skin temperature, at least in part, based on an intermediate estimate of the estimated ambient air temperature or skin temperature from step 506. The second estimate of the skin temperature generated in step 508 may be the product of one or more smoothing or curve-fitting processes on a dataset corresponding to the intermediate estimates of the estimated ambient air temperature or skin temperature. In another example, one or more smoothing or curve-fitting processes on the dataset may account for minor repositioning of the user, such as when the user puts on or takes off a blanket, and other similar behavioral factors (e.g., general user movement). In this example, one or more smoothing or curve-fitting processes may modify the signal of the skin temperature data points so that the signal is less affected by temperature spikes caused by these behavioral factors.

[0055] In another example, generating a second estimate of the user's skin temperature in step 508 may involve modifying the confidence value associated with the second estimate of the skin temperature or the estimated ambient air temperature. In such an example, the confidence value may increase as the temperature difference between the first estimate of the skin temperature or the estimated ambient air temperature and the internal temperature of the wearable device 100 decreases. In this example, a smaller temperature difference between the first estimate of the skin temperature or the estimated ambient air temperature and the internal temperature of the wearable device 100 may be equivalent to greater confidence in the second estimate of the skin temperature.

[0056] In some implementations, at step 508, a second estimate 506 of the skin temperature can be determined by processing a first estimate of the skin temperature from step 504, an intermediate estimate of the skin temperature from step 506, and an estimated ambient temperature from step 506 using a machine learning model. In such implementations, the second estimate of the skin temperature at step 508 can be a predictive output of the machine learning model. For example, the machine learning model can be trained using supervised learning methods on a set of training data that includes pairs of input data and baseline fact labels. For example, for each pair, the input data can include example estimated skin temperature and ambient temperature readings, while the baseline fact label can include a baseline fact skin temperature. For example, this set of training data can be collected over time from cases where the wearable device is operating in conjunction with a more accurate skin temperature sensor. By training on such a training dataset, the machine learning model can be configured to generate refined skin temperature readings (e.g., those that only reflect physiological events rather than external environmental events). In some implementations, this set of training data can be generated using sensors present in a testing chamber that enables control of various temperatures, including temperatures representative of the user's skin and the surrounding external environment. Machine learning models can be any type of model, including various neural networks (e.g., feedforward neural networks, recurrent neural networks, transformer networks, etc.), linear models, support vector machines, clustering models, etc.

[0057] Figure 6 An example physiological event detection system 600 according to an exemplary embodiment of the present disclosure is illustrated. In the example, the physiological event detection system 600 may receive input in the form of additional environmental condition sensor data 602, internal device temperature sensor data 604, skin temperature sensor data 606, and sleep data 608.

[0058] Additional environmental condition sensor data 602 is optional and may include data from location sensors (e.g., GPS), geosensors, weather sensors, motion sensors, altitude sensors, altimeter temperature sensors, ambient light sensors, heart rate sensors, and other physiological sensors (e.g., blood oxygen sensors) contained on or within the wearable device 100. For example, by determining the user's location via the location sensors, information from various databases about local environmental conditions (e.g., temperature or other weather data) can be obtained and used to guide the process. Similarly, data from heart rate and other physiological sensors can help understand physiological rather than environmentally driven temperature changes.

[0059] Internal device temperature sensor data 604 may include data from one or more internal device temperature sensors 302 contained on or inside the wearable device 100. Skin temperature sensor data 606 may include data from one or more skin temperature sensors 402 contained on or inside the wearable device 100 and in thermal contact with the user or substrate 202 of the wearable device 100.

[0060] Sleep data 608 may include data collected by wearable device 100 or input by the user. Examples of sleep data may include data indicating whether the user is awake or asleep, the user's sleep coefficient, and the user's heart rate. In this example, if the user is asleep, sleep data 608 may also include information indicating whether the user is moving and a representation of the type of sleep the user is experiencing (e.g., restless sleep, light sleep, deep sleep, REM sleep).

[0061] In this example of the physiological event detection system 600, data inputs 602 and 608 can be processed by combining the execution of the skin temperature denoising method 500 with the processing of data inputs 604 and 606. As a component 610 of the physiological event detection system 600, a user's physiological events can be detected at least in part based on data inputs 602 and 608 and a second estimate of skin temperature generated by the execution of the denoising method 500. In such an example, physiological events from component 610 may include fever, circadian rhythm, menstrual cycle, ovulation, heat stress, and thermal comfort.

[0062] The detection of a physiological event 610 may include the detection or determination of the condition of the physiological event (e.g., the onset of the physiological event) and / or the prediction of the future condition of the event (e.g., a peak fever expected to occur in the next six hours). In another example, the detection of a physiological event 610 may involve smoothing a second estimated dataset of skin temperature collected after the user's skin temperature has risen.

[0063] In component 612 of the physiological event detection system 600, the system can infer physiological events. Inferring physiological events in component 612 may involve one or more processors contained within device housing 104 to enable display 102 to convey information to the user about the physiological events detected from component 610.

[0064] Figure 7 Another example of a physiological event detection system 600 according to an exemplary embodiment of the present disclosure is illustrated. In this example, the physiological event detection system 600 can receive data from... Figure 6 The physiological event detection system receives the same inputs 602, 604, 606, and 608. Figure 7 The system can also detect and infer related to Figure 6The same physiological events described above.

[0065] Figure 7 The physiological event detection system 600 can be configured to process internal device temperature sensor data 604 and skin temperature sensor data 606 using a machine learning model 702. In this example configuration, the machine learning model 702 may include a linear regression model, a neural network (e.g., a recurrent neural network), or a clustering model. In such an example, the physiological event detection system 600 may receive an estimate of the ambient air temperature or an intermediate estimate of the skin temperature as a prediction output of the machine learning model 702. The estimate of the ambient air temperature or the intermediate estimate of the skin temperature received in this way can also be used to estimate the actual skin temperature of the user wearing the wearable device 100. In component 706, Figure 7 The physiological event detection system 600 can also be configured to detect and monitor transitions (e.g., rate of change) in estimates of ambient air temperature and the user's actual skin temperature. In this configuration, detecting temperature transitions at 706 can help detect physiological events at component 610 and infer physiological events at 612.

[0066] Figure 8 Another example of a physiological event detection system 600 according to an exemplary embodiment of the present disclosure is illustrated. In this example, system 600 can receive data related to events occurring in the presence of... Figure 6 and 7 The physiological event detection system receives the same inputs 602, 604, 606, and 608. Figure 8 The system can also detect and infer the above-mentioned... Figure 6 and Figure 7 The same physiological events are described, including inputs related to the detection of temperature changes from component 706.

[0067] From Figure 8The physiological event detection system 600 may be configured with component 802 to determine the difference between the internal temperature of the wearable device 100 and a first estimate of the skin temperature of a user wearing the wearable device 100. In component 804 of the physiological event detection system 600, the system may estimate an intermediate estimate of the ambient air temperature or skin temperature based on this difference. The intermediate estimate of the ambient air temperature or skin temperature from 804 may also be used to estimate the effect of the ambient air temperature on the user's skin temperature by generating a second estimate of the user's skin temperature. The physiological event detection system 600 may also be configured to modify the confidence value associated with the second estimate of the skin temperature or the estimated ambient air temperature from 804. In this configuration of the physiological event detection system 600, the confidence value may increase as the temperature difference between the first estimate of the skin temperature or the estimated ambient air temperature and the internal temperature of the wearable device decreases. A smaller temperature difference determined at 802 may also mean that the estimated ambient air temperature is unlikely to affect the second estimate of the skin temperature. The second estimate of the skin temperature at 802 may also be used to estimate the user's core temperature. In this configuration, the physiological event detection system 600 can distinguish between core temperature changes caused by physiological factors and core temperature changes caused by external environmental factors.

[0068] Figure 9 An example computing system according to an exemplary embodiment of the present disclosure is illustrated. The computing system may include... Figure 9 The user computing device shown is an example of a wearable device 100. In some embodiments, the user computing device may be connected to a server computing system via a network. (Not shown in...) Figure 9 The image shows the server and network.

[0069] like Figure 9 The user equipment shown may include one or more processors and memory. The one or more processors may be any suitable processing device (e.g., processor core, microprocessor, ASIC, FPGA, controller, microcontroller, etc.) and may be a single processor or multiple processors operatively connected. Memory may include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, disks, etc., and combinations thereof. Memory may store data executed by the processor to enable… Figure 9 The computing system executes operations using data and instructions. Example operations could include the flow of steps in a skin temperature denoising method 500 and data from... Figure 6 , 7 The execution of component tasks described in various configurations of the physiological event detection system of 8.

[0070] Figure 9The user equipment may also include one or more user input components for receiving user input. For example, the user input component may be a touch-sensitive component (e.g., the display 102 of wearable device 100) that is sensitive to the touch of a user input object (e.g., a finger or stylus). The touch-sensitive component can be used to implement a virtual keyboard. Figure 9 The computing system may receive sample user input that may include information about symptoms, sleep patterns, ovulation, menstruation, and other physiological information related to the user's health.

[0071] Figure 9 The user equipment may also include one or more sensors, including one or more skin temperature sensors 402 and one or more internal device temperature sensors 302. The skin temperature sensor 402 and the internal device temperature sensor 302 may include negative temperature coefficient (NTC) thermistors, resistance temperature detectors (RTDs), thermocouples, and semiconductor-based sensors. Figure 9 Additional sensors for the computing system can be location sensors (e.g., GPS), motion sensors, altitude sensors, heart rate sensors, and other physiological sensors (e.g., blood oxygen level sensors). Location sensors can be cellular or satellite-based GPS sensors. Example motion sensors include passive infrared (PIR), microwave, and dual-technology / hybrid motion sensors. Example altitude sensors include barometric altimeters (e.g., non-hydraulic barometers) and radio altimeters. Example heart rate sensors include optically transmitted pulse sensors (e.g., photoplethysmography (PPG)) and electrocardiogram (ECD) sensors.

[0072] Server computing system ( Figure 9 (Not shown) may include one or more processors and memory. The one or more processors can be any suitable processing device (e.g., processor core, microprocessor, ASIC, FPGA, controller, microcontroller, etc.) and may be a single processor or multiple processors operatively connected. Memory may include one or more non-transitory computer-readable storage media, such as RAM, ROM, EEPROM, EPROM, flash memory devices, disks, etc., and combinations thereof. Memory may store data and instructions executed by the processor to enable the server computing system to perform operations.

[0073] In some implementations, a server computing system includes one or more server computing devices or is otherwise implemented by one or more server computing devices. Where the server computing system includes multiple server computing devices, such server computing devices may operate according to a sequential computing architecture, a parallel computing architecture, or some combination thereof.

[0074] A network can be any type of communication network, such as a local area network (e.g., an intranet), a wide area network (e.g., the Internet), or some combination thereof, and can include any number of wired or wireless links. Generally, communication over a network can be carried over any type of wired and / or wireless connection using various communication protocols (e.g., TCP / IP, HTTP, SMTP, FTP), encodings or formats (e.g., HTML, XML), and / or protection schemes (e.g., VPN, Secure HTTP, SSL).

[0075] In some implementations... Figure 9 User equipment may store or include one or more machine learning models. For example, machine learning models may or may not include various machine learning models, such as various neural networks (e.g., feedforward neural networks, recurrent neural networks, transformer networks, etc.), linear models, support vector machines, clustering models, etc.

[0076] In some implementations, one or more machine learning models may be received from a server computing system over a network, stored in the user device's memory, and then used or otherwise implemented by one or more processors. In some implementations, the user computing device may implement multiple parallel instances of a single machine learning model.

[0077] Additionally or alternatively, one or more machine learning models may be included based on the client-server relationship and Figure 9 User equipment communication is performed in a server computing system or otherwise according to the client-server relationship. Figure 9 The server computing system stores and implements user equipment communication. For example, machine learning models can be implemented by the server computing system as part of a network service (e.g., a data augmentation service). Therefore, it is possible to... Figure 9 One or more models may be stored and implemented at the user's device, and / or one or more models may be stored and implemented at the server computing system. Furthermore, some or all of the operations may be performed at one or more locations.

[0078] Example data

[0079] Figure 10 A graph illustrating example temperature data from a wearable device 100 according to an exemplary embodiment of the present disclosure is shown. Figure 10 The figure shows temperature measurements of the wearable device 100 at different times and under different conditions of use by the user of the wearable device 100. Figure 10 It also includes characterization of any fluctuations in temperature data and how these fluctuations are compared to the time of day and whether the user is wearing a wearable device.

[0080] Figure 11An example diagram illustrating temperature data from an internal device temperature sensor 302 and a skin temperature sensor 402 according to an exemplary embodiment of the present disclosure is shown. Figure 11 The diagram illustrates temperature measurements of the internal device temperature sensor 302 and skin temperature sensor 402 under different times and different usage conditions of the wearable device 100 by the user. Figure 11 This indicates temperature data recorded by two sensors, 302 and 402, based on the time of day and whether the user is wearing a wearable device. Figure 11 The illustration further illustrates the time periods of maximum and minimum temperature difference between the measurements of the internal device temperature sensor 302 and the skin temperature sensor 402. The time period of minimum temperature difference between the two temperature sensors 302 and 402 (e.g., when the difference is below a threshold) is shown in [the original text]. Figure 11 The data is marked and shaded. The time periods with the smallest temperature differences were also associated with higher confidence levels in the estimates of user skin temperature, meaning that ambient temperature is unlikely to cause changes in the user's skin temperature from the wearable device 100.

[0081] Figure 12 An example diagram is shown illustrating temperature difference data from an internal device temperature sensor 302 and a skin temperature sensor 402 according to an exemplary embodiment of the present disclosure. Figure 12 and Figure 11 The diagram in the figure is related, but instead plots the temperature difference between the temperature measurements of the internal device temperature sensor 302 and the skin temperature sensor 402. Figure 12 Alternative diagrams are provided (e.g., with...) Figure 11 (For comparison) The graph shows the time periods when the temperature difference is greatest and smallest. Figure 12 The text identifies and marks the time periods when the temperature difference between the two temperature sensors 302 and 402 is at its minimum (e.g., when the temperature difference is below a threshold). Figure 11 As shown, the time period when the temperature difference is smallest is also associated with a higher confidence level of the estimated user skin temperature, which means that ambient temperature is unlikely to cause changes in the skin temperature of the wearable device 100 users.

[0082] Figure 13 The illustration shows a graph of example temperature data from several wearable devices 100 according to exemplary embodiments of the present disclosure. Specifically, Figure 13 The chart plots the temperature measurements of the energy meter and skin temperature sensor 402 contained in the wearable device 100, and how these measurements fluctuate when exposed to a defined ambient temperature for a defined period of time, and whether the wearable device 100 is on the user's wrist.

[0083] Figure 14The illustration shows a graph illustrating an example linear regression model for environmental temperature estimation prediction according to an example embodiment of the present disclosure. Figure 14 In this process, the estimated ambient temperature is compared with temperature sensor data from sources, including environmental and phantom baseline sensors (including altimeter temperature sensor, skin temperature sensor 402, and internal device temperature sensor 302), within a defined time period.

[0084] Figure 15 A graph illustrating example skin temperature changes related to the number of days of fever, according to an example embodiment of this disclosure, is shown. In particular, Figure 15 The changes in average user skin temperature were plotted on real and smoothed temperature data over several days before and after a fever.

[0085] Figure 16 The illustration shows a graph of example temperature changes including sleep data according to an example embodiment of the present disclosure. More specifically, Figure 16 This represents the recording of temperature data from the internal device temperature sensor 302 and skin temperature sensor 402 included in the wearable device 100 within a defined time period, wherein sleep cycles and whether the user is wearing the wearable device 100 are identified and marked with shades. Figure 16 The diagram also shows the count values ​​measured by internal devices and skin temperature sensors during the same time period.

[0086] Additional Publication

[0087] The technologies discussed in this paper refer to servers, databases, software applications, and other computer-based systems, as well as the actions taken and the information sent to and from these systems. The inherent flexibility of computer-based systems allows for a wide variety of possible configurations, combinations, and divisions of tasks and functions among components. For example, the processes discussed in this paper can be implemented using a single device or component, or multiple devices or components working in combination. Databases and applications can be implemented on a single system or distributed across multiple systems. Distributed components can run sequentially or in parallel.

[0088] While the subject matter has been described in detail with respect to various specific example embodiments, each example is provided by way of explanation and not as a limitation of this disclosure. Those skilled in the art, upon gaining an understanding of the foregoing, can readily generate modifications, variations, and equivalents of these embodiments. Therefore, this disclosure does not exclude modifications, variations, and / or additions to the subject matter that will be obvious to those of ordinary skill in the art. For example, features illustrated or described as part of one embodiment may be used in another embodiment to produce yet another embodiment. Therefore, this disclosure is intended to cover such modifications, variations, and equivalents.

[0089] In particular, although Figures 1 to 16 For illustrative and discussion purposes, the steps performed in a specific order are illustrated, but the method of this disclosure is not limited to the specifically illustrated order or arrangement. Without departing from the scope of this disclosure, the individual steps of the skin temperature denoising method 500 and the physiological event detection system 600 may be omitted, rearranged, combined, and / or adjusted in various ways.

Claims

1. A computer-implemented method for providing skin temperature monitoring, the method comprising: A computing system comprising one or more computing devices determines the internal device temperature of the wearable device based on sensor data received from an internal device temperature sensor contained within the wearable device worn by the user. The computing system determines a first estimate of the user's skin temperature based on sensor data received from a skin temperature sensor contained in or within the wearable device; The ambient air temperature is estimated by the computing system based at least in part on a first estimate of the skin temperature and the internal device temperature, wherein estimating the ambient air temperature by the computing system based at least in part on the first estimate of the skin temperature and the internal device temperature includes: The estimated ambient air temperature is adjusted by the computing system based at least in part on sensor data from at least one of a location sensor, geographic sensor, weather sensor, motion sensor, altitude sensor, altimeter temperature sensor, ambient light sensor, or heart rate sensor included in the wearable device. The calculation system adjusts the estimate of the ambient air temperature based at least in part on the user's sleep data collected by the wearable device; and The calculation system refines the first estimate of the skin temperature, at least in part, based on the estimated ambient air temperature, to generate a second estimate of the skin temperature.

2. The computer-implemented method of claim 1, further comprising determining one or more physiological events by the computing system based at least in part on the second estimate of the skin temperature.

3. The computer-implemented method according to claim 2, wherein, The one or more physiological events include at least one of the following: fever, circadian rhythm, menstrual cycle, ovulation, heat stress, or heat comfort.

4. The computer-implemented method according to claim 2, wherein, Determining the one or more physiological events by the computing system based at least in part on the second estimate of the skin temperature includes: The user's core temperature is estimated by the computing system based at least in part on the second estimate of the skin temperature; and The computing system distinguishes between user core temperature changes caused by physiological factors and user core temperature changes caused by external environmental factors.

5. The computer-implemented method according to claim 2, wherein, Determining the one or more physiological events by the computing system based at least in part on the second estimate of the skin temperature includes: The calculation system monitors the rate of change of the second estimate of the skin temperature to detect changes in the second estimate of the skin temperature; and The calculation system determines the one or more physiological events based at least in part on the changes detected in the second estimate of the skin temperature.

6. The computer-implemented method according to claim 1, wherein, Estimating the ambient air temperature by the computing system based at least in part on the first estimate of the skin temperature and the internal device temperature includes: The calculation system determines the difference between the first estimate of the skin temperature and the internal device temperature; and The ambient air temperature is estimated by the calculation system based at least in part on the difference between the first estimate of the skin temperature and the internal device temperature.

7. The computer-implemented method according to claim 1, wherein, Estimating the ambient air temperature by the computing system based at least in part on the first estimate of the skin temperature and the internal device temperature includes: The first estimate of the skin temperature and the internal device temperature are processed using a machine learning model; and The ambient air temperature is received as the prediction output of the machine learning model.

8. The computer-implemented method according to claim 1, wherein, Refining the first estimate of the skin temperature by the calculation system, at least in part, based on the estimated ambient air temperature, to generate a second estimate of the skin temperature includes modifying the confidence value associated with the second estimate of the skin temperature.

9. The computer-implemented method according to claim 8, wherein, As the temperature difference between the first estimate of the skin temperature and the internal device temperature decreases, the confidence value associated with the second estimate of the skin temperature increases.

10. The computer-implemented method according to claim 1, wherein: The internal device temperature sensor included in the wearable device consists of multiple internal device temperature sensors; or The skin temperature sensor included in or within the wearable device consists of a plurality of skin temperature sensors.

11. A wearable device, comprising: A device housing configured to be worn by a user; One or more processors, said one or more processors being included within the device housing; One or more skin temperature sensors, the one or more skin temperature sensors being included on or within the device housing and configured to generate skin temperature sensor data; One or more internal device temperature sensors, the one or more internal device temperature sensors being included within the device housing and configured to generate internal device temperature sensor data; as well as A non-transitory computer-readable storage medium, the non-transitory computer-readable storage medium being included within the device housing and storing instructions that, when executed by the one or more processors, cause the wearable device to perform operations, the operations including: The internal device temperature within the device housing is determined at least in part based on internal device temperature sensor data received from one or more internal device temperature sensors contained within the wearable device; A first estimate of the user's skin temperature is determined based on sensor data received from a skin temperature sensor included on or within the device housing. Estimating the ambient air temperature based at least in part on a first estimate of the skin temperature and the internal device temperature, wherein estimating the ambient air temperature based at least in part on the first estimate of the skin temperature and the internal device temperature includes: The estimate of the ambient air temperature is adjusted based at least in part on sensor data from at least one of a location sensor, geographic sensor, weather sensor, motion sensor, altitude sensor, altimeter temperature sensor, ambient light sensor, or heart rate sensor included in the wearable device. The estimate of the ambient air temperature is adjusted based at least in part on the user's sleep data collected by the wearable device; and The first estimate of the skin temperature is refined, at least in part, based on the estimated ambient air temperature, to generate a second estimate of the skin temperature.

12. The wearable device according to claim 11, wherein, The operation also includes continuously observing the user's core temperature, which is at least in part based on the second estimate of the skin temperature.

Images