CN117015342A - Energy efficient detection and management of atrial fibrillation - Google Patents

Energy efficient detection and management of atrial fibrillation Download PDF

Info

Publication number
CN117015342A
CN117015342A CN202280019778.XA CN202280019778A CN117015342A CN 117015342 A CN117015342 A CN 117015342A CN 202280019778 A CN202280019778 A CN 202280019778A CN 117015342 A CN117015342 A CN 117015342A
Authority
CN
China
Prior art keywords
user
risk
electronic device
monitoring
sensors
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202280019778.XA
Other languages
Chinese (zh)
Inventor
朱立
张晗斌
维斯瓦姆·娜詹
况吉龙
高军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Samsung Electronics Co Ltd
Original Assignee
Samsung Electronics Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from US17/564,198 external-priority patent/US20220369992A1/en
Application filed by Samsung Electronics Co Ltd filed Critical Samsung Electronics Co Ltd
Publication of CN117015342A publication Critical patent/CN117015342A/en
Pending legal-status Critical Current

Links

Landscapes

  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

A method comprising: scheduling, by an electronic device, an Atrial Fibrillation (AF) monitoring period during which the electronic device intermittently monitors AF of a user of the electronic device, wherein the scheduling is based on determining a risk of AF specific to the user; and adjusting, by the electronic device, a time interval between successive AF monitoring periods in response to detecting the change in AF risk.

Description

Energy efficient detection and management of atrial fibrillation
Technical Field
The present disclosure relates generally to devices for monitoring and managing health conditions, and more particularly, to electronic devices for monitoring and managing atrial fibrillation.
Background
Atrial fibrillation is an irregular, often rapid heart rate. During atrial fibrillation, the two upper chambers of the individual's heart (the atria) are chaotic and beat uncoordinated with the two lower chambers of the individual's heart. Paroxysmal atrial fibrillation is characterized by sporadic episodes of atrial fibrillation that are removed, but which can occur frequently and can last for up to a week. Persistent atrial fibrillation is a type of heart of an individual that is unable to restore normal rhythm itself and that requires treatment if the rhythm is to be restored to a normal rate. The long-standing atrial fibrillation is continuous and may last for 12 months or more. For permanent atrial fibrillation, the individual's heart cannot return to normal rhythm, but must be controlled with medications to prevent blood clots and other serious conditions. Atrial fibrillation is associated with an increased risk of stroke, heart failure, and other heart related complications.
Disclosure of Invention
In an example embodiment, a method may include: an atrial fibrillation monitoring period is scheduled, wherein during the atrial fibrillation monitoring period, an electronic device intermittently monitors atrial fibrillation of a user of the electronic device, wherein the scheduling is based on determining a risk of atrial fibrillation specific to the user. During an atrial fibrillation monitoring period, the method may include: one or more signals corresponding to the user are sampled using one or more sensors operatively coupled to the electronic device. The method may comprise: in response to detecting a change in the risk of atrial fibrillation, the time interval between successive atrial fibrillation monitoring cycles is adjusted.
The method further comprises the steps of: in response to detecting AF, switching from intermittently monitoring the AF of the user to continuously monitoring the AF of the user.
Wherein the scheduling comprises: a data structure of an accumulated Normal Sinus Rhythm (NSR) detected by the electronic device is generated and, in response to detecting the NSR, a time interval between AF monitoring periods is changed, wherein the change is based on the AF risk and the number of accumulated NSRs detected.
Wherein the scheduling comprises: generating a data structure of an Undetermined Reading (UR) accumulated by the electronic device, and in response to detecting UR, changing a time interval between AF monitoring periods, wherein the changing is based on the AF risk and the accumulated number of UR.
The method, wherein determining the AF risk comprises: a dynamically evolving risk factor is determined based on at least one of physiological data, medical data, or lifestyle data corresponding to the user.
The method further comprises the steps of: the scheduling of the AF monitoring period is modified in response to determining a time-based reduction in sensitivity of the electronic device to detect AF, wherein the time-based reduction is detected based on a tracking error between a model estimate of the heart rhythm of the user and an observed heart rhythm sensed by the electronic device.
The method further comprises the steps of: during the AF monitoring period, one or more signals corresponding to the user are sampled using one or more sensors operatively coupled to the electronic device, and a sampling rate of at least one of the one or more sensors for monitoring AF of the user is dynamically adjusted, wherein the sampling rate is based on a probability that AF is properly detected by the electronic device.
The method, wherein the probability is determined using a statistical learning model trained to predict whether the electronic device correctly detects AF based on a predetermined set of AF-related factors, wherein the set of AF-related factors includes at least one of: motion detected by a sensor, heart rate, or duration of the AF monitoring period.
In another example embodiment, a system may include: one or more sensors; and one or more processors operatively coupled with the one or more sensors. The one or more processors may be configured to initiate operations. The operations may include: an atrial fibrillation monitoring period is scheduled, wherein during the atrial fibrillation monitoring period the one or more sensors intermittently monitor atrial fibrillation of a user, wherein the scheduling is based on determining a risk of atrial fibrillation specific to the user. The operations may include: in response to detecting a change in the risk of atrial fibrillation, the time interval between successive atrial fibrillation monitoring cycles is adjusted.
The system, wherein the processor is configured to initiate operations further comprising: in response to detecting AF, switching from intermittently monitoring the AF of the user to continuously monitoring the AF of the user.
The system, wherein the scheduling comprises: a data structure of accumulated Normal Sinus Rhythm (NSR) detected by the one or more sensors is generated and, in response to detecting NSR, a time interval between AF monitoring periods is changed, wherein the change is based on the AF risk and the accumulated number of detected NSRs.
The system, wherein the scheduling operation comprises: generating a data structure of the Undetermined Readings (UR) accumulated by the one or more sensors, and in response to detecting UR, changing a time interval between AF monitoring periods, wherein the changing is based on AF risk and an accumulated number of UR.
The system, wherein determining the AF risk comprises: a dynamically evolving risk factor is determined based on at least one of physiological data, medical data, or lifestyle data corresponding to the user.
The system, wherein the processor is configured to initiate operations further comprising: the scheduling of the AF monitoring period is modified in response to determining a time-based reduction in sensitivity of the one or more sensors to detect AF, wherein the time-based reduction is detected based on a tracking error between a model estimate of the heart rhythm of the user and an observed heart rhythm sensed by the one or more sensors.
In another example embodiment, a computer program product includes: one or more computer-readable storage media; and program instructions collectively stored on the one or more computer-readable storage media. The program instructions are executable by one or more processors of the electronic device to initiate operations. The operations may include: an atrial fibrillation monitoring period is scheduled, wherein during the atrial fibrillation monitoring period, one or more sensors of the electronic device intermittently monitor atrial fibrillation of the user. The scheduling may be based on determining an atrial fibrillation risk specific to the user. The operations may include: in response to detecting a change in the risk of atrial fibrillation, the time interval between successive atrial fibrillation monitoring cycles is adjusted.
This summary is provided merely to introduce some concepts and is not intended to identify any critical or essential features of the claimed subject matter. Other features of the inventive arrangements will be apparent from the accompanying drawings and from the detailed description that follows.
Drawings
The arrangement of the invention is shown by way of example in the accompanying drawings. However, the drawings should not be construed as limiting the inventive arrangement to only the particular embodiments shown. Various aspects and advantages will become apparent upon reading the following detailed description and upon reference to the accompanying drawings.
FIG. 1 illustrates an example system for energy efficient monitoring and detection of atrial fibrillation using electronics.
Fig. 2 illustrates an example atrial fibrillation sensing scheduler integrated in the system of fig. 1.
Fig. 3A and 3B illustrate example survey-based determination of atrial fibrillation risk.
Fig. 4 schematically illustrates an example attention model for determining risk of atrial fibrillation.
Fig. 5 illustrates an example atrial fibrillation sampling controller integrated in the system of fig. 1.
Fig. 6A and 6B illustrate an example training process for training a statistical learning model used by the atrial fibrillation sampling controller of fig. 5.
Fig. 7 schematically illustrates an example signal classification by the atrial fibrillation factor analyzer of the sampling controller of fig. 5.
FIG. 8 schematically illustrates an example compressive sensing arrangement for energy efficient detection of atrial fibrillation using electronics.
Fig. 9 illustrates a method for energy efficient detection of atrial fibrillation using an electronic device.
Fig. 10 illustrates an example apparatus for implementing the system of fig. 1.
Detailed Description
While the disclosure concludes with claims defining novel features, it is believed that the various features described herein will be better understood from a consideration of the specification in conjunction with the drawings. The process (es), machine(s), article(s), and any variations thereof described within this disclosure are provided for illustration purposes. Any specific structural and functional details described are not to be interpreted as limiting, but merely as a basis for the claims and as a representative basis for teaching one skilled in the art to variously employ the features described in virtually any appropriately detailed structure. Furthermore, the terms and phrases used within this disclosure are not intended to be limiting but rather to provide an understandable description of the features described.
The present disclosure relates generally to monitoring and managing health conditions, and more particularly to an apparatus for monitoring and managing Atrial Fibrillation (AF). Early detection of individual AF predisposition requires reliable AF screening tools. In order to manage AF in people with paroxysmal, persistent or permanent AF, reliable tools are needed for AF detection and AF burden estimation. An "AF burden estimate" as defined herein measures the amount of time an individual experiences AF per day. The higher the burden, the more likely an individual will experience one or more adverse consequences of AF, such as stroke or heart failure.
Despite many technological advances, portable devices have only limited capabilities for monitoring and managing AF. Their ability is largely limited by the energy required by such devices to perform various AF-related functions. For example, a photoplethysmography (PPG) sensor embedded in a portable device (e.g., smart watch, ear bud) may consume 10mW when continuously monitoring AF of a device user. The microcontroller of the device consumes 1 to 10mW when processing the sensor-generated data. In contrast, an Electrocardiogram (ECG) requires 1mW. The 9-axis Inertial Measurement Unit (IMU) requires only 0.5mW.
Portable devices and applications running on such devices currently only support limited AF monitoring due to the energy required for such monitoring and the inherent power constraints of the portable device. Passive AF monitoring is typically performed only one minute every 2 hour interval and only when the device user is at rest (such as when the user falls asleep at night). Thus, early detection of AF is unlikely and AF burden estimation is inefficient if it is fully achievable with portable devices using conventional techniques.
In accordance with the inventive arrangements described within this disclosure, example methods, systems, and computer program products can use portable or other electronic devices to improve energy saving AF detection and AF burden estimation. One aspect of the disclosed inventive arrangements is to balance the power consumption of an electronic device with the reliability of the AF detection capability of the device. By selectively scheduling AF monitoring based on a determination of AF risk for a device user, the arrangement disclosed herein limits energy consumption of the device to a monitoring period that is most likely to produce reliable AF detection results. Furthermore, the disclosed arrangement enables setting a sampling rate that also balances the energy consumption with the expected correctness of the AF of the user monitored by the identification means.
In one or more example embodiments, the system with the AF sensing scheduler is implemented in software and/or hardware integrated in or operatively coupled with an electronic device. The electronic device may be a portable device such as a smart watch, an ear bud, a smart phone or similar electronic device. The electronics may be assigned one or more sensors, such as PPG sensors, IMUs, and/or other sensors for monitoring AF of the user. The AF sensing scheduler is capable of scheduling AF monitoring periods in response to and based on a determination of AF risk for the user. As defined herein, an "AF risk" is the probability or likelihood that a user will experience AF within a predetermined time frame. The system may determine the risk of AF based on various types of data corresponding to the user, including physiological data (e.g., heart rate, heart rhythm pattern), lifestyle data (e.g., caloric intake, physical activity, alcohol consumption), and applicable medical data.
During an AF monitoring period, the system samples one or more signals corresponding to the user using one or more sensors (e.g., PPG sensors, IMUs) of the electronic device. In one aspect, the AF sensing scheduler is capable of adjusting a time interval between AF monitoring periods in response to detecting a change in AF risk for a user. The AF risk of the user may be detected by the system based on various evolving factors. As defined herein, an "evolution factor" is any user-specific attribute that affects the likelihood that a user will experience AF and that may change over time. The evolution factors may be related to the physiology, lifestyle, and/or medical treatment of the user.
In some arrangements, the system fills in gaps between AF monitoring periods by estimating the heart rate of the user using a statistical or machine learning model. To prevent possible degradation of accuracy of the model in response to an extension of the time interval between AF monitoring periods, the system compares the estimate with sensor-based observations generated during the AF monitoring periods. The system automatically resets the schedule of AF monitoring in response to accumulated error between the estimation and the sensor-based observations exceeding a predetermined threshold.
The system may also balance energy efficiency with AF sensing sensitivity by controlling the sampling rate of the sensor. The system iteratively determines the sampling rate by detecting an AF sensing factor and identifies the effect of the detected factor on AF sensing sensitivity. The identifying is performed by the system using a regression model or a machine learning model for generating the classification. The system sets the sampling rate according to the identification result
Under certain conditions, AF detection is not unduly impaired by decreasing the sampling rate. For example, AF detection algorithms that rely primarily on detecting the signal are robust to certain artifacts, as long as the main peak of the signal generated by the PPG sensor is minimally affected by signal interference. In some embodiments, the AF monitoring sensitivity and energy consumption may be balanced using the novel adaptable sampling strategies described herein. In certain arrangements disclosed herein, compressive sensing below the nyquist sampling rate may be used. The low sampling rate further enhances the energy efficiency of the device, but according to aspects disclosed herein, the AF sensing sensitivity of the device is not sacrificed.
One aspect of the inventive arrangements disclosed herein is to schedule AF sensing and set the sampling rate such that the electronic device consumes only as much energy as is required by the device to accurately detect AF with an acceptable level of confidence. The scheduled sensing is limited to the number of times that the most reliable results are expected to be produced. The sampling rate is limited to just the sampling rate required to have the AF detector correctly determine whether the sensor-generated signal is indicative of adequate assurance of atrial fibrillation in one round. Thus, in a probabilistic sense, the arrangement optimizes or almost optimizes the balance between energy consumption and device sensitivity. The enhanced energy efficiency enables the portable device to operate for a longer time, which makes early detection of AF using the portable device possible, and enables the portable device to perform AF burden estimation, which is two functions that have not been sufficiently achieved with the portable device so far.
Energy efficiency not only enables AF detection and burden estimation in electronic devices such as smart watches, earplugs, or other portable devices, but also enables the electronic device itself to operate more efficiently. For example, the energy consumed to power processes, memory Control Units (MCUs), and other components that would be unnecessarily consumed in processing data that does not increase the probability of correctly detecting AF is released and available for processing data that is not relevant to AF sensing. Thus, the overall efficiency and operability of the electronic device is enhanced for both AF-related and non-AF processing functions.
Other aspects of the inventive arrangements are described in more detail below with reference to the accompanying drawings. For simplicity and clarity of illustration, elements shown in the figures have not necessarily been drawn to scale. For example, the dimensions of some of the elements may be exaggerated relative to other elements for clarity. Further, where considered appropriate, reference numerals have been repeated among the figures to indicate corresponding, analogous, or analogous features.
Fig. 1 illustrates an example system 100 for energy saving monitoring and detection of AF. In various embodiments, each of the components of the described system 100 may be implemented in hardware (e.g., dedicated hardwired circuitry), software (e.g., program code executed by one or more processors), or a combination thereof. The system 100 may be integrated in or operatively coupled with an electronic device having the ability to monitor and/or detect AF of a user of the device. The system 100 may be implemented in an electronic device such as a smart watch, an ear bud, a smart phone, or similar device, such as device 1000 (fig. 10). For example, in an implementation with an apparatus such as apparatus 1000, components of system 100 may include program code that is electronically stored in a memory such as memory 1004 and executed on one or more processors such as processor 1002. The device 1000 may include one or more sensors 1026, such as PPG sensors, IMUs, and/or other such sensors for monitoring physiological properties of a device user.
System 100 illustratively includes an AF sensing scheduler 102, an AF sampling controller 104, a sensor actuator 106, a heart rhythm queue 108, an AF estimator/comparator 110, and a sensing scheduler reset 112. Operationally, the system 100 balances energy consumption with AF sensing sensitivity. The AF sensing scheduler 102 indicates, via the sensor actuator 106, the timing of sensing performed by one or more sensors 114 (e.g., PPG sensors, IMUs) of the electronic device, the sensing schedule being activated only if the probability of AF of the user is greater than a predetermined threshold. The timing of the AF monitoring period is adjusted by the AF sampling controller 104 in response to a change in the risk that a user of the electronic device will experience AF. The risk of AF is determined based on various physiological data. The physiological data may be generated by the AF detector 120 and/or the AF burden estimator 122. The physiological data may be generated by a health-related app running on the electronic device and electronically stored in the database 116. Physiological data may also be received from a network source via interface 118. The AF risk may also be based on lifestyle data 124 corresponding to the user, which is also collected from the app or other source and electronically stored in database 124. During periods when AF sensing is not performed, the AF estimator/comparator 110 predicts the likelihood of AF for the user based on recorded data of the user's past heart rhythm stored in the heart rhythm queue 108. The AF estimator/comparator 110 compares the predicted heart pattern to the sensor-based observations and prompts the sensing scheduler reset 112 to initiate frequent AF monitoring if the deviation between the observations and the predictions is too large.
Fig. 2 shows an example implementation of the AF sensing scheduler 102. Illustratively, the AF sensing scheduler 102 includes an AF risk determiner 202, a machine learning model 204, and an adjustor 206. Operatively, the AF sensing scheduler 102 can indicate the timing of sensing performed with the electronic device by scheduling AF monitoring periods. During each AF monitoring period, the electronic device intermittently monitors AF of a user of the electronic device based on signals generated by one or more sensors 114 (e.g., PPG sensors, IMUs) integrated in or otherwise operatively coupled with the electronic device. The AF sensing scheduler 102 schedules AF monitoring periods based on user-specific AF risk determined by the AF risk determiner 202. The AF risk determiner 202 may determine AF risk based on sensor-generated data corresponding to certain of the physiological attributes of the user (e.g., heart rate, heart rhythm pattern) and/or other AF risk related data specific to the user (e.g., lifestyle attributes).
In the absence of previously acquired data, the sensing scheduler may determine the AF risk based on, for example, a survey assessment that provides a score based on whether the user exhibits certain characteristics that have been statistically determined to be associated with the AF risk from a broad range of individual samples. One such assessment is made by FIG. 3A CHA as shown in evaluation 300 2 DS 2 -a vacc risk criterion. In some arrangements, the assessment 300 presents the CHA to the user via the interface 118 (e.g., GUI) 2 DS 2 The vacc risk criteria, and based on the user's response, the score 302 is determined based on which feature(s) of the characteristics 304 the user exhibits. In other arrangements, the score 302 may be obtained from a remotely located health platform (such as a website maintained by the user's doctor) or determined using a health-related application running on the user's electronic device. In any event, the score 302 may be electronically pre-stored for subsequent use by the system 100 in determining a user-specific risk of AF. In some arrangements, the AF risk determiner 202 will be CHA-based according to equation 1 2 DS 3 -score mapping of the vacc risk criteria to real R AF ∈(0，1)。
[ math 1 ]
R AF ＝1/(1+e (CH-5) )。
Wherein CH is CHA-based 2 DS 2 -score 302 of a vacc risk criterion. As shown in graph 306 in fig. 3B, if R AF Near 1, the user presents a low risk of AF. However, if R AF Near zero, the user presents a high AF risk for AF. If R is AF Near zero within a predetermined ε > O, the system 100 operates as an AF management tool that utilizes an AF sensing scheduler to schedule continuous or near continuous AF sensing for a predetermined interval. Otherwise, the lower the AF risk determined by AF risk determiner 202, the less frequent the scheduled AF sensing, such that system 100 operates as an AF monitoring tool.
When the one or more sensors 114 collect data, the AF risk determiner 202 may determine the AF risk of the user based on sensed physiological data (such as heart rate and heart rhythm pattern of the user). For example, the AF detector 120 can generate cardiac rhythm data based on sampling signals generated by one or more sensors 114. The AF-burden estimator 122 may estimate an AF burden associated with the user, for example, based on sampled signals generated by one or more sensors 114. In some arrangements, the AF sensing scheduler 102 determines a maximum time interval between sensing periods of one or more sensors 114 based on AF risks associated with a plurality of periods in which the user experiences a Normal Sinus Rhythm (NSR). Sinus rhythm is a rhythm that follows depolarization of the heart's myocardium that occurs in a set of cells (so-called "pacemaker" cells) in the right atrial wall of the heart. Sinus rhythm is normal if the rate of "discharge" (in response to depolarized heartbeats) is neither too fast nor too slow, typically defined by a rhythm between sixty and ninety-nine beats per minute.
In some arrangements, the AF sensing scheduler 102 generates a data structure of the accumulated Normal Sinus Rhythm (NSR) detected by one or more sensors 114. The AF risk determiner 202 determines AF risk based on the accumulated NSR, and the AF sensing scheduler 102 determines a maximum time interval between sensing periods of one or more sensors 114 according to equation 2.
[ formula 2 ]
MIN[10×2 (N-1) ×R AF ，M]Minutes).
Where N is the number of normal sinus rhythms detected and m=320 is the maximum number of minutes between monitoring cycles. An "indeterminate reading (undetermined reading, UR)" occurs whenever one or more of the sensors 114 fails to detect NSR or AF. The AF sensing scheduler 102 generates a data structure of accumulated Undetermined Readings (UR) and expands the time interval linearly (rather than exponentially as in the detected NSR) according to equation 3.
[ formula 3 ]
MIN[(N+2)×5×R AF ，K](minutes).
Where N is the number of normal sinus rhythms previously detected and k=30 is the maximum number of minutes between monitoring periods in response to an undetermined detection.
The risk of AF for the user may change over time. The change may be caused by any number of factors. For example, these factors may include the user's lifestyle, such as daily caloric intake, alcohol consumption, amount of exercise, stress level, and the like. For users with paroxysmal, persistent or permanent AF, factors that may affect the risk of AF for the user over time may include, for example, the progression of the disease and the history of treatment, as well as changes in the user's lifestyle. These various factors that relate to both the physiological and lifestyle attributes of the user include evolving factors, and typically change the risk of AF as determined by AF risk determiner 202.
In some arrangements, the system 100 is implemented in an electronic device that includes a health application that tracks and electronically stores data (such as physical activity, diet, sleep, and other lifestyle data) affecting a user's AF risk daily in databases 116 and/or 124. In other arrangements, the system 100 is additionally or alternatively implemented in an electronic device that includes one or more communication subsystems, such as one or more communication subsystems 1024 (fig. 10). Using such a communication subsystem of the electronic device, the system 100 may be communicatively coupled with the communication subsystem of the electronic device to access a remotely located health platform, such as a website, remote healthcare app, or other health-related site maintained by a user's doctor, via a wired or wireless connection to a data communication network (e.g., the internet). The AF sensing scheduler 202 may access the database 116 to retrieve physiological data generated in response to signals generated by one or more sensors 114 or received via the interface 118. Likewise, the AF sensing scheduler 202 may access the database 124 to access data related to the user's lifestyle and, if applicable, to disease progression of the user and/or medical treatment of the user.
The AF risk determiner 202 can determine a user-specific AF risk based on the evolution factors. As described, the evolution factors may include physiological data (e.g., heart rate, blood pressure, respiration rate) and lifestyle data (e.g., diet, alcohol intake, stress), as well as medical data in the case of a user suffering from paroxysmal, persistent, or permanent AF. In some arrangements, the AF risk determiner 202 predicts a user-specific AF risk based on evolving factors using a machine learning model 204.
In some arrangements, the machine learning model 204 is a Recurrent Neural Network (RNN). The RNN includes a plurality of fixed activation functions and algorithmically uses recursive relationships and back propagation to process sequence or order dependent data. In other embodiments, the machine learning model 204 is a long-term memory network (LSTM). LSTM uses different layers of activation functions or gates and maintains internal element state vectors to incorporate past learning and discard irrelevant data. In other arrangements, the machine learning model 204 utilizes an attention mechanism or an attention model. The attention model breaks down the complex processing tasks into smaller processing tasks, processing the input sequentially until the entire dataset is classified.
The machine learning model 204 may generate a function d that adaptively determines a change in user AF risk in response to one or more evolving factors AF (t). In response to a detected change in the risk of AF by the user, the adjustor 206 adjusts the time interval between successive AF monitoring periods. This enables the adjustor 206 of the AF sensing scheduler 102 to increase or decrease the frequency of AF monitoring by the electronic device, the change in frequency commensurate with the change in AF risk for the user.
Fig. 4 schematically shows sequential updating of mathematical formula 4.
[ math figure 4 ]
d AF (t)＝f(HR，BP，RR，SL，CA，MI，R AF ，w)。
Wherein HR is heart rate, BP is blood pressure, RR is respiratory rate, SL is stress level, MI is medical intake, and R AF Is a parameter as defined above. The parameter w is the "window" or calculation d AF Time of (t). Illustratively, d is dynamically learned using the attention model 400 AF (t). Vector 402 (its elements and<HR，BP，RR，SL，CA，MI，R AF >corresponding) indexed by time 404 and sequentially input into the model to generate d AF A time-based value 406 of (t). Once trained, the machine learning model 204 periodically updates the user-specific AF risk in response to one or more evolving factors (such as physiological factors, lifestyle factors, and/or medical-related factors, such as those described).
The adjustment by the adjustor 206 of the AF sensing scheduler 102 can approach a risk-based optimal balance between power consumption of the electronic device in a probabilistic sense and sensitivity of the electronic device to detect AF of the user by the AF detector 120 and to estimate AF burden by the AF burden estimator 122. The AF sensing scheduler 102 again optimizes the balance in a probabilistic sense. Balancing is to limit the power consumption due to the sensing of the electronic device while ensuring that there is an acceptable possibility of detecting whether the user is experiencing AF. Subject to performing sensing with sufficient frequency to achieve a predetermined likelihood of correctly detecting whether a user experiences AF, a probabilistic balance is achieved by maximizing the time interval between sensing cycles of one or more sensors 114.
If the user exhibits a high risk of AF (R AF And O), the system 100 causes sensing with one or more sensors 114 to cause the electronic device to act as an AF monitoring tool. For exhibiting low AF risk (R AF With respect to the user of 1), the system 100 schedules sensing with one or more sensors 114 at a lower frequency such that the time interval between scheduled AF monitoring periods is longer. However, under either condition, the AF sensing scheduler 102 schedules AF sensing such that sensing occurs only as frequently as needed to ensure an acceptable likelihood of accurately detecting AF, thereby optimizing an expected or probabilistic balance between power consumption of the electronic device and AF detection sensitivity.
During the time interval between AF monitoring by the electronic device using one or more sensors 114, the AF estimator/comparator 110 fills the gap by estimating or probabilistically predicting the heart rhythm of the user. In some arrangements, the AF estimator/comparator 100 uses a Hidden Markov Model (HMM) to estimate or predict the heart rhythm. In some arrangements, the HMM may be implemented by the AF burden estimator 122. Based on basic markov assumptions (e.g., P (z t |z t-1 ，z t-2 ，...，z 1 )＝P(z t |z t-1 ) HMM predicts unobserved or "hidden" states (e.g., NSR or AF). As is known in the art, the AF estimator/comparator 110 may predict the true state from a markov chain that includes a probability that reflects the likelihood of transitioning between states or remaining in a state from an initially observed stateA rate matrix.
However, the larger the time interval between AF monitoring periods, the less reliable the predictions generated by the AF estimator/comparator 110 using the HMM or another predictor model. Therefore, the possibility of failing to detect AF of the user increases. Thus, each time an actual sensor-based observation is periodically obtained, the AF estimator/comparator 110 compares the observation with the corresponding prediction or estimate. AF estimator/comparator 110 vs. error E t Accumulation is performed. If the error is accumulatedBeyond a predetermined threshold, the scheduler reset resets the sensing schedule of the AF sensing scheduler 102. The reset causes the AF sensing scheduler 102 to decrease the time interval between monitoring periods. The AF sensing scheduler 102 may return to based on an initial assessment of AF risk for the user (such as based on using CHA 2 DS 2 -R derived from VASc risk criteria AF Or another survey assessment).
Fig. 5 shows an example embodiment of the AF sampling controller 104. The AF sampling controller 104 can improve the energy efficiency of the electronic device by selecting the sampling rate at which the AF detector 120 samples the signals generated by the one or more sensors 114. Specifically, the AF sampling controller 104 sets the sampling rate as low as possible while still maintaining the desired level of accuracy with an acceptable level of confidence when AF is detected by the AF detector 120. By setting the sampling rate to the lowest sampling rate while maintaining the confidence level of sensing the user's AF, the AF sampling controller 104 avoids unnecessary power consumption. Reducing the sampling rate to that required for only the desired accurate sensing not only saves the energy consumed to drive the one or more sensors 114, but also reduces the energy consumption required to power other components of the electronic device that are dedicated to data collection and processing, such as the MCU and memory. Energy is saved by reducing the throughput of the data stream to only that required for accurate AF sensing.
For example, the power consumption of PPG sensors is typically dominated by the LED driver of the sensor. Therefore, the energy cost of AF sensing with PPG sensor is proportional to the operating time of the LEDs of the PPG. Thus, reducing the sampling rate reduces power consumption by reducing the duty cycle of the LEDs of the PPG. Thus, by controlling the sampling rate, the AF sampling controller 104 also reduces the amount of energy consumed by other sensors of the electronic device (e.g., IMU), and limits the energy consumed in signal processing, data processing, memory management, and the like.
The AF sampling controller 104 illustratively includes a feature extractor 502, an AF factor analyzer 504, a sampling strategy selector 506, and a signal reconstructor 508. Operatively, the AF sampling controller 100 predicts the likelihood that the AF detector 120 correctly detects AF of the user (if and when it occurs) based on signals generated by one or more sensors 114 during an AF monitoring period scheduled by the AF sensing scheduler 102 given a particular sampling rate. If the AF sampling controller 104 determines that an incorrect decision is more likely, the AF sampling controller 104 increases the sampling rate. If the AF sampling controller 104 determines that the correct decision is more likely, the AF sampling controller 104 decreases the sampling rate. Iteratively, the AF sample controller 104 may determine the lowest sample rate that is still expected to have a predetermined confidence level for generating a correct AF detection decision by the AF detector 120 based on the sampled signals generated by the one or more sensors 114.
Given a particular sampling rate, the predictive correctness or accuracy of the AF detector 120 is determined by the AF factor analyzer 504. The AF factor analyzer 504 implements a statistical learning model reflecting the capabilities of the AF detector 120. In various arrangements, the statistical learning model implemented by the AF factor analyzer 504 may be a binary or multi-class classifier model or a regression model. The statistical learning model may learn the likelihood of predicting the correct decision (AF or non-AF) of the AF detector 120 by supervised machine learning.
For example, because user motion during an AF monitoring period may be an important AF factor that affects the probability that the AF detector 120 makes an accurate decision, user motion may be an AF factor that is analyzed by the AF factor analyzer 504 to determine the likelihood that the AF detector 120 makes an accurate AF decision. The statistical learning model of the AF factor analyzer 504 may be trained offline using a training data set. Fig. 6A and 6B illustrate an example training arrangement 600 for training three classes of classification models. The example training arrangement 600 of three classes of classification models begins by collecting sensed data from one or more individuals randomly selected from a population. Data may be collected using two sensors, namely an Electrocardiogram (ECG) patch 602 attached to the chest of the individual 604 and a smart watch 606 worn on the wrist of the individual 604. The ECG patches 602 generate an ECG signal 608 and the smartwatch 606 generates a PPG signal 610 and IMU motion data 612. In the case where the PPG signal 610 is sampled at a high sampling rate, the PPG signal 610 is directly input to the AF detector 120 to determine the performance of the AF detector 120. In addition, the PPG signal 610 is converted to a downsampled signal 614 using a uniform duty cycle LED pulse 616 and restored to a PPG signal 618 before being input to the AF detector 120 in order to determine the performance of the AF detector 120 for low sample rate signals. The ECG signal 608 is annotated to provide real data 620 for which the correctness of the decision 622 of the AF detector 120 based on the high sample rate signal and the correctness of the decision 624 of the AF detector 120 based on the low sample rate signal are determined.
The matrix 626 summarizes the different results depending on whether the AF detector 120 correctly decides that the signal segment indicates AF or does not indicate AF and whether the decision is different depending on whether the segment corresponds to a high sampling rate signal or a low sampling rate signal. The decision for each segment is based on the result and one of the tags 628 is assigned accordingly. Class 0 includes segments that may be correctly classified even if the base signal is sampled only at a low sampling rate. Thus, signals classified as belonging to class 0 can be sampled at a low sampling rate that is energy efficient. Class 1 comprises segments that may be correctly classified only if the underlying signal is sampled at a high sampling rate. Class 2 includes segments of a signal that cannot be properly classified regardless of sample rate, and thus, one or more sensors 114 may be powered down to avoid useless energy consumption.
Using the assigned labels 628 that indicate the appropriate sample rates, a statistical learning model of the AF factor analyzer 504 may be trained with one or more AF factors in combination. Examples of AF factors include not only user motion during AF sensing, but also heart rate, monitoring time, and other AF factors that may affect AF sensing. Illustratively, in this context, the AF factor is IMU motion data 612 generated by the smartwatch 606. IMU motion data 612 generated by the smartwatch 606 simultaneously with the PPG signal 610 is synchronized with the PPG signal 610. Thus, the IMU motion data 612 may be vectorized and each vector labeled with a respective one of the labels 628. Each vector and corresponding label may be used as a training example for training a statistical learning model of AF factor analyzer 504. In some arrangements, the model is trained as a Support Vector Machine (SVM), an artificial neural network, or a statistical learning model.
Once the model is trained, the parameters (or weights) of each feature of the statistical learning model of AF factor analyzer 504 represent the effect and are commensurate with the degree to which the feature affects the accuracy of the decision of AF detector 120 in detecting AF. Thus, the statistical learning model of the AF factor analyzer 504 responds to factors that affect the accuracy of the AF detector 120, but is largely unaffected by factors that do not affect the accuracy of the AF detector 120. For example, in this context, the statistical learning model of the AF factor analyzer 504 does not respond to the intensity of motion artifacts, but rather to the type of motion artifacts that may affect the AF detection accuracy of the AF detector 120. Thus, the decision mechanism ensures that only accuracy related factors may affect the sampling rate set by the AF sampling controller 104.
The statistical learning model of the AF factor analyzer 504 learns to determine the type of AF factor (e.g., motion artifact). The type of AF factor indicates a designation (classification) or classification among a plurality of types (multi-category classification), wherein the statistical learning model learns to classify the AF factor based on the expected effect of the AF factor on the AF detector 120 according to the designation or classification. For example, for motion artifacts, conventional techniques tend to rely on the amplitude or power of signals generated by IMU sensors or accelerometers as a representation of the motion intensity. In contrast, the statistical learning model learning of the AF factor analyzer 504 classifies AF factors affecting the signal input to the AF detector 120. For example, motion artifacts generated by a user running quickly or sitting in a strongly vibrating vehicle are classified by a statistical learning model as high intensity type motion outside the normal range used by the AF detector 120 based on different frequency characteristics of the motion. By identifying the type of AF factor (schematically, motion artifact), the system 100 may effectively initiate signal processing (e.g., filtering) to remove motion artifact from the signal input to the AF detector 120. In case of motion artifact removal, energy is saved by sampling the signal using a lower sampling rate. Otherwise, a higher sampling rate will be enforced, resulting in unnecessary energy consumption. Then, in balance, the strategy of determining the type of AF factor (e.g., motion artifact) enables the AF sampling controller 104 to set the sampling rate accordingly, thereby improving the overall energy efficiency of the electronic device.
Fig. 7 schematically illustrates an example signal classification 700 made by the AF factor analyzer 504. Illustratively, IMU features 702 are extracted from signals 704 by feature extractor 502, and feature extractor 502 vectorizes the extracted IMU features for input to AF factor analyzer 504. Graph 706 graphically represents classification categories, category 0, category 1, and category 2, separated by decision boundaries 708 and 710, respectively. For illustration purposes only, the vector generated by feature extractor 502 and corresponding to the extracted IMU feature is a 2-tuple, < Acc-x, acc-y >. Each vector corresponds to a signal segment and is classified according to the position of the vector on the graph (in category 0, category 1, or category 2). The classification of the signal determines the sampling rate set by the sampling strategy selector 506 of the AF sampling controller 104.
The statistical learning model used by the AF factor analyzer 504 is pre-trained, and has been trained prior to deployment of the system 100 in an electronic device, such as a smart watch, earplug, smart phone, or other such device. The classification performed by the AF factor analyzer 504 using the pre-trained model is lightweight and can be performed in real-time. Thus, at the beginning of a scheduled AF monitoring period, classification may be made at the beginning of the monitoring, and sampling strategy selector 506 sets the sampling rate for the duration of the monitoring period.
The signal sampled at the low sampling rate may be recovered by a signal reconstructor 508 and transferred to the AF detector 120. In one embodiment, the signal reconstructor 508 performs spline interpolation to reconstruct the original signal (e.g., the PPG sensor signal). Cubic spline interpolation may mitigate the likelihood of overfitting that may occur with higher order interpolation. Furthermore, cubic spline interpolation achieves high accuracy with less computational overhead than higher order interpolation.
In another embodiment, the signal reconstructor 508 reconstructs the original signal by compressed sensing reconstruction. Using compressed sensing reconstruction, the original signal can theoretically be recovered with fewer samples than specified according to the nyquist theorem. The sampling process using compressive sensing can be formulated as:
[ formula 5 ]
y m×1 ＝Φ m×n x n×1 。
The mx1 vector corresponds to a signal (compressed signal) generated from compressed sensing. The vector y has a low dimension (m < n). Φ is an mxn sensing matrix. The n x 1 vector x is the original signal.
From y m×1 Reconstructing x n×1 X is required to be sufficiently sparse so that
[ formula 6 ]
y m×1 ＝Φ m×n Ψ n×p s p×1 。
Wherein phi is m×n Is a measurement matrix, ψ n×p Is a sparse basis and s p×1 Is a sparse vector (multiple elements are zero). The basis typically includes n independent vectors (normalized) that can be combined to express each vector of an n-dimensional vector space. If the sparse base ψ can be found n×p So that x is n×1 Sufficiently sparse, s can be determined based on the following optimizations p×1 ：
[ formula 7 ]
The reconstructed signal (e.g., PPG sensor signal) is then reconstructed according to:
[ math figure 8 ]
The reconstruction signal is sufficiently sparse by learning the sparse basis ψ (a set of supervised learning training examples) using a dictionary training algorithm to overcome some of the obstacles in determining the sparse basis ψ. The dictionary found by the dictionary training algorithm is the framework of training data to provide sparse representation ("dictionary"). The dictionary may be obtained by solving an optimization problem.
[ formula 9 ]
Where λ is the regularization parameter, K is the size of the training example set, and r i Sparse coefficient x, which is the ith training example i . The optimization problem may be solved iteratively. Sparse coding fixes the basis ψ and determines the sparse vectorUpdating the dictionary by r as learned from the previous step i Update the base ψ.
Fig. 8 schematically illustrates an example compressed sensing arrangement 800 that effectively combines the AF sampling controller 104 and the AF detector 120. During the offline learning phase, one or more sensors 114 (e.g., PPG sensors) perform sensing according to the nyquist theorem. According to equation 9 above, the sparse basis 802 is obtained by dictionary learning and passed to the signal reconstructor 508 for performing signal reconstruction in an online phase. The offline phase ends with CS reconstruction and training of model 804 according to equations 7 and 8 above. In an online phase of the system 100 implemented in an electronic device (e.g., smart watch, earplug), one or more sensors 114 (e.g., PPG sensors) perform transmitting the sensed signals as one or more compression sensors to the signal reconstructor 508. The reconstructed signal is transmitted to an AF detector 120, which AF detector 120 determines whether the signal indicates that the user is experiencing AF based on a classification of the signal according to a pre-trained model. The AF detector 120 makes AF or non-AF decisions in real time.
Thus, the sampling rate with compressive sensing may be lower than the nyquist sampling rate, which further improves the energy efficiency of the electronic device. An additional advantage is that the AF detector 120 may use an AF detection algorithm that is substantially dependent on peaks in the sensed signal (e.g., generated by the PPG sensor), and that the information lost with the recovered signal may be substantially well tolerated without significant loss of AF detection sensitivity.
Fig. 9 illustrates an example method 900 of energy efficient detection of atrial fibrillation using electronics. Method 900 may be performed by an electronic device comprising a system such as system 100 described herein. The electronic device and the system integrated therein or operatively coupled thereto are collectively referred to as a "system" for performing the method 900.
At block 902, the system schedules an AF monitoring period during which the system intermittently monitors the user's AF. Scheduling may be based on the system determining a user-specific AF risk.
At block 904, during an AF monitoring period, the system samples one or more signals corresponding to the user. The system may perform sampling using one or more sensors operatively coupled with or integrated in the system.
At block 906, the system adjusts the time interval between successive AF monitoring periods. The system may adjust the time interval in response to detecting a change in the risk of AF. In response to detecting the AF, the system may switch from intermittently monitoring the AF of the user to continuously monitoring the AF of the user.
In some arrangements, scheduling may include generating a data structure of the accumulated NSR detected by the electronic device. The system may change the time interval between AF monitoring periods in response to detecting NSR. The system may change the time based on the AF risk and the number of accumulated NSRs detected. The system may generate a data structure of URs and, in response to detecting URs, change the time interval between AF monitoring periods, the change being based on AF risk and the number of URs accumulated.
In some arrangements, the system determines one or more dynamically evolving risk factors based on at least one of physiological data, medical data, or lifestyle data corresponding to the user. The system may determine the risk of AF based on one or more dynamically evolving risk factors.
In other arrangements, the system may modify the schedule AF monitoring period in response to determining a time-based reduction in sensitivity of the system to detect AF. The system may detect the time-based reduction based on a tracking error between a model estimate of the heart rhythm of the user and an observed heart rhythm sensed by the system.
In other arrangements, the system may dynamically adjust the sampling rate of at least one of the one or more sensors used to monitor the AF of the user. The system may adjust the sampling rate based on the probability that the system correctly detects AF. The system may determine the probability using a statistical learning model that is trained to predict whether the system properly detects AF based on a predetermined set of AF-related factors. The predetermined set of AF-related factors may include the motion detected by the sensor, the heart rate, and/or the duration of the AF monitoring period.
Fig. 10 illustrates an example apparatus 1000 in which the system 100 may be implemented. The apparatus 1000 includes one or more processors 1002 coupled to a memory 1004 through interface circuitry 1006. The apparatus 1000 stores computer readable instructions (also referred to as "program code") within a memory 1004, the memory 1004 being an example of a computer readable storage medium. The one or more processors 1002 execute program code that is accessed from memory 1004 via interface circuitry 1006.
For example, the memory 1004 may include one or more physical memory devices, such as local memory 1008 and mass storage 1010. The local memory 1008 is implemented as one or more non-persistent memory devices, which are typically used during actual execution of program code. The local memory 1008 is an example of runtime memory. Examples of local memory 1008 include any of various types of RAM suitable for execution of program code by the processor. Mass storage this 1010 is implemented as a permanent data storage device. Examples of mass storage such 1010 include a Hard Disk Drive (HDD), a Solid State Drive (SSD), flash memory, read Only Memory (ROM), erasable Programmable Read Only Memory (EPROM), electrically Erasable Programmable Read Only Memory (EEPROM), or other suitable memory. The device 1000 may also include one or more cache memories (not shown) that provide temporary storage of at least some program code in order to reduce the number of times program code must be retrieved from mass storage during execution.
Examples of interface circuitry 1006 include, but are not limited to, input/output (I/O) subsystems, I/O interfaces, bus systems, and memory interfaces. For example, the interface circuit 1006 may be implemented as any of a variety of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus, and/or a combination of bus structures.
In one or more example implementations, the processor 1002, the memory 1004, and/or the interface circuit 1006 are implemented as separate components. The processor 1002, memory 1004, and/or interface circuitry 1006 may be integrated in one or more integrated circuits. For example, various components in the device 1000 may be coupled by one or more communication buses or signal lines (e.g., interconnect lines and/or wires). The memory 1004 may be coupled to the interface circuit 1006 via a memory interface, such as a memory controller or other memory interface (not shown).
The device 1000 may include one or more displays. Illustratively, for example, the device 1000 includes a display 1012 (e.g., screen). The display 1012 may be implemented as a touch-sensitive or touch screen display capable of receiving touch input from a user. The touch sensitive display and/or touch sensitive panel can detect contacts, movements, gestures, and contact discontinuities using any of a variety of available touch sensitive technologies. Example touch sensitive technologies include, but are not limited to, capacitive, resistive, infrared, and surface acoustic wave technologies, as well as other proximity sensor arrays or other elements for determining one or more points of contact with a touch sensitive display and/or device.
The device 1000 may include a camera subsystem 1014. The camera subsystem 1014 may be coupled to the interface circuit 1006 directly or through a suitable input/output (I/O) controller. The camera subsystem 1014 may be coupled to an optical sensor 1016. The optical sensor 1016 may be implemented using any of a variety of technologies. Examples of optical sensor 1016 may include, but are not limited to, a Charge Coupled Device (CCD) or a Complementary Metal Oxide Semiconductor (CMOS) optical sensor. The optical sensor 1016 may be, for example, a depth sensor. The camera subsystem 1014 and the optical sensor 1016 can perform camera functions such as recording or capturing images and/or recording video.
The device 1000 may include an audio subsystem 1018. The audio subsystem 1018 may be coupled to the interface circuit 1006 directly or through a suitable input/output (I/O) controller. The audio subsystem 1018 can be coupled to a speaker 1020 and a microphone 1022 to facilitate voice-enabled functions, such as voice recognition, voice replication, digital recording, and telephony functions.
The device 1000 may include one or more communication subsystems 1024, each communication subsystem 1024 may be coupled to the interface circuit 1006 directly or through a suitable I/O controller (not shown). Each of the one or more communication subsystems 1024 can facilitate communication functions. For example, communication subsystem 1024 may include one or more wireless communication subsystems such as, but not limited to, radio frequency receivers and transmitters and optical (e.g., infrared) receivers and transmitters. The specific design and implementation of the communication subsystem 1024 may depend on the particular type of device 1000 implemented and/or the communication network on which the device 1000 is intended to operate.
As an illustrative and non-limiting example of a wireless communication system, one or more communication subsystems 1024 can be designed to operate on one or more mobile networks, wiFi networks, short-range wireless networks (e.g., bluetooth), and/or any combination of the foregoing. The communications subsystem 1024 may implement a hosting protocol such that the device 1000 may be configured as a base station for other devices.
The apparatus 1000 may include one or more sensors 1026 of various types, each of the one or more sensors 1026 may be coupled to the interface circuit 1006 directly or through a suitable I/O controller (not shown). The one or more sensors 1026 may include sensors particularly adapted to detect and/or measure physiological properties, such as a user's heart rate. For example, the one or more sensors 1026 may include a PPG sensor. PPG sensors use a light source and a photodetector to measure the volume change of the blood circulation of a user. Thus, if the device 1000 is, for example, an ear bud with a PPG sensor integrated therein, the PPG sensor can estimate the skin blood flow of the user by emitting infrared light in the user's ear canal and detecting the reflected infrared light. In other embodiments, device 1000 may be another type of wearable device (e.g., a smart watch) with a PPG sensor, or may be a device such as a smart phone with a PPG sensor. PPG sensors can measure heart rate, blood pressure, oxygen saturation, and other physiological attributes. The one or more sensors 1026 may include an IMU that detects movement of the user. The device 1000 may be a smart watch, an earplug, or other wearable device with an integrated IMU. In other embodiments, the device 1000 may be a smart phone or other such device that is integrated with an IMU.
Other examples of one or more sensors 1026 that may be included in the device 1000 include, but are not limited to, proximity sensors that facilitate orientation, illumination, and proximity functions of the device 1000, respectively. Other examples of sensors 1026 may include, but are not limited to, a position sensor (e.g., a GPS receiver and/or processor) capable of providing geolocation sensor data, an electronic magnetometer (e.g., an integrated circuit chip) capable of providing sensor data that may be used to determine the direction of magnetic north for directional navigation, an accelerometer capable of providing data indicative of changes in speed and direction of movement of device 1000 in 3D, and an altimeter (e.g., an integrated circuit) capable of providing data indicative of altitude.
The device 1000 may also include one or more input/output (I/O) devices 1028 coupled to the interface circuit 1006. One or more I/O devices 1028 may be coupled to interface circuit 1006 directly or through an intermediate I/O controller (not shown). Examples of I/O devices 1028 include, but are not limited to, a track pad, a keyboard, a display device, a pointing device, one or more communication ports (e.g., universal Serial Bus (USB) ports), a network adapter, and buttons or other physical controls. Network adapter refers to circuitry that enables device 1000 to be coupled to other systems, computer systems, remote printers, and/or remote storage devices through intervening private or public networks. Modems, cable modems, ethernet interfaces, and wireless transceivers that are not part of wireless communication subsystem 1024 are examples of the different types of network adapters that may be used by device 1000. One or more of the I/O devices 1028 may be adapted to control the function of one or more or all of the sensors 1026 and/or one or more of the wireless communication subsystems 1024.
The memory 1004 stores program codes. Examples of program code include, but are not limited to, routines, programs, objects, components, logic, and other data structures. For purposes of illustration, memory 1004 stores an operating system 1030 and applications 1032. In addition, the memory 1004 may store energy saving AF monitoring and detection program code 1034 for implementing a system, such as system 100.
The apparatus 1000 is provided for purposes of illustration and not limitation. Devices and/or systems configured to perform the operations described herein may have architectures different than those shown in fig. 10. The architecture may be a simplified version of the architecture described in connection with fig. 10, including a memory capable of storing instructions and a processor capable of executing instructions. In this regard, the apparatus 1000 may include fewer components than are shown in fig. 10 or additional components not shown in fig. 10 depending on the particular type of apparatus implemented. Furthermore, the particular operating system and/or applications included may vary depending on the type of device, such as the type of I/O device that may be included. Further, one or more of the illustrative components may be incorporated into or otherwise form a portion of another component. For example, the processor may include at least some memory.
The apparatus 1000 may be implemented as a data processing system, a communication device, or other suitable system adapted to store and/or execute program code. The apparatus 1000 may be implemented as an edge apparatus. Example implementations of the apparatus 1000 may include, but are not limited to, a computing apparatus. Computing devices include, for example, computers (e.g., desktop computers, laptop computers, tablet computers), televisions, entertainment consoles, XR systems, or other devices capable of operating cooperatively as a display device (e.g., HMD, AR glasses) or in conjunction with a source device (e.g., smart phone, console, computer) that operates in conjunction with an electronic display device, as described herein.
The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. Nonetheless, several definitions will now be presented throughout this application.
As defined herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
The term "about" means nearly correct or precise, with values or amounts approaching but not being precise. For example, the term "about" may mean that a listed characteristic, parameter, or value is within a predetermined amount of the exact characteristic, parameter, or value.
As defined herein, unless specifically stated otherwise, the terms "at least one of … …," "one or more of … …," and/or "are open-ended expressions that are both connected and separated in operation. For example, each of the expressions "at least one of A, B and C", "at least one of A, B or C", "one or more of A, B and C", "one or more of A, B or C" and "A, B and/or C" means a alone, B alone, C, A and B together, a and C together, B and C together, or A, B and C together.
As defined herein, the term "automatically" means without human intervention.
The term "computer-readable storage medium" as defined herein refers to a storage medium that contains or stores program code for use by or in connection with an instruction execution system, apparatus, or device. As defined herein, a "computer-readable storage medium" is not a transitory propagating signal per se. The computer readable storage medium may be, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing storage devices. As described herein, different types of memory are examples of computer-readable storage media. A non-exhaustive list of more specific examples of the computer readable storage medium would include: portable computer magnetic disks, hard disks, random Access Memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), static Random Access Memories (SRAM), portable compact disc read-only memories (CD-ROM), digital Versatile Discs (DVD), memory sticks, floppy disks, and the like.
As defined herein, the term "if" means "when … …" or "at … …" or "responsive" or "response", depending on the context. Thus, depending on the context, the phrase "if determined … …" or "if detected (a stated condition or event)" may be interpreted to mean "upon determining … …" or "in response to determining … …" or "upon detecting (a stated condition or event)" or "in response to detecting (a stated condition or event)".
As defined herein, the term "processor" means at least one hardware circuit. The hardware circuitry may be configured to execute instructions contained in the program code. The hardware circuit may be an integrated circuit. Examples of processors include, but are not limited to, central Processing Units (CPUs), array processors, vector processors, digital Signal Processors (DSPs), field Programmable Gate Arrays (FPGAs), programmable Logic Arrays (PLAs), application Specific Integrated Circuits (ASICs), programmable logic circuits, and controllers.
As defined herein, the term "respond" and similar language as described above (e.g., "if … …," "when … …," or "at … …") means to respond or react quickly to an action or event. The response or reaction is performed automatically. Thus, if a second action is performed "in response to" a first action, there is a causal relationship between the occurrence of the first action and the occurrence of the second action. The term "response" indicates causal relationships.
As defined herein, "real-time" means the level of processing responsiveness of a user or system to a particular process to be performed or to determine a sufficiently immediate sensing or to enable a processor to keep up with a certain external process.
The term "substantially" means that the listed characteristics, parameters or values need not be exactly achieved, but may deviate or vary by an amount that does not preclude the effect that the characteristics are intended to provide, including, for example, tolerances, measurement errors, measurement accuracy limitations and other factors known to those of skill in the art.
The terms "user" and "individual" refer to humans.
The terms first, second, etc. may be used herein to describe various elements. When these terms are only used to distinguish one element from another, the element should not be limited by the terms unless otherwise specified or the context clearly indicates otherwise.
The computer program product may include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to perform aspects of the present invention. Within this disclosure, the term "program code" may be used interchangeably with the term "computer-readable program instructions". The computer readable program instructions described herein may be downloaded from a computer readable storage medium to a corresponding computing/processing device or to an external computer or external storage device via a network (e.g., the internet, LAN, WAN, and/or wireless network). The network may include copper transmission cables, optical transmission fibers, wireless transmissions, routers, firewalls, switches, gateway computers and/or edge devices including edge servers. The network adapter card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing/processing device.
Computer readable program instructions for performing operations described herein for the inventive arrangements can be assembly instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, or source code or object code written in any combination of one or more programming languages, including an object-oriented programming language and/or a procedural programming language. The computer readable program instructions may specify state setting data. The computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network (including a LAN or a WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider). In some cases, electronic circuitry, including, for example, programmable logic circuitry, FPGA, or PLA, can personalize the electronic circuitry by utilizing state information of the computer readable program instructions to execute the computer readable program instructions in order to perform aspects of the inventive arrangements described herein.
Certain aspects of the inventive arrangements are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer-readable program instructions (e.g., program code).
These computer readable program instructions may be provided to a processor of a computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. In this manner, operatively coupling a processor to program code instructions converts the processor's machine into a special-purpose machine for executing the instructions of the program code. These computer-readable program instructions may also be stored in a computer-readable storage medium that can direct a computer, programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer-readable storage medium having the instructions stored therein includes an article of manufacture including instructions which implement the aspects of the operations specified in the flowchart and/or block diagram block or blocks.
The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operations to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer, other programmable apparatus or other devices implement the functions/acts specified in the flowchart and/or block diagram block or blocks.
The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various aspects of arrangements of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified operations. In some alternative implementations, the operations noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may be executed substantially concurrently or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The corresponding structures, materials, acts, and equivalents of all means or step plus function elements in the claims below are intended to include any structure, material, or act for performing the function in combination with other claimed elements as specifically claimed.
The description of the embodiments provided herein is for purposes of illustration and is not intended to be exhaustive or limited to the disclosed forms and examples. The terminology used herein was chosen to explain the principles of the inventive arrangements, the practical application, or technical improvements in the technology found in the marketplace, and/or to enable others of ordinary skill in the art to understand the embodiments disclosed herein. Modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described inventive arrangements. Accordingly, reference should be made to the appended claims, rather than to the foregoing disclosure, as indicating the scope of such features and embodiments.

Claims (15)

1. A method, comprising:
scheduling, by an electronic device, an atrial fibrillation AF monitoring period, wherein during the AF monitoring period the electronic device intermittently monitors AF of a user of the electronic device, wherein the scheduling is based on determining a risk of AF specific to the user; and
The time interval between successive AF monitoring periods is adjusted by the electronic device in response to detecting the change in AF risk.
2. The method of claim 1, further comprising:
in response to detecting AF, switching from intermittently monitoring the AF of the user to continuously monitoring the AF of the user.
3. The method of claim 1, wherein
The scheduling includes:
generating a data structure of the accumulated normal sinus rhythm NSR detected by the electronic device; and
in response to detecting an NSR, a time interval between AF monitoring periods is changed, wherein the change is based on the AF risk and the number of NSRs detected that are accumulated.
4. The method of claim 1, wherein
The scheduling includes:
generating a data structure of the undetermined readings UR accumulated by the electronic device; and
in response to detecting a UR, changing a time interval between AF monitoring periods, wherein the changing is based on the AF risk and the number of accumulated URs.
5. The method of claim 1, wherein
Determining the AF risk includes: a dynamically evolving risk factor is determined based on at least one of physiological data, medical data, or lifestyle data corresponding to the user.
6. The method of claim 1, further comprising:
the scheduling of the AF monitoring period is modified in response to determining a time-based reduction in sensitivity of the electronic device to detect AF, wherein the time-based reduction is detected based on a tracking error between a model estimate of the heart rhythm of the user and an observed heart rhythm sensed by the electronic device.
7. The method of claim 1, further comprising:
during the AF monitoring period, sampling one or more signals corresponding to the user using one or more sensors operatively coupled to the electronic device; and
dynamically adjusting a sampling rate of at least one of the one or more sensors for monitoring AF of the user, wherein the sampling rate is based on a probability that the electronic device correctly detects AF.
8. The method of claim 7, wherein
The probability is determined using a statistical learning model trained to predict whether the electronic device properly detects AF based on a predetermined set of AF-related factors, wherein the set of AF-related factors includes at least one of: motion detected by a sensor, heart rate, or duration of the AF monitoring period.
9. A system, comprising:
one or more sensors; and
a processor operably coupled with the one or more sensors, wherein the processor is configured to initiate operations comprising:
scheduling an atrial fibrillation AF monitoring period, wherein during the AF monitoring period the one or more sensors intermittently monitor AF of a user, wherein the scheduling is based on determining a risk of AF specific to the user; and
in response to detecting the change in AF risk, the time interval between successive AF monitoring periods is adjusted.
10. The system of claim 9, wherein the processor is configured to initiate operations further comprising:
in response to detecting AF, switching from intermittently monitoring the AF of the user to continuously monitoring the AF of the user.
11. The system of claim 9, wherein
The scheduling operation includes:
generating a data structure of the accumulated normal sinus rhythm NSR detected by the one or more sensors; and
in response to detecting an NSR, a time interval between AF monitoring periods is changed, wherein the change is based on the AF risk and the number of NSRs detected that are accumulated.
12. The system of claim 9, wherein
The scheduling includes:
generating a data structure of the uncertainty readings UR accumulated by the one or more sensors; and
in response to detecting a UR, changing a time interval between AF monitoring periods, wherein the changing operation is based on the AF risk and the accumulated number of URs.
13. The system of claim 9, wherein
Determining the risk of AF includes determining a dynamically evolving risk factor based on at least one of physiological data, medical data, or lifestyle data corresponding to the user.
14. The system of claim 9, wherein the processor is configured to initiate operations further comprising:
the scheduling of the AF monitoring period is modified in response to determining a time-based reduction in sensitivity of the one or more sensors to detect AF, wherein the time-based reduction is detected based on a tracking error between a model estimate of the heart rhythm of the user and an observed heart rhythm sensed by the one or more sensors.
15. A computer program product, the computer program product comprising:
one or more computer-readable storage media; and
Program instructions collectively stored on the one or more computer-readable storage media, the program instructions being executable by a processor of an electronic device to cause the processor to initiate operations comprising:
scheduling an atrial fibrillation AF monitoring period, wherein during the AF monitoring period, one or more sensors of the electronic device intermittently monitor AF of a user of the electronic device, wherein the scheduling is based on determining a risk of AF specific to the user; and
in response to detecting the change in AF risk, the time interval between successive AF monitoring periods is adjusted.
CN202280019778.XA 2021-04-30 2022-03-11 Energy efficient detection and management of atrial fibrillation Pending CN117015342A (en)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US63/182,155 2021-04-30
US17/564,198 US20220369992A1 (en) 2021-04-30 2021-12-28 Energy efficient detection and management of atrial fibrillation
US17/564,198 2021-12-28
PCT/KR2022/003428 WO2022231121A1 (en) 2021-04-30 2022-03-11 Energy efficient detection and management of atrial fibrillation

Publications (1)

Publication Number Publication Date
CN117015342A true CN117015342A (en) 2023-11-07

Family

ID=88573116

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202280019778.XA Pending CN117015342A (en) 2021-04-30 2022-03-11 Energy efficient detection and management of atrial fibrillation

Country Status (1)

Country Link
CN (1) CN117015342A (en)

Similar Documents

Publication Publication Date Title
CN107773214B (en) Method, computer readable medium and system for optimal wake-up strategy
US20200303046A1 (en) Wi-fi-based condition monitoring
US20180008191A1 (en) Pain management wearable device
EP3367883A1 (en) Monitoring activities of daily living of a person
JP2020536623A (en) Continuous monitoring of user health using mobile devices
US11580421B2 (en) Weakly supervised learning for improving multimodal sensing platform
US20220110546A1 (en) System and methods for tracking behavior and detecting abnormalities
CA3055187A1 (en) Medical adverse event prediction, reporting, and prevention
Vildjiounaite et al. Unsupervised stress detection algorithm and experiments with real life data
US20230081657A1 (en) System and method for determining and predicting of a misstep
US20230263400A1 (en) System and method for filtering time-varying data for physiological signal prediction
JP2024513618A (en) Methods and systems for personalized prediction of infections and sepsis
EP4025120A1 (en) Systems and methods for detecting sleep activity
US20220031250A1 (en) Toothbrush-derived digital phenotypes for understanding and modulating behaviors and health
CN117015342A (en) Energy efficient detection and management of atrial fibrillation
US20220369992A1 (en) Energy efficient detection and management of atrial fibrillation
WO2019115432A1 (en) Determining whether a hypothesis concerning a signal is true
JP7211441B2 (en) Arousal level estimation device, arousal level estimation method, and program
WO2021127566A1 (en) Devices and methods for measuring physiological parameters
US10380882B1 (en) Reconfigurable hardware platform for processing of classifier outputs
KR102551856B1 (en) Electronic device for predicting emotional state of protected person using walking support device based on deep learning based prediction model and method for operation thereof
US20230109079A1 (en) Data collection system, data collection method. and data collection device
US20220296133A1 (en) Methods Circuits Devices Systems and Machine Executable Code for Glucose Event Detection
US20230108162A1 (en) Data collection system, data collection device, data acquisition device, and data collection method
Pijl Tracking of human motion over time

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination