Method, device and system for evaluating reliability of estimation indicator relating to physiological state of user

By establishing biomechanical and physiological boundaries between physiological indicators, cross-checking and evaluating their reliability, the problem of large errors in the prior art is solved, and the accuracy of physiological state assessment and the reliability of health monitoring are improved.

CN118891004APending Publication Date: 2024-11-01KONINKLIJKE PHILIPS NV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202380028499.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2022-03-22
Filing Date
2023-03-15
Publication Date
2024-11-01

AI Technical Summary

Technical Problem

There are errors in existing algorithms for estimating physiological and anatomical indicators, resulting in incorrect assessment of patient status, affecting the accuracy of clinical decision-making and health monitoring.

Method used

By cross-checking based on biomechanical and physiological boundaries between different physiological indicators, the indicators are independently estimated and their reliability is evaluated using multiple methods, and reliability ratings are output to improve the accuracy of the estimation results.

Benefits of technology

It effectively reduces estimation errors, improves the reliability of physiological status assessment, and ensures the accuracy of clinical decision-making and the accuracy of health monitoring.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118891004A_ABST
    Figure CN118891004A_ABST
Patent Text Reader

Abstract

The invention relates to a method, a device and a system for evaluating the reliability of an estimation indicator related to a physiological state of a user. The method comprises: obtaining one or more sensor signals related to the physiological state of the user; estimating a first indicator and a second indicator independently of each other using different methods and / or algorithms based on the one or more sensor signals, where the first indicator is associated with the second indicator by a first mathematical function; determining a first value of the first mathematical function; comparing the first value to a boundary of the first value; evaluating a reliability rating for the estimation of the first indicator and / or the second indicator based on the comparison; and outputting the evaluated reliability rating.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a method, a device and a system for evaluating the reliability of an estimated indicator related to a user's physiological state. Background Art

[0002] In order to improve patient care, reduce hospital stays and improve patient outcomes, continuous, covert and remote monitoring of activity and vital signs is becoming increasingly popular. Ambulatory speed (i.e., walking speed) is considered the sixth vital sign and can be used to assess the patient's physical performance and determine frailty and associated health risks in clinical practice. Activity level (e.g., in the form of step counting) can assess the patient's total level of movement in the hospital and help contextualize the physiological response to activity. All of these indicators become particularly relevant in ambulatory patient monitoring devices to enhance the hospital's decision-making process (e.g., decision-making processes related to discharge preparation); to date, in most cases, discharge preparation relies on nurses' subjective input of activity and patient status.

[0003] There are several algorithms for estimating physiological and anatomical indices, especially ambulatory indices. Most of these algorithms use features of acceleration and cardiovascular signals (e.g., obtained by on-body accelerometers, photoplethysmography (PPG), and electrocardiogram (ECG)) and data analysis techniques (e.g., linear regression and machine learning) to estimate indices (e.g., walking speed, step count, and walking economy).

[0004] US2013 / 0138007A1 describes a technology for measuring a user's heart rate. The disclosed heartbeat measurement device includes: an information storage unit that stores specific information about a user's specific heart rate or information about a heartbeat detected in the user; and a determination unit that determines whether the user's heartbeat measurement state is normal based on the specific information or information about the heartbeat stored in the information storage unit.

[0005] Other methods for estimating physiological or anatomical indicators are described in US 2015 / 0099945 A1 and US 2013 / 0080255 A1.

[0006] For each algorithm, the estimated value may deviate from the actual value due to unforeseen data or inherent limitations of the features and data analysis techniques used. In fact, the estimated value of the walking index is always characterized by a certain amount of error. However, it may not always be possible to improve the estimation results of the mentioned estimation methods, and such improvement does not guarantee that estimation errors will not occur on new data. Unfortunately, these estimation errors may lead to an erroneous assessment of the patient's state. In addition to the aforementioned serious consequences in clinical decision-making, an erroneous estimation of the user's physiological or anatomical indicators may lead to an erroneous assessment of the user's physical fitness level, resulting in an inappropriate training plan. In addition, such a misestimation may lead to an inappropriate or even harmful diet plan.

[0007] Therefore, accurate estimation of physiological and anatomical indices (particularly ambulatory indices) is important for assessing the patient's state or the health of a person with high (or at least known) reliability. Therefore, there is a need to minimize these errors. In particular, assessing reliability becomes crucial to guarantee the quality of the estimated physiological indices. Summary of the invention

[0008] The object of the present invention is to improve the reliability of an estimated indicator related to a user's physiological state, thereby more accurately assessing the user's health state. In particular, the object of the present invention is to improve the reliability of an estimated indicator related to a user's walking ability.

[0009] In a first aspect of the present invention, a method for evaluating the reliability of an estimation indicator related to a user's physiological state is proposed, the method comprising:

[0010] obtaining one or more sensor signals related to the physiological state of the user;

[0011] estimating a first indicator and a second indicator independently of each other using different methods and / or algorithms based on the one or more sensor signals, wherein the first indicator is related to the second indicator by a first mathematical function;

[0012] determining a first value of the first mathematical function;

[0013] comparing the first value to a boundary of the first value;

[0014] evaluating a reliability rating of an estimation result for the first indicator and / or the second indicator based on the comparison; and

[0015] Output the assessed reliability rating.

[0016] In a further aspect of the present invention, a device for evaluating the reliability of an estimated indicator related to a user's physiological state is proposed, the device comprising:

[0017] a sensor input configured to obtain one or more sensor signals related to the physiological state of the user;

[0018] A processor configured to:

[0019] estimating a first indicator and a second indicator independently of each other using different methods and / or algorithms based on the one or more sensor signals, wherein the first indicator is related to the second indicator by a first mathematical function;

[0020] determining a first value of the first mathematical function;

[0021] comparing the first value to a boundary of the first value; and

[0022] evaluating a reliability rating of an estimation result for the first indicator and / or the second indicator based on the comparison; and

[0023] An information output unit is configured to output the evaluated reliability rating.

[0024] In another aspect of the present invention, a system for evaluating the reliability of an estimated indicator related to a user's physiological state is provided, the system comprising:

[0025] a sensor configured to measure one or more sensor signals related to the physiological state of the user;

[0026] The apparatus disclosed herein; and

[0027] A user interface is configured to publish the reliability rating of the device.

[0028] In yet another aspect of the present invention, a corresponding computer program is provided, which includes a program code unit. When the computer program is executed on a computer, the program code unit is used to enable the computer to perform the steps of the method disclosed in this article; a non-volatile computer-readable recording medium is also provided, in which a computer program product is stored, and when the computer program product is run by a processor, the method disclosed in this article is executed.

[0029] Preferred embodiments of the invention are defined in the dependent claims. It should be understood that the claimed devices, systems, computer programs and media have similar and / or identical preferred embodiments as the claimed methods, in particular as defined in the dependent claims and disclosed herein.

[0030] Conventional methods for estimating physiological or anatomical indices (particularly ambulatory indices) are not as accurate as required. In particular, estimates that differ greatly from the true value are often not accepted. This is the case for a variety of physiological and anatomical parameter estimates, from acceleration estimates to cardiac signal estimates.

[0031] The present invention proposes a method in which the reliability of two or more estimated indicators is estimated based on the biomechanical boundaries and physiological boundaries of the relationship of two or more estimated indicators related to each other. In other words, it is proposed to evaluate the reliability of the estimation results of physiological (or anatomical) indicators (especially ambulatory indicators) by cross-checking the two indicators based on the biomechanical boundaries, physiological boundaries or anatomical boundaries of the relationship of different indicators related to each other.

[0032] For many clinical decision-making processes, determining when an estimate is correct is crucial, since decisions related to the user's health should be based on reliable indicators. Moreover, only if the reliability of an indicator (i.e., the output of the estimation algorithm used, or in other words, the estimated result of an activity or vital parameter) is precisely known, can indicators that deviate significantly from the true value be rejected (i.e., singled out) and ignored from further consideration. Furthermore, knowing the reliability of an estimated indicator can help improve the algorithm to obtain more reliable results.

[0033] In the proposed method, user activities (and / or vital parameters) are monitored and at least two boundary-related physiological (or anatomical) indicators are calculated (or more precisely, estimated) based on the signals of the monitoring device. In order to estimate these indicators, one or more methods (in particular algorithms) can be used. Then, it is evaluated whether an indicator complies with the boundaries specified by (one or more) other indicators (and preferably, vice versa). In the case that (one or more) boundaries are not complied with, the (one or more) corresponding indicators can be marked / labeled in a specific way so that such (one or more) unreliable indicators can be ignored in further considerations. For example, an indicator that does not comply with the boundaries of another indicator can be labeled as "unreliable" and / or a specific reliability score can be provided for the indicator.

[0034] More precisely, the first indicator A can be estimated using method g, and the second indicator B can be estimated using method h, where g and h are methods (or algorithms) using one or more features of one or more sensor signals. The feature sets of functions g and h can be the same or different. In particular, some features of each feature set can be the same. However, these indicators can usually be estimated independently of each other. The one or more features mentioned may include, for example, any of the following: the periodicity of the acceleration signal, the standard deviation of the acceleration signal, the slope of the pressure signal (e.g., from a barometer), the average heart rate, the maximum heart rate, and the respiratory rate range.

[0035] The first indicator A and the second indicator B can be related to each other via a first mathematical function. For example, A and B can be related to each other via a third indicator f(A, B) such that f(A, B)=A / B. By inserting the indicator value into the function f(A, B), a first value C can be determined. Then, the first value C is compared with the bounds for the value (or more precisely, the bounds of the possible values ​​of the function), and based on the result of the comparison, it is evaluated whether the first indicator and / or the second indicator is reliable.

[0036] The proposed idea stems from the fact that the values ​​of different physiological indicators may delimit each other (i.e., one indicator may define the minimum and / or maximum of another indicator). Instead of defining the boundaries of each of these indicators based on the maximum and minimum values ​​of the first value C, it is possible to obtain an estimated first value C via f(A, B)=A / B and delimit it to obtain the reliability of indicators A and B.

[0037] For example, two walking indicators, walking speed and step frequency (i.e., step rhythm), are associated with each other based on the step length; walking speed [t] = measurement_step frequency [t] * measurement_step length. In addition, these two indicators are bounded by each other. These indicators can be extracted from one or more sensor signals (as examples of the first indicator and the second indicator, respectively). For example, the walking speed can be estimated based on a signal provided by an inertial sensor (e.g., an accelerometer or a gyroscope). At the same time, the user's step frequency can be estimated based on the same sensor signal or another sensor signal (using the same or other signal features). Since the first indicator of walking speed is associated with the second indicator of step frequency through a step length function, the value of the step length function (i.e., the first value) can be calculated based on the estimated results of these two indicators. Since the step length represents a bounded value (bounded above and below), the estimated (i.e., calculated) step length can be compared with its boundaries.

[0038] In general, the comparison of the first value with its boundary may include a direct comparison as well as an indirect comparison.

[0039] In a direct comparison, it may be checked whether the first value complies with its boundary, i.e., whether the first value is below, equal to, or above the boundary. The direct comparison may also include calculating the difference between the first value and the boundary. For example, in the above example, it may be checked whether the estimated step value is less than an upper limit for the step size. Similarly, the difference between the upper limit for the step size and the estimated step value may be calculated.

[0040] On the other hand, in an indirect comparison, the first value may not be compared directly to the boundary, but rather the first value may be compared indirectly by comparing it to a function of the boundary. For example, the function of the boundary may be a function of the difference between the boundary and the first value. In the above example, a function including the difference between the step size estimate and its boundary may be calculated. In addition, the evaluation result of the function may also be compared to the boundary.

[0041] The result of the comparison may then be used to evaluate the reliability rating of the estimation result regarding the first indicator and / or the second indicator. The reliability rating may be output directly or filtered before output. For example, reliability ratings below (or above) a predefined limit value may be classified as unreliable, and thus these reliability ratings may not be output. On the other hand, only reliability ratings below (or above) a predefined limit value may be output.

[0042] Typically, different methods / algorithms can be used to determine (the values ​​of) multiple first indicators and / or second indicators. Therefore, these indicators can be used to determine multiple first values ​​of the mathematical function, and these multiple first values ​​can be compared with the boundaries of the first values. Therefore, multiple reliability ratings of the estimation results for the first indicator and / or the second indicator can be evaluated based on the comparison. These multiple reliability ratings can then be analyzed, and methods for determining these indicators can be selected based on these reliability ratings. In particular, only those indicator estimation methods that have made the reliability rating of the corresponding indicator higher than a predetermined threshold will be used. Therefore, the reliability of the indicator estimation results can be further improved.

[0043] The main applications of the invention relate to health monitoring in hospitals and homes, with a particular focus on sleep and respiratory care, vital signs sensing and associated sensing.

[0044] In an embodiment of the method, comparing the first value to the boundary of the first value comprises determining whether the first value satisfies a relationship between the first mathematical function and the boundary.

[0045] The relationship between the first mathematical function and the boundary may include an equation or an inequality. Therefore, it is checked whether the first value satisfies the equation or inequality. For example, the first value may be A / B, where A is the first indicator and B is the second indicator, and it may be checked whether A / B < C, where C represents the boundary value. Typically, the relationship between the first mathematical function may include any one of the following: the difference between the first mathematical function and the boundary, the sum of the first mathematical function and the boundary, the fraction of the first mathematical function and the boundary, and the product of the first mathematical function and the boundary.

[0046] Boundaries may be time-dependent or user-dependent.User dependencies may include dependencies on any of the following: age, weight, height, gender, leg length, foot size, or any other physiological or anatomical characteristic of the user.

[0047] In an embodiment, the boundary is a lower limit and / or an upper limit.

[0048] For example, the first indicator may be walking speed, and the second indicator may be cadence. The two indicators are associated via stride length. The stride length value of the user is bounded by physiological / anatomical limitations, such as the leg length of the user (or a person in general). Therefore, the boundary of the first value may be an upper limit of the stride length. In a similar manner, a lower limit or another first value of the stride length value may be defined.

[0049] In another example, the first indicator may be the user's heart rate, and the second indicator may be the user's activity count (during a particular activity). The relationship between these indicators is highly dependent on the user's physical fitness level. Physical fitness can be represented by VO2max (i.e., the maximum oxygen consumption rate measured during incremental exercise), which is bounded. The lower and upper limits of VO2max can be determined based on the user's physical characteristics (e.g., weight, BMI, fat-free mass, and age).

[0050] As an example, the lower limit may be set to 0, and the upper limit may be set to (positive) infinity.

[0051] In another embodiment, comparing the first value to the boundary of the first value comprises determining whether the first value is above the lower limit and / or below the upper limit.

[0052] Likewise, it may be determined whether the first value is equal to or below a lower limit and / or whether the first value is equal to or above an upper limit. For example, the estimated step value may be compared to a lower limit, including determining whether the estimated step value is above the lower limit.

[0053] In another embodiment, the boundary of the method is predetermined, or the method further comprises: estimating the boundary based on a previous first indicator, a previous second indicator, a previous first value, or any one of the one or more sensor signals.

[0054] The boundaries (particularly the lower and / or upper limits) can be determined from the literature. For example, as known from the literature, during walking, a person's minimum step length may be 30 cm and the maximum step length may be 90 cm. However, during running, these boundaries may be different, as shown in the literature data.

[0055] However, the boundaries may also be determined empirically. In particular, the boundaries may depend on one or more signals obtained and / or an estimated indicator based on these signals. In fact, the lower step length limit and the upper step length limit may vary based on the accelerometer signal obtained or based on an indicator estimated therefrom. The boundaries may be inferred as statistical data based on the recorded data (e.g., for user A, the score 99% (time walking speed / cadence) < N, then N may be the upper limit). In fact, based on empirical knowledge, the fixed thresholds may be inferred by common sense or based on previously collected data.

[0056] Furthermore, the boundaries may depend on one or more predefined parameters.

[0057] In any case, in particular for the estimated (ie empirically determined) lower and upper bounds, possible estimation errors may be further taken into account. Thus, if the step size is mainly overestimated, the bounds on the step size (in particular the upper bound) can be lowered.

[0058] In yet another embodiment of the method, any of the first indicator, the second indicator and the first mathematical function may be time-dependent. For example, the relationship between the first indicator (i.e., the estimated walking speed) and the second indicator (i.e., the estimated step frequency) may change depending on the time of day or the time that has passed during walking.

[0059] Furthermore, any one of the first indicator, the second indicator and the first mathematical function may depend on a physiological parameter of the user.

[0060] In a further embodiment of the method, further comprising: determining a second value of a second mathematical function associating the first mathematical function with the boundary, wherein evaluating the reliability rating of the estimation result for the first indicator and / or the second indicator is based on the second value.

[0061] The second mathematical function may specifically include the difference between the bounds of the first value and the value of the first mathematical function (when the first indicator and the second indicator are inserted). For example, if the first mathematical function represents a step size, the second mathematical function may include the difference between the lower step size limit MIN_SL and the estimated step size value. Similarly, the second mathematical function may include the difference between the upper step size limit MAX_SL and the estimated step size value. For example, the second function may be:

[0062]

[0063] or

[0064]

[0065] It can be seen from this example that the second mathematical function can be time-dependent. In particular, the second mathematical function can be a function of the time-dependent boundary of the first value. However, the second mathematical function can also include the time-dependent first mathematical function. However, the second mathematical function can also be independent of time.

[0066] In another embodiment, evaluating the reliability rating comprises assigning a score on a binary scale or a continuous scale to the estimation results of the first indicator and / or the second indicator.

[0067] For example, the reliability rating of the indicator can be expressed as a binary score "reliable" and "unreliable". For example, in the case of two indicators, estimated speed and estimated cadence, a first function can be defined as the estimated step length. The first value of the step length can be compared with a minimum step length MIN_SL and a maximum step length MAX_SL. In other words, it can be checked whether the following relationship is satisfied:

[0068] MIN_SL<First value <MAX_SL

[0069] If the estimated step size is less than the minimum step size, and if the estimated step size is greater than the maximum step size, both estimation indicators may be assigned a score of "unreliable." In any other case, the estimation indicators may be assigned a score of "reliable."

[0070] On the other hand, if the first value is above the upper limit of the value, the reliability may be set to 0, otherwise to 1. Furthermore, if the first value is below the lower limit of the value, the reliability may be set to 0, otherwise to 1, wherein a reliability rating of 0 may correspond to an unreliable estimated first value, and a reliability rating of 1 may correspond to a reliable estimated first value.

[0071] As another example, two second functions may be determined, in particular a first and a second function representing the “next step reliability”

[0072]

[0073] And the second function representing the "upper step reliability"

[0074]

[0075] In the case where the first second function or the second second function is less than 0, the value of the first second function or the second second function may be set to 0, respectively. On the other hand, if the first second function or the second second function is higher than 1, the value of the first second function or the second second function may be set to 1, respectively. The reliability rating may be set to the smaller value of the first second function value and the second second function value, that is, the reliability rating may be set to:

[0076]

[0077] Therefore, the reliability rating can be expressed as a score ranging from 0 to 1.

[0078] In yet another (different) example, the reliability rating may be any score between 0 and 1 (or any other number), which score, for example, depends on the difference between the first value and a boundary. If the difference between the first value and the upper limit is equal to the difference between the first value and the lower limit (i.e., if the first value is in the middle between the lower limit and the upper limit), the reliability may be set to 1. On the other hand, the farther the first value is from the lower limit and the closer it is to the upper limit, the lower the score of the reliability rating may be, and vice versa, wherein if the first value is below the lower limit or above the upper limit, the score of the reliability rating is set to 0, which means that such a first value is unreliable.

[0079] In another alternative example, any first value between a lower limit and an upper limit may be considered reliable, and first values ​​outside of either boundary may be assigned a reliability rating that decreases with increasing distance from the boundary.

[0080] In another embodiment, the method further includes: adjusting or changing the method for estimating the first indicator and / or the second indicator based on the reliability rating. In other words, the reliability estimation result can be used to update the estimation result of the first indicator and / or the second indicator. Generally, more than one estimation method can be used to estimate one (or both) of these indicators. For example, method A can be used to estimate walking speed, while methods A and B can be used to estimate cadence, or even methods X and Y can be used for speed estimation, and methods A and B (and C, etc.) can be used to estimate cadence. In this case, the corresponding reliability rating can be used to select which predicted (ie, estimated) indicator (e.g., the cadence estimation result of algorithm A or B) or a combination of estimated indicators (e.g., the cadence estimation result of algorithm A and the speed of algorithm X) to use for further consideration, especially further reliability evaluation. In addition, based on the reliability rating, it can be determined which indicator can be directly output or used as further input for health assessment. This can improve the accuracy of the indicator while maintaining a high coverage.

[0081] As another example, algorithm A1 can be used to perform speed estimation, thereby obtaining an estimated speed a1; algorithm A2 can be used to perform speed estimation, thereby obtaining an estimated speed a2; algorithm A3 can be used to perform speed estimation, thereby obtaining an estimated speed a3; and so on. Assuming that these algorithms are different from each other, not all of these algorithms can perform equally well for various possible walking speeds or walking patterns to which the algorithms are to be applied. For example, for low speeds, algorithm A1 may be the most accurate algorithm, while for higher speeds, algorithm A2 may be more suitable, and for walking with a walker, etc., algorithm A3 may be the best choice. If the speed estimation result a1 is unreliable (based on the above method), then it is obviously not used to evaluate the user. Therefore, for speed estimation, algorithm A1 can be rejected (excluded). Excluding algorithm A1 and therefore excluding the estimation result a1 reduces the coverage of the speed estimation results (i.e., resulting in one less data point). However, this can be overcome by using other algorithms (A2, A3, ..., An) in parallel. Check the reliability of these methods, and use the method with the best reliability instead of algorithm A1.

[0082] In addition, after excluding the speed estimation result a1, the speed estimation results of other algorithms can be merged (for example, by averaging the speed estimation results). In fact, although the walking speed and other parameters representing the physiological characteristics or anatomical characteristics of the user show similar boundaries and relationships with each other when measured within and across individuals and are therefore usually kept within certain boundaries, there are usually specific parameters that specify the type of relationship. For example, Latt, MD, Menz, HB, Fung, VS and Lord, SR's "Walking speed, cadence and step length are selected to optimize the stability of head and pelvis accelerations" (Experimental brain research, Vol. 184, No. 2, pp. 201-209) discloses various relationships between self-selected gait speed (i.e., the gait speed experienced by the user) and several step-related indicators. It can be seen from the document that the relationship between the measured speed and the step frequency is linear when it is as high as approximately 2.0 m / s (corresponding to a step frequency of approximately 150 steps / minute). However, the error bar at this time is very large. In fact, for faster walking speeds, the relationship between these two indicators changes. This is because: at this point, the user's activity changes from walking to jogging. The relationship between these two indicators is mediated by several person-dependent biomechanical characteristics and background variables. A major factor is the variability of step length. In fact, for the same speed between people and within people, the step length is not constant. Therefore, the speed estimation via the step frequency may have a considerable error (even for speeds below 2m / s). This "weak" relationship does not allow the step frequency to be accurately calculated based on the speed by knowing the step length (and vice versa). However, it allows boundaries to be set for each of these indicators based on the minimum possible step length value and the maximum possible step length value. In addition, by using the disclosed data, the boundaries of the reliability estimation results can be adjusted. In addition, the algorithm used can be changed according to the value of the previously estimated indicator (or more precisely, according to the parameter on which the indicator depends, or according to the first value).

[0083] In yet another embodiment, the one or more sensor signals include any of the following: acceleration signals, gyroscope signals, heart rate signals, breathing signals, oxygen saturation (SpO2) signals, motion signals, force signals, pressure signals, electromyography (EMG) signals, temperature signals, calorie consumption signals, global positioning system signals, camera signals, radar signals, speed measurement signals, torque measurement signals, electrocardiogram (ECG) signals, electroencephalogram (EEG) signals, photoplethysmography (PPG) signals, barometer signals, muscle tension signals, sweat level signals, pressure level signals, skin conductance signals, galvanic skin response signals, inertial measurement signals, and blood pressure signals, and / or, wherein the first indicator and / or the second indicator include any of the following: walking speed, cadence, step length, heart rate, energy expenditure, and activity type.

[0084] Any other indicator related to the physiological or anatomical state of the user is also contemplated.

[0085] In an embodiment of the system, the sensor includes any one of the following: an accelerometer, a gyroscope, a heart rate sensor, a respiration sensor, an oxygen saturation (SpO2) sensor, a motion sensor, a force sensor, a pressure sensor, an electromyography (EMG) sensor, a temperature sensor, a calorie consumption sensor, a global positioning system sensor, a camera, a radar, a speed measurement sensor, a torque measurement sensor, an electrocardiogram (ECG) sensor, an electroencephalogram (EEG) sensor, a photoplethysmography (PPG) sensor, a barometer, a muscle tension sensor, a sweat level sensor, a pressure level sensor, a skin conductance sensor, a skin galvanic response sensor, an inertial measurement sensor, and a blood pressure sensor.

[0086] In another embodiment of the system, the sensor is a remote sensor or a wearable sensor. The sensor may be a contact sensor in particular. For example, the remote sensor includes a camera (recording the user's activities) or a shoe sensor. The wearable sensor may be integrated into a wearable object (e.g., a pendant or a trouser pocket), or may be attached to any of a wrist, a fingertip, an earlobe, an in-ear part, a torso, etc., and may include, for example, any of a smart phone, a smart watch, and an in-ear device.

[0087] The system may include one or more sensors. In the case where the system includes multiple sensors, one or more sensors may be wearable sensors. Wearable sensors provide a high degree of concealment and are particularly convenient for long-term use. In particular, the walking index can be estimated by using wearable sensors. BRIEF DESCRIPTION OF THE DRAWINGS

[0088] These and other aspects of the invention will become apparent and elucidated with reference to the embodiment(s) described hereinafter. In the accompanying drawings:

[0089] Figure 1 A flow chart showing an embodiment of a method according to the present invention;

[0090] Figure 2A A graph showing the relationship between measured step length and walking speed is shown;

[0091] Figure 2B A graph showing the relationship between the measured stride length to height ratio and walking speed is shown;

[0092] Figure 2C A plot of estimated step length versus walking speed is shown;

[0093] Figure 2D A plot of the estimated step length to height ratio versus walking speed is shown;

[0094] Figure 2E A graph showing the relationship between estimated walking speed and measured walking speed is shown;

[0095] Figure 2F A graph showing the relationship between the estimated cadence and the measured cadence is shown;

[0096] Figure 3 shows a schematic diagram of an embodiment of the device according to the invention; and

[0097] Figure 4 A schematic diagram of an embodiment of a system according to the invention is shown. DETAILED DESCRIPTION

[0098] Figure 1 A flow chart of an embodiment of a method according to the present invention is shown. The steps of the method may be performed by the device 10, wherein the main steps of the method are performed by the processor 14. The method may be implemented, for example, as a computer program running on a computer or a processor.

[0099] In the first step S2, a plurality of sensor signals are obtained from a plurality of sensors. These sensor signals are related to the physiological state of the user. In particular, these signals are related to the user's ability to walk. In the second step S4, a first index and a second index are extracted from these signals. In other words, the first index and the second index are estimated by using the sensor signals. In this embodiment, the first index is estimated using algorithm A and algorithm B, and the second index is estimated using algorithm A and algorithm B. Therefore, two values ​​of the first index and two values ​​of the second index are actually estimated. The first index and the second index are selected so that they are associated with each other by numerically bounded mathematical functions. In the next step S6, the first value of the mathematical function is derived (calculated) by inserting the first index and the second index estimated by algorithm A. In addition, another first value of the mathematical function is derived by inserting the first index and the second index estimated by algorithm B. Then, in step S8, the first value is compared with the boundaries of these values ​​respectively. In this embodiment, the boundaries are predetermined. According to the result of the comparison, in step S10, the reliability rating of each estimated first value is evaluated and the estimation results of (one or more) first indexes and (one or more) second indexes are therefore evaluated. Therefore, two reliability ratings are obtained. One reliability rating is for an indicator based on algorithm A, and the other reliability rating is for an indicator based on algorithm B. In this embodiment, the reliability ratings are binary. Either the first indicator is marked as "reliable" or the first indicator is marked as "unreliable". The same applies to the second indicator. In an optional step S12, the obtained reliability ratings are filtered and a specific rating is selected. More precisely, the estimated indicator corresponding to the "unreliable" rating is selected. In step S14, the evaluated reliability rating (and corresponding indicator) is output, wherein only the "reliable" rating (and indicator) is output if the "unreliable" indicator has been selected in step S12.

[0100] The reliability rating evaluated for the first value (and the corresponding indicator) in step S10 may also (optionally) be used as input in step S4. In particular, if the reliability rating of the indicator estimated using algorithm A is higher than the reliability rating of the indicator estimated using algorithm B, future estimation using algorithm B may be omitted.

[0101] Figure 2A and Figure 2B The (actual) relationship between walking speed and step length measured for the first data set and the (actual) relationship between the ratio of walking speed and step length to the height of the user are shown, respectively, where height means the height of the user. In other words, the data points depicted are points corresponding to the measurement results, not points estimated using the algorithm. In particular, Figure 2AThe graph shows the user's measured step length (in meters) on the y-axis and the user's measured walking speed (in meters per second) on the x-axis. Figure 2B The graph of shows the ratio of the measured step length to the user's height on the y-axis and the measured walking speed of the user in meters per second on the x-axis. Based on these (reference) measurements, a likelihood bound is determined. Figure 2A and Figure 2B The solid line shown in indicates the 95% estimated interval of linear regression, that is, the true best fit line of the data points has a 95% probability of being within the confidence interval marked by the solid line. In this embodiment, the 95% estimated interval of linear regression is used as a boundary related to speed / cadence.

[0102] Similarly, the dashed line indicates the upper limit of the data shown. In particular, from Figure 2A It can be seen from the figure that the upper limit of the absolute estimated step length can be set to 0.9m (Max_SL = 90cm). Similarly, the upper limit of the ratio of step length to user height can be set to 0.5 (Max_SL / height = 0.5) (see Figure 2B ).

[0103] Figure 2C and Figure 2D The relationship between predicted (i.e., estimated) speed (using an algorithm that takes accelerometer data (i.e., accelerometer signals) as input) and predicted (i.e., estimated) step length (also using an algorithm that takes accelerometer data as input) and the ratio of step length to height are shown, respectively, for a second data set (different from the first data set used to determine the boundary). In particular, Figure 2C The graph shows an estimated user's step length (in meters) on the y-axis and an estimated user's walking speed (in meters per second) on the x-axis. Figure 2B The graph shows the ratio of estimated stride length to user height on the y-axis and the estimated user's walking speed (in meters per second) on the x-axis.

[0104] The estimation of step length and step length to height ratio can be based on the fact that the estimated step length can be expressed as the ratio of estimated speed to estimated cadence. In this embodiment, the accelerometer used to measure walking speed and step length to height ratio is a suspension accelerometer. Figure 2C and Figure 2D The boundaries shown correspond to Figure 2A and Figure 2B The boundaries of the measured data are shown. Figure 2A and Figure 2B The comparison from Figure 2C and Figure 2DAs can be seen in the graph: apart from some outliers, the estimated values ​​have a similar relationship to the measured values, which can be considered an indicator that the algorithm for estimating step length and step length to height ratio (or more precisely, the algorithm for walking speed and cadence) is working as expected. On the other hand, if the predicted values ​​have a very different relationship to each other than the measured values ​​have to each other, this would indicate that at least one of the algorithms used (either the algorithm for walking speed estimation or the algorithm for cadence estimation) will generate incorrect predictions / estimations and will therefore be defective.

[0105] exist Figure 2E and Figure 2F In , the estimated speed and estimated cadence are plotted against the (reference) speed and cadence, respectively. More precisely, Figure 2E A graph showing estimated walking speed (in meters per second) (y-axis) versus measured walking speed (in meters per second) (x-axis), and Figure 2F The graph shows the estimated cadence (in steps / second) (y-axis) versus the measured cadence (in steps / second) (x-axis). Figure 2E and Figure 2F The graph shown is based on a second data set.

[0106] If the algorithm for estimating walking speed and cadence works as expected, all displayed data points will be on the solid black line. Therefore, the farther the data points are from the black line, the worse the estimate. Figure 2A and Figure 2B The determined boundaries are marked as unreliable data points that do not comply with the boundaries. In other words, in this embodiment, a binary reliability rating method is applied to the walking activity indicator.

[0107] For example, in Figure 2C to Figure 2F In the figure, the data points indicated by hollow circles correspond to data points corresponding to step lengths that are too large (i.e., outside the upper limit of the step length). Estimated data points corresponding to estimated step length to height ratios that are too large are indicated by triangles. Estimated data points corresponding to step length estimation results that exceed their boundaries are indicated by crosses, and estimated data points corresponding to estimated step length to height ratios that exceed their boundaries are indicated by asterisks. In this way, four boundaries related to step length can be used to determine unreliable data points.

[0108] from Figure 2E and Figure 2F It can be seen that the data points marked as unreliable for estimating cadence are (in most cases) far from the black line and are therefore indeed inaccurate, while the data points marked as unreliable for estimating walking speed are less clear. This result is also confirmed by the performance indicators shown in the following table:

[0109]

[0110] The table shows the results of speed and cadence estimation with all data retained (third row) and after removing data points outside the respective imposed boundaries (fourth to seventh rows). The abbreviation "MAE" stands for "mean absolute error", the abbreviation "RMSE" stands for "root mean square error", and the abbreviation "R" stands for "Pearson correlation coefficient" (here, "R" represents the correlation coefficient between the measured value and the predicted / estimated value).

[0111] It can be seen from the table that when the data based on the estimated stride length (ratio) are excluded, the reliability of the indicators improves, especially for the stride frequency (see the right part of the table). In fact, if all the predictions / estimates were completely correct, all the errors shown in the table would be 0 and "R" would be 1. If only random points were removed, the accuracy of the data would not be significantly improved. Although this improvement comes at the cost of not estimating the results of a certain part of the data, it can be seen that this effect is relatively limited, at most only 4% (see the first column).

[0112] Although in the above example, if the first values ​​of stride length and the ratio of stride length to height are not within their imposed boundaries, then both estimated indicators (i.e., estimated stride length and estimated stride frequency) are considered unreliable, other embodiments are conceivable in which the estimated indicator that is considered to be most reliable (e.g., based on previously calculated results) can be used as a guide for other indicators, so that if the boundaries are not respected, then only the estimated stride frequency is considered unreliable. For example, in the above example, the speed estimate result is on average more accurate than the stride frequency estimate result. Therefore, only the estimated stride frequency may be considered unreliable, so that the estimated speed may be retained. In addition, the stride frequency estimate result can be updated based on the speed estimate result and the stride length boundary and / or the stride length to height ratio boundary.

[0113] Figure 3 A schematic diagram of an embodiment of a device 10 according to the invention is shown. The device 10 comprises a sensor input 12, a processor 14 and an information output 16.

[0114] The sensor input 12 is configured to obtain at least one sensor signal from at least one sensor, the sensor signal being related to a physiological state of the user. The sensor input 12 may be directly coupled or connected to the sensor(s), or may obtain (i.e., retrieve or receive) the sensor signal from a storage device, a buffer, a network or a bus, etc. Thus, the sensor input 12 may be, for example, a (wired or wireless) communication interface or data interface, such as a Bluetooth interface, a WiFi interface, a LAN interface, an HDMI interface, a direct cable connection, or any other suitable interface that allows signal transmission to the device 10.

[0115] The processor 14 applies at least one algorithm to the signal and its features to estimate the first indicator and the second indicator, which are related to each other through a mathematical relationship. In particular, the processor 14 is configured to determine a first value of the first mathematical function by inserting the estimated indicator into the first mathematical function. The first value is then compared to a boundary of the first value. Based on the result of the comparison, the processor 14 evaluates a reliability rating of the estimation result for the first indicator and / or the second indicator.

[0116] The information output 16 then provides the assessed reliability rating for further processing. The information output 16 may generally be any interface that provides the assessed reliability rating (e.g., transmits it (to another device) or provides it for retrieval (by another device (e.g., smartphone, computer, tablet, etc.)). Thus, it may generally be any (wired or wireless) communication or data interface.

[0117] Figure 4A schematic diagram of an embodiment of a system 100 according to the present invention is shown. The system includes two sensors 20a and 20b, a device 10 and a user interface 40. Both sensors 20a and 20b are configured to measure at least one sensor signal related to a user's physiological state. In this embodiment, sensor 20a is a smart watch configured to be worn by a user, and sensor 20b is a smart phone configured to be worn by a user. Sensor 20a is configured to measure at least one signal related to the user's heart rate, and sensor 20b is configured to measure a signal related to the user's activity count during a specific activity. Device 10 is configured to obtain the signals of sensors 20a and 20b, and estimate the user's heart rate and activity count based on these signals. The relationship between the heart rate index and the activity count index depends on the person's physical fitness level. This physical fitness can be expressed by VO2max, which generally cannot be higher or lower than certain boundary values. Therefore, by using the boundary of VO2max, the device 10 of system 100 is configured to assign reliability ratings to the estimated indicators of heart rate and activity count. The reliability rating may then be provided to a user interface 40 that is configured to publish the reliability rating of the device 10. The user interface 40 may generally be any unit that outputs the reliability rating in a visual or auditory form (e.g., in text form, images or charts, sounds or spoken words, etc.). For example, the user interface 40 may be a display, a speaker, a touch screen, a computer monitor, a screen of a smartphone or tablet, etc.

[0118] Although the invention has been described and illustrated in detail in the drawings and the foregoing description, such illustration and description are to be considered illustrative or exemplary rather than restrictive; the invention is not limited to the disclosed embodiments. Other variations of the disclosed embodiments may be understood and effected by those skilled in the art in practicing the claimed invention by studying the drawings, the disclosure, and the appended claims.

[0119] In the claims, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" does not exclude a plurality. A single element or other unit may fulfill the functions of several items recited in the claims. The fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0120] The computer program may be stored / distributed on suitable non-transitory media, such as optical storage media or solid-state media provided with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems.

[0121] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A method for evaluating the reliability of an estimated indicator related to a user's physiological state, the method comprising: obtaining one or more sensor signals related to the physiological state of the user; estimating a first indicator and a second indicator independently of each other using different methods and / or algorithms based on the one or more sensor signals, wherein the first indicator is related to the second indicator by a first mathematical function; determining a first value of the first mathematical function; comparing the first value to a boundary of the first value; evaluating a reliability rating of an estimation result for the first indicator and / or the second indicator based on the comparison; and Output the assessed reliability rating.

2. The method according to claim 1, in, Comparing the first value to the boundary of the first value includes determining whether the first value satisfies a relationship between the first mathematical function and the boundary.

3. The method according to claim 1 or 2, in, The boundary is a lower limit and / or an upper limit.

4. The method according to claim 3, in, Comparing the first value to the boundaries of the first value includes determining whether the first value is above the lower limit and / or below the upper limit.

5. The method according to any one of the preceding claims, in, The boundary is predetermined, or wherein the method further comprises estimating the boundary based on any of a previous first indicator, a previous second indicator, a previous first value, or the one or more sensor signals.

6. The method according to any one of the preceding claims, in, Any of the first indicator, the second indicator and the first mathematical function may be time dependent.

7. The method according to any one of the preceding claims, Also includes: determining a second value of a second mathematical function relating the first mathematical function to the boundary, Therein, evaluating the reliability rating of the estimation result for the first indicator and / or the second indicator is based on the second value.

8. The method according to any one of the preceding claims, in, Evaluating the reliability rating includes assigning a score on a binary scale or a continuous scale to the estimation results of the first indicator and / or the second indicator.

9. The method according to any one of the preceding claims, Also includes: A method for estimating the first indicator and / or the second indicator is adjusted or changed based on the reliability rating.

10. The method according to any one of the preceding claims, in, The one or more sensor signals include any of the following: acceleration signals, gyroscope signals, heart rate signals, breathing signals, oxygen saturation signals, motion signals, force signals, pressure signals, electromyography signals, temperature signals, calorie consumption signals, global positioning system signals, camera signals, radar signals, speed measurement signals, torque measurement signals, electrocardiogram signals, electroencephalogram signals, photoplethysmography signals, barometer signals, muscle tension signals, sweat level signals, stress level signals, skin conductance signals, galvanic skin response signals, inertial measurement signals, and blood pressure signals, and / or The first indicator and / or the second indicator include any one of the following: walking speed, cadence, step length, heart rate, energy consumption, and activity type.

11. A device (10) for evaluating the reliability of an estimated indicator related to a user's physiological state, the device comprising: a sensor input unit (12) configured to obtain one or more sensor signals related to the physiological state of the user; A processor (14) configured to: estimating a first indicator and a second indicator independently of each other using different methods and / or algorithms based on the one or more sensor signals, wherein the first indicator is related to the second indicator by a first mathematical function; determining a first value of the first mathematical function; comparing the first value to a boundary of the first value; evaluating a reliability rating of an estimation result for the first indicator and / or the second indicator based on the comparison; and an information output section (16) configured to output the evaluated reliability rating.

12. A system (100) for evaluating the reliability of an estimated indicator related to a user's physiological state, the system comprising: a sensor (20) configured to measure one or more sensor signals related to the physiological state of the user; The device (10) according to claim 11; and A user interface (40) configured to publish the reliability rating of the device (10).

13. The system according to claim 12, in, The sensor (20) includes any one of the following: an accelerometer, a gyroscope, a heart rate sensor, a respiration sensor, an oxygen saturation (SpO2) sensor, a motion sensor, a force sensor, a pressure sensor, an electromyography (EMG) sensor, a temperature sensor, a calorie consumption sensor, a global positioning system sensor, a camera, a radar, a speed measurement sensor, a torque measurement sensor, an electrocardiogram (ECG) sensor, an electroencephalogram (EEG) sensor, a photoplethysmography (PPG) sensor, a barometer, a muscle tension sensor, a sweat level sensor, a pressure level sensor, a skin conductance sensor, a skin galvanic response sensor, an inertial measurement sensor, and a blood pressure sensor.

14. The system according to claim 12 or 13, in, The sensor (20) is a remote sensor or a wearable sensor.

15. A computer program comprising program code means for causing a computer to perform the steps of the method according to claims 1 to 10 when the computer program is executed on the computer.