FI120619B - Device and method for monitoring performance - Google Patents

Description

Device and method for monitoring performance

The invention relates to methods and apparatus for monitoring the performance of individuals during exercise. In particular, the invention relates to wrist-worn devices.

5

Reliable heart rate measurement has been a long-term goal in the heart rate monitor industry. Many methods have been developed to detect heart rate, the most important of which is the capacitive 'listening' of the EMFI (Electromechanical Film), the monitoring of changes in blood impedance or inductance (bioimpedance / bioinductance) and the mechanical recognition of heart rate. Ultrasound echo of heart rate and infrared light reflection, scattering or absorption can also be used to detect heart rate. In addition, acoustic heart rate listening has been tried.

The problem with the known methods and devices is the poor quality of the signal and hence the reliability of the heart rate data. Heart rate sensors are typically located on the wrist of the device, since both the electrical and mechanical heart rate signals on the volar (palm side) of the wrist are at their strongest due to the proximity of the blood vessels. However, even these signals are much more difficult to pick up than, for example, the electrical measurement of the 20 heart QRS complex by means of two skin electrodes. It is known that there are no devices on the market that are capable of successfully measuring heart rate, particularly on the dorsal side of the wrist, that is, the side typically carried by the central unit of wrist equipment.

Devices utilizing volumetric hand measurement have been described e.g. WO 00/28892, US 2006/0047208 and Design and Evaluation of Handheld Impedance Plethysmography for Measuring Heart Rate Variability, Medical & Biological Engineering & Computing 2005, Vol. 43. The dorsal pulse electrode pair of the wrist is mentioned in US Pat. 2006/0122521.

Methods are also known (eg, EP 0575984, US 6982930), in which measurement is performed only half-hand, whereby the signal is measured between at least two limbs. In the wrist unit application, the user is also required to touch the electrode or electrodes of the wrist unit with one hand, which is not pleasant or even possible during operation.

2

Indeed, resources in the field have been continuously devoted to the development of increasingly sensitive wrist sensors to dispense with commonly used chest-based heart rate belts and the like during heart rate monitoring. So far, however, heart rate belts are superior in reliability. Reliable heart rate measurement is generally considered to be such an essential function that consumer devices utilizing wrist measurement alone are not yet on the market.

Although wrist-measuring technology has advanced, it may not be possible in the foreseeable future to measure heart rate reliably on the wrist alone, not necessarily on the wrist dorsal side of the wrist, at least not with portable wristcorders or other similar wristbands.

It is an object of the invention to provide a wrist unit suitable for the evaluation of the training effect, which utilizes the measurement of the hemodynamic signal from the wrist in a novel and reliable way.

It is also an object of the invention to provide a new method for assessing training effect.

The invention is based on the finding that a hemodynamic signal for which heart rate cannot be determined can generally identify respiratory rate or other respiratory parameters. The respiratory rate, in turn, is directly applicable, for example, to the calculation of energy consumption during exercise. Information on energy consumption is again sufficient for a large number of users, and even more interesting than the absolute value of heart rate. Thus, the invention provides a reliable method for evaluating the effect of exercise training, and an apparatus that can be manufactured inexpensively and reliably for this purpose.

In the method of the invention, the physical performance of a subject is detected during the exercise by collecting a hemodynamic signal on a wrist sensor and transmitting the hemodynamic signal to a data processing unit in which at least one respiratory physiological parameter is derived from the signal. Based on this, at least one training effect parameter dependent on the person and performance is further calculated.

3

The wrist wearable device of the invention comprises a central unit having a display panel and sensor means for collecting a hemodynamic signal from the wrist. In addition, the device has a data processing unit operatively coupled to the sensing means, adapted to derive at least one respiratory physiological parameter from the hemodynamic signal and further to calculate at least one person and performance-dependent training effect parameter.

More specifically, the process of the invention is characterized by what is said in the characterizing part of claim 1. The device according to the invention is in turn characterized by what is said in the characterizing part of claim 16.

We have surprisingly found that sensors mounted on the wrist unit, even on the back panel of the wrist unit, especially electrical electrodes or an electret film (such as an EMF membrane), can be used to determine respiratory rate and further energy consumption to collect sufficient signal.

Weakness and poor quality of the signal have proven to be a problem, especially in the area of wrist detection, and a good way to improve the signal has not yet been developed. Chest heart rate recognition is usually measured electronically by the QRS complex of the heart, 20 as this has proven to be a superior technique in terms of reliability. However, it is not possible to measure the clear QRS complex from the limb during exercise. However, according to the present invention, by eliminating the need for heart rate detection and by recognizing respiratory rate, even a lower quality signal will reveal what the end user is often most interested in, i.e., the hitting effect. Thus, it is not necessary for the invention to detect (or even be able to recognize) the heart rate of the signal. However, in certain embodiments, it is advantageous if at least part of the heart rate is detected, as will be described in more detail below.

According to a preferred embodiment of the invention, the hemodynamic signal is collected from the dorsal side of the wrist. The corresponding device has a back panel opposite the display panel where the sensor means are located. Although the signal strength is clearly weaker on the dorsal side than on the volar due to the distal position of the large blood vessels, we have found dorsal measurement for the present use, i.e., assessment of the effect of exercise on breathing, sufficient.

4

More generally, the invention is not limited to any particular method of detecting a hemodynamic signal. In principle, any non-invasive method that is sensitive to respiratory-induced periodic vascular changes is suitable for use with the invention. However, some specific methods can provide specific advantages, for example, in terms of device configuration, price, power consumption, or where the device can be used in sports. For example, in swimming and running, the hemodynamic signal may not be reliably detected by the same method. Particularly preferably, measurement methods that are more sensitive to the respiratory component of the signal than the heart rate component are suitable for measuring within the scope of the invention.

The invention generally achieves considerable advantages in the manufacture of equipment. Namely, it avoids the use of heart rate collars or similar remote sensors to evaluate the training effect. In particular, positioning the sensors on the back panel of the device, against the dorsal side of the hand, avoids sensors mounted on the bracelet and hence the contact manufacturing problems between the wrist unit's central unit and the bracelet. It is known that elastic and durable electrical contacts between the bracelet and the central unit of the device are difficult and expensive.

The invention is also advantageous for the end user, since the user only needs one device to perform the training effect assessment. There has been a need in the industry for many years for such new applications because it makes it as easy as possible for a user to enter the performance being evaluated, which further encourages the user to improve performance and analyze performance.

25

By the dorsal side of the wrist, we mean the surface on the palm of the plane defined by the large wrists.

In the following, various embodiments of the invention will be explained in more detail with reference to the accompanying drawings, in which: Figure 1 is a flow chart of typical process steps of the invention; effect.

Each heart rate causes a small bulge in the blood vessels to flow and further to the vessel, which is illustrated in Figure 3 by reference numeral 32. Plethysmographic measurements are known to dilate blood vessels not only by the effect of the heart rate but also by the rate of respiration. Thus, breathing cyclically changes the vascular system pressure. This is because when inhaled, and thus the thoracic muscles expand, the pressure in the blood vessels in the limbs decreases, and when exhaled in reverse. The "suction" caused by Hen-10 grafting thus produces a pressure change which is well transmitted in the fluid, in this case the blood. The heart rate of a healthy person is many times higher than the respiratory rate. Thus, breathing is seen as a lower-frequency modulation of blood flow, i.e., a cyclic change in heart rate pulse. This is illustrated in Figure 4. The hemodynamic signal collected from the wrist according to the invention identifies this modulation, which we have found to be more reliably extracted from a very noisy signal.

Due to the non-spherical nature of the blood cells 36 (mainly red cells), the degree of organization of the blood cells in the bulge 32 changes during heart rate. The change in the degree of organization is reflected as a change in the electrical conductivity (impedance) of the veins and thus the tissue as a whole. This feature can be utilized in some embodiments, as will be explained in more detail below. In an article entitled "Detection of Pulse and Respiratory Signals from the Wrist Using Dry Electrodes, Biomedical Instrumentation and Technology," by Farag et al, July / August 1994, one exemplary measuring system utilizing volumetric measurement of the wrist is described.

In Figure 1, signal acquisition is denoted by the reference numeral 10. The signal is transferred to a data processing unit (processing unit) in step 12. In the data processing unit, a respiratory rate is identified in step 14. Some possible methods for achieving this are described below.

According to a preferred embodiment, the measured signal is low-pass filtered so that the respiratory rate can be calculated based on the periodicity of the filtered signal, typically the maximum (or minimum) time information. Thus, in this embodiment, the identification or detection of individual heartbeats 6 is not at all necessary, and thus is not typically performed to measure respiratory rate.

Instead of determining the respiratory rate over time, a back-to-back assay may also be used. The collected hemodynamic signal is then converted to the frequency domain by discrete Fourier Transform (DFT) and the respiratory component extracted from the converted signal. In signal processing, e.g. known coordinate conversion, filtering and / or pulse detection techniques.

Generally speaking, the frequency of respiratory modulation is most preferably detected in a hemodynamic signal independent of heart rate modulation. Thus, also or only the signal collected from the signal periods (heart rate intervals) between the heart rate pulses is used to determine the respiratory rate. The embodiments described above fall into this category.

It is to be noted that although the invention does not require the detection of heart rate, or even individual heart rates, from the signal, they may be monitored, or at least attempted, to provide alternative or particularly advantageous embodiments. Especially in the future, if the sensitivity of the measurement methods is improved, it may be possible to reliably determine the individual heart rate or heart rate from the dorsal side of the wrist. Some possible embodiments are described below.

According to one embodiment, to determine the respiratory rate, individual heartbeats and further the height variations caused by them in the signal are identified. The cycle of height variations is further derived from the frequency of respiratory modulation. An advantage of the known solutions is that even if individual heartbeats are not recognized, the frequency of the lower frequency respiratory modulation can still be calculated from the measurement. However, it is advantageous if the heart rate pulses are successfully detected with an average frequency twice the respiratory rate. However, this frequency can be significantly lower, at least half, and even lower than the heart rate. Thus, at typical respiratory and pulse rates, the respiratory rate can be determined even if, on average, every other heart rate is not recognized. According to one embodiment, a breathing parameter is derived from the hemodynamic signal, even if the signal quality is insufficient for reliable heart rate detection.

7

According to one embodiment, the respiratory rate is determined, at least in part, by the periodic variation of the heart rate information contained in the heart rate signal (i.e., so-called heart rate interval noise). The interval between heart rate noise can be calculated via frequency conversion or preferably directly in the time domain. In the time domain, periodicity is preferably determined by time stamps based on the heart rate signal. Typically, the time stamps are formed into a series of consecutive time points, the sequence period is determined, and the respiratory rate is determined based on the sequence period. The sequence of a series can be further determined by calculating the second derivative of the series and looking for its zeroes. The advantage of time-domain determination over frequency-domain analysis is that it requires less computation. Thus, the computation is fast 10 can be performed with a small processor and program memory capacity, thus also reducing the power consumption and making the device more economical.

However, heart rate-based respiratory measurement methods always produce a small error at a specified respiratory rate, so heart rate-independent '' direct 'methods as described above are most preferably used.

If the quality of the collected hemodynamic signal is sufficient to detect all heart rates, the heart rate is preferably determined from the signal in the frequency domain by a discrete Fourier transform (DFT).

20

In order to improve the reliability of the respiratory rate, it may be advantageous to try to identify also the heart rate or some correlated quantity from the he-modynamic signal and to further compare this magnitude with the respiratory rate. An example of an application where such a comparison is useful is the detection of a stress state. When the body receives a stress response, ad-25 renalin enters the blood and the heart rate rises. According to current knowledge, the increase in heart rate is due to the fact that the bloodstream is made to deliver adrenaline more intensively, thereby increasing muscular fitness. However, the respiratory rate does not increase in the stress response until you actually start to metabolize. That's the idea behind measuring stress. According to a preferred embodiment 30, stress measurement based on a comparison of respiratory rate and heart rate may be implemented by utilizing the invention to observe respiration and noise in the measured signal, or the heart rate correlated or calculated from it. A simple increase in signal noise level can indicate increased cardiac activity. If an increase in noise is observed without an increase in respiratory rate, one may conclude that there is some stress response. It should be noted that in this embodiment too, the absolute value of the heart rate 8 need not be known, but its relative value (such as signal noise level, etc.) is sufficient. Generally speaking, an indication of an abnormally rapid change in heart rate relative to a change in respiratory rate is sufficient to indicate a change in stress state.

5

In some embodiments, the prediction data, for example, from normal ranges of heart rate and / or respiratory rate, or from typical runtime correlation, may also be used to calculate respiratory rate. This can further improve the reliability of the method in the wrist environment.

10 Using the defined respiratory rate, we further calculate a training effect parameter in step 16. By a hitting effect parameter, we mean a performance and performance-dependent quantity that indicates the strenuousness of the exercise, its energy consumption, or other effect on the performer's physical condition, general condition, or recovery. Typically, the simulated exercise effect parameter is the instantaneous or cumulative energy consumption or EPOC (Excess Post-Excercise Oxygen Consumption), or any of these further derived. Such a measure can be, for example, the '' Training Effect '' which describes fitness, aerobic performance, healing effect.

EPOC describes the amount of oxygen that is required after exercise to stimulate the body to return to normal, homeostasis. According to one embodiment of the present invention, the accumulation of EPOC during exercise is estimated, at least in part, by the respiratory rate determined according to the invention. The use of EPOC to control training is described in more detail, for example, in US Publication 25 2006/0004265 and other publications by Firstbeat Technologies Oy (e.g., White Papers May 2005 and September 2005).

According to one preferred embodiment, the respiratory parameter is used to calculate the energy consumption during the run. Typically, at least one prior knowledge of either the person being measured and / or the species performed by him or her will be used to assist. Preliminary data may include person-measured data such as VO2max. However, according to a preferred embodiment, information that can be determined directly from tests or data not directly related to oxygen uptake is used, which may include, for example, a person's activity class, weight, height, or sex, or the nature of the species. The nature of the species primarily refers to whether it is a Sprint type or a durability type. The activity class (typically on a scale of 1-10) can again be determined e.g. based on a person's training volume without physical examinations. Other personal or species-specific information may also be used. The calculation of energy consumption or other training effect 5 is performed on the basis of the information used and the measured breathing parameter. According to a particularly preferred embodiment, the pre-selected data is used directly as scaling factors for the breathing parameter or parameters, which simplifies and speeds up the calculation. Different weights can be assigned to different pre-data. Preferably, the final result is converted into absolute instantaneous units of energy consumption (e.g., kcal / min). The cumulative energy consumption of performance 10 can also be calculated. Consumption may also be reported as some relative values.

Especially at the beginning or end of the exercise or at any other change in the rhythm of breathing, respiratory rate generally does not directly correlate with current energy consumption or other training effects. When a person begins to exercise, his or her breathing does not immediately reach a level comparable to that of instantaneous energy. On the other hand, after exercise or breaks, respiratory rate remains high even though physical exertion is over. However, these factors can be taken into account by observing changes in the measurable quantity of the respiratory rate, heart rate, or any other measure of change in rhythm. If a predetermined change in such a quantity is observed within a given 20 time intervals, the respiratory rate can be corrected computationally toward a respiratory rate value that better corresponds to the actual training effect. Real-time correction may occur, for example, by storing momentary respiratory rates in a buffer memory for a predetermined interval, and comparing the latest obtained respiratory rate with historical values of respiratory rate. It will be appreciated by those skilled in the art from the foregoing that the desired effect calculation can be accomplished in various ways.

The training effect correction is preferably performed in an enhanced manner. This means that estimates of energy consumption are corrected more proportionally to how much the change in the rhythm of the performance changes. For example, this compensates for a slow change in breathing or heart rate relative to the instantaneous intensity of the exercise. Of course, the measure of rhythm change may also be information obtained through an accelerometer, for example, whereby an enhanced correction may not be necessary.

10

Referring now to Figure 2, by way of example, a wrist unit by means of which the above process steps can be performed will be described.

Preferably, the wrist unit comprises a central unit 20 having a display panel 25 and a rear panel on the opposite side thereof. A wristband 23 attached to or coupled to the central unit. Sensor means 28 for collecting a hemodynamic signal from the wrist are located substantially on the rear panel 24, typically on or partially embedded therein, so that they can be contacted with the skin when the device is worn. The wrist unit also comprises a data processing unit 26 operably coupled to the sensor means 28, in which a breathing parameter is derived from the hemodynamic signal, from which at least one personal and performance dependent training effect parameter can be derived from the data processing unit 26. An exemplary calculation of the training effect parameter will be described in more detail later.

Preferably, sensing 28 is used which comprises a plurality, preferably four, electric bioimpedance measuring electrodes. A four-electrode sensor typically has a first pair of electrodes for supplying current to the wrist and a second pair of electrodes for detecting breath-modulated bioimpedance of the wrist. Most preferably, the electrodes are arranged sequentially / side by side so that the pair of electrodes formed by the outermost electrodes supplies power to the tissue. The identification can thus be carried out, for example, with a similar electrode arrangement as disclosed in WO 00/28892 but mounted on the back panel of the wrist unit.

Alternatively, sensor arrangements based on mechanical pulse detection of heart rate pulses may be used as sensing 28. Examples of such are sensors based on EMF membranes or similar electret films, mechanical capacitive sensors, and mechanical sensors based on springs and gels and liquids. The advantage of the EMF film is its high sensitivity and lightness.

Other known and unknown sensors, including optical (especially infrared) sensors, ultrasonic sensors and acoustic sensors, can also be used for sensing. As an example of infrared measurement 30, the method described in US 6080110 for actively measuring heart rate in the ear canal by means of infrared light reflections is provided. This principle can also be applied to measurements on the dorsal side of the wrist. Passive infrared detection can also be used to monitor the passage of light from any external light source into the tissue.

11

The sensor may extend or substantially fill a portion of the area facing the skin of the back panel. Generally, it is advantageous if the base facing the wrist of the device, in particular the base sensing portion, is elevated relative to the other device parts (particularly the peripheral portions of the watch case). This ensures that when the device is attached to the wrist by means of a wrist strap, the sensor is securely connected to the skin and is thus capable of reliably transmitting a signal.

The central unit of the wrist unit is preferably manufactured to be relatively lightweight, whereby the back panel of the wrist unit is better attached to the skin during exercise, i.e., bouncing is reduced.

10 Once the desired exercise effect parameter or parameters have been determined, the result can be displayed to the user on the display board in analogue or digital form. Data can also be stored in device memory for later analysis.

The sensing can also be completely incorporated into the wristband bracelet, whereby the strength of the collected signal may be slightly improved. In this case, electrical contacts are made between the bracelet and the central unit. Depending on the type of sensor, it can also be distributed with part of it on the back panel of the central unit and part on the bracelet.

The exemplary embodiments described above are not limiting of the invention and can be freely combined and modified. The claims are to be construed in their entirety with reference to equivalence interpretation.

Claims (30)

A method for monitoring a person's physical performance during exercise, comprising: 5. collecting a hemodynamic signal by a sensor, - transmitting the hemodynamic signal to a data processing unit, wherein at least one physiological parameter is derived from the signal, characterized by: - using a sensor and data processing unit located in the same portable wrist unit, - in the data processing unit, at least one physiological parameter describing respiration is derived from the hemodynamic signal, and based on this, at least one person and performance-dependent training effect parameter is calculated. Method according to claim 1, characterized in that the hemodynamic signal is collected by a sensor located on the dorsal side of the wrist. Method according to claim 1 or 2, characterized in that a sensor comprising at least two, preferably four, electrode impedance measuring electrodes is used. 20 A method according to claim 3, characterized in that a sensor is used which comprises a pair of electrodes for supplying current to the wrist and a pair of electrodes for detecting respiratory modulated bioimpedance. A method according to any one of the preceding claims, characterized in that a sensor for sensing pressure on the skin surface, such as an electret film, preferably an EMF film, is used. 6. A method according to any one of the preceding claims, characterized in that the data processing unit detects from the hemodynamic signal the frequency of the respiratory modulation to determine the respiratory rate. A method according to claim 6, characterized in that the frequency of the respiratory modulation is detected in a manner independent of the heart rate modulation. Method according to claim 6 or 7, characterized in that the frequency of the respiratory modulation is detected by low-pass filtering of the hemodynamic signal and by examining the periodicity of the filtered signal. 9. A method according to any one of claims 6 to 9, characterized in that the data processing unit also detects the heart rate modulation frequency or some correlated quantity thereof to determine the absolute or relative heart rate and further comprises the step of comparing heart rate and respiratory rate. A method according to claim 9, characterized in that it comprises the step of detecting temporal changes in absolute or relative heart rate relative to changes in respiratory rate to assess a person's stress state. A method according to claim 6, characterized in that the frequency of the respiratory modulation 15 is detected by detecting a cyclic change in the amplitude of the heart rate pulses. 12. A method according to claim 6, characterized in that the frequency of the respiratory modulation is detected by observing the periodicity of the heart rate noise. A method according to any one of the preceding claims, characterized in that a breathing parameter is derived from the hemodynamic signal, even if the signal quality is insufficient for reliable heart rate detection. A method according to any one of the preceding claims, characterized in that said dissipation effect parameter is energy consumption or a derivative thereof. The method according to any one of the preceding claims, characterized in that said training effect parameter is EPOC (excess post-excercise oxygen consumption) or a derivative thereof. 30 A wrist unit for monitoring the physical performance of a person during exercise, the wrist unit comprising: - a central processing unit having a display panel and a wrist strap attached thereto, signal, characterized in that - the data processing unit is adapted to derive at least one respiratory physiological parameter from the hemodynamic signal and further to calculate at least one person and performance-dependent training effect parameter. The wrist unit of claim 16, characterized in that the sensor means is disposed on a back panel opposite the central display panel. 10 The wrist unit according to claim 16 or 17, characterized in that the sensor means comprise at least two, preferably four, electrode impedance measuring electrodes. The wrist unit of claim 18, wherein the sensor means comprises a pair of electrodes for supplying current to the wrist and a pair of electrodes for detecting pulsed and respiratory modulated bioimpedance. A wrist unit according to any one of claims 16 to 19, characterized in that the sensor means comprise means for sensing pressure on the skin surface, such as an electret film, preferably an EMF film. A wrist unit according to any one of claims 16 to 20, characterized in that the data processing unit is adapted to detect, from a hemodynamic signal, the frequency of respiratory modulation to determine the respiratory rate. 25 A wrist unit according to claim 21, characterized in that the data processing unit is adapted to detect, from a hemodynamic signal, the frequency of respiratory modulation in a manner independent of heart rate modulation. A wrist unit according to claim 22, characterized in that the data processing unit is adapted to detect, from the hemodynamic signal, the frequency of respiratory modulation by lowpass filtering the hemodynamic signal and examining the periodicity of the filtered signal. A wrist unit according to any one of claims 16 to 23, characterized in that the data processing unit is adapted to detect also from the hemodynamic signal the heart rate modulation frequency to determine heart rate and to further compare heart rate and respiratory rate. 5 The wrist unit of claim 24, characterized in that the data processing unit is adapted to detect temporal changes in absolute or relative heart rate relative to changes in respiratory rate and to further evaluate the individual's stress state based on these changes. 10 The wrist unit of claim 21, characterized in that the data processing unit is adapted to detect, from a hemodynamic signal, the frequency of respiratory modulation by detecting a cyclic change in the amplitude of heart rate pulses. The wrist unit of claim 21, characterized in that the data processing unit is adapted to detect the frequency of the respiratory modulation from the hemodynamic signal by observing the periodicity of the heart rate noise. A wrist unit according to any one of claims 16 to 27, characterized in that the data processing unit 20 is adapted to derive a parameter describing hemodynamic signal respiration even if the signal quality is insufficient for reliable heart rate detection. A wrist unit according to any one of claims 16 to 28, characterized in that the dissipation effect parameter is energy consumption or a derivative thereof. 25 The wrist unit according to any one of claims 16 to 29, characterized in that the training effect parameter is EPOC (excess post-excercise oxygen consumption) or a derivative thereof.