[go: up one dir, main page]

WO2024260794A1 - Apparatus for determining the heart rate of a person in motion - Google Patents

Apparatus for determining the heart rate of a person in motion Download PDF

Info

Publication number
WO2024260794A1
WO2024260794A1 PCT/EP2024/066053 EP2024066053W WO2024260794A1 WO 2024260794 A1 WO2024260794 A1 WO 2024260794A1 EP 2024066053 W EP2024066053 W EP 2024066053W WO 2024260794 A1 WO2024260794 A1 WO 2024260794A1
Authority
WO
WIPO (PCT)
Prior art keywords
motion
heart rate
output signal
person
sensor output
Prior art date
Application number
PCT/EP2024/066053
Other languages
French (fr)
Inventor
Alexander Gaiduk
Patrick Celka
Robert Alcock
Amedeo SORIA
Original Assignee
Ams-Osram Ag
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
Application filed by Ams-Osram Ag filed Critical Ams-Osram Ag
Publication of WO2024260794A1 publication Critical patent/WO2024260794A1/en

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording for evaluating the cardiovascular system, e.g. pulse, heart rate, blood pressure or blood flow
    • A61B5/024—Measuring pulse rate or heart rate
    • A61B5/02416—Measuring pulse rate or heart rate using photoplethysmograph signals, e.g. generated by infrared radiation
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Measuring devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor or mobility of a limb
    • A61B5/1118—Determining activity level
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/6802—Sensor mounted on worn items
    • A61B5/681—Wristwatch-type devices
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A61B5/7207—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal of noise induced by motion artifacts
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording for evaluating the cardiovascular system, e.g. pulse, heart rate, blood pressure or blood flow
    • A61B5/024—Measuring pulse rate or heart rate
    • A61B5/02438—Measuring pulse rate or heart rate with portable devices, e.g. worn by the patient
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
    • A61B5/7207—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal of noise induced by motion artifacts
    • A61B5/721—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal of noise induced by motion artifacts using a separate sensor to detect motion or using motion information derived from signals other than the physiological signal to be measured

Definitions

  • the disclosure relates to an apparatus for determining the heart rate of a person in motion . Furthermore , the disclosure relates to a method for determining the heart rate of a person in motion .
  • the heart rate of a person can be measured by a variety of methods .
  • One of these methods is based on photoplethysmography, an optical method that uses light to measure changes in blood volume .
  • determining heart rate using photoplethysmography is di f ficult when the person is in motion, because the estimation of the heart rate depends on the activity of the subj ect .
  • Accurate heart rate estimation by evaluating a photoplethysmogram recorded by a PPG (photoplethysmography) sensor is di f ficult without additional hardware components and suitable algorithms and leads to erroneous measurement results .
  • An apparatus for determining the heart rate of a person with high accuracy, when the person performs an activity, for example a movement is speci fied in claim 1 .
  • the proposed apparatus for determining the heart rate of a person in motion comprises an optical sensor, a motion sensor and a processing circuit to provide an output signal representing the heart rate of the person .
  • the optical sensor is configured to provide an optical sensor output signal suitable for creating a photoplethysmogram of the person .
  • the motion sensor is configured to detect a resting state or at least one motion state of the person, and to provide a motion sensor output signal in response to the detected resting or at least one motion state of the person .
  • the processing circuit is configured to provide the output signal in response to the evaluation of the photoplethysmogram and the motion sensor output signal , when the at least one motion state is detected for the person .
  • the processing circuit is further configured to provide the output signal in response to the evaluation of the photoplethysmogram, when the resting state is detected for the person .
  • the optical sensor may be configured as a PPG sensor .
  • the motion sensor may be embodied as an ACC ( accelerometer ) .
  • the processing circuit evaluates the motion sensor output signal for performing a classi fication of a motion of the person prior to the heart rate analysis . I f the evaluation of the motion sensor output signal shows that the person is in a resting state , i . e . not performing any movement , the heart rate is determined by the processing circuit only by evaluating the optical sensor output signal or the photoplethysmogram and is thus determined without taking the motion sensor output signal into account. On the other hand, if the evaluation of the motion sensor output signal shows that the person is in at least one motion state, i.e.
  • the heart rate is determined by the processing circuit by evaluating the optical sensor output signal or the photoplethysmogram as well as the motion sensor output signal.
  • a heart rate model that is dedicated to a specific motion class is applied to provide an approximate value of the heart rate.
  • the heart rate analysis is based on different models as classified from motion classes.
  • the motion classification may be improved by gaining motion information from more than one motion sensor, if available, located at different body positions and calibrated for a particular position.
  • the processing circuit is configured to determine a respective one of several motion classes assigned to the person in dependence on the motion sensor output signal.
  • Each motion class represents a respective kind of motion of the person, i.e. a motion state or also a resting state of the person.
  • the motion sensor output signal may be subjected to filtering.
  • filtering By applying a filter function to the motion sensor output signal, spurious signal components are filtered out of the motion sensor output signal.
  • the motion sensor output signal provided, for example, by an accelerometer is a three-dimensional vector signal having components in different spatial directions. By normalizing this three-dimensional vector signal , a one-dimensional signal is determined that indicates , for example , the amplitude of the vector signal .
  • the power spectrum density ( PSD) of the motion sensor output signal is determined .
  • a frequency analysis of the power spectrum density is performed to estimate a cadence and/or other parameters , for example an entropy of the signal .
  • the cadence indicates the rhythm of motion of a person .
  • the motion class is determined depending on the normali zed motion sensor output signal and the cadence .
  • the processing circuit is configured to determine a first one of the motion classes when the processing circuit detects , based on the evaluation of the motion sensor output signal , that the person is in the resting state .
  • the processing circuit is further configured to determine at least a second one of the motion classes when the processing circuit detects , based on the evaluation of the motion sensor output signal , that the person is engaged in an activity .
  • the at least one second motion class is thus determined in dependence on a detected kind of movement of the person .
  • a second motion class is assigned to the person when, for example , the person is performing a movement with low rhythm .
  • the person is assigned higher motion classes , for example a third, fourth or an even higher motion class .
  • the processing circuit is configured to determine the output signal in a time domain by evaluating the optical sensor output signal , when the processing circuit detects that the person is in the resting state .
  • the output signal indicating the heart rate can be determined directly in the time domain from the optical sensor output signal without considering the motion sensor output signal .
  • the photoplethysmogram is evaluated to determine the heart rate in the resting state of the person .
  • the heart rate is determined in the time domain, for example , by evaluating inter-beat-intervals ( IBI ) .
  • IBI inter-beat-intervals
  • the heart rate is determined by evaluating both of the optical sensor output signal and the motion sensor output signal .
  • the processing circuit is configured to determine a heart rate zone in dependence on the determined motion class . Depending on the determined intensity of the motion and thus the motion class , fixed heart rate zones are selected for further estimation of the heart rate .
  • the processing is configured to provide an approximate value of the heart rate by evaluating a heart rate model dedicated to the determined motion class .
  • the heart rate model may be dependent from at least a determined energy, the determined cadence of the power spectral density of the motion sensor output signal and the determined heart rate zone .
  • the energy may be determined from the normali zed motion sensor output signal .
  • the output of the heart rate model i . e . the approximate value of the heart rate calculated by applying the heart rate model , is a gross estimate of the heart rate of a person in motion .
  • the processing circuit is configured to determine a first heart rate estimation value in a frequency domain by evaluating a power spectral density of the optical sensor output signal in dependence on the approximate value of the heart rate model .
  • the first heart rate estimation value is calculated in the frequency domain by the processing circuit , i f the processing circuit detects that the person is in a motion state di f ferent from the resting state , i . e . the person is engaged in an activity and a motion class greater than one has been assigned to the person .
  • the power spectral density of the optical sensor output signal is determined, for example from the filtered optical sensor output signal .
  • frequency selection is applied to the power spectral density of the optical sensor output signal .
  • only certain frequency ranges are selected from the power spectral density of the optical sensor output signal for further analysis in the frequency domain .
  • the processing circuit is further configured to determine a second heart rate estimation value in a time domain by evaluating the optical sensor output signal in dependence on the approximate value of the heart rate model .
  • the second heart rate estimation value is calculated in the time domain by the processing circuit , i f the processing circuit detects that the person is in a motion state di f ferent from the resting state , i . e . the person is engaged in an activity and a motion class greater than one has been assigned to the person .
  • the selection function which depends on the output of the heat rate , i . e . the approximate value of the heart rate determined by the heart rate model , is applied to the optical sensor output signal in the time domain . Subsequently, filtering is performed to filter out spurious signal components .
  • the heart rate can then be determined in the time domain by evaluating, for example , the inter-beatinterval .
  • the processing circuit is configured to determine the ( final ) output signal in dependence on a fusion of the first heart rate estimation value and the second heart rate estimation value .
  • the final output signal representing the heart rate value is calculated by the processing circuit by a fusion of the first heart rate estimation value calculated in the frequency domain and the second heart rate estimation value calculated in the time domain, i f the processing circuit detects that the person is in a motion state di f ferent from the resting state , i . e . the person is engaged in an activity and a motion class greater than one has been assigned to the person .
  • two heart rate estimations are available , one from analyzing the power spectral density and one coming from the inter-beat-interval .
  • a fusion of the first heart rate estimation value determined in the frequency domain by evaluating the power spectral density, and the second heart rate estimation value determined in the time domain from the inter-beat-interval is performed .
  • the processing circuit is configured to calculate a signal quality index value of the optical sensor output signal , when the person is engaged in an activity .
  • the signal quality index value can be extracted from the raw output of the optical sensor or the filtered optical sensor output signal , when the person is in motion .
  • An embodiment of a method for reliably determining the heart rate of a person in motion with high accuracy is speci fied in claim 11 .
  • an optical sensor output signal suitable for creating a photoplethysmogram of a person is provided .
  • a motion sensor output signal is provided in response to a detected resting or at least one motion state of the person .
  • the photoplethysmogram and the motion sensor output signal are evaluated, and an output signal representing the heart rate of the person is provided in response to the evaluation of the photoplethysmogram and the motion sensor output signal , when the at least one motion state is detected for the person .
  • the photoplethysmogram is evaluated, and the output signal representing the heart rate is provided in response to the evaluation of the photoplethysmogram, when the resting state is detected for the person .
  • the proposed method thus allows the heart rate of a person to be determined i f the person is in motion by evaluating an output signal from an optical sensor, particularly a PPG sensor, and a motion sensor, particularly an accelerometer, i f it is detected that the person is in motion .
  • an optical sensor particularly a PPG sensor
  • a motion sensor particularly an accelerometer
  • the heart rate estimation is performed based on the evaluation of the optical sensor output signal only .
  • a respective one of several motion classes assigned to the person is determined in dependence on the motion sensor output signal .
  • Each motion class represents a respective kind of motion of the person .
  • the output signal of the motion sensor may be filtered to remove spurious components from the motion sensor output signal .
  • the motion sensor output signal in particular the output signal from the accelerometer, which is a three-dimensional vector signal is normali zed to obtain a one-dimensional signal .
  • an analysis in the frequency domain is carried out by calculating a power spectral density from the normali zed motion sensor output signal .
  • further parameters for example a cadence , are extracted .
  • the cadence indicates the frequency of a movement of the person .
  • the motion class is determined based on the normali zed one-dimensional motion sensor output signal ' s features and the cadence .
  • the output signal representing the heart rate is determined in the time domain by evaluating the optical sensor output signal , when a first motion class representing a resting state of the person is determined . That means that in the case that the person is in a resting state , the output signal representing the heart rate is determined by only evaluating the optical sensor signal , i . e . the photoplethysmogram obtained from the optical sensor output signal .
  • the output signal representing the heart rate may be determined in the time domain by evaluating an inter-beat-interval from the photoplethysmogram .
  • an approximate value of the heart rate is provided by evaluating a heart rate model dedicated to the determined motion class .
  • a heart rate zone is assigned to the person depending on the detected state of movement and in particular the intensity of the motion .
  • an estimate of the heart rate is determined .
  • the approximate value determined with the heart rate model is dependent on the heart rate zone , a cadence determined from the power density spectrum of the optical sensor output signal and the energy of the optical sensor output signal .
  • a first heart rate estimation value is determined in a frequency domain by evaluating a power spectral density of the optical sensor output signal in dependence on the approximate value of the heart rate model .
  • a selection function which is dependent on the output approximate value of the heart rate model , is applied to the power spectral density of the motion sensor output signal . To determine the first heart rate estimation value , only certain frequency ranges of the power density spectrum which are determined by the selection function are considered .
  • a second heart rate estimation value is determined in a time domain by evaluating the optical sensor output signal in dependence on the approximate value of the heart rate model .
  • the selected regions of the power density spectrum of the optical sensor output signal are trans formed back into the time domain .
  • the second heart rate estimation value is then determined in the time domain by evaluating the inter-beatinterval .
  • the ( final ) output signal representing the heart rate is determined in dependence on a fusion of the first heart rate estimation value and the second heart rate estimation value .
  • a signal quality index value of the optical sensor output signal is calculated when the person is engaged in an activity .
  • the signal quality index value is calculated from the raw output of the optical sensor or the filtered optical sensor output signal .
  • Figure 1 shows an embodiment of an apparatus for determining the heart rate of a person in motion
  • Figure 2 illustrates the schematics of a method/algorithm for determining the heart rate of a person in motion
  • Figure 3 illustrates the schematics of a heart rate analysis and classi fication including multiple sensors ; and Figure 4 shows the schematics of possible positions for sensors used for determining the heart rate of a person in motion .
  • an embodiment of an apparatus 1 for determining the heart rate of a person in motion comprises an optical sensor 10 , a motion sensor 20 and a processing circuit 30 to provide an output signal HR representing the heart rate of the person .
  • the optical sensor 10 and the motion sensor 20 are each configured to be attached to the person whose heart rate is to be determined .
  • the optical sensor 10 provides an optical sensor output signal OS from which the processing circuit 30 can create a photoplethysmogram of the person to be examined .
  • the optical sensor may be embodied as a PPG (photoplethysmography) sensor .
  • the PPG sensor comprises at least one pair of an emitter device and a photodetector device for PPG measurements in reflection or transmission configuration .
  • the motion sensor 20 is configured to provide a motion sensor output signal MS in response to a detected resting state or a detected at least one motion state of the person to be examined .
  • the motion sensor 20 may be configured as an accelerometer which provides a three-dimensional vector output signal MS .
  • the processing circuit 30 is configured to evaluate the optical sensor output signal OS and the motion sensor output signal MS based on an algorithm/method for determining the heart rate of the person in motion illustrated in Figure 2 .
  • the data of the optical sensor output signal and the data of the motion sensor output signal may be co-sampled, i . e . simultaneously sampled at the same rate or di f ferent rates or time aligned, with a following trans fer to pre-processing and processing units .
  • the processing circuit 30 provides the output signal HR representing the heart rate of the person in response to the evaluation of the photoplethysmogram and the motion sensor output signal MS , when the at least one motion state is detected for the person, i . e . the person performs a movement .
  • the processing circuit 30 is further configured to provide the output signal HR in response to the evaluation of the photoplethysmogram only, i . e . without considering the optical sensor output signal , when it is detected by the processing circuit that the person is in the resting state .
  • the optical sensor 10 particularly the PPG sensor, provides the optical sensor output signal OS and the motion sensor 20 , particularly the accelerometer, provides the motion sensor output signal MS in response to a detected motion of the person .
  • the evaluation of the optical sensor output signal OS and the motion sensor output signal MS is performed by the processing circuit 30 .
  • the processing circuit 30 calculates a signal quality index value SQI of the optical sensor output signal .
  • the processing circuit 30 is configured to calculate the signal quality index value SQI by evaluating the optical sensor output signal OS , when the person to be examined is engaged in an activity .
  • the optical sensor output signal OS is subj ected to filtering by a bandpass filter .
  • the filtering eliminates spurious or unwanted frequencies from the optical sensor output signal OS that are not needed to determine the heart rate value HR or the SQI value .
  • the signal quality index value SQI can be extracted from the output of the bandpass filter . The signal quality index value is thus determined from the raw output of the optical sensor .
  • the motion sensor output signal MS provided by the motion sensor 20 is filtered by a bandpass filter to also remove spurious signal components which are not needed for further evaluation .
  • the motion sensor output signal MS may be configured as a three- dimensional signal , for example a vector having three dimensions .
  • the amplitude of the motion sensor output signal MS is calculated by normali zation . After normali zation, the result is a one-dimensional signal that is then further evaluated by the processing circuit 30 .
  • the processing circuit 30 determines a motion class assigned to the person in dependence on the motion sensor output signal MS or the normali zed motion sensor output signal .
  • the algorithm may classi fy motion according to the intensity and the rhythmicity of the motion signals . This allows to further di f ferentiate between stable rhythmical activities such as walking, running cycling from unstable arrhythmical situations such as burst walking, running, and cycling for example .
  • Each motion class k represents a respective kind of motion or resting state of the person .
  • the processing circuit 30 is further configured to determine a higher motion class , for example a second, third, fourth, etc .
  • motion class k 2 , 3 , 4 , etc . when the processing circuit 30 detects , based on the evaluation of the motion sensor output signal MS or the normali zed motion sensor output signal , that the person is engaged in an activity .
  • the proposed apparatus and method for determining the heart rate of a person in motion enables precise determination of the heart rate when a person is moving or even when no movement is being performed .
  • the heart rate can be determined in the time domain by the processing circuit 30 by evaluating an inter-beat-interval in the optical sensor output signal OS or the filtered optical sensor output signal .
  • the inter-beat-interval is extracted from the filtered optical sensor output signal .
  • all unreliable intervals caused, for example , by arrhythmia of the heart are removed by a subsequent filtering operation .
  • the heart rate HR may be determined in the time domain as an average of the previously determined inter-beat-intervals .
  • the motion sensor output signal MS is further analyzed to determine the heart rate .
  • the processing circuit 30 determines a heart rate zone HZ in dependence on the previously determined motion class .
  • the heart rate zone HZ indicates the heart beats per minute to be expected for a speci fic movement of the person .
  • the amplitude or rhythm of the motion sensor output signal OS is changing, the heart rate zone will also be changed .
  • a transition matrix may be provided to initiate the change of motion from class to class .
  • the transition matrix defines the transition from one motion state to another motion state , when the heart rate goes up or down .
  • heart rate zones may be defined to enhance the search for the heart rate in low PPG SQI cases .
  • the signal quality index SQI determined from the optical sensor output signal can trigger additional activation of heart rate zones searched .
  • a power spectral density and an energy of the signal are determined based on the normali zed motion sensor output signal . Then, a frequency analysis is performed from the power spectral density to estimate a cadence which indicates a rhythm of motion .
  • the algorithm estimates the cadence of the user while in either of the motion classes , for example , walking, j ogging, running, or cycling .
  • the cadence can be further used for the estimation of the energy expenditure and to adj ust the heart rate zones , as will be explained below . For example , a lower value of the cadence indicates a slower movement , while a higher value of the cadence indicates a higher rhythm of the movement .
  • further parameters for example an entropy of the signal , can be extracted .
  • the processing circuit 30 provides an approximate value AV of the heart rate by evaluating a heart rate model dedicated to the determined motion class .
  • the algorithm makes use of a biomechanical heart rate estimation model which accounts for a typical heart rate curve mani fested during aerobic exercises , but can be extended to other anaerobic sports .
  • the heart rate model is dependent from at least the energy of the motion sensor output signal , the cadence of the power spectral density of the motion sensor output signal and the determined heart rate zone HZ .
  • the heart rate model is thus based on the motion sensor output signal .
  • the heart rate model From the heart rate zone , the energy and the cadence , the heart rate model provides a reasonable estimation of the heart rate when the person is moving . In particular, no extraction of the heart rate in the time domain is possible when the signal quality of the optical sensor output signal OS is low . However, due to the heart rate model using the defined heart rate zones , it is possible to determine a speci fic zone for the heart rate .
  • the heart rate model is used as a guidance in di f ferent heart rate zones for di f ferent activity classes such as , but not limited to walking, running, and cycling and also for a resting state (including sleeping) .
  • a first heart rate estimation value HRF is determined by the processing circuit 30 in the frequency domain
  • a second heart rate estimation value HRT is determined by the processing circuit 30 in a time domain .
  • the power spectral density of the optical sensor output signal OS or the filtered optical sensor output signal is determined .
  • the first heart rate estimation value HRF is determined by the processing circuit 30 in the frequency domain by evaluating the power spectral density of the optical sensor output signal or the filtered optical sensor output signal in dependence on the approximate value HV of the heart rate model .
  • the output of the HR model is used for refining the heart rate zone search to search for the best zone in the power spectral density of the optical sensor output signal or the filtered optical sensor output signal .
  • the particular zone in the frequency spectrum of the optical sensor output signal or the filtered optical sensor output signal is selected by using a selection function .
  • the selection allows some frequency bands in the power spectral density to be eliminated based on the output of the heart rate model .
  • the highest frequency peak in the selected zone of the power spectral density is determined as the first heart rate estimation value HRF in the frequency domain .
  • the algorithm has a smart initiali zation phase based on frequency estimation and a reliability index of the PPG and motion spectral content while in resting class .
  • the optical sensor output signal OS or the filtered optical sensor output signal is evaluated in dependence on the approximate value AV of the heart rate model .
  • the selected zone of the power density spectrum is trans formed in the time domain .
  • an inter-beat-interval can be determined, and after a subsequent filtering the second heart rate estimation value HRT is determined in the time domain .
  • the processing circuit 30 determines the output signal HR representing the heart rate with high accuracy in dependence on a fusion of the first heart rate estimation value HRF which has been determined in the frequency domain, and the second heart rate estimation value HRT which has been determined in the time domain .
  • a heart rate tracking functionality may be applied to more precisely determine the selected zone in the power spectral density of the optical sensor output signal or the filtered optical sensor output signal .
  • the heart rate tracking thus allows the zone of the determined heart rate to be limited from one observation window/ frame to another observation window/ frame .
  • the heart rate estimation in the frequency domain is performed by a selective peak spectrum search from both the PPG and motion spectrum .
  • the bandwidth for the heart rate zone search varies in function of the activity class and the signal quality index SQI .
  • the tracking is finally fine-tuned for maximum accuracy using an exponential narrowing of the heart rate zone search during sports activities , for example during aerobic walking, after a certain time being in a speci fic activity class .
  • the fine exponential tuning is reset when a change of class is detected, as for example from walking to j ogging to running and back .
  • a class transition matrix can be used to further improve the heart rate tracking in adapting the heart rate zone search .
  • the algorithm uses an adaptive Q notch filter bank technique which allows for an accurate tracking of the heart rate depending on the signal quality index SQI . Additionally, the algorithm uses a high pass filter on the PPG sensor output signal based on previous reliable (high SQI ) HR (heart rate ) estimation .
  • the algorithm uses an explicit frequency overlap flag that indicates whether the potential HR candidate is close or far from any motion harmonics .
  • the minimum distance between motion and HR frequencies to flag an overlap is adj ustable to avoid the notch filters to remove the HR information during the motion artefact management .
  • the signal quality index is used not only in the later fusion strategy of the HR estimations , but to modulate the HR estimations and to avoid HR estimation sudden unphysiological j umps .
  • the algorithm for determining the heart rate is described by using two sensors , i . e . an optical sensor 10 , particularly a PPG sensor, and a motion sensor 20 , particularly an accelerometer . Nevertheless, the method for determining the heart rate of a person in motion described above with reference to Figure 2 is not limited to the use of only two sensors .
  • other sensors 40 such as ambient light sensors , spectral surface sensors , humidity sensors , pressure sensors , temperature sensors , bio impedance sensors , galvanic skin response sensors , gyrometers , etc . can be used in addition to the optical sensor/PPG sensor 10 and the motion sensor/accelerometer 30, as shown in Figure 3.
  • a signal processing and parameter estimation is carried out by evaluating the output signals of the respective sensors. Based on the signal processing of the output signals of the motion sensor 20/accelerometer and the other sensors 40, a respective classification (shown in Figure 3 as 'classification 1' and 'classification 2' ) is carried out. Subsequently, the heart rate is estimated depending on the 'classification 1' and the 'classification 2’ and the signal processing of the optical sensor output signal by means of a suitable heart rate model. After an additional classification (shown in Figure 2 as 'classification 3' ) , a signal post-processing and an analysis are performed to determine the final output value of the heart rate.
  • a signal processing and parameter estimation is carried out by evaluating the output signals of the respective sensors. Based on the signal processing of the output signals of the motion sensor 20/accelerometer and the other sensors 40, a respective classification (shown in Figure 3 as 'classification 1' and 'classification 2' ) is carried out. Subsequently, the heart
  • Figure 4 shows general schematics with different positions Pl, ..., P8 for the individual sensors 10, 20 and 40 on the body of a person.
  • the position of the motion sensor/accelerometer influences the performance of the heart rate-in-motion algorithm of Figure 2.
  • the information on the exact position of the sensors can be used to improve the motion classifications and consequently the heart rate model and the heart rate analysis.
  • the respective signal quality index of the optical sensor output signal (PPG SQI) and the signal quality index of the inter-beat-interval ( IBI SQI ) may be used for further classi fication and detection .
  • Possible embodiments are classi f ication/detection of heart dys function or respiration and circulatory dys function .
  • the activity classi bomb together with the transition matrix and the HR model as well as the information from other sensors or history about user activities may be used for fatigue detection and thus prevention of inj uries during exercise .
  • the design of the apparatus and the method for determining the heart rate of a person in motion is not limited to the disclosed embodiments , and gives examples of many alternatives as possible for the features included in the embodiments discussed .
  • any modi fications , equivalents and substitutions of the disclosed concepts be included within the scope of the claims which are appended hereto .

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Veterinary Medicine (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • Physics & Mathematics (AREA)
  • Public Health (AREA)
  • Pathology (AREA)
  • Physiology (AREA)
  • Signal Processing (AREA)
  • Cardiology (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Psychiatry (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

An apparatus (1) for determining a heart rate of a person in motion comprises an optical sensor (10) being configured to provide an optical sensor output signal (OS) suitable for creating a photoplethysmogram of the person, and a motion sensor (20) being configured to detect a resting or at least one motion state of the person. The apparatus (1) further comprises a processing circuit (30) to provide an output signal (HR) representing the heart rate of the person. The processing circuit (30) is configured to provide the output signal (HR) in response to the evaluation of the photoplethysmogram and a motion sensor output signal (MS), when the at least one motion state is detected for the person, and to provide the output signal (HR) in response to the evaluation of the photoplethysmogram, when the resting state is detected for the person.

Description

Description
APPARATUS FOR DETERMINING THE HEART RATE OF A PERSON IN MOTION
Technical Field
The disclosure relates to an apparatus for determining the heart rate of a person in motion . Furthermore , the disclosure relates to a method for determining the heart rate of a person in motion .
Background
The heart rate of a person can be measured by a variety of methods . One of these methods is based on photoplethysmography, an optical method that uses light to measure changes in blood volume . However, determining heart rate using photoplethysmography is di f ficult when the person is in motion, because the estimation of the heart rate depends on the activity of the subj ect . Accurate heart rate estimation by evaluating a photoplethysmogram recorded by a PPG (photoplethysmography) sensor is di f ficult without additional hardware components and suitable algorithms and leads to erroneous measurement results .
It would be welcome in the art to provide an apparatus for reliably determining the heart rate of a person in motion with high accuracy . There is also a desire to provide a method for reliably determining the heart rate of a person in motion with high accuracy .
Summary An apparatus for determining the heart rate of a person with high accuracy, when the person performs an activity, for example a movement , is speci fied in claim 1 .
The proposed apparatus for determining the heart rate of a person in motion comprises an optical sensor, a motion sensor and a processing circuit to provide an output signal representing the heart rate of the person . The optical sensor is configured to provide an optical sensor output signal suitable for creating a photoplethysmogram of the person . The motion sensor is configured to detect a resting state or at least one motion state of the person, and to provide a motion sensor output signal in response to the detected resting or at least one motion state of the person . The processing circuit is configured to provide the output signal in response to the evaluation of the photoplethysmogram and the motion sensor output signal , when the at least one motion state is detected for the person . The processing circuit is further configured to provide the output signal in response to the evaluation of the photoplethysmogram, when the resting state is detected for the person .
The optical sensor may be configured as a PPG sensor . The motion sensor may be embodied as an ACC ( accelerometer ) . The processing circuit evaluates the motion sensor output signal for performing a classi fication of a motion of the person prior to the heart rate analysis . I f the evaluation of the motion sensor output signal shows that the person is in a resting state , i . e . not performing any movement , the heart rate is determined by the processing circuit only by evaluating the optical sensor output signal or the photoplethysmogram and is thus determined without taking the motion sensor output signal into account. On the other hand, if the evaluation of the motion sensor output signal shows that the person is in at least one motion state, i.e. performing any movement, the heart rate is determined by the processing circuit by evaluating the optical sensor output signal or the photoplethysmogram as well as the motion sensor output signal. In this case, a heart rate model that is dedicated to a specific motion class is applied to provide an approximate value of the heart rate.
The heart rate analysis is based on different models as classified from motion classes. The motion classification may be improved by gaining motion information from more than one motion sensor, if available, located at different body positions and calibrated for a particular position.
According to a possible embodiment of the apparatus for determining the heart rate of a person in motion, the processing circuit is configured to determine a respective one of several motion classes assigned to the person in dependence on the motion sensor output signal. Each motion class represents a respective kind of motion of the person, i.e. a motion state or also a resting state of the person.
In order to determine the motion class, the motion sensor output signal may be subjected to filtering. By applying a filter function to the motion sensor output signal, spurious signal components are filtered out of the motion sensor output signal.
The motion sensor output signal provided, for example, by an accelerometer, is a three-dimensional vector signal having components in different spatial directions. By normalizing this three-dimensional vector signal , a one-dimensional signal is determined that indicates , for example , the amplitude of the vector signal .
After normali zing the filtered motion sensor output signal , the power spectrum density ( PSD) of the motion sensor output signal is determined . Then, a frequency analysis of the power spectrum density is performed to estimate a cadence and/or other parameters , for example an entropy of the signal . The cadence indicates the rhythm of motion of a person . The motion class is determined depending on the normali zed motion sensor output signal and the cadence .
According to a possible embodiment of the apparatus for determining the heart rate of a person, the processing circuit is configured to determine a first one of the motion classes when the processing circuit detects , based on the evaluation of the motion sensor output signal , that the person is in the resting state . The processing circuit is further configured to determine at least a second one of the motion classes when the processing circuit detects , based on the evaluation of the motion sensor output signal , that the person is engaged in an activity . The at least one second motion class is thus determined in dependence on a detected kind of movement of the person .
Assuming that the first motion class is assigned to the person when the person is in the resting state , a second motion class is assigned to the person when, for example , the person is performing a movement with low rhythm . Depending on the determined movement rhythm, in particular the intensity of the movement rhythm, the person is assigned higher motion classes , for example a third, fourth or an even higher motion class .
According to a possible embodiment of the apparatus for determining the heart rate of a person, the processing circuit is configured to determine the output signal in a time domain by evaluating the optical sensor output signal , when the processing circuit detects that the person is in the resting state . Thus , i f it is determined that the person is in the resting state , i . e . the assigned motion class is , for example , k = 1 , the output signal indicating the heart rate can be determined directly in the time domain from the optical sensor output signal without considering the motion sensor output signal .
In the case of an optical sensor being configured as a PPG sensor, for example , the photoplethysmogram is evaluated to determine the heart rate in the resting state of the person . After filtering the optical sensor output signal to remove spurious components from the signal , the heart rate is determined in the time domain, for example , by evaluating inter-beat-intervals ( IBI ) . Moreover, in the case of no motion, the proposed apparatus of fers an option to apply classi fication of a PPG signal according to physiologic or medical aspects ( disease-related, etc . ) .
On the other hand, i f the person is in motion and has been assigned a motion class greater than one , the heart rate is determined by evaluating both of the optical sensor output signal and the motion sensor output signal . In this case , according to a possible embodiment of the apparatus for determining the heart rate of a person in motion, the processing circuit is configured to determine a heart rate zone in dependence on the determined motion class . Depending on the determined intensity of the motion and thus the motion class , fixed heart rate zones are selected for further estimation of the heart rate .
According to an embodiment of the apparatus for determining the heart rate of a person in motion, the processing is configured to provide an approximate value of the heart rate by evaluating a heart rate model dedicated to the determined motion class . The heart rate model may be dependent from at least a determined energy, the determined cadence of the power spectral density of the motion sensor output signal and the determined heart rate zone . The energy may be determined from the normali zed motion sensor output signal . The output of the heart rate model , i . e . the approximate value of the heart rate calculated by applying the heart rate model , is a gross estimate of the heart rate of a person in motion .
According to a possible embodiment of the apparatus for determining the heart rate of a person in motion, the processing circuit is configured to determine a first heart rate estimation value in a frequency domain by evaluating a power spectral density of the optical sensor output signal in dependence on the approximate value of the heart rate model . The first heart rate estimation value is calculated in the frequency domain by the processing circuit , i f the processing circuit detects that the person is in a motion state di f ferent from the resting state , i . e . the person is engaged in an activity and a motion class greater than one has been assigned to the person .
For this purpose , the power spectral density of the optical sensor output signal is determined, for example from the filtered optical sensor output signal . Depending on the approximate value of the heart rate resulting from the heart rate model , frequency selection is applied to the power spectral density of the optical sensor output signal . As a result , only certain frequency ranges are selected from the power spectral density of the optical sensor output signal for further analysis in the frequency domain .
According to an embodiment of the apparatus for determining the heart rate of a person in motion, the processing circuit is further configured to determine a second heart rate estimation value in a time domain by evaluating the optical sensor output signal in dependence on the approximate value of the heart rate model . The second heart rate estimation value is calculated in the time domain by the processing circuit , i f the processing circuit detects that the person is in a motion state di f ferent from the resting state , i . e . the person is engaged in an activity and a motion class greater than one has been assigned to the person .
For this purpose , the selection function, which depends on the output of the heat rate , i . e . the approximate value of the heart rate determined by the heart rate model , is applied to the optical sensor output signal in the time domain . Subsequently, filtering is performed to filter out spurious signal components . The heart rate can then be determined in the time domain by evaluating, for example , the inter-beatinterval .
According to an embodiment of the apparatus for determining the heart rate of a person in motion, the processing circuit is configured to determine the ( final ) output signal in dependence on a fusion of the first heart rate estimation value and the second heart rate estimation value . The final output signal representing the heart rate value is calculated by the processing circuit by a fusion of the first heart rate estimation value calculated in the frequency domain and the second heart rate estimation value calculated in the time domain, i f the processing circuit detects that the person is in a motion state di f ferent from the resting state , i . e . the person is engaged in an activity and a motion class greater than one has been assigned to the person .
In conclusion, based on the proposed algorithm, two heart rate estimations are available , one from analyzing the power spectral density and one coming from the inter-beat-interval . For determining the accurate heart rate value , a fusion of the first heart rate estimation value determined in the frequency domain by evaluating the power spectral density, and the second heart rate estimation value determined in the time domain from the inter-beat-interval is performed .
According to an embodiment of the apparatus for determining the heart rate of a person in motion, the processing circuit is configured to calculate a signal quality index value of the optical sensor output signal , when the person is engaged in an activity . The signal quality index value can be extracted from the raw output of the optical sensor or the filtered optical sensor output signal , when the person is in motion .
An embodiment of a method for reliably determining the heart rate of a person in motion with high accuracy is speci fied in claim 11 . According to the proposed method, an optical sensor output signal suitable for creating a photoplethysmogram of a person is provided . Furthermore , a motion sensor output signal is provided in response to a detected resting or at least one motion state of the person . The photoplethysmogram and the motion sensor output signal are evaluated, and an output signal representing the heart rate of the person is provided in response to the evaluation of the photoplethysmogram and the motion sensor output signal , when the at least one motion state is detected for the person . Alternatively, the photoplethysmogram is evaluated, and the output signal representing the heart rate is provided in response to the evaluation of the photoplethysmogram, when the resting state is detected for the person .
The proposed method thus allows the heart rate of a person to be determined i f the person is in motion by evaluating an output signal from an optical sensor, particularly a PPG sensor, and a motion sensor, particularly an accelerometer, i f it is detected that the person is in motion . On the other hand, i f it is detected that the person is in a resting state , the heart rate estimation is performed based on the evaluation of the optical sensor output signal only .
According to a possible embodiment of the method for determining the heart rate of a person in motion, a respective one of several motion classes assigned to the person is determined in dependence on the motion sensor output signal . Each motion class represents a respective kind of motion of the person .
To determine the motion class , the output signal of the motion sensor may be filtered to remove spurious components from the motion sensor output signal . The motion sensor output signal , in particular the output signal from the accelerometer, which is a three-dimensional vector signal is normali zed to obtain a one-dimensional signal . After normali zation, an analysis in the frequency domain is carried out by calculating a power spectral density from the normali zed motion sensor output signal . From this , further parameters , for example a cadence , are extracted . The cadence indicates the frequency of a movement of the person . Based on the normali zed one-dimensional motion sensor output signal ' s features and the cadence , the motion class is determined .
According to an embodiment of the method for determining the heart rate of a person in motion, the output signal representing the heart rate is determined in the time domain by evaluating the optical sensor output signal , when a first motion class representing a resting state of the person is determined . That means that in the case that the person is in a resting state , the output signal representing the heart rate is determined by only evaluating the optical sensor signal , i . e . the photoplethysmogram obtained from the optical sensor output signal . The output signal representing the heart rate may be determined in the time domain by evaluating an inter-beat-interval from the photoplethysmogram .
According to an embodiment of the method for determining the heart rate of a person in motion, an approximate value of the heart rate is provided by evaluating a heart rate model dedicated to the determined motion class . When it is detected that the person is in a moving state , a heart rate zone is assigned to the person depending on the detected state of movement and in particular the intensity of the motion . Using the heart rate model , an estimate of the heart rate is determined . The approximate value determined with the heart rate model is dependent on the heart rate zone , a cadence determined from the power density spectrum of the optical sensor output signal and the energy of the optical sensor output signal .
A first heart rate estimation value is determined in a frequency domain by evaluating a power spectral density of the optical sensor output signal in dependence on the approximate value of the heart rate model . A selection function, which is dependent on the output approximate value of the heart rate model , is applied to the power spectral density of the motion sensor output signal . To determine the first heart rate estimation value , only certain frequency ranges of the power density spectrum which are determined by the selection function are considered .
A second heart rate estimation value is determined in a time domain by evaluating the optical sensor output signal in dependence on the approximate value of the heart rate model . To determine the second heart rate estimation value , the selected regions of the power density spectrum of the optical sensor output signal are trans formed back into the time domain . The second heart rate estimation value is then determined in the time domain by evaluating the inter-beatinterval .
The ( final ) output signal representing the heart rate is determined in dependence on a fusion of the first heart rate estimation value and the second heart rate estimation value .
According to a possible embodiment of the method for determining a heart rate of a person in motion, a signal quality index value of the optical sensor output signal is calculated when the person is engaged in an activity . The signal quality index value is calculated from the raw output of the optical sensor or the filtered optical sensor output signal .
Additional features and advantages of the apparatus and the method for determining the heart rate of a person in motion are set forth in the detailed description that follows . It is to be understood that both the foregoing general description and the following detailed description are merely exemplary, and are intended to provide an overview or framework for understanding the nature and character of the claims .
Brief Description of the Drawings
The accompanying drawings are included to provide further understanding, and are incorporated in, and constitute a part of , the speci fication . As such, the disclosure will be more fully understood from the following detailed description, taken in conj unction with the accompanying figures in which :
Figure 1 shows an embodiment of an apparatus for determining the heart rate of a person in motion;
Figure 2 illustrates the schematics of a method/algorithm for determining the heart rate of a person in motion;
Figure 3 illustrates the schematics of a heart rate analysis and classi fication including multiple sensors ; and Figure 4 shows the schematics of possible positions for sensors used for determining the heart rate of a person in motion .
Detailed Description of the Drawings
Referring to Figure 1 , an embodiment of an apparatus 1 for determining the heart rate of a person in motion comprises an optical sensor 10 , a motion sensor 20 and a processing circuit 30 to provide an output signal HR representing the heart rate of the person . The optical sensor 10 and the motion sensor 20 are each configured to be attached to the person whose heart rate is to be determined . The optical sensor 10 provides an optical sensor output signal OS from which the processing circuit 30 can create a photoplethysmogram of the person to be examined . The optical sensor may be embodied as a PPG (photoplethysmography) sensor . The PPG sensor comprises at least one pair of an emitter device and a photodetector device for PPG measurements in reflection or transmission configuration .
The motion sensor 20 is configured to provide a motion sensor output signal MS in response to a detected resting state or a detected at least one motion state of the person to be examined . The motion sensor 20 may be configured as an accelerometer which provides a three-dimensional vector output signal MS .
The processing circuit 30 is configured to evaluate the optical sensor output signal OS and the motion sensor output signal MS based on an algorithm/method for determining the heart rate of the person in motion illustrated in Figure 2 . The data of the optical sensor output signal and the data of the motion sensor output signal may be co-sampled, i . e . simultaneously sampled at the same rate or di f ferent rates or time aligned, with a following trans fer to pre-processing and processing units .
The processing circuit 30 provides the output signal HR representing the heart rate of the person in response to the evaluation of the photoplethysmogram and the motion sensor output signal MS , when the at least one motion state is detected for the person, i . e . the person performs a movement . The processing circuit 30 is further configured to provide the output signal HR in response to the evaluation of the photoplethysmogram only, i . e . without considering the optical sensor output signal , when it is detected by the processing circuit that the person is in the resting state .
The general algorithm architecture/method used by the apparatus 1 for determining the heart rate of a person in motion is explained in the following with reference to Figure 2 .
The optical sensor 10 , particularly the PPG sensor, provides the optical sensor output signal OS and the motion sensor 20 , particularly the accelerometer, provides the motion sensor output signal MS in response to a detected motion of the person . The evaluation of the optical sensor output signal OS and the motion sensor output signal MS is performed by the processing circuit 30 .
The processing circuit 30 calculates a signal quality index value SQI of the optical sensor output signal . In particular, the processing circuit 30 is configured to calculate the signal quality index value SQI by evaluating the optical sensor output signal OS , when the person to be examined is engaged in an activity . For this purpose , the optical sensor output signal OS is subj ected to filtering by a bandpass filter . The filtering eliminates spurious or unwanted frequencies from the optical sensor output signal OS that are not needed to determine the heart rate value HR or the SQI value . According to the algorithm architecture , the signal quality index value SQI can be extracted from the output of the bandpass filter . The signal quality index value is thus determined from the raw output of the optical sensor .
The motion sensor output signal MS provided by the motion sensor 20 , for example an accelerometer, is filtered by a bandpass filter to also remove spurious signal components which are not needed for further evaluation . The motion sensor output signal MS may be configured as a three- dimensional signal , for example a vector having three dimensions . In a subsequent step, the amplitude of the motion sensor output signal MS is calculated by normali zation . After normali zation, the result is a one-dimensional signal that is then further evaluated by the processing circuit 30 .
According to the proposed apparatus for determining the heart rate of a person in motion or the method for determining the heart rate of the person in motion, the processing circuit 30 determines a motion class assigned to the person in dependence on the motion sensor output signal MS or the normali zed motion sensor output signal . The algorithm may classi fy motion according to the intensity and the rhythmicity of the motion signals . This allows to further di f ferentiate between stable rhythmical activities such as walking, running cycling from unstable arrhythmical situations such as burst walking, running, and cycling for example .
Depending on the detected moving, especially the intensity of the movement , or a detected resting of the person di f ferent motion classes k are assigned to the person . Each motion class k represents a respective kind of motion or resting state of the person .
The processing circuit 30 determines , for example , a first motion class k = 1 , when the processing circuit 30 detects , based on the evaluation of the motion sensor output signal MS or the normali zed motion sensor output signal , that the person is in a resting state . The processing circuit 30 is further configured to determine a higher motion class , for example a second, third, fourth, etc . motion class k = 2 , 3 , 4 , etc . when the processing circuit 30 detects , based on the evaluation of the motion sensor output signal MS or the normali zed motion sensor output signal , that the person is engaged in an activity .
That means that the processing circuit 30 is configured to determine the motion class k assigned to the person in dependence on a detected kind of motion of the person . For example , i f it is determined, by the processing circuit 30 by evaluating the motion sensor output signal or the normali zed motion sensor output signal , that the person is performing a relatively slow movement , the person is assigned motion class k = 2 . As the intensity of the movement increases , the person is assigned classes of higher order, for example a third, fourth or even higher motion class , by evaluating the motion sensor output signal or the normali zed motion sensor output signal . The proposed apparatus and method for determining the heart rate of a person in motion enables precise determination of the heart rate when a person is moving or even when no movement is being performed . I f , for example , the first class k = 1 is determined during the step of assigning a motion class to the person, i . e . the person is not performing any activity, according to the proposed approach, the heart rate is determined in the time domain . The heart rate can be determined in the time domain by the processing circuit 30 by evaluating an inter-beat-interval in the optical sensor output signal OS or the filtered optical sensor output signal .
Referring to Figure 2 , the inter-beat-interval is extracted from the filtered optical sensor output signal . In a subsequent step all unreliable intervals caused, for example , by arrhythmia of the heart , are removed by a subsequent filtering operation . The heart rate HR may be determined in the time domain as an average of the previously determined inter-beat-intervals .
When it is determined by the processing circuit 30 that the person is in motion or performs a movement , i . e . the motion class k is determined to be greater than 1 , the motion sensor output signal MS is further analyzed to determine the heart rate . Referring to Figure 2 , the processing circuit 30 determines a heart rate zone HZ in dependence on the previously determined motion class . The heart rate zone HZ indicates the heart beats per minute to be expected for a speci fic movement of the person . When the amplitude or rhythm of the motion sensor output signal OS is changing, the heart rate zone will also be changed . In particular, a transition matrix may be provided to initiate the change of motion from class to class . The transition matrix defines the transition from one motion state to another motion state , when the heart rate goes up or down . Furthermore , heart rate zones may be defined to enhance the search for the heart rate in low PPG SQI cases . Moreover, the signal quality index SQI determined from the optical sensor output signal can trigger additional activation of heart rate zones searched .
Referring to the algorithm shown in Figure 2 , based on the normali zed motion sensor output signal , a power spectral density and an energy of the signal are determined . Then, a frequency analysis is performed from the power spectral density to estimate a cadence which indicates a rhythm of motion . The algorithm estimates the cadence of the user while in either of the motion classes , for example , walking, j ogging, running, or cycling . The cadence can be further used for the estimation of the energy expenditure and to adj ust the heart rate zones , as will be explained below . For example , a lower value of the cadence indicates a slower movement , while a higher value of the cadence indicates a higher rhythm of the movement . Using the analysis of the power density spectrum, further parameters , for example an entropy of the signal , can be extracted .
The processing circuit 30 provides an approximate value AV of the heart rate by evaluating a heart rate model dedicated to the determined motion class . The algorithm makes use of a biomechanical heart rate estimation model which accounts for a typical heart rate curve mani fested during aerobic exercises , but can be extended to other anaerobic sports . The heart rate model is dependent from at least the energy of the motion sensor output signal , the cadence of the power spectral density of the motion sensor output signal and the determined heart rate zone HZ . The heart rate model is thus based on the motion sensor output signal .
From the heart rate zone , the energy and the cadence , the heart rate model provides a reasonable estimation of the heart rate when the person is moving . In particular, no extraction of the heart rate in the time domain is possible when the signal quality of the optical sensor output signal OS is low . However, due to the heart rate model using the defined heart rate zones , it is possible to determine a speci fic zone for the heart rate . The heart rate model is used as a guidance in di f ferent heart rate zones for di f ferent activity classes such as , but not limited to walking, running, and cycling and also for a resting state ( including sleeping) .
I f the person under investigation has been assigned a motion class greater than 1 and it has thus been determined that the person is performing a movement , a first heart rate estimation value HRF is determined by the processing circuit 30 in the frequency domain, and a second heart rate estimation value HRT is determined by the processing circuit 30 in a time domain .
In order to determine the first heart rate estimation value HRF, the power spectral density of the optical sensor output signal OS or the filtered optical sensor output signal is determined . As illustrated in Figure 2 , the first heart rate estimation value HRF is determined by the processing circuit 30 in the frequency domain by evaluating the power spectral density of the optical sensor output signal or the filtered optical sensor output signal in dependence on the approximate value HV of the heart rate model . The output of the HR model is used for refining the heart rate zone search to search for the best zone in the power spectral density of the optical sensor output signal or the filtered optical sensor output signal .
The particular zone in the frequency spectrum of the optical sensor output signal or the filtered optical sensor output signal is selected by using a selection function . The selection allows some frequency bands in the power spectral density to be eliminated based on the output of the heart rate model . According to a possible embodiment of the apparatus or the method for determining the heart rate of the person in motion, the highest frequency peak in the selected zone of the power spectral density is determined as the first heart rate estimation value HRF in the frequency domain .
According to a possible embodiment , the algorithm has a smart initiali zation phase based on frequency estimation and a reliability index of the PPG and motion spectral content while in resting class .
To determine the second heart rate estimation value HRT in the time domain, the optical sensor output signal OS or the filtered optical sensor output signal is evaluated in dependence on the approximate value AV of the heart rate model . For this purpose , the selected zone of the power density spectrum is trans formed in the time domain . Then, an inter-beat-interval can be determined, and after a subsequent filtering the second heart rate estimation value HRT is determined in the time domain .
The processing circuit 30 then determines the output signal HR representing the heart rate with high accuracy in dependence on a fusion of the first heart rate estimation value HRF which has been determined in the frequency domain, and the second heart rate estimation value HRT which has been determined in the time domain .
According to a possible embodiment of the apparatus and the method for determining the heart rate of a person in motion, a heart rate tracking functionality may be applied to more precisely determine the selected zone in the power spectral density of the optical sensor output signal or the filtered optical sensor output signal . The heart rate tracking thus allows the zone of the determined heart rate to be limited from one observation window/ frame to another observation window/ frame .
Since the heart rate should not coincide with any motion harmonics , the heart rate estimation in the frequency domain is performed by a selective peak spectrum search from both the PPG and motion spectrum . The bandwidth for the heart rate zone search varies in function of the activity class and the signal quality index SQI . The tracking is finally fine-tuned for maximum accuracy using an exponential narrowing of the heart rate zone search during sports activities , for example during aerobic walking, after a certain time being in a speci fic activity class . The fine exponential tuning is reset when a change of class is detected, as for example from walking to j ogging to running and back . A class transition matrix can be used to further improve the heart rate tracking in adapting the heart rate zone search .
The algorithm uses an adaptive Q notch filter bank technique which allows for an accurate tracking of the heart rate depending on the signal quality index SQI . Additionally, the algorithm uses a high pass filter on the PPG sensor output signal based on previous reliable (high SQI ) HR (heart rate ) estimation .
According to a possible embodiment , the algorithm uses an explicit frequency overlap flag that indicates whether the potential HR candidate is close or far from any motion harmonics . The minimum distance between motion and HR frequencies to flag an overlap is adj ustable to avoid the notch filters to remove the HR information during the motion artefact management .
According to a possible embodiment , the signal quality index is used not only in the later fusion strategy of the HR estimations , but to modulate the HR estimations and to avoid HR estimation sudden unphysiological j umps .
In Figure 2 , the algorithm for determining the heart rate is described by using two sensors , i . e . an optical sensor 10 , particularly a PPG sensor, and a motion sensor 20 , particularly an accelerometer . Nevertheless , the method for determining the heart rate of a person in motion described above with reference to Figure 2 is not limited to the use of only two sensors . In principle , other sensors 40 , such as ambient light sensors , spectral surface sensors , humidity sensors , pressure sensors , temperature sensors , bio impedance sensors , galvanic skin response sensors , gyrometers , etc . can be used in addition to the optical sensor/PPG sensor 10 and the motion sensor/accelerometer 30, as shown in Figure 3.
Referring to Figure 3, after pre-processing, for example filtering, the output signals of the individual sensors 10, 20, 40, a signal processing and parameter estimation is carried out by evaluating the output signals of the respective sensors. Based on the signal processing of the output signals of the motion sensor 20/accelerometer and the other sensors 40, a respective classification (shown in Figure 3 as 'classification 1' and 'classification 2' ) is carried out. Subsequently, the heart rate is estimated depending on the 'classification 1' and the 'classification 2’ and the signal processing of the optical sensor output signal by means of a suitable heart rate model. After an additional classification (shown in Figure 2 as 'classification 3' ) , a signal post-processing and an analysis are performed to determine the final output value of the heart rate.
Figure 4 shows general schematics with different positions Pl, ..., P8 for the individual sensors 10, 20 and 40 on the body of a person. In particular, the position of the motion sensor/accelerometer influences the performance of the heart rate-in-motion algorithm of Figure 2. According to an advantageous embodiment, the information on the exact position of the sensors can be used to improve the motion classifications and consequently the heart rate model and the heart rate analysis.
In addition to the use for the determination of the heart rate, the respective signal quality index of the optical sensor output signal (PPG SQI) and the signal quality index of the inter-beat-interval ( IBI SQI ) may be used for further classi fication and detection . Possible embodiments are classi f ication/detection of heart dys function or respiration and circulatory dys function . The heart-related classi fication can be performed in the class "resting" ( k = 1 ) . Moreover, the activity classi fier together with the transition matrix and the HR model as well as the information from other sensors or history about user activities may be used for fatigue detection and thus prevention of inj uries during exercise .
The embodiments of the apparatus and the method for determining the heart rate of a person in motion disclosed herein have been discussed for the purpose of familiari zing the reader with novel aspects of the apparatus and the method . Although preferred embodiments have been shown and described, many changes , modi fications , equivalents and substitutions of the disclosed concepts may be made by one having skill in the art without unnecessarily departing from the scope of the claims .
In particular, the design of the apparatus and the method for determining the heart rate of a person in motion is not limited to the disclosed embodiments , and gives examples of many alternatives as possible for the features included in the embodiments discussed . However, it is intended that any modi fications , equivalents and substitutions of the disclosed concepts be included within the scope of the claims which are appended hereto .
Features recited in separate dependent claims may be advantageously combined . Moreover, reference signs used in the claims are not limited to be construed as limiting the scope of the claims .
Furthermore , as used herein, the term "comprising" does not exclude other elements . In addition, as used herein, the article "a" is intended to include one or more than one component or element , and is not limited to be construed as meaning only one . This patent application claims the priority of German patent application with application No . 10 2023 116 551 . 8 , the disclosure content of which is hereby incorporated by reference .
References
1 apparatus for determining a heart rate of a person in motion 10 optical sensor/PPG sensor
20 motion sensor/accelerometer
30 processing circuit
40 other sensors
OS optical sensor output signal MS motion sensor output signal
HZ heart rate zone
AV approximate value of the heart rate model
HRF first heart rate estimation value
HRT second heart rate estimation value SQI signal quality index value
IBI inter-beat-interval
Pl , P8 sensor position

Claims

Claims
1. Apparatus for determining a heart rate of a person in motion, comprising:
- an optical sensor (10) being configured to provide an optical sensor output signal (OS) suitable for creating a photoplethysmogram of the person,
- a motion sensor (20) being configured to detect a resting or at least one motion state of the person, and to provide a motion sensor output signal (MS) in response to the detected resting or at least one motion state of the person,
- a processing circuit (30) to provide an output signal (HR) representing the heart rate of the person,
- wherein the processing circuit (30) is configured to provide the output signal (HR) in response to the evaluation of the photoplethysmogram and the motion sensor output signal (MS) , when the at least one motion state is detected for the person,
- wherein the processing circuit (30) is configured to provide the output signal (HR) in response to the evaluation of the photoplethysmogram, when the resting state is detected for the person.
2. The apparatus of claim 1, wherein the processing circuit (30) is configured to determine a respective one of several motion classes (k) assigned to the person in dependence on the motion sensor output signal (MS) , each motion class (k) representing a respective kind of motion of the person.
3. The apparatus of claim 2,
- wherein the processing circuit (30) is configured to determine a first one of the motion classes (k) , when the processing circuit (30) detects, based on evaluation of the motion sensor output signal (MS) , that the person is in the resting state,
- wherein the processing circuit (30) is configured to determine at least a second one of the motion classes (k) , when the processing circuit (30) detects, based on evaluation of the motion sensor output signal (MS) , that the person is engaged in an activity.
4. The apparatus of claim 3, wherein the processing circuit (30) is configured to determine the output signal (HR) in a time domain by evaluating the optical sensor output signal (OS) , when the processing circuit (30) detects that the person is in the resting state.
5. The apparatus of any of the claims 2-4, wherein the processing circuit (30) is configured to determine a heart rate zone (HRZ) in dependence on the determined motion class.
6. The apparatus of claim 5, wherein the processing circuit (30) is configured to provide an approximate value (AV) of the heart rate by evaluating a heart rate model dedicated to the determined motion class.
7. The apparatus of claim 6, wherein the processing circuit (30) is configured to determine a first heart rate estimation value (HRF) in a frequency domain by evaluating a power spectral density of the optical sensor output signal (OS) in dependence on the approximate value (AV) of the heart rate model.
8 . The apparatus of claim 6 or 7 , wherein the processing circuit ( 30 ) is configured to determine a second heart rate estimation value (HRT ) in a time domain by evaluating the optical sensor output signal ( OS ) in dependence on the approximate value (AV) of the heart rate model .
9 . The apparatus of claim 8 , wherein the processing circuit ( 30 ) is configured to determine the output signal (HR) in dependence on a fusion of the first heart rate estimation value (HRF) and the second heart rate estimation value (HRT ) .
10 . The apparatus of any of the claims 1- 9 , wherein the processing circuit ( 30 ) is configured to calculate a signal quality index value ( SQI ) of the optical sensor output signal , when the person is engaged in an activity .
11 . A method for determining a heart rate of a person in motion, comprising :
- providing an optical sensor output signal ( OS ) suitable for creating a photoplethysmogram of the person;
- providing a motion sensor output signal (MS ) in response to a detected resting or at least one motion state of the person,
- evaluating the photoplethysmogram and the motion sensor output signal , and providing an output signal (HR) representing the heart rate of the person in response to the evaluation of the photoplethysmogram and the motion sensor output signal , when the at least one motion state is detected for the person, or, - evaluating the photoplethysmogram, and providing the output signal (HR) in response to the evaluation of the photoplethysmogram, when the resting state is detected for the person .
12 . The method of claim 11 , determining a respective one of several motion classes ( k) assigned to the person in dependence on the motion sensor output signal ( OS ) , each motion class ( k) representing a respective kind of motion of the person .
13 . The method of claim 12 , determining the output signal (HR) in a time domain by evaluating the optical sensor output signal ( OS ) , when a first motion class representing a resting state of the person is determined .
14 . The method of claim 12 or 13 ,
- providing an approximate value (AV) of the heart rate by evaluating a heart rate model dedicated to the determined motion class ( k) ,
- determining a first heart rate estimation value (HRF) in a frequency domain by evaluating a power spectral density of the optical sensor output signal ( OS ) in dependence on the approximate value (AV) of the heart rate model ,
- determining a second heart rate estimation value (HRT ) in a time domain by evaluating the optical sensor output signal ( OS ) in dependence on the approximate value (AV) of the heart rate model ,
- determine the output signal (HR) in dependence on a fusion of the first heart rate estimation value (HRF) and the second heart rate estimation value (HRT ) .
15. The method of any of the claims 11-14, comprising: calculating a signal quality index value (SQI) of the optical sensor output signal (OS) , when the person is engaged in an activity .
PCT/EP2024/066053 2023-06-23 2024-06-11 Apparatus for determining the heart rate of a person in motion WO2024260794A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE102023116551.8 2023-06-23
DE102023116551 2023-06-23

Publications (1)

Publication Number Publication Date
WO2024260794A1 true WO2024260794A1 (en) 2024-12-26

Family

ID=91539999

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP2024/066053 WO2024260794A1 (en) 2023-06-23 2024-06-11 Apparatus for determining the heart rate of a person in motion

Country Status (1)

Country Link
WO (1) WO2024260794A1 (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8945017B2 (en) * 2012-06-22 2015-02-03 Fitbit, Inc. Wearable heart rate monitor
US10278647B2 (en) * 2015-06-09 2019-05-07 University Of Connecticut Method and apparatus for removing motion artifacts from biomedical signals
WO2021252768A1 (en) * 2020-06-10 2021-12-16 Whoop, Inc. Wearable infection monitor

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8945017B2 (en) * 2012-06-22 2015-02-03 Fitbit, Inc. Wearable heart rate monitor
US10278647B2 (en) * 2015-06-09 2019-05-07 University Of Connecticut Method and apparatus for removing motion artifacts from biomedical signals
WO2021252768A1 (en) * 2020-06-10 2021-12-16 Whoop, Inc. Wearable infection monitor

Similar Documents

Publication Publication Date Title
US7727159B2 (en) Apparatus and method for detecting blood flow signal free from motion artifact and stress test apparatus using the same
US11540748B2 (en) Method and system for gait detection of a person
WO2012051300A2 (en) Method and system for detection and rejection of motion/noise artifacts in physiological measurements
US9386931B2 (en) System and method for receiving an indication of proper body locations of sensors on a patient
CN103034837B (en) Characteristic parameter is associated with pulse condition key element
CN105496416A (en) Human motion state recognition method and device
Gupta et al. Higher order derivative-based integrated model for cuff-less blood pressure estimation and stratification using PPG signals
US20120123689A1 (en) Methods and systems for discriminating bands in scalograms
CN106994010A (en) A kind of heart rate detection method and system based on PPG signals
US20100081898A1 (en) Detecting A Probe-Off Event In A Measurement System
KR101704491B1 (en) Heart rate measufing apparatus and method, recording medium for performing the method
TWI535415B (en) Method for monitoring a heart rate
CN105997043A (en) Pulse frequency extraction method based on wrist type wearable device
Ahmadi et al. Heart Rate monitoring during physical exercise using wrist-type photoplethysmographic (PPG) signals
Kumar et al. STSR: Spectro-temporal super-resolution analysis of a reference signal less photoplethysmogram for heart rate estimation during physical activity
KR20200066290A (en) Method and device for detecting stress using ECG function between beats
US11452483B2 (en) Apparatus for eliminating motion artifacts by using PPG signal and method thereof
JP2001198094A (en) Pulse rate detector
John et al. A generalized signal quality estimation method for IoT sensors
Carbonaro et al. Wearable biomonitoring system for stress management: A preliminary study on robust ECG signal processing
KR102342106B1 (en) Non-contact biosignal analysis system for body condition determination
Hao et al. Adaptive R-peak detection algorithm based on brown exponential smoothing model
Cherif et al. Detection of artifacts on photoplethysmography signals using random distortion testing
CN113274008B (en) A three-wavelength blood oxygen saturation monitoring method based on wearable devices
WO2024260794A1 (en) Apparatus for determining the heart rate of a person in motion

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 24732875

Country of ref document: EP

Kind code of ref document: A1