KR102214686B1 - Method and apparatus for accurately and consistently estimating a heart rate based on a finite state machine - Google Patents

Abstract

The present invention discloses a method and apparatus for estimating an accurate and consistent heart rate based on a finite state machine. An apparatus for estimating an accurate and consistent heart rate based on a finite state machine according to an aspect of the present invention may reduce a heart rate estimation error rate by estimating a heart rate based on a finite state machine, and provide an accurate and consistent heart rate.

Description

A method for estimating an accurate and consistent heart rate based on a finite state machine, and an apparatus for the same {METHOD AND APPARATUS FOR ACCURATELY AND CONSISTENTLY ESTIMATING A HEART RATE BASED ON A FINITE STATE MACHINE}

The present invention relates to a method and apparatus for estimating an accurate and consistent heart rate based on a finite state machine, and more particularly, to a finite state machine that provides accurate heart rate results by applying a finite state machine (FSM). It relates to a method and apparatus for estimating an accurate and consistent heart rate.

The heart rate is the most basic information among human health information, and changes in the body can be monitored through heart rate monitoring in a situation where interest in recent health is continuously increasing. At this time, the wearable device is a new and potential tool for home medical devices with a function that can measure vital signs anytime, anywhere. Such a wearable device may be provided in the form of a watch or a band, and a reflective PPG (PhotoPlethysmoGraphy) sensor that can be simply worn on a wrist is representative. As described above, when measuring heart rate using a wearable medical device such as a wearable device, accuracy is one of the most important issues. However, while typical wearable medical devices provide accurate heart rate measurements, they are limited only to certain conditions of rest, walking and low-intensity exercise. In addition, wearable medical devices employing reflective PPG sensors may be limited due to the low signal-to-noise ratio (SNR) caused by the large volume of skin, muscle, and fat and relatively small pulsating components of arterial blood. Is damaged under the influence of various motion artifacts (MA), resulting in inaccurate heart rate measurement results.

Korean Patent Registration No. 10-1850803 (announced on April 20, 2018)

The present invention has been proposed to solve the above problems, and reduces the heart rate estimation error rate by estimating the heart rate based on a finite state machine, and provides an accurate and consistent heart rate based on a finite state machine that provides an accurate and consistent heart rate. An object of the present invention is to provide an estimation method and an apparatus therefor.

Other objects and advantages of the present invention can be understood by the following description, and will be more clearly understood by an embodiment of the present invention. In addition, it will be easily understood that the objects and advantages of the present invention can be realized by the means shown in the claims and combinations thereof.

An apparatus for estimating an accurate and consistent heart rate based on a finite state machine according to an aspect of the present invention for achieving the above object includes a 2-channel optical blood flow analyzer signal measured by a reflective optical blood flow analyzer sensor, a 3-axis A signal acquisition unit acquiring an acceleration signal and a 1-channel ECG signal; A signal filtering unit for filtering the two-channel optical blood flow meter signal and the three-axis acceleration signal obtained by the signal acquisition unit using a Butterworth BPF; A signal normalizing unit for normalizing a two-channel optical blood flow meter signal among the filtered signals to a zero average having a unit variance and averaging using an average and a standard deviation; A signal down-sampling unit for down-sampling the normalized and averaged optical blood flow meter signal and the 3-axis acceleration signal to 25 hertz (Hz); A motion artifact removal unit that removes motion artifacts from a two-channel optical blood flow meter signal using the down-sampled acceleration signal; And generating a state transition according to a specific condition using a finite state machine (FSM) from the signal from which the motion artifact has been removed, and only the heart rate when the state is in a stable state to the user. It includes; heart rate information providing unit to provide.

The specific condition in which the state transition occurs is determined according to a crest factor and a value of a change in heart rate measured in a continuous window.

The motion artifact removal unit may apply a Wiener filter to estimate a clean optical blood flow analyzer signal spectrum by subtracting a noise signal spectrum from a measurement signal spectrum.

The two-channel optical blood flow meter signal is characterized in that it is measured from the user's wrist by two pulse oximeters having green LEDs.

The 3-axis acceleration signal is characterized by being measured from the user's wrist through a 3-axis accelerometer.

The ECG signal is characterized in that it is measured from the user's chest using a wet ECG sensor.

A method for estimating heart rate in a device for estimating an accurate and consistent heart rate based on a finite state machine according to another aspect of the present invention for achieving the above object is, a two-channel measurement by a reflective optical blood flow meter sensor. A signal acquisition step of acquiring an optical blood flow meter signal, a 3-axis acceleration signal, and a 1-channel ECG signal; A signal filtering step of filtering the two-channel optical blood flow meter signal and the three-axis acceleration signal obtained in the signal acquisition step by using a Butterworth BPF; A signal normalization step of normalizing a two-channel optical blood flow meter signal among the filtered signals to a zero average having a unit variance and averaging using the average and standard deviation; A signal down-sampling step of down-sampling the normalized and averaged optical blood flow meter signal and 3-axis acceleration signal to 25 hertz (Hz); A motion artifact removal step of removing motion artifacts from a two-channel optical blood flow meter signal using the down-sampled acceleration signal; And generating a state transition according to a specific condition using a finite state machine (FSM) from the signal from which the motion artifact has been removed, and only the heart rate when the state is in a stable state to the user. It includes; providing heart rate information providing step.

The specific condition in which the state transition occurs is determined according to a crest factor and a value of a change in heart rate measured in a continuous window.

The motion artifact removal step may include estimating a clean optical blood flow analyzer signal spectrum using a value obtained by subtracting a noise signal spectrum from a measurement signal spectrum by applying a Wiener filter.

According to an aspect of the present invention, there is an effect of remarkably reducing the heart rate estimation error rate.

The effects obtainable in the present invention are not limited to the above-mentioned effects, and other effects not mentioned will be clearly understood by those of ordinary skill in the technical field to which the present invention belongs from the following description. .

The following drawings attached to the present specification illustrate preferred embodiments of the present invention, and serve to further understand the technical idea of the present invention along with specific details for carrying out the present invention. It should not be interpreted as being limited to the information described.
1 is a schematic configuration diagram of an apparatus for estimating heart rate according to an embodiment of the present invention;
2 is a diagram schematically illustrating a state transition based on a finite state machine according to an embodiment of the present invention;
3 is a diagram illustrating a flow of a method for estimating a heart rate according to an embodiment of the present invention.

The above-described objects, features, and advantages will become more apparent through the following detailed description in connection with the accompanying drawings, whereby those of ordinary skill in the technical field to which the present invention pertains can easily implement the technical idea of the present invention. There will be. In addition, in describing the present invention, if it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the subject matter of the present invention, a detailed description thereof will be omitted. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Throughout the specification, when a certain part "includes" a certain component, it means that other components may be further included rather than excluding other components unless otherwise stated. In addition, “… A term such as “sub” means a unit that processes at least one function or operation, which may be implemented by hardware or software, or a combination of hardware and software.

1 is a schematic configuration diagram of an apparatus for estimating heart rate according to an embodiment of the present invention.

In describing the present embodiment, the heart rate during exercise may be monitored by using a wrist-type reflective PPG (PhotoPlethysmoGraphy) signal.

Referring to FIG. 1, the apparatus for estimating heart rate according to the present embodiment includes a signal acquisition unit 110, a signal filtering unit 120, a signal normalization unit 130, a signal downsampling unit 140, and a motion artifact generator. It includes a rejection 150 and a heart rate information providing unit 160.

The signal acquisition unit 110 acquires a 2-channel optical blood flow analyzer signal, a 3-axis acceleration signal, and a 1-channel ECG signal measured by a reflective optical blood flow analyzer sensor. At this time, the two-channel optical blood flow meter signal can be measured from the user's wrist by two pulse oximeters having green LEDs. The green LED has a wavelength of 515nm, and the 3-axis acceleration signal can be measured from the user's wrist through a 3-axis accelerometer. Here, the pulse oximeter and the 3-axis accelerometer (x, y, and z axes) may be sensors included in wearable medical devices such as a wearable device worn on a user's wrist. Accordingly, the user may wear a wearable medical device such as a wearable device on the wrist and measure a signal through the sensor. In addition, the ECG signal may be measured from the user's chest using a wet ECG sensor. Meanwhile, all of the acquired signals may be sampled at 125 Hz and may be transmitted to a server (not shown) through wireless communication such as Bluetooth.

The signal filtering unit 120 filters the two-channel optical blood flow meter signal and the three-axis acceleration signal acquired by the signal acquisition unit 110 using a Butterworth BPF. For example, using two PPG signals S1(i), S2(i) and three acceleration signals Ax(i), Ay(i), Az(i) in the i-th window, cutoffs of 0.4 Hz and 4 Hz Five signals can be filtered with a fourth-order Butterworth band pass filter (BPF) having a frequency.

The signal normalization unit 130 normalizes the two-channel optical blood flow meter signal among the filtered signals to a zero average having a unit variance, and averages them using the average and standard deviation. In other words, after signal filtering, S1(i) and S2(i) may be normalized with a zero average having a unit variance, and the signal may be averaged by Equation 1 below.

[Equation 1]

와

는 평균과 표준 편차 연산자이다. here,

Wow

Is the mean and standard deviation operator.

The signal down-sampling unit 140 may down-sample the normalized and averaged optical blood flow meter signal and the 3-axis acceleration signal. In other words, the normalized (normalized) and averaged PPG signal and Ax(i), Ay(i) and Az(i) can be downsampled to 25 Hz.

The motion artifact removal unit 150 may remove motion artifacts from the two-channel optical blood flow meter signal by using the down-sampled acceleration signal. In the present embodiment, the motion artifact may be motion information that interferes with signal measurement, such as wrist twist and fist trembling/unfolding. The removal of motion artifacts of the PPG using an acceleration signal may be one of the most important steps to greatly increase the accuracy of heart rate measurement. The motion artifact removal unit 150 may estimate a clean optical blood flow analyzer signal spectrum by subtracting the noise signal spectrum from the measurement signal spectrum by applying a Wiener filter. The Wiener filter estimates a clean signal spectrum without motion artifacts through linear time-invariant filtering on the PPG signal damaged by motion artifacts with noise signals from the acceleration signal.

The heart rate information providing unit 160 generates a state transition according to a specific condition using a finite state machine (FSM) on the signal from which the motion artifact has been removed, and the state is a stable state. Only the heart rate of the case can be provided to the user. In this case, a specific condition in which the state transition occurs may be determined according to a crest factor and a value of a change in heart rate measured in a continuous window. The crest factor can be obtained by using a fast Fourier transform to find the maximum peak and frequency corresponding to the maximum power. According to the present embodiment, the heart rate information providing unit 160 used a fast Fourier transform (FFT) of 2048 points in a range of 0.6 Hz to 3.3 Hz (about 40 bpm to 200 bpm) to find a frequency corresponding to the maximum power, You can estimate the minimum and maximum heart rates for all ages in rest and intensive activity. On the other hand, using the estimated heart rate in each window, it is possible to evaluate the performance accuracy with an absolute error (AE) as shown in Equation 2 below.

[Equation 2]

과

는 i번째 창에서 측정된 실제 심박수(bpm)일 수 있다. 전체 평가를 위해 아래의 수학식 3, 4, 5와 같이 bpm으로 절대 오류의 평균(AAE)을, bpm으로 절대 오류의 표준 편차(SdAE)를, 퍼센트(%)로 상대 절대 오류(ARE)의 평균을 사용할 수 있다. here,

and

May be the actual heart rate (bpm) measured in the i-th window. For the overall evaluation, as shown in Equations 3, 4, and 5 below, bpm is the mean of absolute error (AAE), bpm is the standard deviation of absolute error (SdAE), and percentage (%) is the relative absolute error (ARE). Average can be used.

[Equation 3]

[Equation 4]

[Equation 5]

Here, N may be the total number of windows considered for heart rate measurement.

In describing this embodiment, the finite state machine (FSM) is a mathematical model of computation and an abstract machine (which stays in one state of a finite state number at a given time). Finite state machines can be applied in various fields such as hardware digital systems, software engineering and electrical engineering. The finite state machine moves from one state to another in response to external inputs. On the other hand, in order to use the reflective optical blood flow meter signal in a wearable medical device, the measured heart rate must be accurate and consistent. Here, "consistency" may mean that the measured heart rate must be consistently accurate. The crest factor is a measurement value of a signal indicating the ratio of the peak value to the effective value, and indicates how extreme the peak signal appears. In more detail, the crest factor can be expressed as Equation 6 below.

[Equation 6]

와

는 각각 원근 평균 제곱값(root mean squared value)과 주기도의 피크값이다. 가속도 신호를 사용하여 PPG에서 모션 아티팩트를 제거한 후, 주기도를 얻은 이후 가장 많은 결과를 얻은 주기도는 실제 심박수에 해당하는 지배적인 피크(dominant peak)를 갖는다. 지배적인 피크 주파수가 다른 주파수 피크 전력보다 훨씬 높은 전력을 갖는다면, 신호는 모션 아티팩트에 의해 덜 손상된 것으로 간주할 수 있다. 즉, 파고율(crest factor) 값이 크다면, 심박수에 대한 신뢰도가 높다고 판단할 수 있다.here,

Wow

Is the root mean squared value and the peak value of the periodogram, respectively. After removing motion artifacts from the PPG using an acceleration signal, the periodogram that obtained the most results after obtaining the periodogram has a dominant peak corresponding to the actual heart rate. If the dominant peak frequency has a much higher power than the other frequency peak power, the signal can be considered less damaged by motion artifacts. That is, if the crest factor value is large, it may be determined that the reliability of the heart rate is high.

)는 유효하다고 할 수 있다. 복구 상태는 심박수 추정 결과가 주어진 시간에 중간 정도의 신뢰도를 가지고 있으며, 안정 상태로 이동할 수 있는지 여부를 조사할 수 있다. 경고 상태는 주어진 시간에 심박수 추정 결과가 낮은 신뢰도를 가짐을 나타낸다. 불확실 상태는 심박수 추정 결과가 매우 낮은 신뢰도를 가짐을 나타낸다. 본 발명에서, 상태 천이는 파고율과 심박수 변화를 외부 입력으로 결정한다. 한편, 본 실시예에 따른 유한 상태 머신 기반의 상태 천이에 대해 도 2에 개시되어 있다. 도 2는 본 발명의 일 실시예에 따른 유한 상태 머신 기반의 상태 천이에 대해 개략적으로 설명한 도면이다. 도 2를 참조하면, 매 2초마다 심박수(

)가 측정된다(따라서, i=2, 4, … i(측정 지속 시간 i초)). 처음(i=2)에 유한 상태 머신은 안정된 상태를 유지하고,

는 전처리 단계(신호 필터링, 신호 정규화 및 신호 다운 샘플링), 위너 필터를 사용한 모션 아티팩트 제거 및 0.6~3.3Hz 사이의 범위에서 주기도(periodogram)로부터 주요 피크 선택에 의해 얻어진다. 다음 측정(i=4, 6, … i)에서 유한 상태 머신은 이전 상태를 기반으로 각기 다른 데이터 처리 흐름을 수행한다. 시간 i에서, 시간 i-2에서의 이전 상태가 안정 상태에 있다면, 파고율 및 심박수 변화의 두 조건이 체크될 수 있다. 이는 아래의 수학식 7과 같이 나타낼 수 있다.Meanwhile, according to the present embodiment, a state that is transitioned in the finite state machine FSM may be defined as follows. The four states for a finite state machine can be a stable state, a recovery state, an alert state, and an uncertain state. The steady state indicates that the heart rate estimation result at a given time has high reliability. Therefore, in the present invention, the measured heart rate (

) Can be said to be valid. In the recovery state, it is possible to investigate whether the heart rate estimation result has intermediate reliability at a given time and can move to a stable state. A warning state indicates that the heart rate estimation result at a given time has low confidence. The uncertain state indicates that the heart rate estimation result has very low reliability. In the present invention, the state transition determines the crest rate and heart rate changes as external inputs. Meanwhile, a state transition based on a finite state machine according to the present embodiment is disclosed in FIG. 2. 2 is a diagram schematically illustrating a state transition based on a finite state machine according to an embodiment of the present invention. Referring to Figure 2, the heart rate every 2 seconds (

) Is measured (hence, i=2, 4,… i (measurement duration i seconds)). Initially (i=2) the finite state machine remains stable,

Is obtained by preprocessing steps (signal filtering, signal normalization and signal downsampling), motion artifact removal using a Wiener filter, and selection of major peaks from a periodogram in the range of 0.6 to 3.3 Hz. In the next measurement (i=4, 6,… i), the finite state machine performs different data processing flows based on the previous state. At time i, if the previous state at time i-2 is in a stable state, then two conditions can be checked: crest rate and heart rate change. This can be expressed as Equation 7 below.

[Equation 7]

는 파고율에 대한 임계값이다. 첫번째 용어는 파고율 조건이고, 두번째 용어는 심박수 변경 조건이다. 만약, 두 조건이 만족되면, 시점 i에서의 상태는 안정 상태로 유지되고,

는 유효하다고 선언될 수 있다. 그렇지 않으면, 시간 i의 상태는 경고 상태로 천이하고,

는 무효라고 선언할 수 있다. 시간 i-2에서의 이전 상태가 경고 상태에 있다면, 파고율 조건

만 체크된다.

가 무효로 선언되었기 때문에 심박수 변경 조건

을 고려하지 않는다. 만약,

조건이 만족되면, 상태는 시간 i에서 복구 상태로 천이한다. 그렇지 않으면, 상태는 경고 상태로 유지된다. 또한, 시간 i-2의 이전 상태가 경고 상태에 있을 때, 경고 대기 시간

에 의해 정의된 일정 시간 동안 상태가 경고 상태에 계속 머물러 있으면, 상태는 불확실 상태로 천이한다. 잠시 동안 경고 상태에서 지속적으로 유지되면, 광혈류측정기 신호가 제대로 측정되지 않고 심박수가 신호에 반영되지 않는다는 것을 나타낸다. 일단, 상태가 불확실 상태로 천이되면, 상태는 불확실한 대기 시간

에 의해 정의되는 또 다른 특정 지속 시간 동안 불확실한 상태로 유지된다.

시간 동안, 파고율 조건

이 지속적으로 충족되면 상태는 경고 상태로 상승한다. 그렇지 않으면, 상태는 불확실한 상태로 유지된다. 시간 i-2의 이전 상태가 복구 상태에 있을 때, 유일한 대기 시간 조건인

도 복구 대기 시간

에 의해 정의된 다른 지속 시간 동안 확인된다.

시간 동안,

조건이 계속적으로 충족되지 않으면, 즉시 상태가 경고 상태로 전환된다.

기간 동안 계속적으로

조건을 만족하면, 세 가지 조건이 확인된다. 이는 아래의 수학식 8과 같이 나타낼 수 있다.here,

Is the threshold for the crest factor. The first term is a crest rate condition, and the second term is a heart rate change condition. If both conditions are satisfied, the state at time i remains stable,

May be declared valid. Otherwise, the state of time i transitions to the warning state,

Can be declared invalid. If the previous state at time i-2 is in the warning state, the crest factor condition

Only is checked.

Heart rate change condition because is declared invalid

Do not take into account. if,

If the condition is satisfied, the state transitions from time i to the recovery state. Otherwise, the state remains in the warning state. Also, when the previous state of time i-2 is in the warning state, the warning waiting time

If the state remains in the warning state for a period of time defined by, the state transitions to an uncertain state. If it remains in a warning state for a while, it indicates that the photometer signal is not being properly measured and the heart rate is not reflected in the signal. Once a state transitions to an uncertain state, the state is an uncertain wait time

It remains in an uncertain state for another specific duration defined by.

Over time, crest factor condition

If this is continuously met the state rises to a warning state. Otherwise, the state remains uncertain. When the previous state of time i-2 is in the recovery state, the only waiting time condition

Road recovery waiting time

Is checked for a different duration defined by.

For hours,

If the condition is not continuously met, the state immediately switches to the warning state.

Continuously throughout the period

If the condition is satisfied, three conditions are checked. This can be expressed as Equation 8 below.

[Equation 8]

의 최근 시간 사이에 심박수가 변화하는지를 확인하기 위한 장기 심박수 변화 조건이다. 수학식 8의 조건을 만족하면, 상태는 안정 상태로 상승하고,

은 유효하다고 선언될 수 있다. 그렇지 않으면, 상태는 복구 상태로 유지되고, 기간

동안 파고율 조건이 반복된다. 모든 측정된 심박수는 안정 상태를 제외하고는 유효하지 않다. 상술한 대기 시간 파라미터

,

및

는 임의로 조정 가능하다. 이러한 파라미터는 심박수가 광혈류측정기 신호에 반영되는 대기 시간을 제공한다. 본 실시예에서는

=5,

=3,

=4로 설정했다. Here, the first and second terms are the crest rate and heart rate change conditions as in Equation 7, and the last term is the current time i and the steady state

It is a long-term heart rate change condition to determine if the heart rate changes between the last time in When the condition of Equation 8 is satisfied, the state rises to a stable state,

May be declared valid. Otherwise, the state remains in the recovery state, and the period

During the crest factor condition is repeated. All measured heart rates are not valid except at rest. Waiting time parameter described above

,

And

Can be arbitrarily adjusted. These parameters provide the waiting time for the heart rate to be reflected in the photogrammetry signal. In this example

=5,

=3,

I set it to =4.

Referring to FIG. 2, an example will be described as follows.

Initially, the device operates in a state assuming that the state of the finite state machine (FSM) is a stable state.

First, the device determines a state transition by comparing a difference between a heart rate predicted 2 seconds ago and a current predicted heart rate and a crest factor with a preset threshold. Here, a threshold value for comparing the heart rate difference may be 5 bpm, and a threshold value for comparing the crest factor may be 2.4. In this case, the set threshold may be arbitrarily preset by the user. For example, if the difference between the predicted heart rate two seconds ago and the current predicted heart rate is less than 5 bpm, and the crest rate exceeds 2.4, it is determined that the device is in a stable state.

If the device does not satisfy the above-described condition, the device transitions to an alert state. When it transitions to the alert state, it remains in the transitioned state for a preset alert time. After the alert state, the device predicts the heart rate and, if the crest factor is higher than the threshold value (2.4), it transitions to the recovery state. When transitioning to the recovery state, it stays for a preset recovery time as in the alert state, and it is continuously checked whether the crest factor is higher than the threshold value (2.4). If the crest factor condition is not satisfied during the recovery time, it immediately transitions to the alert state, and if the crest factor condition is continuously satisfied during the recovery time ( That is, if the crest rate exceeds the threshold (2.4), the difference between the predicted heart rate 2 seconds before and the current predicted heart rate 2 seconds later is lower than 5 bpm (threshold value), and the crest factor is the threshold value. The difference between the heart rate, which is higher and the most recent stable state, and the current heart rate, is calculated and then compared with the threshold value of the statistic value, and if it is lower, it transitions to a stable state.

On the other hand, after the alert time in the alert state, if the condition where the crest factor is greater than the threshold value is not satisfied, it transitions to an uncertain state. When the device enters an uncertain state, it waits for an uncertain time and, while waiting, transitions to an alert state if the condition where the crest factor is higher than the threshold value is continuously satisfied, and the intermediate If the condition is not satisfied, the uncertain time is reset and starts again from the beginning.

3 is a diagram illustrating a flow of a method for estimating a heart rate according to an embodiment of the present invention.

Referring to FIG. 3, first, the apparatus for estimating heart rate acquires a two-channel optical blood flow analyzer signal, a 3-axis acceleration signal, and a 1-channel electrocardiogram signal measured by a reflective optical blood flow analyzer sensor (S310).

Next, the apparatus for estimating the heart rate filters the obtained two-channel optical blood flow meter signal and the three-axis acceleration signal using a Butterworth band pass filter (Butterworth BPF) (S320).

Next, the apparatus for estimating heart rate normalizes the two-channel optical blood flow meter signal among the filtered signals to a zero average having a unit variance, and averages it using the average and standard deviation (S330).

Next, the apparatus for estimating the heart rate down-samples the normalized and averaged optical blood flow meter signal and the 3-axis acceleration signal to 25 Hertz (Hz) (S340).

Next, the apparatus for estimating the heart rate removes motion artifacts from the two-channel optical blood flow meter signal by using the down-sampled acceleration signal. The apparatus for estimating the heart rate may estimate a clean optical blood flow analyzer signal spectrum by subtracting the noise signal spectrum from the measurement signal spectrum by applying a Wiener filter (S350).

Next, the apparatus for estimating the heart rate generates a state transition according to a specific condition using a signal from which motion artifacts have been removed using a finite state machine (FSM), and the state is a stable state. Only the heart rate in the case of is provided to the user. In this case, the specific condition in which the state transition occurs may be determined according to a crest factor and a value of a change in heart rate measured in a continuous window (S360).

The methods according to the embodiments of the present invention may be implemented as an application or in the form of program instructions that can be executed through various computer components and recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded in the computer-readable recording medium may be specially designed and configured for the present invention, and may be known and usable to those skilled in the computer software field. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magnetic-optical media such as floptical disks. media) and hardware devices specially configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of the program instructions include not only machine language codes such as those produced by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules to perform processing according to the present invention, and vice versa.

While this specification includes many features, such features should not be construed as limiting the scope or claims of the invention. In addition, features described in separate embodiments of the present specification may be combined and implemented in a single embodiment. Conversely, various features described in a single embodiment of the present specification may be individually implemented in various embodiments, or may be properly combined and implemented.

Although the operations have been described in a specific order in the drawings, it should not be understood that such operations are performed in a specific order as shown, or as a series of consecutive sequences, or that all described operations are performed to obtain a desired result. Multitasking and parallel processing can be advantageous in certain environments. In addition, it should be understood that classification of various system components in the above-described embodiments does not require such classification in all embodiments. The above-described app components and systems may generally be implemented as a package in a single software product or multiple software products.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the technical spirit of the present invention for those of ordinary skill in the technical field to which the present invention belongs. It is not limited by the drawings.

110: signal acquisition unit
120: signal filtering unit
130: signal normalization unit
140: signal down-sampling unit
150: motion artifact removal unit
160: heart rate information providing unit

Claims (9)

A signal acquisition unit that acquires a 2-channel optical blood flow analyzer signal, a 3-axis acceleration signal, and a 1-channel electrocardiogram signal measured by the reflective optical blood flow analyzer sensor;
A signal filtering unit for filtering the two-channel optical blood flow meter signal and the three-axis acceleration signal obtained by the signal acquisition unit using a Butterworth BPF;
A signal normalizing unit for normalizing a two-channel optical blood flow meter signal among the filtered signals to a zero average having a unit variance and averaging using an average and a standard deviation;
A signal down-sampling unit for down-sampling the normalized and averaged optical blood flow meter signal and the 3-axis acceleration signal to 25 hertz (Hz);
A motion artifact removal unit that removes motion artifacts from a two-channel optical blood flow meter signal using the down-sampled acceleration signal; And
Using a finite state machine (FSM) from the motion artifact removed signal, a state transition is generated according to a condition determined according to a crest factor and a value of a heart rate change measured in a continuous window, and the state is A device for estimating an accurate and consistent heart rate based on a finite state machine including; a heart rate information providing unit that provides only a heart rate in a stable state to a user. delete The method of claim 1,
The motion artifact removal unit,
An apparatus for estimating an accurate and consistent heart rate based on a finite state machine, characterized in that, by applying a Wiener filter to subtract a noise signal spectrum from a measurement signal spectrum, a clean optical blood flow analyzer signal spectrum is estimated. The method of claim 1,
The two-channel optical blood flow meter signal,
A device for estimating accurate and consistent heart rate based on a finite state machine, characterized in that it is measured from the user's wrist by two pulse oximeters with green LEDs. The method of claim 1,
The three-axis acceleration signal,
A device for estimating an accurate and consistent heart rate based on a finite state machine, characterized in that it is measured from a user's wrist through a 3-axis accelerometer. The method of claim 1,
The ECG signal is,
An apparatus for estimating an accurate and consistent heart rate based on a finite state machine, characterized in that measured from the user's chest using a wet electrocardiogram sensor. In the method of estimating heart rate in a device for estimating accurate and consistent heart rate based on a finite state machine,
A signal acquisition step of acquiring a 2-channel optical blood flow analyzer signal, a 3-axis acceleration signal, and a 1-channel ECG signal measured by the reflective optical blood flow analyzer sensor;
A signal filtering step of filtering the two-channel optical blood flow meter signal and the three-axis acceleration signal obtained in the signal acquisition step by using a Butterworth BPF;
A signal normalization step of normalizing a two-channel optical blood flow meter signal among the filtered signals to a zero average having a unit variance and averaging using the average and standard deviation;
A signal down-sampling step of down-sampling the normalized and averaged optical blood flow meter signal and 3-axis acceleration signal to 25 hertz (Hz);
A motion artifact removal step of removing motion artifacts from a two-channel optical blood flow meter signal using the down-sampled acceleration signal; And
Using a finite state machine (FSM) from the motion artifact removed signal, a state transition is generated according to a condition determined according to a crest factor and a value of a heart rate change measured in a continuous window, and the state is A method for estimating heart rate comprising a; heart rate information providing step of providing only the heart rate in the case of a stable state to the user. delete The method of claim 7,
The motion artifact removal step,
And estimating a clean optical blood flow analyzer signal spectrum by applying a Wiener filter to a value obtained by subtracting the noise signal spectrum from the measurement signal spectrum.

Images