KR20230077605A - Apparatus for estmating respiration rate using doppler velocity based on FMCW radar and method thereof - Google Patents

Abstract

The present invention relates to a device and a method for estimating a respiratory rate using the Doppler velocity based on an FMCW radar. According to the present invention, the method comprises the steps of: receiving a signal reflected from a target after being transmitted through an FMCW radar in accordance with time; obtaining Doppler magnitude data in accordance with a distance and velocity by performing a second discrete Fourier transform after performing a first discrete Fourier transform on the signal received from the FMCW radar; calculating the sum of absolute values of Doppler magnitudes for each distance observed in a corresponding velocity component for each velocity within a measurement velocity range of the FMCW radar; deriving a velocity value of the target at corresponding time by averaging the sums of the absolute values of the Doppler magnitudes for each distance obtained for each velocity; and applying a window for each hour to velocity data in accordance with time to estimate a respiratory rate of the target for each time from a frequency component analyzed in the window. According to the present invention, a respiratory rate of a target can be accurately estimated based on the Doppler velocity of the target extracted through an FMCW radar.

Description

Apparatus for estmating respiration rate using doppler velocity based on FMCW radar and method thereof}

The present invention relates to an apparatus and method for estimating a respiratory rate using a Doppler velocity based on an FMCW radar, and more particularly, to a Doppler capable of estimating a respiratory rate of a target based on a Doppler velocity value of a target extracted through an FMCW radar. It relates to an apparatus and method for estimating respiratory rate using velocity.

Along with the recent increase in the aging population, people's desire for health promotion is increasing due to COVID-19. As a result, many studies are being conducted to periodically check one's health even in daily life and measure vitals to prevent diseases in advance. Among them, it is important to monitor respiratory signals during vitality because respiratory dysfunction can also be attributed to cardiovascular, neurological, and psychiatric problems.

Existing respiration monitoring research methods use contact sensors. The respiratory signal is extracted with the pressure signal obtained by wearing a wearable belt or attaching a single pressure sensor to the body. Alternatively, fiber optic sensors are attached to the body to measure respiration using optical interference caused by chest movement.

However, the commonly used sensor contact method for measuring respiration is subject to restrictions on activity during the measurement time, and has disadvantages that may cause unpleasant feelings because the measurement sensor contacts the user's body.

In order to overcome these disadvantages, research on respiration measurement using Frequency Modulated Continuous Wave (FMCW) radar is being actively conducted. FMCW radar has the advantage of non-contact extraction of the distance to the target and displacement information with the frequency difference extracted between the transmitted and received radio waves, and since it uses millimeter waves, it uses low power and has a small sensor size, so it has a wide range of applications.

Existing studies have measured respiration using only the magnitude and phase of the demodulation signal of radar transmission and reception. In this study, even in invisible micro-movements of the human body that occur in a standing position without leaning back, a distorted signal is output due to the signal combined with the motion of the chest and the micro-movement, resulting in a greatly reduced detection accuracy.

In order to expand the scope of limited use of non-contact breathing research by mitigating the random body movement (RBM) phenomenon that causes this phenomenon, research is being conducted to monitor breathing in moving situations. There is a study to extract the breathing frequency shifted by the amount of frequency shift modulated using Motion Direction Detection with the DC offset extracted through the phase of the signal changed due to the RBM phenomenon. However, since this is to estimate the frequency range distorted according to the degree of movement, there is a limit to accurate breathing frequency extraction.

In addition, a study using the range-bin alignment method by combining two-way distance information extracted by two radars and a method of discarding respiration signals that have energy above the average energy value of the respiration signal in a fixed range-bin are used to determine the RBM phenomenon. There are studies that compensate for the disadvantages caused by this. However, the method of mitigating the RBM phenomenon through range-bin range selection increases computational complexity as well as two radar structures.

The background technology of the present invention is disclosed in Korean Patent Registration No. 10-2091974 (announced on March 24, 2020).

An object of the present invention is to provide an apparatus and method for estimating a respiratory rate using a Doppler velocity based on an FMCW radar capable of estimating the respiratory rate of a target based on the Doppler velocity value of the target extracted from a received signal of the FMCW radar.

In the present invention, in a respiration rate estimation method performed by an FMCW radar-based respiration rate estimation device, the steps of receiving a signal reflected from a target according to time after transmission through the FMCW radar, and receiving a signal received from the FMCW radar by 1 Acquiring Doppler amplitude data according to distance and speed by performing second discrete Fourier transform after differential discrete Fourier transform, Calculating the sum of the absolute values, averaging the sum of the absolute values of the Doppler magnitudes for each distance obtained for each velocity, and deriving the velocity value of the target at the corresponding time, and applying a window for each hour to the velocity data according to time It provides a respiratory rate estimation method comprising the step of estimating the respiratory rate of the target by time from the frequency components analyzed within the window.

In addition, the obtaining of the Doppler magnitude data may include applying the first and second discrete Fourier transforms to an intermediate frequency signal generated by applying a low-pass filter to a combined signal of a transmission signal and a reception signal of the FMCW radar. Doppler magnitude data can be obtained.

In addition, in the step of deriving the velocity value of the target, the velocity value D _v (t) of the target at the corresponding time t may be derived using the following equation.

Here, N _chirp is the number of chirps, v ₀ is each velocity component within the measured velocity range, and D(t,v ₀ ) is the absolute value of the Doppler magnitude for each distance at time t obtained from the corresponding velocity component v ₀ represents the sum.

In addition, the D(t,v ₀ ) may be defined by the following equation.

Here, Z(t,r _k ,v ₀ ) is the Doppler magnitude for each distance observed in the corresponding velocity component v ₀ , and r _k represents each distance component corresponding to the measured distance range.

In addition, in the step of estimating the respiratory rate, a plurality of peak frequencies detected by fast Fourier transforming the data on the window applied at the current time point (t) are separated from the tracking frequency value determined from the window at the previous time point (t-1). After the comparison, determining the peak frequency with the minimum frequency deviation as the tracking frequency value at the current time point (t), and respiratory rate (RR) at the current time point (t) through the determined tracking frequency value It may include the step of determining.

In addition, the step of estimating the number of breaths per minute may calculate the number of breaths per minute (RR) of the target at the current time point t by applying the tracking frequency value determined at the current time point t to the following equation. .

는 현재 시점(t)에서 결정된 추적 주파수 값을 나타낸다.here,

represents the tracking frequency value determined at the current point in time (t).

In addition, in the present invention, in the FMCW radar-based respiratory rate estimation device, a signal acquisition unit for obtaining a signal reflected from a target over time after being transmitted through the FMCW radar, and a signal received from the FMCW radar as a primary A signal processing unit that obtains Doppler magnitude data according to distance and speed by performing a second discrete Fourier transform after discrete Fourier transform, and a Doppler magnitude for each distance observed at a corresponding speed component for each speed within the measurement speed range of the FMCW radar A calculation unit that calculates the sum of absolute values and averages the sum of the absolute values of the Doppler magnitudes for each distance obtained for each velocity to derive the velocity value of the target at the corresponding time, and applies a window for each hour to the velocity data according to time Provided is a respiration rate estimating device including an estimator for estimating the respiration rate of a target by time from frequency components analyzed within a window.

According to the present invention, the target's respiratory rate can be accurately estimated based on the Doppler velocity value of the target extracted through the FMCW radar.

1 is a diagram showing the configuration of an apparatus for estimating respiratory rate according to an embodiment of the present invention.
2 is a diagram showing the concept of a respiratory rate estimation technique through FMCW radar.
Figure 3 is a diagram explaining a method for estimating respiratory rate using the device of FIG.
Figure 4 is a diagram comparing M (t, r _k ) and D _v (t) used in estimating respiratory rate in the conventional technique and the technique of the present invention.
5 is a diagram explaining the principle of extracting a respiratory rate from Doppler velocity data in an embodiment of the present invention.
6 is a diagram showing four experimental environments used to verify the performance of a respiratory rate estimation technique according to an embodiment of the present invention.
7 to 10 are diagrams comparing the respiration rate estimation results of the present invention for Experiments 1 to 4 with those of conventional techniques.

Then, with reference to the accompanying drawings, an embodiment of the present invention will be described in detail so that those skilled in the art can easily practice it. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is said to be "connected" to another part, this includes not only the case where it is "directly connected" but also the case where it is "electrically connected" with another element interposed therebetween. . In addition, when a certain component is said to "include", this means that it may further include other components without excluding other components unless otherwise stated.

The present invention measures a biosignal with a single channel using FMCW radar and measures respiration by processing the speed change according to the movement of the human body in terms of the Doppler velocity rather than the magnitude and phase change of the signal due to the movement of the human body. Suggest.

The extracted speed information includes the speed change frequency according to the movement of the chest that occurs during the breathing process. Therefore, unlike previous studies that estimate the breathing frequency by tracking the change in distance according to the movement, we present a new approach that can extract the breathing frequency by processing the speed information caused by the chest movement.

In addition, unlike existing research methods in which movement is corrected and respiration is measured by setting a range-bin range to mitigate the RBM phenomenon, the method proposed in the present invention does not require setting a measurement range.

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

As shown in FIG. 1, the respiratory rate estimation device 100 includes a signal acquisition unit 110, a signal processing unit 120, a calculation unit 130, and an estimation unit 140, and further includes the FMCW radar 10. can include Here, the operation of each unit 110 to 140 and the flow of data between each unit may be controlled by a controller (not shown).

The signal acquisition unit 110 acquires the received signal reflected from the target after transmission through the FMCW radar 10 over time. The FMCW radar 10 may transmit a linearly modulated frequency signal, receive a signal reflected from a target (target), and transmit the signal to the signal acquisition unit 110 .

The signal processing unit 120 obtains Doppler magnitude data according to distance and speed by performing a first discrete Fourier transform (DFT) on the received signal of the acquired radar and then a second discrete Fourier transform again. Such Doppler magnitude data can be obtained using a time-distance Doppler Map.

Here, the signal processing unit 120 applies first- and second-order discrete Fourier transforms to an intermediate frequency (IF) generated by applying a low-pass filter to the combined signal of the radar transmission signal and the received signal, and then applies the second-order discrete Fourier transform for each time. Acquire Doppler magnitude data according to distance and speed.

To this end, the signal processing unit 120 includes a mixer for combining the transmitted signal and the received signal, a filter for applying a low-pass filter to the combined signal, and an ideal Fourier transform for performing discrete Fourier transform on the intermediate frequency signal obtained by applying the filter. It may consist of parts.

The calculation unit 130 calculates the sum of the absolute values of the Doppler amplitudes for each distance observed in the corresponding velocity component for each velocity within the radar measurement velocity range. The sum of the absolute values of the Doppler magnitude obtained for each velocity is averaged over all velocities to derive the velocity value of the target for a corresponding time.

That is, the calculation unit 130 averages the sum of the absolute values of the Doppler magnitudes for each distance obtained for each speed to obtain the speed value of the target at the corresponding time t. The calculation unit 130 may transmit the speed value obtained by time to the estimation unit 140 in real time. Through this, the speed value of the target is derived for each time, and speed data according to time is obtained.

The estimator 140 estimates the respiration rate of the target hourly from frequency components analyzed within the window by applying a window every hour to the speed data according to time.

After applying a window of the set time size for each hour to the speed data over time, fast Fourier transform (FFT) is performed to extract the breathing frequency for each time. Assume that the value is extracted. This is because one breath includes inhalation and exhalation, reflecting both of these characteristics.

Here, the respiratory rate of the target estimated in the embodiment of the present invention may mean the respiratory rate per minute (RR).

The estimator 140 may convert the respiratory frequency value extracted by time into a respiratory rate (RR) per minute and provide the result of estimating the respiratory rate of the target in real time.

2 is a diagram showing a concept of respiration rate estimation through FMCW radar. 2 (a) shows the concept of the conventional technique and (b) the technique of the present invention. In the conventional case, the respiration rate was measured using the data of the distance bin r _k selected from the entire range of detectable distance bins, but the respiration rate estimation technique proposed in the present invention can measure the respiration rate without setting a distance bin.

Figure 3 is a diagram explaining a method for estimating respiratory rate using the device of FIG.

First, the signal acquisition unit 110 acquires the received signal reflected from the target after transmission from the FMCW radar every hour (S310).

Thereafter, the signal processing unit 120 obtains Doppler size data according to distance and speed by performing a first order discrete Fourier transform and then a second order discrete Fourier transform on the received signal of the radar (S320).

At this time, the signal processing unit 120 mixes the transmitted signal and the received signal and passes the mixed signal through a low-pass filter to generate an intermediate frequency signal of Equation 1 below.

Here, x(t,n,m) can be seen as a signal of a target moving at a speed of v ₀ at time t, m is the number of chirps, and n is the sample index of the chirp. .

In Equation 1, f _r is a frequency corresponding to a distance r, M ₀ is a reflected power at a distance r, and P is a time delay of a radio wave reflected at r. f _r and P are expressed as in Equation 2 below.

는 시간 지연의 상수, F_s는 샘플링 주파수를 나타낸다.Here, BW is the frequency bandwidth, c is the speed of light, T _c is the duration of the chirp,

is the constant of the time delay, and F _s represents the sampling frequency.

In the conventional technique shown in (a) of FIG. 2, X(t,r _k ,m) of Equation 3 below obtained by first-order DFT of Equation 1 is used to obtain power magnitude M(t,r _k ) and phase P (t, r _k ) was extracted as in Equation 4, and the respiratory rate was measured using the power magnitude M (t, r _k ).

이고 k=0, …, N-1이다.

는 호흡으로 인한 흉부 움직임을 표현한다. 물체가 거리 r_k에서

의 미세 움직임이 존재한다고 가정하면, 수학식 4는 다음 수학식 5와 같이 표현된다.here,

and k = 0, ... , N-1.

represents the movement of the chest due to breathing. object at distance r _k

Assuming that there is a fine motion of , Equation 4 is expressed as Equation 5 below.

However, if 'r _k + t' in Equation 5 includes the displacement due to the movement of the human body as well as the minute movement due to respiration, M(t,r _k ) and P(t) corresponding to the distance r _k , r _k ), a distorted IF signal is extracted.

Therefore, in the case of the embodiment of the present invention, unlike the conventional technique that utilizes the magnitude M(t, r _k ) calculated from the signal obtained through the first DFT, the time-distance Doppler obtained by performing the second DFT after the first DFT Using the data, the velocity values (Doppler velocity) of the target are extracted, and the respiratory rate is derived from these velocity values.

First, when Equation 1 is firstly DFTed to obtain Equation 3, and then second-order DFT is applied again, the result of Equation 6 is obtained.

Here, N _chirp is the number of chirps, and k is the distance (distance bin) index. (In Equation 1 above, m is the number of chirps and n is defined as the sample index of the chirp. Please check whether the definition of N _chirp is correct.)

Z(t, r _k , v _o ) extracted by Equation 6 appears in proportion only to the reflected power M ₀ at r _k when the object has a velocity v _o at the r _k position.

In this way, if the radar reception signal received from the target is subjected to the second DFT again after the first DFT process, Doppler magnitude data (distance-velocity Doppler map) according to the distance and speed at the corresponding time t can be obtained.

A Doppler map is composed of a two-dimensional map having, for example, a distance axis for the entire measurable distance (distance bin) range and a velocity axis for the entire measurable velocity (velocity bin) range, and each coordinate point in the map is A Doppler magnitude corresponding to a corresponding speed value and a corresponding distance value may be expressed as color information. Here, the measurement speed range and the measurement distance range of the radar may be predefined by product standards or preset by a user.

Next, the calculation unit 130 derives the velocity value of the target at the corresponding time t using the Doppler magnitude data according to the distance and velocity obtained through the first and second DFT processing (S330).

Specifically, the calculation unit 130 calculates the sum D(t,v ₀ ) of the absolute values of the Doppler magnitudes for each distance observed in the corresponding speed component for each speed within the radar measurement speed range as in Equation 7 below. .

Here, Z(t,r _k ,v ₀ ) is the Doppler magnitude by distance at time t observed at the corresponding velocity component v ₀ , r _k is each distance component within the measured distance range, and |·| is the absolute value indicates

As such, the present invention proposes a respiratory rate estimation technique approached from the viewpoint of the Doppler velocity, and utilizes the absolute value of the Doppler magnitude. In addition, since Doppler data in all distance bins within the measurement distance range is used, it is not necessary to select a specific distance bin as in the prior art.

Then, the calculation unit 130 averages the sum of the absolute values of the Doppler magnitudes for each distance obtained for each speed with respect to all velocities, and calculates the speed value D _v (t) of the target at the corresponding time t as shown in Equation 8 below. derive in a way

Here, N _chirp is the number of chirps, v ₀ is each velocity component within the measured velocity range, and D(t, v ₀ ) represents the sum of absolute values of Doppler magnitudes for each distance obtained from the corresponding velocity component v ₀ .

That is, as shown in Equation 8, the speed value of the target at the corresponding time t is calculated by summing up all converted data for each distance in all speed bins within the measured speed range, dividing by N _chirp and averaging.

As such, in the case of the embodiment of the present invention, D(t,v ₀ ) obtained for each speed through the Doppler map is substituted into Equation 8 to derive an average value for the measured speed range. An embodiment of the present invention uses this average value as a feature value for measuring the target's respiratory rate at time t.

Figure 4 is a diagram comparing M (t, r _k ) and D _v (t) used in estimating respiratory rate in the conventional technique and the technique of the present invention. Figure 4 (a) is a result of observation in a stationary state without movement of the target (person), and (b) is a result of observation in a state in which the target is moving.

In each case, the top figure represents the ground truth repeating a pattern in which the power increases during inhalation and decreases during exhalation during the breathing process.

The middle figure shows M(t,r _k ) over time t measured by the conventional method. The lower figure is the D _v (t) value according to time t measured by the technique of the present invention. In actual exhalation, unlike in inhalation, D _v (t) has a negative value, but it represents the result of inverting through the absolute value. Therefore, when using the Doppler velocity perspective approach, each inhalation and exhalation is represented as one cycle as shown in the figure below.

Respiration measurement studies using existing FMCW radars have extracted respiration signals by looking only at M(t,r _k ) in the range of distances where the target exists. In this method, even when the subject (target) is standing still with their backs against the wall, a distorted signal is generated by adding minute movements of the human body to the displacement due to chest movements according to Equation 5. Due to this, inaccurate breathing is extracted when the human body is not fixed.

As shown in FIG. 4, in the environment in which the human body is in a fixed state (a), both the proposed method and the existing research method output accurate breathing signals with a regular cycle. In the experimental environment (b) that includes minute movements of the human body, D _v (t) according to the technique of the present invention extracts signals for each set period, whereas M(t,r _k ) according to the existing technique It shows that a distorted signal is extracted due to

Since the method proposed in the present invention derives only a result proportional to the speed approaching the radar by Equation 6, an improved respiratory rate output compared to the existing research method is possible even with minute movements.

Referring back to FIG. 3 , the estimator 140 estimates the respiration rate of the target by time from the frequency components analyzed within the window by applying a window every hour to the speed data according to time (S340).

Here, of course, the window means a time window, and the target's respiration rate at the current time point is estimated by applying a window of a set time length to past data including the current time point for each hour and analyzing frequency components within the window. Of course, a sliding window method can be applied for real-time respiratory rate detection. In this case, a fast Fourier transform (FFT) may be used for frequency analysis.

Human respiration usually appears between 0.1 Hz and 0.4 Hz. In the Doppler velocity perspective approach, each inhalation and exhalation are represented as one cycle by Equation 7. Accordingly, it is assumed that when the signal extracted through Equation 8 is divided by the window and the frequency is extracted through FFT, the breathing frequency is extracted at 0.2 to 0.8 Hz, which is twice the normal breathing frequency of humans.

5 is a diagram explaining the principle of extracting a respiratory rate from Doppler velocity data in an embodiment of the present invention.

The upper figure of FIG. 5 corresponds to the true value of the respiration signal, the middle figure shows Doppler velocity data D _v (t) obtained by the technique according to an embodiment of the present invention, and the lower figure shows the data on the window applied at time t Shows an example of the result of FFT conversion of . As a result of FFT conversion, peaks (marked with *) are detected at several frequencies.

)를 이전 시점(t-1)에서의 윈도우로부터 결정된 추적 주파수 값(

)과 개별 비교한 후에 주파수 편차가 최소(min)인 피크 주파수를 현재 시점(t)에서의 추적 주파수 값(

)으로 결정한다. Here, the estimator 140, as shown in Equation 9 below, a plurality of peak frequencies detected by fast Fourier transforming the data on the window applied at the current time point t (

) to the tracking frequency value determined from the window at the previous time point (t-1) (

) and individual comparisons, the peak frequency with the minimum frequency deviation (min) is the tracking frequency value (

) to determine

을 다음의 수학식 10에 적용하여 현재 시점(t)에서의 타겟의 분당 호흡수(RR)를 연산하여 결정한다.Then, the tracking frequency value determined at the current time point (t)

is applied to the following Equation 10 to calculate and determine the respiratory rate per minute (RR) of the target at the current time point (t).

는 현재 시점(t)에서 결정된 추적 주파수 값을 나타낸다.here,

represents the tracking frequency value determined at the current point in time (t).

집합에서 수학식 9를 통하여 이전 호흡 주파수인

와 가장 가까운 peak를

로 추적 주파수 값을 결정하고, 추출한 주파수 값을 수학식 10을 통하여 Respiratory rate(RR)로 계산하여 호흡수를 최종 출력한다.It is assumed that human breathing does not suddenly increase or decrease. Therefore, having a peak value through FFT

In the set, the previous breathing frequency through Equation 9

the peak closest to

The tracking frequency value is determined by , and the extracted frequency value is calculated as the respiratory rate (RR) through Equation 10 to finally output the respiratory rate.

The following describes the results of verifying the performance of the respiratory rate estimation technique proposed in the present invention.

Human respiration has a frequency of about 0.1 to 0.4 Hz. The existing research method using M(t,r _k ) of data obtained from FMCW radar and applying a 0.1~0.4Hz Band-pass filter and the method using only D _v (t) proposed in the embodiment of the present invention are described in the following 4 Compare through experiments.

6 is a diagram showing four experimental environments used to verify the performance of a respiratory rate estimation technique according to an embodiment of the present invention.

As shown in (a) to (d) of FIG. 6, experiment 1 in which the movement of the human body was minimized by fixing the body against the wall, experiment 2 in which the subject stood still without leaning back, experiment 3 in which the subject moved back and forth while sitting on a chair , Experiment 4 in which the subject sits on a chair and moves left and right, compares the accuracy of the respiratory rate estimation between the existing method and the proposed study for each situation.

Table 1 shows the parameters and specifications of the FMCW radar used in the experiment.

Parameter Value Center frequency 60 GHz Detection range ~6.4m Field of view (Azimuth) ±65° Field of view (Elevation) ±30° Frequency Bandwidth 3GHz Chirp duration 256 μs Sampling rate 1MHz Scan interval 50ms Number of transmit antennas (Tx antenna) 2 Number of receiving antennas (Rx antenna) 4 Rx antenna spacing 0.5λ

7 to 10 are diagrams comparing the respiration rate estimation results of the present invention for Experiments 1 to 4 with those of conventional techniques. At this time, FIGS. 7 to 9 show the results for 6 subjects together, and FIG. 10 shows the results for 2 subjects together.

As shown in FIG. 6(a), the experiment 1 was a situation in which the subject leaned against the wall and minimized the movement of the human body, showing only chest movement due to respiration. As a result, as shown in FIG. 7 , in the case of Experiment 1, accurate values are obtained when breathing is extracted by the movement of the chest due to breathing without the movement of the human body using radar in both methods.

For accuracy calculation, Equation 11 below was used.

Here, d refers to the total experiment time, and the respiratory rate G _t (t) (correct answer value) extracted from the contact respiration sensor directly mounted on the subject through Equation 11 in each experimental method and the respiratory rate RR (t) estimated in the present invention ) to extract the accuracy.

Table 2 compares the accuracy of respiratory rate through the technique of the present invention and the conventional technique, and in the case of the present invention, it shows an average accuracy of 95% and shows higher accuracy than the conventional technique.

test subject Conventional Proposed technique A 99.03% 99.36% B 90.26% 93.55% C 94.93% 98.50% D 76.02% 92.77% E 85.22% 95.34% F 76.29% 94.79% average 86.96% 95.72%

Next, as shown in FIG. 6(b), in the case of Experiment 2, chest movements including minute movements of the human body that occur while the subject is standing still are shown.

Referring to the experimental results of FIG. 8 , when measuring respiration of a subject standing upright without leaning back as in Experiment 2, the existing method is inaccurate due to a difference in displacement caused by a fine movement of the body. However, since the proposed method uses the speed due to micro-movement, it is possible to extract the frequency according to the specific speed of the chest, and the breath extraction accuracy is high. Table 3 shows the accuracy calculation results for Experiment 2.

test subject Conventional Proposed technique A 64.56% 92.97% B 72.80% 81.62% C 86.36% 97.08% D 73.57% 90.81% E 77.68% 89.14% F 66.56% 75.05% average 73.59% 87.78%

Next, as shown in (c) of FIG. 6, in the case of Experiment 3, the chest movement generated by the subject sitting on a chair and moving back and forth at a speed lower than the breathing rate is shown.

As a result, when measuring the respiration of a subject moving back and forth, as shown in FIG. 9, the conventional method has less effect of chest movement displacement due to respiration due to a change in distance from the radar. Also, it is inaccurate because the movement of breathing is measured out of the set range-bin. On the other hand, the proposed method extracts the specific velocity of the chest due to respiration regardless of the measurement range, and since the movement velocity frequency exists individually, it is possible to extract the exact respiratory rate in the movement situation compared to the existing method. Table 4 below shows the accuracy calculation results for Experiment 3.

test subject Conventional Proposed technique A 69.62% 89.57% B 69.71% 84.06% C 71.13% 93.98% D 80.95% 88.14% E 59.57% 85.97% F 65.59% 92.66%

Next, as shown in (d) of FIG. 6, in the case of Experiment 4, chest movements generated when the subject moves left and right while sitting on a chair are shown.

As a result, when measuring respiration of a subject moving left and right as shown in FIG. 10 , in the conventional method, the effect of chest movement displacement due to respiration is reduced due to a change in distance from the radar. Also, it is inaccurate because the movement of breathing is measured out of the set range-bin. On the other hand, the proposed method extracts the specific velocity of the chest due to respiration regardless of the measurement range, and since the movement velocity frequency exists individually, it is possible to extract the exact respiratory rate in the movement situation compared to the existing method. Table 5 below shows the accuracy calculation results for Experiment 4.

test subject Conventional Proposed technique A 84.66% 94.63% B 67.09% 93.04% average 75.88% 93.82%

The experimental results show that the existing research method and the method proposed in this paper are compared in two experiments accompanied by motion, and respiration estimation is dramatically improved in the presence of motion.

As a result, in the case of the present invention, a method for detecting the microvelocity of the chest in terms of Doppler velocity was proposed. The proposed method confirmed improved breathing accuracy through experiments and algorithm simulations compared to existing research methods in terms of breathing accuracy in the presence of motion.

The existing research method showed 73.59% accuracy in respiration measurement of subjects standing up without leaning back, whereas the proposed method showed 87.78% accuracy. In measuring the respiration of subjects moving back and forth, the existing method showed 69.43% accuracy, but the proposed method showed 89.06% respiration rate estimation accuracy. In addition, the existing method showed 75.88% accuracy in measuring the respiration of subjects moving left and right, while the proposed method showed 93.82% respiration rate estimation accuracy.

Through this, the proposed method in terms of Doppler velocity proved that it is possible to extract the specific velocity of the chest due to respiration in the presence of motion regardless of the measurement range. This research method, which alleviates the RBM phenomenon, can be used in a wide range of areas, such as measuring the respiration of occupants in vehicles with many movements and medical devices using radar.

Although the present invention has been described with reference to the embodiments shown in the drawings, this is merely exemplary, and those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. Therefore, the true technical scope of protection of the present invention should be determined by the technical spirit of the appended claims.

10: FMCW radar 100: respiratory rate estimation device
110: signal acquisition unit 120: signal processing unit
130: calculation unit 140: estimation unit

Claims (12)

In the respiratory rate estimation method performed in the FMCW radar-based respiratory rate estimation device,
Receiving a signal transmitted through the FMCW radar and then reflected from a target according to time;
Acquiring Doppler magnitude data according to distance and speed by performing a first order discrete Fourier transform and then a second order discrete Fourier transform on the signal received from the FMCW radar;
calculating a sum of absolute values of Doppler amplitudes for each distance observed at a corresponding velocity component for each velocity within the measurement velocity range of the FMCW radar;
deriving a velocity value of the target at a corresponding time by averaging the sum of absolute values of Doppler magnitudes for each distance obtained for each velocity; and
Respiratory rate estimation method comprising the step of estimating the respiration rate of the target for each hour from the frequency component analyzed within the window by applying a window every hour to the velocity data according to time. The method of claim 1,
Obtaining the Doppler magnitude data,
Respiratory rate estimation method for obtaining the Doppler magnitude data by applying the first and second discrete Fourier transforms to the intermediate frequency signal generated by applying a low-pass filter to the combined signal of the transmission signal and the reception signal of the FMCW radar. The method of claim 1,
The step of deriving the speed value of the target,
Respiratory rate estimation method for deriving the velocity value D _v (t) of the target at the time t using the following equation:

Here, N _chirp is the number of chirps, v ₀ is each velocity component within the measured velocity range, and D(t,v ₀ ) is the absolute value of the Doppler magnitude for each distance at time t obtained from the corresponding velocity component v ₀ represents the sum. The method of claim 3,
The D (t, v ₀ ) is a respiratory rate estimation method defined by the following equation:

Here, Z(t,r _k ,v ₀ ) is the Doppler magnitude for each distance observed in the corresponding velocity component v ₀ , and r _k represents each distance component corresponding to the measured distance range. The method of claim 1,
Estimating the respiratory rate,
A plurality of peak frequencies detected by fast Fourier transforming the data on the window applied at the current time point (t) are individually compared with the tracking frequency values determined from the window at the previous time point (t-1), and then the peak frequency with the minimum frequency deviation is determined. determining a tracking frequency value at a current point in time (t); and
Respiratory rate estimation method comprising the step of determining the respiratory rate (Respiratory Rate, RR) per minute at the current time point (t) through the determined tracking frequency value. The method of claim 5,
The step of estimating the number of breaths per minute,
Respiratory rate estimation method for calculating the target's respiratory rate per minute (RR) at the current time point (t) by applying the tracking frequency value determined at the current time point (t) to the equation below:

here,

represents the tracking frequency value determined at the current point in time (t). In the FMCW radar-based respiratory rate estimation device,
a signal acquiring unit receiving a signal transmitted through the FMCW radar and then reflected from a target according to time;
a signal processing unit for obtaining Doppler magnitude data according to distance and speed by performing a first order discrete Fourier transform on the signal received from the FMCW radar and then a second order discrete Fourier transform;
For each speed within the measurement speed range of the FMCW radar, the sum of the absolute values of the Doppler magnitude for each distance observed in the corresponding velocity component is calculated, and the sum of the absolute values of the Doppler magnitude for each distance obtained for each speed is averaged to a calculation unit for deriving a speed value of a target; and
A respiratory rate estimation apparatus including an estimator for estimating the respiratory rate of the target for each hour from the frequency components analyzed within the window by applying a window for each hour to velocity data according to time. The method of claim 7,
The signal processing unit,
Respiratory rate estimation device for obtaining the Doppler magnitude data by applying the first and second discrete Fourier transforms to the intermediate frequency signal generated by applying a low-pass filter to the combined signal of the transmission signal and the reception signal of the FMCW radar. The method of claim 7,
The calculation unit,
Respiratory rate estimation device for deriving the velocity value D _v (t) of the target at the corresponding time t using the following equation:

Here, N _chirp is the number of chirps, v ₀ is each velocity component within the measured velocity range, and D(t,v ₀ ) is the absolute value of the Doppler magnitude for each distance at time t obtained from the corresponding velocity component v ₀ represents the sum. The method of claim 9,
The D (t, v ₀ ) is a respiratory rate estimation device defined by the following equation:

Here, Z(t,r _k ,v ₀ ) is the Doppler magnitude for each distance observed in the corresponding velocity component v ₀ , and r _k represents each distance component corresponding to the measured distance range. The method of claim 7,
The estimator,
A plurality of peak frequencies detected by fast Fourier transforming the data on the window applied at the current time point (t) are individually compared with the tracking frequency values determined from the window at the previous time point (t-1), and then the peak frequency with the minimum frequency deviation is determined. After determining the tracking frequency value at the current time point (t),
Respiratory rate estimating device for determining the respiratory rate (Respiratory Rate, RR) per minute at the current time point (t) through the determined tracking frequency value. The method of claim 11,
The estimator,
Respiratory rate estimator for calculating the target's respiratory rate per minute (RR) at the current time point (t) by applying the tracking frequency value determined at the current time point (t) to the equation below:

here,

represents the tracking frequency value determined at the current point in time (t).

Images