TW200820939A - Method for measuring heart rate - Google Patents

Abstract

A method for measuring a heart rate is disclosed. The method comprises steps of: (a) collecting signals in a predetermined signal collecting period by a heart rate monitor, (b) calculating intervals between any two signals, (c) calculating heart rates through reciprocals of the intervals, (d) performing a majority voting algorithm by voting the heart rates into ballot boxes they belong, wherein the ballot boxes represent different heart rate ranges, (e) determining the ballot box having the most votes, and (f) calculating the true heart rate. Through the majority voting algorithm for measuring the heart rate, it is possible to extract the heartbeat signals from a mixture of pulses and noises and further obtain an accurate heart rate value.

Description

200820939 IX. Invention description: • [Technical field to which the invention belongs] This case relates to a heart rate measurement method, especially a high reliability heart rate measurement method. [Prior Art] Heart rate is an important parameter of human health, and it has important human health indicators, whether it is during exercise or at rest. For example, studies have shown that abnormal exercise heart rate and abnormal heart rate recovery rate are dangerous signs of sudden death, in which too low exercise heart rate has the highest risk of sudden death; and the higher the resting heart rate, or exercise The slower the heart rate recovery, the higher the risk index. The best way to measure heart rate is to perform an electrocardiogram (ECG) measurement and then use the R-R interval of the ECG to calculate the heart rate. However, the use of an electrocardiogram to measure heart rate is too costly and too laborious to be feasible. A variety of heart rate monitors have been developed on the market to monitor heart rate during exercise. The heart rate monitor comprises at least two parts, a pulse sensing unit and a heart rate display unit, wherein the pulse sensing unit comprises an electrode (or a transducer) and a filter circuit. However, pulse sensing technology is more susceptible to human interference than ECG technology. In order to accurately sense the pulse signal, the contact between the electrode of the heart rate monitor and the skin must be very stable, and this often depends on the way the user uses it. However, when measuring the heart rate during exercise, the pulse sensing unit 6 200820939 Vibration can easily cause noise. Commercially available sensing units include a range of components to sense pulse, such as band pass filters, peak detectors, and radio frequency circuits. Due to the lack of precise signal processing, the heart rate calculation unit may treat the received noise as a pulse signal. Therefore, these unfiltered noises may cause the heart rate display unit to display an incorrect heart rate value. In response to such a deficiency, the inventors have conceived and improved the method, and the invention has finally invented the method of the present invention. SUMMARY OF THE INVENTION The main purpose of the present invention is to provide a heart rate measurement method that utilizes a majority vote process to filter out noise to obtain an accurate heart rate rate value. To achieve the above objectives, one of the broader aspects of the present invention provides a heart rate measurement method that includes the following steps: (a) using a heart rate monitor to collect signals during a predetermined signal collection period; (b) calculating The interval between the two signals; (c) calculating the heart rate value by the reciprocal of the interval; (d) voting the heart rate value calculated in the foregoing step, and respectively inputting them into the ballot box representing the different heart rate ranges; e) find the ballot box with the highest number of votes; and (f) calculate the heart rate. According to the above concept of the present invention, the signal collection period is preferably 2 to 8 seconds. According to the above concept of the present invention, the information collected in the step (a) 7 200820939 includes a pulse signal and a noise signal. According to the above concept of the present invention, the step (b) includes calculating the interval between the signal and the last signal during the last signal collection period. According to the above concept of the present invention, in step (d), only the heart rate value within a reasonable heart rate range is voted, and the heart rate value not within the reasonable heart rate range is not voted. The reasonable heart rate range is 50-240. In one embodiment, wherein the heart rate range of the ballot box is obtained by dividing the reasonable R-R interval into a plurality of intervals and converting to a heart rate range, and the reasonable R-R interval is 0.25-1.2 seconds. In another embodiment, wherein the heart rate range of the ballot box is obtained by dividing the reasonable heart rate range into a plurality of intervals at an equal increase rate. According to the above concept of the present invention, the ballot box with the highest number of votes found in the step (e) is a single ballot box or two adjacent ballot boxes with the highest number of votes after the sum of the votes. According to the above concept in the present case, in step (e), if it is impossible to find the ballot box with the highest number of votes, the cumulative number of votes to the next signal collection period, repeat steps (a)-(d) to find the number of votes. The southernmost ballot box. According to the above concept of the present invention, step (f) calculates the heart rate by calculating the average of the heart rate values of the ballot box with the highest number of votes. Another broad aspect of the present invention provides a heart rate measurement method comprising the steps of: (a) utilizing a heart rate monitor to collect signals during a predetermined signal collection period; and (b) calculating between any two of the signals. (c) vote at the interval calculated by the preceding steps, and put them into ballot boxes representing different intervals; (d) find the ballot box with the highest number of votes 8 200820939; and (e) calculate Out of heart rate. According to the above concept in the present case, in the step (c), only voting is performed for the interval within the reasonable R-R interval, and voting is not performed at intervals not within the reasonable r_r interval. The Acacia R_R interval is

〇·25-1/2 seconds, and the interval time range of the ballot box is obtained by dividing the reasonable R_R interval into a plurality of intervals. ... According to the above concept of the present invention, the t-step (4) is obtained by counting the average of the heart rate values by the reciprocal of the interval between the two ballot boxes. [Embodiment] Some typical embodiments embodying the characteristics of the fish in the description of this case will be described in the following paragraphs. Scope: and the description and the heart rate of the custom = to limit the case. The wrong heartbeat rate value is obtained; the filtering: the 'easy method' uses the majority of the decision I I ^ readable heart rate to measure the accurate heart rate value. However, it is known to reconstruct the R_R interval to get in the hair; in some cases, it cannot be distinguished, but they

The variation of the interval is very different. For example, in a short period of time, R-R is more regular, while heterogeneous phase is no; in other words, the interval between R-R intervals is more irregular. Therefore, according to this 9 200820939 kinds of cases, a voting algorithm is proposed to accurately calculate the heart rate. The voting algorithm of this case will be further explained below. First, the voting algorithm used to measure heart rate in this case is performed by a microprocessor that performs signal detection through a commercially available heart rate monitor kit. The heart rate monitor includes a chest strap, a wrist display unit, and a receiving unit 'where the receiving unit is coupled to the microprocessor. Please refer to the first figure, which is a schematic diagram of the collection of signals for this case. If the _ figure does not 'tl5 l ···, tn denotes the time point when the signal collection period (SCP) receives the signal s ι,&,...,& respectively, during a predetermined signal collection period (SCP), the signal includes The pulse signal (equivalent to the heartbeat signal) and the noise signal, wherein the line is a pulse signal, and the dotted line is a noise signal. In addition, ~ is the last signal received before the start of the signal collection period. S〇' is the last signal received during the last signal collection period. By the intervention of noise, the RR interval has been cut into several segments by one or more noises, so it is not possible to directly use these signal intervals to reconstruct the RR interval. Therefore, the interval between any two signals must be further calculated to find out the true The pulse signal, and then the heartbeat rate value after filtering the noise. Taking the first picture as an example, during the signal collection period, signals Si, S2, I, S4, and S5 appearing at five time points, such as ti, t2, t3, and q ts, are collected, and the interval between the two signals is calculated. ,include:

^ ~~^〇h h-to U ~~ t5 -tQ i2 - q t3 —tx t4 —tx t5 —tl G —'2 h—h XA t5-13 h 10 200820939 and calculate the heart rate by the reciprocal of the interval, then Heart rate numbers within a reasonable heart rate range (50-240) are voted into a ballot box representing a range of heart rate ranges (eg, a reasonable RR interval (0.25 seconds - 1.2 seconds) is divided into a plurality of interval ranges, and then calculated These ranges represent the heart rate range), or the heart rate range ballot box that distinguishes the reasonable heart rate range by equal increase rate (for example, in the case of an increase rate of 1%, the heart rate range of the ballot box is [50-55], [ 56-62], [63-69], [70-77]···, and so on), but not heart rate numbers within a reasonable heart rate are not voted. At the end of the signal collection period to be sampled, the voting results are reviewed to find the ballot box with the highest number of votes, and the ballot box with the highest number of votes is considered to be the true heart rate, and the signal generating these heart rates contains the real pulse. . However, if the true heart rate is at the boundary of the ballot box heart rate range, it may result in the number of votes in the adjacent ballot box being equal without a single maximum number of votes. To avoid this situation, in a preferred embodiment, two phases are The number of votes in the adjacent ballot box is added to find the two adjacent ballot boxes with the highest number of votes, and the heart rate value of the two adjacent ballot boxes is the true heart rate, so the average heart rate can be calculated. In some cases, such as too much noise or a pulse signal not being detected, it may result in the inability to find the two adjacent ticket boxes with the highest number of votes in a single signal collection period. If the two adjacent ticket boxes cannot generate the highest ticket, the last signal is used as the SO during the next signal collection period, and the number of votes is accumulated until the next signal collection period, and the two adjacent ticket boxes with the highest number of votes are found. . As mentioned above, the variation of RR interval 200820939 is more limited than the interval between noises. Therefore, the two adjacent ballot boxes with the highest number of votes are more likely to be voted by the heart rate number calculated by the real RR interval. Got it. Therefore, the heartbeat rate can be accurately calculated by finding the true pulse signal in a majority decision by the calculation of the interval between the two signals and the voting process. Preferably, the signal collection period is between 2 and 8 seconds, and the shorter the signal collection period, the higher the resolution. The voting time can be calculated every time a signal is received, and it can be voted on, or it can be voted at the end of the signal collection period. According to another embodiment of the present invention, in addition to voting in the heart rate range, it is of course possible to vote in the interval of the signal. In this embodiment, the interval between the reasonable RR intervals (0.25 seconds - 1.2 seconds) is voted into the ballot box of different interval time ranges, for example, 0.25 seconds - 1.2 seconds is equally divided into 30 intervals. Box to vote, and find the two adjacent ballot boxes with the highest number of votes, you can find the real pulse signal. Similarly, the heart rate is calculated by calculating the heart rate value by the reciprocal of the interval between the two adjacent ballot boxes, and calculating the average of the heart rate values. An example of applying the voting algorithm of this case to calculate the heart rate will be described below. In this example, the signal collection period is set to 4 seconds, and a total of 13 signals are received during the signal collection period, namely Sb S2, ..., S13, and the last signal during the last signal collection period is S〇. . Step 1: Calculate the interval between signals: The first line of the table below shows the interval between S〇 to Si (0.25 seconds), S〇 to S2 (0.50 12 200820939 seconds)..., S〇S S13 (3.92 seconds). The second line lists the interval between s] to S2 (0·25 seconds), S! to S3 (0·33 seconds)..., S! to S13 (3.66 seconds), and so on, up to 312 to S13 (0.56) Seconds interval between noses. (Unit: 0.01 seconds)

To \ So S1 S2 s3 s4 s5 S7 S8 S9 S 10 Sn S 12 Si 25 - - S2 50 25 - - S3 58 33 8 - - -* S4 100 75 50 42 - - S5 111 86 61 53 11 - - - - - - - - 126 101 75 67 25 14 - - - - - - - s7 165 140 115 107 65 54 39 - - - - - - S8 200 175 150 142 100 88 74 34 - - - - - s9 222 197 171 163 121 110 96 56 21 - - - - S i〇279 254 229 221 179 167 153 113 79 57 - - - Sn 292 267 242 234 192 180 166 126 92 70 12 - - Sl2 335 310 285 277 235 224 209 169 135 113 56 43 - s 1 3 392 366 341 333 291 280 265 226 191 169 112 99 56

Step 2·· Calculate the heart rate by the reciprocal of the interval, as shown in the following table: ^N^rom To So s, s2 S3 s4 s5 s6 S7 s8 S9 S1 〇S11 S12 Si 240 - - S2 120 240 - - s3 103 181 750 - - S4 60 80 120 142 - - - - - - - - - S5 54 69 98 113 545 - - - - - - - - S6 47 59 80 89 240 428 - - - - - - - S7 36 42 52 56 92 111 153 - - - - - - S8 30 34 40 42 60 68 81 176 - - - - - s9 27 30 35 36 49 54 62 107 285 - - - - S 10 21 23 26 27 33 35 39 53 75 105 - - - Sn 20 22 24 25 31 33 36 47 65 85 500 - - s 1 2 17 19 21 21 25 26 28 35 44 53 107 139 - s 13 15 16 17 18 20 21 22 26 31 35 53 60 107 13 200820939 Step 3 · · Vote the above heart rate numbers within a reasonable range (5〇~24〇) to the following range: 240-214:3 147-138:2 106-102:2 213-193:0 137-129:0 101- 97:1 192-!75;2 128-121:0 96-92:1 174-161:0 120-113:3 91-88:1 160-148:1 112-107:4 87-84:1 83 -81:1 68-67:1 58-57:0 80-77:2 66-65:1 56-55:1 76-74:1 64-62:1 54-54:3 73-72:0 61 -61:1 53-52:3 71-69:1 60-59:3 51-50:0 where The heart of such a ballot box on behalf of, the Department of rate ranges reasonable range R_R

(0.25 sec - 1.2 sec.) is divided into 3 interval time ranges, and the heart rate range is calculated by reciprocal of the interval time. ^^1 At the end of the signal collection period to be sampled, examine the two adjacent tickets to find the two adjacent ballot boxes with the highest number of votes. In the above example, the highest votes are [120-113] and [112_1G7] two adjacent ballot boxes, so it is possible to draw the bottom line in the range of 107-12G w in the second table. The pulse signal, so the pulse signal in this example should be included in 0 2 s3 s4 s5, s7, s9, s10, s12, s13. That is, the true heart rate values collected during the 4-second signal collection should be 12〇, 120, 107, 107, in, 111, ill, 107, so the average heart rate can be taken. Of course, if the two adjacent ticket boxes cannot produce the highest ticket, the last signal S13 is used as the next-signal collection period _Sq, and the number of votes is accumulated to the next-signal collection period, and steps _ to step four are repeated. . S〇 during the signal collection period, and ##五··_ with the last signal as the next 14 200820939 Clear the number of votes, and then start from the first step. In this way, the real-time heart rate can be continuously detected for a continuous period of time. In order to further test the accuracy of the voting algorithm used in heart rate measurement, the heart rate value measured in this case is compared with the heart rate value measured by the electrocardiogram. Due to too much noise during exercise, it is impossible to read the RR interval from the three-lead ECG. Therefore, the voting algorithm of this case will be used to measure the resting heart rate and the recovery heart rate after exercise and the measurement using the three-lead electrocardiogram. The heart rate is compared. The second graph shows the resting heart rate measured by the voting algorithm and the three-lead ECG, respectively, and the third graph shows the heart rate after exercise using the voting algorithm and the three-lead ECG, respectively. The representative uses the heart rate measured by the voting algorithm in this case, while the dotted line represents the heart rate measured using the three-lead ECG. As can be seen from the second and third figures, the heart rate measured by the low-cost heart rate monitor combined with the majority voting algorithm of the case is quite close to the heart rate measured by the three-lead electrocardiogram, and the error is about 3%. It is within acceptable tolerances. Therefore, the majority voting algorithm in this case can filter out the noise and get a more accurate heart rate value. In addition, a part of the error (circle) is displayed in the third figure, and the estimated heart rate value is about one-half of the actual heart rate value. However, the error may be due to the fact that the intermediate pulse signal is not detected, so that the two are included. The interval between the RR intervals is added to the vote and becomes a majority, so that only the measured value of only half of the actual heart rate value is obtained, so the error is caused by the missing pulse signal, not by the noise signal. Furthermore, as can be seen from the figure, the majority of the decision in this case 15 200820939 ticketing algorithm can quickly restore the correct heart rate value estimation, and the above error containing two R-R intervals can be easily filtered out. Therefore, the majority voting algorithm in this case is a fairly reliable heart rate measurement method. The fourth graph shows the heart rate measured by the voting algorithm and the 12-lead ECG, respectively. The 12-lead ECG instrument is used in conjunction with the treadmill's five stages of exercise, sampling every 10 minutes for 10 seconds. Heart rate value. Similarly, the solid line represents the heart rate measured using the voting algorithm of the case, and the dotted line represents the heart rate measured using the 12-lead ECG. The heart rate of the second to fifth stages of the exercise is shown enlarged, and it can be seen that the voting algorithm in this case has a fairly good accuracy. In summary, this case proposes a high-reliability heart rate measurement method, which uses the majority voting process to reconstruct the R-R interval to avoid noise interference and obtain accurate heart rate values. In addition, this case only needs to cooperate with the low-cost heart rate monitor to detect the signal, without the need of high cost and high manpower ECG instrument, you can get the heart rate value with the accuracy and ECG, which can provide an important reference for human health indicators. . Therefore, this case has a very high creative value and is submitted in accordance with the law. This case has been modified by people who are familiar with the technology, but it is not intended to be protected by the scope of the patent application. 16 200820939 [Simple description of the diagram] The first picture: is a schematic diagram of the collection of signals for this case. Figure 2: shows the rest time heart rate measured by the voting algorithm and the three-lead ECG. The third picture shows the heart rate after exercise measured by the voting algorithm and the three-lead ECG. Figure 4: shows the heart rate during exercise measured by the voting algorithm and the 12-lead ECG. [Main component symbol description]

17

Claims (1)

200820939 X. Patent application scope: 1. A method for measuring heartbeat I includes the following steps: (8) Using a heart rate monitor to collect signals during the scheduled signal collection period; ^ (b) Calculating between any two of the signals Interval time; (4) calculating the heart rate value by the reciprocal of the interval time; (d) The above steps are used to calculate the heart rate values of the 屮 and / 矸 矸 , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , . The heart rate measurement method, wherein the information is as described in the second paragraph of the patent application, preferably between 2 and 8 seconds. 3. The heart rate measurement method described in the application for full-time covenant i, in which (4) is used in step (8) to reduce the amount of information. 4. The method of heart rate measurement as described in claim i, wherein step (8) comprises calculating an interval between the signal and the last signal of the last _ signal collection period. Take the method of measuring the rate as described in the scope of the patent application, wherein in step = (4), only the heart rate value within the range of reasonable heart rate is voted, and the heart rate value not within the range of reasonable heart rate is not voted. 6. The method of heart rate measurement according to claim 5, wherein the reasonable heart rate range is 50-240. The heart rate measurement method described in item 1 of the patent scope, wherein the heart rate range of the 200820939 ballot box is obtained by dividing the reasonable R-R interval into a plurality of intervals and converting them into a heart rate range. 8. The method of heart rate measurement according to claim 7, wherein the reasonable R-R interval is 0.25-1.2 seconds. 9. The heart rate measurement method according to claim 1, wherein the heart rate range of the voting box is obtained by dividing the reasonable heart rate range into a plurality of intervals by an equal increase rate. 10. The method of heart rate measurement according to claim 1, wherein the ballot box with the highest number of votes found in step (e) is a single ballot box. 11. The heart rate measurement method according to item 1 of the patent application scope, wherein the ballot box with the highest number of votes found in step (e) is the sum of the votes and the two highest number of votes. Neighboring ballot box. 12. The method for measuring heart rate according to item 1 of the patent application, wherein in step (e), if the ballot box with the highest number of votes cannot be found, the cumulative number of votes is added to the next signal collection period, and the steps are repeated. (a)-(d) to find the ballot box with the highest number of votes. 13. The heart rate measurement method according to claim 1, wherein the step (1) calculates the heart rate by calculating an average value of the heart rate value of the ballot box with the highest number of votes. 14. A heart rate measurement method comprising the steps of: (a) using a heart rate monitor to collect signals during a predetermined signal collection period; (b) counting the interval between two signals; 19 200820939, (c) vote at the interval between the uranium steps and the ballot boxes representing the different intervals; (d) find the ballot box with the highest number of votes; and (e) calculate the heart rate. 15. The heart rate measurement method described in claim 14, wherein the signal collection period is preferably 2 to 8 seconds. 16. The heart rate measurement method of claim 14, wherein the signal collected in the step (a) includes a pulse signal and a noise signal. 17. The heart rate measurement method of claim 14, wherein the step (b) comprises calculating an interval between the signal and the last signal during the last signal collection period. ^ 18. = The heart rate measurement method described in claim 14 of the patent application, wherein in = (C), only the interval between the reasonable corrections is performed, and the interval is not within the reasonable RR interval. Then do not vote. The heart rate measurement method described in claim 18, wherein the reasonable R-R interval is 0.25-1.2 seconds. The heart rate measurement method according to claim 19, wherein the interval time range of the ballot box is obtained by dividing the reasonable R-R interval into a plurality of intervals. 21. For the heart rate measurement method described in item 14 of the scope of patent application, the ballot box with the highest number of votes found in =(d) is a single and two ballot boxes. 22 & The heart rate measurement method according to claim 14, wherein the ballot box with the highest number of votes found in the step (4) is the sum of the votes 20 200820939 the highest number of votes Neighboring ballot box. 23. The heart rate measurement method according to claim 14, wherein in step (d), if the ballot box having the highest number of votes cannot be found, the cumulative number of votes is repeated until the next signal collection period, and the steps are repeated ( a)-(c) to find the ballot box with the highest number of votes. 24. The heart rate measurement method according to claim 14, wherein the step (e) calculates the heart rate value by the reciprocal of the interval time of the ballot box with the highest number of votes, and calculates the average of the heart rate values. The value is derived from the heart rate. twenty one