JP2016168149A - Pulse measuring device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a pulse measuring device that can measure a pulse with high accuracy in a comparatively simple structure, and uses a photoelectric pulse sensor.SOLUTION: An FFT64 being a Fourier transform processing unit finds a Fourier calculation mean value M50 being a mean value of the Fourier calculation results obtained by performing Fourier transform processing on each of segment signals S51-S55. A pulse detection unit 65 finds a peak frequency being a frequency component of the peak position from the Fourier calculation mean value M50, and performs pulse detection processing for outputting a BPM value corresponding to the peak frequency as pulse detection results D65. The pulse detection unit 65 performs the pulse detection processing by assuming only a frequency domain in the reference pulse width instructed by reference pulse width information D67 as a detection target, in a body motion recognition period when a body motion recognition signal S26 instructs a body motion state.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a pulse measuring device, and more particularly to a pulse measuring device using a photoelectric pulse signal from a photoelectric pulse sensor.

  In recent years, development of watch-type wearable health care products using photoelectric pulse sensors has been progressing. In the photoelectric pulse sensor, light is emitted from the light emitting element to the arm, finger, etc., the reflected or transmitted light is detected by the light receiving element, and the received light signal is converted into an electric signal to obtain a photoelectric pulse signal (photoelectric pulse signal). The pulse measured based on the photoelectric pulse signal was displayed on an LCD or the like. Such a type of photoelectric pulse clock exists on the market.

  A pulse measuring device using a photoelectric pulse sensor is disclosed in, for example, Patent Document 1 and Patent Document 2.

JP 2014-54448 A JP 2011-50745 A

  When a pulse measuring device using a photoelectric pulse sensor is wearable, the photoelectric pulse sensor is always attached to an arm or the like to measure the pulse, and body movement such as arm movement causes damage such as noise to the photoelectric pulse signal. I will give it. When the user of the pulse measuring device performs exercise or the like and damages the photoelectric pulse signal, there is a problem that the pulse cannot be measured accurately.

  On the other hand, in order to accurately measure the pulse, it is conceivable to perform complicated and sophisticated arithmetic processing on the photoelectric pulse signal. However, when a pulse measuring device is provided as a wearable device, since downsizing is required, it is practically difficult to employ the above-described complicated and advanced arithmetic processing that tends to increase power consumption. there were.

  The present invention has been made to solve the above problems, and an object of the present invention is to obtain a pulse measuring device using a photoelectric pulse sensor that can accurately measure a pulse with a relatively simple configuration.

  The pulse measuring device according to claim 1 according to the present invention is a photoelectric pulse sensor that detects a pulse of a person to be measured and obtains a photoelectric pulse signal, and acquires a measurement pulse of the person to be measured based on the photoelectric pulse signal, A pulse measurement unit that outputs pulse information that indicates the measurement pulse, the pulse measurement unit includes a pulse calculation unit that executes a calculation process for obtaining the measurement pulse, and the pulse calculation unit includes the photoelectric pulse Based on the signal, a pulse signal storage unit that stores a predetermined number of time-series pulse signals along a time series, and a plurality of sample pulse signals obtained from the predetermined number of time-series pulse signals are subjected to a Fourier transform process. A Fourier transform processing unit for obtaining the Fourier computation average value, which is an average value of the plurality of Fourier computation results, and the Fourier computation average value. Based, and a measured pulse acquisition unit for obtaining the measured pulse.

  According to a sixth aspect of the present invention, there is provided a pulse measuring device according to a sixth aspect of the present invention, a photoelectric pulse sensor for obtaining a photoelectric pulse signal by detecting a pulse of the measured person, acquiring the measured pulse of the measured person based on the photoelectric pulse signal, A pulse measurement unit that outputs pulse information that indicates the measurement pulse, the pulse measurement unit includes a pulse calculation unit that executes a calculation process for obtaining the measurement pulse, and the pulse calculation unit includes the photoelectric pulse A sample pulse signal based on the signal is sequentially received, a Fourier transform process is performed on the sample pulse signal to obtain a Fourier computation result, and a Fourier transform processing unit that executes the computation process, and a pulse detection result is obtained based on the Fourier computation result A pulse detection unit that executes a pulse detection process; and the pulse detection results sequentially obtained from the pulse detection unit are received as a plurality of pulse detection results in time series, and the plurality of pulse detection results Based on the average value of the beat detection result, and a measured pulse determination unit for obtaining the measured pulse.

  The pulse calculation unit in the pulse measurement device of the present invention according to claim 1, after performing a Fourier transform process on each of a plurality of sample pulse signals obtained from a predetermined number of time-series pulse signals by a Fourier transform processing unit, By executing a calculation process for obtaining an average value to obtain a Fourier calculation average value, a highly accurate measurement value can be obtained for a measurement pulse.

  As a result, a highly accurate measurement pulse can be obtained from the pulse calculation unit based on the Fourier arithmetic mean value. Furthermore, the pulse measuring device can be realized with a relatively simple configuration by using only the Fourier transform processing unit as the main component for performing the above arithmetic processing in the pulse measuring unit.

  The pulse calculation unit in the pulse measurement device according to the present invention described in claim 6 is based on an average value of a plurality of pulse detection results sequentially obtained from the pulse detection unit using the Fourier calculation result, and the measurement pulse from the measurement pulse determination unit. Have gained.

  As a result, the pulse measuring device of the present invention can obtain a highly accurate measured pulse based on the average value of a plurality of pulse detection results. Furthermore, the pulse measuring device can be realized with a relatively simple configuration by using only the Fourier transform processing unit as the main component for performing the above arithmetic processing in the pulse measuring unit.

It is a block diagram which shows the whole structure of the pulse measuring device which is embodiment of this invention. It is a block diagram which shows the internal structure of the signal processing part shown in FIG. 3 is a block diagram illustrating an internal configuration of a pulse calculation unit according to Embodiment 1. FIG. It is a graph for the effect explanation using the Fourier arithmetic mean value. It is a graph for the effect explanation using the Fourier arithmetic mean value. It is a graph for the effect explanation using the Fourier arithmetic mean value. It is a graph which shows a time-dependent change of the Fourier-calculation result based on a photoelectric pulse signal. It is a graph which shows the Fourier-calculation result based on the photoelectric pulse signal in a rest period. It is a graph which shows the Fourier-calculation result based on the photoelectric pulse signal in the exercise period. It is a graph which shows the effect by the setting of a reference pulse width. 6 is a block diagram showing an internal configuration of a pulse calculation unit in Embodiment 2. FIG.

(1. Overall configuration)
FIG. 1 is a block diagram showing an overall configuration of a pulse measuring device according to the present embodiment (Embodiment 1 and Embodiment 2) of the present invention. As shown in the figure, the present embodiment is composed of a PPG sensor 1 that is a photoelectric pulse sensor, a pulse measuring unit 11, a display 9, an acceleration sensor 21, and a speaker 27.

  The PPG sensor 1 irradiates the arm, finger, etc. with light from the light emitting element, detects the reflected light or transmitted light with the light receiving element, converts the received light signal into an electric signal, and detects the pulse, thereby detecting the photoelectric pulse signal S1. (PPG signal) is output to the pulse measuring unit 11. The sampling frequency of the photoelectric pulse signal S1 of the PPG sensor 1 is about 100 to 300 Hz.

  On the other hand, the acceleration sensor 21 detects a movement of an arm or the like by a measurement subject wearing the pulse measurement device of the present embodiment, and outputs an acceleration signal S21 that indicates an acceleration component in three axes.

  The pulse measuring unit 11 performs arithmetic processing including Fourier arithmetic processing based on the photoelectric pulse signal S1, measures the measured pulse, and indicates the pulse information HR indicating the measured pulse and the calorie consumption indicating the consumed calorie calculated from the measured pulse. Information CC is output. The display 9 displays the measured pulse indicated by the pulse information HR and the calorie consumption visually indicated by the calorie consumption information CC.

  Furthermore, the pulse measuring unit 11 recognizes whether or not the movement of the measurement subject wearing the pulse measuring device of the present embodiment is a body movement state that affects the measured pulse based on the acceleration signal S21. A motion detection process is executed, and based on the result of the body motion detection process, a body motion recognition signal S26 that indicates the presence or absence of a body motion state is output. Furthermore, the pulse measuring unit 11 outputs a speaker drive signal S27 that instructs sound output to the speaker 27 at a predetermined timing in a period in which the body motion recognition signal S26 indicates the body motion state.

  When the speaker drive signal S27 from the pulse measuring unit 11 instructs to output sound, the speaker 27 moves the body movement to the user (measured person) of the pulse measuring device according to the present embodiment, as will be described in detail later. Audio output for notifying that it is in a state is performed.

  The pulse measuring device of the present embodiment having such a configuration has the function of the pulse measuring unit 11 in an arithmetic unit including a CPU or the like that can be attached to an arm or the like on which the PPG sensor 1, the acceleration sensor 21, the speaker 27, and the display 9 are mounted. For example, it can be implemented as a wristwatch-type compact wearable device.

(2. Pulse measurement unit)
The pulse measurement unit 11 includes an ADC 2, a buffer 3, a buffer 4, a signal processing unit 5, a pulse calculation unit 6 (16), a calorie consumption calculation unit 7, a timing control unit 8, an ADC 22, a signal processing unit 25, and a body motion recognition unit 26. Consists of When realizing a wearable pulse measuring device, each of the components 2 to 8, 22, 25, and 26 in the pulse measuring unit 11 is executed at least partially by a program process using a CPU based on software.

  The ADC 2 outputs to the buffer 3 a pulse signal S2 obtained by digitizing the photoelectric pulse signal S1 by A / D conversion. The buffer 3 buffers the pulse signal S2 with a period of 1 to 5 seconds and outputs the pulse signal S3. The buffer 4 buffers the pulse signal S3 at a period of 30 to 60 seconds and outputs the pulse signal S4 to the signal processing unit 5. The signal processing unit 5 performs various signal processing on the pulse signal S4 and outputs the pulse signal S5 to the pulse calculation unit 6.

  The pulse calculation unit 6 or the pulse calculation unit 16 calculates a measurement pulse by performing a calculation process mainly including a Fourier calculation process based on the pulse signal S5. At this time, the pulse calculation unit 6 uses the pulse signal S4 buffered in the buffer 4 as an execution unit of the calculation process, and executes the calculation process once for the pulse signal S3 unit buffered in the buffer 3. Update cycle. Hereinafter, in this specification, for convenience of explanation, the buffer 3 is buffered at a 5-second cycle that is an update cycle, and the buffer 4 is buffered at a 30-second cycle that is a unit cycle of Fourier arithmetic processing according to the first embodiment described later. It demonstrates as a structure to ring.

  Then, the pulse calculation unit 6 outputs pulse information HR instructing the measurement pulse to the calorie consumption calculation unit 7. The calorie consumption calculation unit 7 calculates the calorie consumption of the measurement subject based on the measured pulse indicated by the pulse information HR, the calorie consumption information CC indicating the calculated calorie consumption, and the pulse information HR obtained from the pulse calculation unit 6 Is output to the display 9.

  The ADC 22 outputs an acceleration signal S22 obtained by digitizing the acceleration signal S21 by A / D conversion to the signal processing unit 25. Note that buffers corresponding to the buffers 3 and 4 may be inserted between the ADC 22 and the signal processing unit 25.

  Similar to the signal processing unit 5, the signal processing unit 25 performs various signal processing on the acceleration signal S <b> 22 and outputs the acceleration signal S <b> 25 to the body motion recognition unit 26.

  Based on the acceleration signal S25, the body motion recognition unit 26 obtains an acceleration vector V (n), which will be described in detail later, and executes a body motion detection process for detecting whether the acceleration vector V (n) exceeds the body motion reference value. When the body motion reference value is exceeded, a body motion recognition signal S26 for instructing the body motion state is output to the internal pulse calculation unit 6.

  That is, the body motion recognition unit 26 performs body motion detection processing for detecting whether or not the measurement subject is in a body motion state that may affect the measurement pulse based on the acceleration signal S21 obtained from the acceleration sensor 21. The body motion recognition signal S26 is generated to instruct the presence or absence of the body motion state based on the detection result of the body motion detection process. Hereinafter, the period in which the body movement recognition signal S26 indicates the body movement state is referred to as “body movement recognition period”.

  Further, the body motion recognition unit 26 obtains an acceleration spectrum (value) obtained by classifying the acceleration intensity of the acceleration vector V (n) by the pulse component (frequency component), and the body motion spectrum data D26 indicating the content of the acceleration spectrum is determined as the pulse. Output to the calculation unit 6. The acceleration spectrum is obtained, for example, by performing a Fourier transform process (FFT process) on the acceleration vector V (n).

  At this time, the body motion spectrum data D26 is output to the pulse calculation unit 6 (pulse detection unit 65 described later) so as to be synchronized with the Fourier arithmetic average value M50 under the control of the timing control unit 8. That is, under the control of the timing control unit 8, the Fourier calculation average value M50 based on the photoelectric pulse signal S1 and the body motion spectrum data D26 based on the acceleration signal S21 in the same period are supplied to the pulse detection unit 65 in the pulse calculation unit 6. The timing is controlled so as to be output.

  Hereinafter, the operation of the body movement recognition unit 26 will be described in detail. Since the acceleration sensor 21 outputs an acceleration signal S21 indicating acceleration in the x, y, and z-axis directions, the vector value of the x, y, and z axes that are three axes is taken and a vector is obtained for a certain period, for example, 10 seconds. If the value is equal to or greater than a certain body motion reference value, it can be recognized that the body motion is intense.

  If the acceleration signal S21 is x (y), y (n) and z (n) obtained in the x, y and z directions obtained from the acceleration signal S25 via the signal processing unit 25, the acceleration vector V (n) is The following equation (1) or equation (2) obtained by simplifying equation (1) is obtained. Note that n is an index in the time direction.

  As a preparatory step for accurately obtaining the measured pulse in the body movement state, first, the acceleration vector V (n) is obtained by the equation (1) or the equation (2), and the following equation (3) is obtained over a predetermined average period. The average acceleration vector MV (n) is obtained as shown in FIG.

  In Expression (3), for example, when the data rate of the acceleration signal S25 is 25 Hz and the predetermined average period is 10 seconds, N = 250.

  When the average acceleration vector MV (n) obtained by the expression (3) exceeds the body motion reference value prepared in advance, the body motion is recognized as a severe body motion state, and the body motion state is instructed. A body movement recognition signal S26 is output. Thereafter, the pulse measurement for the reference pulse width setting process, which will be described in detail later, is performed while waiting for a resting state after the average acceleration vector MV (n) falls below the body motion reference value.

  Thus, the body motion recognition unit 26 calculates the acceleration vector V (n) and the average acceleration vector MV (n) from the acceleration signal S25 (x (n), y (n), and z (n)), and calculates the average. By comparing the acceleration vector MV (n) with the body motion reference value, a body motion detection process is performed to detect whether or not the measurement subject is in a body motion state that may affect the measured pulse. A body movement recognition signal S26 that indicates the presence or absence of the state is generated.

  That is, the body movement recognition unit 26 sequentially obtains the acceleration signal S25, the acceleration vector V (n), and the average acceleration vector MV (n) based on the acceleration signal S21 from the acceleration sensor 21, and the person to be measured takes the measurement pulse. A body motion detection process for detecting whether there is a body motion state that may affect the body motion is performed, and a body motion recognition signal S26 that indicates the presence or absence of the body motion state is generated based on the detection result of the body motion detection process. doing. As described above, the period in which the body movement recognition signal S26 indicates the body movement state is defined as the body movement recognition period.

  Furthermore, the body motion recognition unit 26 outputs a speaker drive signal S27 (notification signal) that instructs voice output (notification) at a predetermined timing to the speaker 27 during the body motion recognition period. Note that, as the predetermined timing, for example, a timing in which the body motion recognition signal S26 is a voice output timing in a period from the start of the body motion state instruction to the passage of several seconds can be considered.

  On the other hand, the speaker 27 that has received the speaker drive signal S27 instructing voice output functions as a body motion state notification unit that notifies the subject of the occurrence of a body motion state by performing sound output (first sound output). To do.

  Upon receiving the body movement recognition signal S26 instructing body movement, the pulse calculation unit 6 recognizes that it is a body movement recognition period, and does not deteriorate the accuracy of the measured pulse even during the body movement recognition period due to the influence of body movement. Various processes are performed. This point will be described in detail later.

  In addition, when the setting of the reference pulse width, which will be described in detail later, is completed, the pulse calculation unit 6 outputs a setting completion signal R27 instructing that to the speaker 27.

  The timing control unit 8 controls the timing of the output of the pulse signal S3 from the buffer 3 and the output of the pulse information HR of the pulse calculation unit 6, and the pulse calculation unit 6 and the body motion recognition unit 26 as described above. The timing is controlled so as to operate in synchronization with each other.

(2-1. Signal processing unit)
FIG. 2 is a block diagram showing an internal configuration of the signal processing unit 5 shown in FIG. FIG. 2A shows the internal configuration of the signal processing unit 5.

  First, the internal configuration of the signal processing unit 5 will be described with reference to FIG. The signal processing unit 5 includes an offset removing unit 51, a resampling unit 52, a baseline removing unit 53, an LPF (Low-pass filter) 54, and an HPF (High-pass filter) 55.

  The offset removal unit 51 performs an offset removal process so that the first input level of the pulse signal S4 becomes “0”, and outputs the pulse signal S41 after the offset removal to the resampling unit 52. The resampling unit 52 performs resampling processing so that the pulse signal S41 has a data rate suitable for pulse calculation, and outputs a pulse signal S42. The baseline removal unit 53 removes a signal wave serving as a baseline from the pulse signal S42 and outputs the pulse signal S43 to the LPF 54. The LPF 54 performs a filtering process on the pulse signal S43 to remove a component having a frequency higher than the cutoff frequency, and outputs the pulse signal S44. The HPF 55 performs a filtering process for removing a component having a frequency lower than the cutoff frequency on the pulse signal S44, and outputs the pulse signal S5.

<Embodiment 1>
(2-2. Pulse calculation unit)
FIG. 3 is a block diagram showing an internal configuration of the pulse calculation unit 6 in the first embodiment. As shown in the figure, the pulse calculation unit 6 includes segment storage units 61 to 63, an FFT (Fast Fourier Transform circuit) 64, a pulse detection unit 65, a reference pulse width setting unit 67, a smoothing processing unit 68, and a signal switching unit 69. Consists of

  Each of the three segment storage units 61 to 63 constituting the pulse signal storage unit can store a pulse signal S5 (time-series pulse signal) for 10 seconds, and each stores a pulse signal S5 storage unit for 5 seconds. It has a combination of partial storage units 61a to 63a and 61b to 63b.

  The pulse signals S5 for a total of 30 seconds are stored in the segment storage units 61 to 63, the newest pulse signal S5 for 5 seconds is stored in the partial storage unit 61a, and the oldest pulse for 5 seconds is stored in the partial storage unit 63b. The signal S5 is stored. Thereafter, at intervals of 5 seconds as the update period, the latest pulse signal S5 is stored in the partial storage unit 61a, the storage signal of the partial storage unit 61a is stored in the partial storage unit 61b, the storage signal of the partial storage unit 61b is stored in the partial storage unit 62a. The storage signal of the partial storage unit 62a is transferred to the storage unit 62b, the storage signal of the partial storage unit 62b is transferred to the partial storage unit 63a, the storage signal of the partial storage unit 63a is transferred to the partial storage unit 63b, and the storage signal of the partial storage unit 63b is discarded. Is done.

  Then, the segment signal S51 from the partial storage units 61a and 61b (stored signal thereof), the segment signal S52 from the partial storage units 61b and 62a, the segment signal S53 from the partial storage units 62a and 62b, and the segment signal S54 from the partial storage units 62b and 63a. The segment signal S55 is output to the signal switching unit 69 from the partial storage units 63a and 63b.

  The signal switching unit 69 performs a switching operation to sequentially output all the segment signals S51 to S55 (a plurality of sample pulse signals) to the FFT 64 within the update period of 5 seconds.

  The FFT 64 that is a Fourier transform processing unit performs Fourier transform processing (FFT processing) on each of the segment signals S51 to S55 to obtain Fourier computation results D51 to D55, and then is an average value of the Fourier computation results D51 to D55. A Fourier arithmetic average value M50 is obtained. Then, the FFT 64 outputs the Fourier calculation average value M50 to the pulse detection unit 65 and the reference pulse width setting unit 67, which are main parts constituting the measurement pulse acquisition unit.

  The pulse detector 65 detects the entire frequency region (substantially about 0 to 14 Hz) in a normal period that is not a body movement recognition period, that is, a period in which the body movement recognition signal S26 does not indicate a body movement state, A pulse frequency detection process is performed in which a peak frequency that is a frequency component at the peak position is obtained from the Fourier calculation average value M50, and a BPM (Beats Per Minute) value corresponding to the peak frequency is output as a pulse detection result D65. Since the frequency (Hz) and the pulse rate (BPM) are in a similar relationship that 60 times the frequency is the pulse rate, the “pulse component” is a common concept for the frequency (Hz) and the pulse rate (BPM) below. Sometimes called.

  The reference pulse width setting unit 67 receives the Fourier calculation average value M50 and the body movement recognition signal S26, and the period during which the measured person stops the body movement state and the body movement recognition signal S26 does not indicate the body movement state is a predetermined continuation. In a resting state after a lapse of a period, a reference pulse width is obtained based on the Fourier calculation average value M50, and reference pulse width information D67 indicating the reference pulse width is output to the pulse detecting unit 65.

  As the above-described predetermined duration, for example, a period in which the body movement recognition signal S26 does not indicate a body movement state is set to 30 seconds or more, which is a period in which the stored contents of the pulse signal S5 in all the segment storage units 61 to 63 are switched. It can be considered.

  As a result, even if the reference pulse width setting unit 67 is immediately after the body motion recognition period, the reference pulse width setting unit 67 is based on the body motion recognition signal S26 and the Fourier calculation average value M50 in the resting state in the pulse component range of the measurement subject in the resting state. A reference pulse width can be set, and reference pulse width information D67 indicating the reference pulse width can be output to the pulse detector 65.

  As a method for setting the reference pulse width, the reference pulse component corresponding to the pulse component (frequency or BPM) at the peak position detected from the Fourier arithmetic average value M50 in the resting state is set as the center (reference) in each of the positive direction and the negative direction. A method such as adding a predetermined pulse component width is conceivable.

  The pulse detector 65 detects only the frequency region (pulse component range) within the reference pulse width indicated by the reference pulse width information D67 during the body movement recognition period, and the frequency component at the peak position from the Fourier calculation average value M50. A pulse detection process is performed in which a certain peak frequency is obtained and a BPM value corresponding to the peak frequency is output as a pulse detection result D65.

  The pulse detector 65 further receives the body motion spectrum data D26, and in the body motion recognition period, the pulse component (frequency or pulse rate) at the peak position (first peak position) of the Fourier calculation average value M50 within the reference pulse width. When the pulse component at the peak position of the acceleration spectrum data coincides, invalid processing is executed that does not adopt the first peak position as the detection target and does not adopt it as the pulse detection result D65.

  Thereafter, the pulse detection unit 65 excludes the first peak position, and the frequency component at the peak position of the Fourier calculation average value M50 within the reference pulse width (the second peak position, the peak position having the next highest spectrum intensity after the first peak position). The BPM corresponding to (pulse component) is set as a pulse detection result D65.

  The smoothing processing unit 68 uses the pulse detection result D65 obtained in the past based on the pulse detection result D65 obtained from the pulse detection unit 65 as a measurement pulse, and uses the pulse obtained as a measurement pulse to indicate the measurement pulse. Information HR is output.

  The pulse detector 65, the reference pulse width setting unit 67, and the smoothing processing unit 68 described above function as a measurement pulse acquisition unit that obtains a measurement pulse based on the Fourier arithmetic average value M50.

(3. Effect)
(3-1. Effect of Fourier arithmetic average value M50)
4 to 6 are graphs for explaining the effect of obtaining the Fourier calculation average value M50 by the segment storage units 61 to 63, the signal switching unit 69, and the FFT 64. FIG. In these figures, the horizontal axis is BPM and the vertical axis is frequency spectrum intensity. The frequency spectrum intensity is shown as a relative magnitude.

  In the above example, a pulse signal S5 for a total of 30 seconds, which is a pulse of 100 BPM, is generated in a pseudo manner while mixing 59 BPM components as noise due to body movement.

  FIG. 4 shows a Fourier calculation result D100 obtained by collectively performing a Fourier transform process on the pulse signal S5 for 30 seconds. In this case, a detected pulse with a peak position PK100 to 59 BPM is obtained.

  5 and 6 divide the 30-second pulse signal S5 used in FIG. 4 into segment storage units 61 to 63 in units of 10 seconds and store the segment signals S51 to S61 obtained from the segment storage units 61 to 63. The content obtained by performing Fourier transform processing on S55 to obtain Fourier computation results D51 to D55 and further obtaining the Fourier computation average value M50 of the Fourier computation results D51 to D55 is shown.

  5A to 5C show Fourier calculation results D51 to D53, FIGS. 6A and 6B show Fourier calculation results D54 and D55, and FIG. 6C shows the Fourier calculation average value. M50 is shown. When the detected pulse is obtained independently for each of the Fourier calculation results D51 to D55, it is detected with a relatively large fluctuation in the range of 59 to 100 BMP from the peak positions PK51 to PK55. Specifically, 100 BPM is detected from peak positions PK51 and PK52, peak positions PK53 to 94 BPM are detected, and peak positions PK54 and PK55 to 59 BPM are detected.

  However, 100 BPM can be accurately detected from the Fourier calculation average value M50 of the Fourier calculation results D51 to D55.

  As described above, when a relatively strong spectrum intensity is detected in the vicinity of 59 BMP due to the influence of body movement or the like, when the Fourier transform processing is performed on the pulse signal S5 for 30 seconds (see FIG. 4), When the Fourier transform process is performed with the pulse signal S5 alone (FIGS. 5A and 5B), the possibility of erroneous detection is relatively high.

  On the other hand, by making the influence of noise caused by body motion relatively weak by the Fourier calculation average value M50, a more correct peak position PK50M can be recognized, and an accurate 100 BPM can be set as a detected pulse.

  As described above, the pulse calculation unit 6 in the pulse measurement device according to the first embodiment stores the three (predetermined number) of time series pulse signals in the segment storage units 61 to 63 by the FFT 64 that is a Fourier transform processing unit. After performing a Fourier transform process on each of the five (plural) segment signals S51 to S55 (sample pulse signals) obtained from the signal, a calculation process for obtaining an average value is executed to obtain a Fourier calculation average value M50. .

  As a result, as shown in the examples of FIGS. 4 to 6, the pulse measuring device according to the first embodiment can obtain a highly accurate measured pulse based on the highly accurate Fourier calculation average value M50. Furthermore, by using only the FFT 64 as the main component for performing the Fourier transform process, which is the calculation process in the pulse calculator 6, it is possible to realize a pulse measuring device with a relatively simple configuration.

  In addition, among the segment signals S51 to S55 that are a plurality of sample pulse signals, the segment signals S5i and S5 (i + 1) (i = 1 to 4) that are a pair of sample pulse signals that are temporally adjacent to each other correspond to 5 seconds. The time zones (stored signals of the partial storage units 61b, 62a, 62b, 63a) are overlapped. For this reason, as a result of improving temporal continuity between the segment signals S5i and S5 (i + 1) and smoothing the signal change, the pulse measuring device of the first embodiment obtains a measured pulse with higher accuracy. be able to.

(3-2. Removal of influence of body movement state)
Pulse measurement using the PPG sensor 1 is easily affected by body movement. In particular, body movements such as walking and jogging where the limb moves periodically change the blood flow according to the period of body movement. The periodic signal of body movement may affect the photoelectric pulse signal S1 of the PPG sensor 1, and the pulse period signal existing in the photoelectric pulse signal S1 may be buried.

  FIG. 7 is a graph showing the temporal change of the Fourier calculation result based on the photoelectric pulse signal S1. In the figure, the vertical axis indicates the frequency spectrum intensity, the horizontal axis indicates the number of samples, and the unit of the sampling frequency 100 Hz of the photoelectric pulse signal S1 is shown as (ms (milliseconds) × 10). As shown in the figure, it can be seen that the spectrum intensity greatly fluctuates during the exercise period T1 compared to the rest period T0, and the body movement affects the photoelectric pulse signal S1 of the PPG sensor 1.

  FIG. 8 is a graph showing a Fourier calculation result based on the photoelectric pulse signal S1 in the rest period T0 of FIG. In the figure, the horizontal axis represents frequency (Hz) and the vertical axis represents frequency spectrum intensity.

  As shown in FIGS. 9A and 9B, in any case, the normal peak position PK1 can be detected, and thus the measurement pulse can be obtained with high accuracy during the rest period T0.

  FIG. 9 is a graph showing a Fourier calculation result based on the photoelectric pulse signal S1 in the exercise period T1 of FIG. In the figure, the horizontal axis represents frequency (Hz), and the vertical axis represents spectral intensity.

  As shown in FIGS. 9A to 9D, in any case, the body motion peak position BK1 has a higher spectral intensity than the normal peak position PK1, and therefore the body motion peak position BK1 is set as the normal peak position PK1. If it is detected erroneously, it can be seen that an accurate measurement pulse cannot be obtained.

  In the pulse measurement device according to the first embodiment, the influence of the body movement state described above can be removed by the pulse detection unit 65 and the reference pulse width setting unit 67.

  FIG. 10 is a graph showing the effect of the reference pulse width WS. Since it is affected by body movement in the exercise period T1, when the pulse detection process for obtaining the peak position of the Fourier arithmetic mean value M50 is executed with all frequency regions as detection targets, as shown in FIG. There is a high possibility that BK1 is erroneously detected as the peak position for pulse determination.

  Therefore, in the body motion recognition period, the pulse detection unit 65 performs the above-described pulse detection processing with the reference pulse width WS indicated by the reference pulse width information D67 from the reference pulse width setting unit 67 as a detection target, as shown in FIG. By executing the above, it is possible to obtain the pulse detection result D65 corresponding to the frequency component (pulse component) of the normal peak position PK1, not the body movement peak position BK1.

  In other words, there is slight movement within the pulse period for several seconds in the body movement state, but there is no significant blurring. Therefore, by making the detection target within the reference pulse width WS set in the resting state, it exists outside the reference pulse width WS Thus, it is possible to easily exclude the body movement peak position BK1 to be obtained and to accurately acquire the measured pulse.

  As described above, the pulse detection unit 65 executes the pulse detection process for the detection target only within the reference pulse width in the Fourier calculation average value M50 during the body movement recognition period to obtain the pulse detection result D65, so that the measurement subject can Even in a body movement state due to exercise or the like, the pulse detector 65 can obtain a pulse detection result D65 with high accuracy, and as a result, the pulse calculator 6 having the pulse detector 65 accurately calculates the measured pulse. be able to.

  Furthermore, even when the body motion peak position BK1 exists within the reference pulse width WS, the pulse detection unit 65 performs the invalidation process described above, thereby causing an erroneous pulse detection result D65 accompanying the erroneous detection of the body motion peak position BK1. Can be avoided.

  That is, as described above, the pulse detector 65 detects the pulse component (average pulse component) at the peak position (first peak position) of the Fourier calculation average value M50 within the reference pulse width and the body motion during the body movement recognition period. When the pulse component (acceleration pulse component) at the peak position of the acceleration spectrum indicated by the spectrum data D26 coincides, the average pulse component is invalidated without being set as the pulse detection result D65, that is, the first peak position is determined. Invalid processing to be excluded from detection is executed.

  Then, the pulse detection unit 65 uses the pulse component at the peak position (second peak position) of the Fourier calculation average value M50 with the reference pulse width as a detection region, except for the first peak position, as the pulse detection result D65. Therefore, even if the body motion peak position BK1 exists within the reference pulse width WS, the pulse detector 65 can obtain an accurate pulse detection result D65.

  As described above, the pulse detector 65 in the pulse measuring device according to the first embodiment matches the acceleration pulse component at the peak position of the acceleration spectrum with the average pulse component at the peak position of the Fourier arithmetic average value M50 in the body motion recognition period. When invalidating, the invalidation processing for invalidating the average value pulse component is executed.

  For this reason, the pulse detection unit 65 accurately obtains the pulse detection result D65 without erroneous detection even if the measurement subject is in a body movement state due to exercise or the like and the acceleration pulse component is present within the reference pulse width. be able to.

(3-3. Reference pulse width setting part)
A predetermined duration (30) for measuring the reference pulse width by the sound output (first sound output) by the speaker 27 executed in response to the sound output (notification) instruction of the speaker drive signal S27 (notification signal). Second or more), the person to be measured can be prompted to maintain a resting state in which the body motion recognition signal S26 does not indicate a body motion state.

  The reference pulse width setting unit 67 performs the above-described reference pulse width setting process in a resting state in which the body movement recognition signal S26 does not indicate a body movement state for a predetermined duration, and after setting the reference pulse width, A setting completion signal R27 for instructing the completion of setting is output to the speaker 27.

  The speaker 27 that has received the body movement recognition unit 26 instructing the completion of the setting of the reference pulse width notifies the measurement subject of the completion of the setting of the reference pulse width by performing an audio output (second audio output). The measurement subject can be informed that the constraint has been released from the resting condition, so that the measurement subject can resume the exercise state.

  Since the reference pulse width setting unit 67 has received the body movement recognition signal S26, the state in which the body movement recognition signal S26 does not indicate the body movement state has been determined to be a resting state. The reference pulse width setting process is executed with the pulse component at the peak position of the Fourier arithmetic mean value M50 as the central reference. For this reason, since the pulse measuring device according to the first embodiment can always obtain a highly accurate reference pulse width in a resting state even when the subject is exercising, the pulse measuring device is highly accurate and is not affected by body movements. A measurement pulse can be obtained.

  Furthermore, the person to be measured can surely recognize the body movement state by the notification by the first sound output from the speaker 27 which is the body movement state notification unit, so that the person to be measured can rest immediately after the body movement state is recognized. Thus, the reference pulse width setting process by the reference pulse width setting unit 67 can be led to an executable state at an early stage.

<Embodiment 2>
(4-1. Overall configuration)
FIG. 11 is a block diagram showing an internal configuration of the pulse calculation unit 16 in the second embodiment. As shown in the figure, the pulse calculation unit 16 includes a pulse signal storage register 70, an FFT 74, a pulse detection unit 75, a reference pulse width setting unit 77, a smoothing processing unit 78, a pulse detection result storage register 80, an addition unit 87, and a division. Part 88.

  The overall configuration of the pulse measurement device according to the second embodiment is mainly different from the overall configuration shown in FIG. 1 in that the pulse calculation unit 6 in the pulse measurement unit 11 is replaced with the pulse calculation unit 16. Further, the following points are different from the first embodiment in the pulse measurement unit 11.

  The buffer 14 provided in place of the buffer 4 is different from the buffer 4 buffering at a cycle of 30 seconds, in that it is buffered at a cycle of 10 seconds, which is a unit cycle of Fourier calculation processing in the second embodiment.

  The body movement recognition unit 26 has substantially the same function as that of the first embodiment, and therefore has the same reference numerals as those of the first embodiment. However, the only difference is that the body motion spectrum data D26 is output to the pulse calculation unit 6 (pulse detection unit 75 described later) so as to be synchronized with the pulse detection result D80 under the control of the timing control unit 8. That is, under the control of the timing control unit 8, the pulse detection result D80 based on the photoelectric pulse signal S1 and the body motion spectrum data D26 based on the acceleration signal S21 in the same period are output to the pulse detection unit 75 in the pulse calculation unit 6. This is different from the first embodiment in that the timing is controlled as described above.

  The pulse measuring apparatus according to the second embodiment has the same configuration and operation as those of the pulse measuring apparatus according to the first embodiment except for the above in the overall configuration.

(4-2. Internal configuration of pulse calculation unit)
Hereinafter, the internal configuration of the pulse calculation unit 16 in the pulse measurement device according to the second embodiment will be described with reference to FIG.

  The two partial storage units 70a and 70b constituting the pulse signal storage register 70 can each store a pulse signal S5 (time-series pulse signal) for 5 seconds, and store a pulse signal S5 for a total of 10 seconds.

  Between the partial storage units 70a and 70b, the pulse signal S5 for the new 5 seconds is stored in the partial storage unit 70a, and the pulse signal S5 for the old 5 seconds is stored in the partial storage unit 70b. Thereafter, at every 5-second interval that is the update period, the latest pulse signal S5 is transferred to the partial storage unit 70a, the storage signal of the partial storage unit 70a is transferred to the partial storage unit 70b, and the storage signal of the partial storage unit 70b is discarded. .

  Then, a register storage signal S70 (sample pulse signal), which is a storage signal of the partial storage units 70a and 70b, is sequentially output to the FFT 74 every update period of 5 seconds.

  The FFT 74, which is a Fourier transform processing unit, performs a Fourier transform process (FFT process) on one unit of the register storage signal S70 to obtain a Fourier operation result D70, and then uses the Fourier operation result D70 as the pulse detection unit 75 and the reference pulse width. Output to the setting unit 77.

  The pulse detection unit 75, like the pulse detection unit 65 of the first embodiment, detects all frequency regions in the normal period that is not the body movement recognition period, and the peak frequency that is the frequency component at the peak position from the Fourier calculation result D70. And the pulse detection process for outputting the BPM value corresponding to the peak frequency as the pulse detection result D80 is executed.

  The reference pulse width setting unit 77 receives the Fourier calculation result D70 and the body movement recognition signal S26, and the period during which the measured person stops the body movement state and the body movement recognition signal S26 does not indicate the body movement state is a predetermined duration. In the resting state that has elapsed, the reference pulse width is obtained based on the Fourier calculation result D70, and the reference pulse width information D77 that indicates the reference pulse width is output to the pulse detector 75.

  As the above-mentioned predetermined duration, for example, the period in which the body movement recognition signal S26 does not indicate the body movement state is set to 10 seconds or more, which is the period in which the stored content of the pulse signal S5 in the pulse signal storage register 70 is switched. Can be considered.

  As a result, the reference pulse width setting unit 77 is the pulse component range of the measurement subject in the resting state based on the body motion recognition signal S26 and the Fourier calculation result D70 in the resting state even immediately after the body motion recognition period. The reference pulse width information D77 that sets the reference pulse width and indicates the reference pulse width can be output to the pulse detector 75.

  As a method for setting the reference pulse width, a setting method centering on the reference pulse component corresponding to the pulse component at the peak position detected from the Fourier calculation result D70 in the resting state can be considered as in the first embodiment.

  The pulse detector 75 detects only the frequency region (pulse component range) within the reference pulse width indicated by the reference pulse width information D77 in the body movement recognition period, and is a frequency component at the peak position from the Fourier calculation result D70. A pulse detection process for obtaining a peak frequency and outputting a BPM value corresponding to the peak frequency as a pulse detection result D80 is executed.

  The pulse detector 75 further receives the body motion spectrum data D26, and in the body motion recognition period, the pulse component at the peak position (first peak position) of the Fourier calculation result D70 within the reference pulse width and the peak position of the acceleration spectrum data. When the pulse component matches, the invalidation process is executed that does not adopt the first peak position as a detection target and excludes it as the pulse detection result D80.

  Thereafter, the pulse detection unit 75 excludes the first peak position, and calculates the BPM corresponding to the frequency component (pulse component) at the peak position (second peak position) of the Fourier calculation result D70 within the reference pulse width as the pulse detection result D80. To do.

  The pulse detection result storage register 80 includes five pulse detection result storage units 81 to 85. The pulse detection result D80 sequentially obtained from the pulse detection unit 75 is used as five (a plurality of) stored pulse detection results D81 in time series. It functions as a time-series pulse detection result storage unit that stores as D85 (pulse detection result).

  That is, among the pulse detection result storage units 81 to 85, the latest pulse detection result D80 is stored as the stored pulse detection result D81 in the pulse detection result storage unit 81, and the oldest pulse detection result D80 is stored in the pulse detection result storage unit 85. Stored as stored pulse detection result D85.

  Thereafter, each time a pulse detection result D80 is newly obtained, the latest pulse detection result D80 is stored in the pulse detection result storage unit 81, and the stored pulse detection result D81 (in the pulse detection result storage unit 81) is stored in the pulse detection result storage unit 82. The stored pulse detection result D82 is transferred to the pulse detection result storage unit 83, the stored pulse detection result D83 is transferred to the pulse detection result storage unit 84, and the stored pulse detection result D84 is transferred to the pulse detection result storage unit 85, and stored in the pulse detection result storage unit 85. The pulse detection result D85 is discarded.

  The adder 87 adds the stored pulse detection results D81 to D85 to obtain a pulse detection result total T80 that is the total value of the stored pulse detection results D81 to D85. The division unit 88 divides the pulse detection result total T80 by “5” to obtain a pulse detection result average value M80.

  The smoothing processing unit 68 uses the pulse detection result average value M80 obtained in the past based on the pulse detection result average value M80 obtained from the division unit 88 as a measurement pulse using the value obtained by performing the smoothing process as the measurement pulse. The instructing pulse information HR is output.

  The pulse detection result storage register 80, the addition unit 87, the division unit 88, and the smoothing processing unit 78 described above store the pulse detection results D80 obtained sequentially from the pulse detection unit 75 into five stored pulse detection results D81 to D85 in time series. Based on the average value of the stored pulse detection results D81 to D85 received as (a plurality of pulse detection results) and stored in the pulse detection result storage register 80, it functions as a measurement pulse determining unit that obtains a measurement pulse.

(5. Effect)
(5-1. Effects of Pulse Detection Result Average Value M80)
In the example shown in FIGS. 4 to 6, 100, 100, 94, 59, and 59 BPM are obtained as the stored pulse detection results D81 to D85. That is, when the detection pulse is obtained independently for each of the Fourier calculation results D51 to D55 (corresponding to the Fourier calculation result D70), the detection is performed with a relatively large fluctuation in the range of 59 to 100 BMP from the peak positions PK51 to PK55.

  However, 82.4 BPM closer to 100 BPM can be detected from the pulse detection result average value M80, which is the average value of the stored pulse detection results D81 to D85.

  As described above, when a relatively strong spectrum intensity is detected in the vicinity of 59 BMP due to the influence of body movement or the like, when the Fourier transform processing is performed on the pulse signal S5 for 30 seconds (see FIG. 4), When the Fourier transform process is performed with the pulse signal S5 alone (FIGS. 5A and 5B), the possibility of erroneous detection is relatively high.

  On the other hand, by making the influence of noise caused by body motion relatively weak by the pulse detection result average value M80, it is possible to obtain a detected pulse from 59 BPM with a more accurate measured pulse close to 100 BPM.

  As described above, the pulse calculation unit 16 in the pulse measurement unit 11 of the pulse measurement device according to the second embodiment uses a plurality of pulse detection results D80 (D81 to D81) sequentially obtained from the pulse detection unit 75 using the Fourier calculation result D70. Based on the average value of D85), the measurement pulse is obtained from the measurement pulse determination unit (pulse detection result storage register 80, addition unit 87, division unit 88, and smoothing processing unit 78).

  As a result, the pulse measuring device according to the present embodiment can obtain a measured pulse with relatively high accuracy based on the pulse detection result average value M80 in which the influence of noise is weakened. Furthermore, by using only the FFT 74 as the main component for performing the Fourier transform process, which is a calculation process in the pulse calculator 16, it is possible to realize a pulse measuring device with a relatively simple configuration.

  Further, by changing the register storage signal S70 stored in the pulse signal storage register 70 with an update period of 5 seconds, the pair of register storage signals S70 that the FFT 74 receives in time continuously has a time zone of 5 seconds. Duplicate. That is, assuming that the register storage signal S70 that is temporally continuous is S701 and S702, the storage data of the partial storage unit 70a in the register storage signal S701 and the storage signal of the partial storage unit 70b in the register storage signal S702 are the same.

  For this reason, as a result of improving the continuity of the temporally continuous register storage signal S70 and smoothing the signal change, the pulse measuring device of the second embodiment can obtain a more accurate measured pulse. .

(5-2. Elimination of effects of body movement status)
Like the pulse detection unit 65 of the first embodiment, the pulse detection unit 75 detects the pulse within the reference pulse width indicated by the reference pulse width information D77 from the reference pulse width setting unit 77 in the body movement recognition period. Detection processing can be executed to obtain a pulse detection result D80.

  As described above, the pulse detection unit 75 executes the pulse detection process for the detection target only within the reference pulse width in the Fourier calculation result D70 in the body motion recognition period to obtain the pulse detection result D80, so that the measurement subject exercises. Even if the body is in a body motion state, the pulse detector 75 can obtain a highly accurate pulse detection result D80, and as a result, the pulse calculator 16 having the pulse detector 75 can accurately calculate the measured pulse. Can do.

  Furthermore, even when the body motion peak position exists within the reference pulse width, the pulse detection unit 75 performs the above-described invalidation process, thereby detecting an erroneous pulse detection result D80 accompanying the erroneous detection of the body motion peak position. Can be avoided.

  That is, as described above, the pulse detection unit 75 performs the pulse component (calculation result pulse component) at the peak position (first peak position) of the Fourier calculation result D70 within the reference pulse width and the body motion spectrum in the body motion recognition period. When the pulse component (acceleration pulse component) at the peak position of the acceleration spectrum indicated by the data D26 coincides, the calculation result pulse component is invalidated without setting it as the pulse detection result D80, that is, the first peak position is detected. Invalid processing to be excluded is being executed.

  Then, the pulse detection unit 75 uses the pulse component at the peak position (second peak position) of the Fourier calculation result D70 with the reference pulse width as a detection region, except for the initial peak position, as the pulse detection result D80. Therefore, even if the body motion peak position exists within the reference pulse width, the pulse detector 75 can obtain an accurate pulse detection result D80.

  As described above, the pulse detection unit 75 in the pulse measurement device according to the second embodiment matches the acceleration pulse component at the peak position of the acceleration spectrum and the calculation result pulse component at the peak position of the Fourier calculation result D70 during the body movement recognition period. At the time, the invalidation process for invalidating the pulse result pulse component is executed.

  For this reason, the pulse detection unit 75 accurately obtains the pulse detection result D80 without erroneous detection even when the measurement subject is in a body movement state due to exercise or the like and the acceleration pulse component is present within the reference pulse width. be able to.

(5-3. Reference pulse width setting part)
As in the first embodiment, the body motion recognition signal S26 does not indicate the body motion state for a predetermined duration (10 seconds or more) in order to measure the reference pulse width by the sound output (first sound output) from the speaker 27. The subject can be encouraged to remain at rest.

  The reference pulse width setting unit 77 performs the above-described reference pulse width setting process when the state in which the body movement recognition signal S26 does not indicate the body movement state is in a resting state after a predetermined duration, and after the reference pulse width is set, Similar to the reference pulse width setting unit 67 of the first mode, a setting completion signal R27 for instructing completion of setting of the reference pulse width is output to the speaker 27.

  The speaker 27 that has received the body movement recognition unit 26 instructing the completion of the setting of the reference pulse width notifies the measurement subject that the setting of the reference pulse width has been completed by outputting a sound, so that the measurement subject is in a resting state. Can be informed that the restriction has been released, so that the subject can resume the exercise state.

  Since the reference pulse width setting unit 77 has received the body motion recognition signal S26, the state where the body motion recognition signal S26 does not indicate the body motion state is determined to be a resting state, and the resting state is determined. The reference pulse width setting process is executed with the pulse component at the peak position of the Fourier calculation result D70 as the center reference. For this reason, the pulse measuring device according to the second embodiment, like the first embodiment, can always obtain a highly accurate reference pulse width in the resting state even during the exercise of the measurement subject. A highly accurate measurement pulse that is not affected can be obtained.

  Further, as in the first embodiment, the measurement subject can surely recognize the body movement state by the notification by the first sound output from the speaker 27, so that the measurement subject can quickly rest after the body movement state is recognized. Thus, the reference pulse width setting process by the reference pulse width setting unit 77 can be brought into an executable state at an early stage.

(5-4. Comparison with Embodiment 1)
In the first embodiment, the measured pulse is obtained by using the pulse detection result D65 detected based on the Fourier calculation average value M50 that is the average value of the five Fourier calculation results, whereas in the second embodiment, each measurement pulse is obtained. Is obtained by using a pulse detection result average value M80, which is an average value of five pulse detection results D80 (D81 to D85) detected based on a single Fourier calculation result.

  They are similar to each other in that they take an average value of five intermediate measurement values (Fourier calculation result in the first embodiment, and pulse detection result in the second embodiment), and the intermediate measurement values to be averaged are different. ing.

  Therefore, in the example shown in FIGS. 5 and 6, when the effect of the pulse detection process and the invalid process in the reference pulse width is ignored, the Fourier calculation that is an average value of five Fourier calculation results instead of being independent There is an advantage that the accuracy of the measurement pulse of the first embodiment is higher than that of the second embodiment by the amount based on the average value M50.

  However, in the second embodiment, by using a pulse detection process or an invalid process that is targeted for detection within the reference pulse width, the accuracy as an intermediate measurement value of the Fourier calculation result in the body motion recognition period can be increased. It is considered that there is no great difference in the accuracy of the finally obtained measurement pulse between the first embodiment and the second embodiment.

  On the other hand, in the update period of the pulse signal S5, it is necessary to perform five Fourier computations in the first embodiment, but in the second embodiment, one Fourier computation can be performed. Therefore, the pulse calculation unit 16 according to the second embodiment is implemented in that the processing burden associated with five Fourier calculation processes within the update cycle and the hardware load for holding the five Fourier calculation results can be reduced. It has the predominance which can be implement | achieved by a simple structure from the pulse calculation part 6 of the form 1.

<6. Other>
In the above-described embodiment (Embodiment 1 and Embodiment 2), the processing contents of the pulse detector 65 or the pulse detector 75 in the body movement recognition period in which the body movement recognition signal S26 indicates the body movement state. Although an example in which (the pulse detection process and invalid process within the reference pulse width) is different from the normal period has been shown, the same process as the body movement recognition period may be executed in the normal period. In this case, the pulse detector 65 (75) does not need to receive the body movement recognition signal S26. However, the reference pulse width setting unit 67 (77) must always output the reference pulse width information D67 (D77) indicating the reference pulse width to the pulse detection unit 65 (75), and the body motion recognition unit 26 always outputs the body pulse. It is necessary to output the dynamic spectrum data D26 to the pulse detector 65 (75).

  In the above-described embodiment, the example in which the reference pulse width setting process by the reference pulse width setting unit 67 (77) is executed at the timing after the first sound output from the speaker 27 is shown. For example, the reference pulse width setting process may be executed at a different timing (for example, when the use of the pulse measuring device of the above-described embodiment is started).

  In the first embodiment, the configuration in which the five segment signals S51 to S55 obtained from the three storage signals stored in the segment storage units 61 to 63 are output to the FFT 64 has been described. You may comprise so that it may output to FFT64 as three segment signals. In other words, a predetermined number of time-series pulse signals (stored signals) may be directly used as a plurality of sample pulse signals (segment signals).

  Similarly, in the second embodiment, the update processing of the register storage signal S70 in the pulse signal storage register 70 is performed in units of 5 seconds. However, the update processing is performed in units of 10 seconds, and the data storage between the register storage signals S70 that are temporally continuous is performed. You may make it the structure which does not produce duplication.

  In the present invention, the embodiments can be appropriately modified and omitted within the scope of the invention.

DESCRIPTION OF SYMBOLS 1 PPG sensor 3, 4, 14 Buffer 5,25 Signal processing part 6,16 Pulse calculation part 7 Calorie consumption calculation part 8 Timing control part 9 Display 11 Pulse measurement part 21 Acceleration sensor 26 Body motion recognition part 27 Speaker 61-63 Segment Storage unit 64, 74 FFT
65, 75 Pulse detection unit 67, 77 Reference pulse width setting unit 68, 78 Smoothing processing unit 69 Signal switching unit 70 Pulse signal storage register 80 Pulse detection result storage register 87 Addition unit 88 Division unit

Claims (10)

A photoelectric pulse sensor that detects the pulse of the subject and obtains a photoelectric pulse signal;
A pulse measurement unit that obtains a measurement pulse of the measurement subject based on the photoelectric pulse signal, and outputs pulse information indicating the measurement pulse;
The pulse measurement unit includes a pulse calculation unit that performs calculation processing to obtain the measurement pulse,
The pulse calculation unit
Based on the photoelectric pulse signal, a pulse signal storage unit that stores a predetermined number of time series pulse signals along the time series;
Each of the plurality of sample pulse signals obtained from the predetermined number of time-series pulse signals is subjected to Fourier transform processing to obtain a plurality of Fourier calculation results, and then a Fourier calculation average value that is an average value of the plurality of Fourier calculation results is obtained. A Fourier transform processing unit for executing the calculation processing to be obtained;
A measurement pulse acquisition unit for obtaining the measurement pulse based on the Fourier arithmetic mean value,
Pulse measuring device. The pulse measuring device according to claim 1,
A pair of sample pulse signals that are temporally adjacent to each other among the plurality of sample pulse signals are characterized in that at least some of the time zones overlap.
Pulse measuring device. The pulse measuring device according to claim 1 or 2,
An acceleration sensor that detects the movement of the measurement subject and obtains an acceleration signal;
Based on the acceleration signal, a body motion detection process is performed to detect whether or not the subject is in a body motion state that may affect the measurement pulse, and based on the detection result of the body motion detection process A body motion recognition unit that generates a body motion recognition signal that indicates the presence or absence of a body motion state, and a period during which the body motion recognition signal indicates a body motion state is defined as a body motion recognition period,
The measurement pulse acquisition unit,
A pulse detection unit that receives the Fourier calculation average value and performs a pulse detection process in which a pulse component at a peak position of the Fourier calculation average value is a pulse detection result;
A reference pulse width setting unit that obtains a reference pulse width that is a pulse component range in a resting state of the subject and outputs a reference pulse width information that indicates the reference pulse width. ,
The pulse detection unit further receives the body movement recognition signal and the reference pulse width information, and executes the pulse detection process with the reference pulse width in the Fourier calculation average value as a detection target in the body movement recognition period. ,
Pulse measuring device. The pulse measuring device according to claim 3,
The body motion recognition unit calculates an acceleration spectrum classified by a pulse component based on the acceleration signal, and further outputs body motion spectrum data indicating the content of the acceleration spectrum,
The pulse detection unit further receives the body motion spectrum data in the body motion recognition period, and includes an acceleration pulse component that is a pulse component at a peak position in the acceleration spectrum and a pulse component at a peak position of the Fourier calculation average value. When the average pulse component matches, the invalidation processing is performed to invalidate the average pulse component without using the pulse detection result.
Pulse measuring device. The pulse measurement device according to claim 3 or 4, wherein
The body movement recognition unit outputs a notification signal instructing notification at a predetermined timing in the body movement recognition period,
The pulse measuring device is
A body movement state notification unit for notifying the measurement subject of the occurrence of a body movement state when the notification signal indicates notification;
The reference pulse width setting unit further receives the body motion recognition signal and the Fourier calculation average value, and determines that the state in which the body motion recognition signal does not indicate a body motion state continues for a predetermined time as the rest state, Performing the reference pulse width setting process with reference to the pulse component at the peak position of the Fourier arithmetic mean value during the rest state,
Pulse measuring device. A photoelectric pulse sensor that detects the pulse of the subject and obtains a photoelectric pulse signal;
A pulse measurement unit that obtains a measurement pulse of the measurement subject based on the photoelectric pulse signal, and outputs pulse information indicating the measurement pulse;
The pulse measurement unit includes a pulse calculation unit that performs calculation processing to obtain the measurement pulse,
The pulse calculation unit
A Fourier transform processing unit that sequentially receives a sample pulse signal based on the photoelectric pulse signal, performs a Fourier transform process on the sample pulse signal, and obtains a Fourier computation result; and
Based on the Fourier calculation result, a pulse detection unit that executes a pulse detection process for obtaining a pulse detection result;
Including a measurement pulse determination unit that receives the pulse detection results sequentially obtained from the pulse detection unit as a plurality of pulse detection results in time series, and obtains the measurement pulse based on an average value of the plurality of pulse detection results. ,
Pulse measuring device. The pulse measuring device according to claim 6, wherein
Between the pair of sample pulse signals that the Fourier transform processing unit receives continuously in time, at least some of the time zones overlap,
Pulse measuring device. The pulse measuring device according to claim 6 or 7,
An acceleration sensor that detects the movement of the measurement subject and obtains an acceleration signal;
Based on the acceleration signal, a body motion detection process is performed to detect whether or not the subject is in a body motion state that may affect the measurement pulse, and based on the detection result of the body motion detection process A body motion recognition unit that generates a body motion recognition signal that indicates the presence or absence of a body motion state, and a period during which the body motion recognition signal indicates a body motion state is defined as a body motion recognition period,
The pulse detection unit performs the pulse detection process using the pulse component at the peak position of the Fourier calculation result as the pulse detection result,
The pulse calculation unit
A reference pulse width setting unit that executes a reference pulse width setting process that obtains a reference pulse width that is a pulse component range when the subject is in a resting state and outputs reference pulse width information that indicates the reference pulse width. ,
The pulse detection unit further receives the body movement recognition signal and the reference pulse width information in addition to the Fourier calculation result, and detects the inside of the reference pulse width in the Fourier calculation result in the body movement recognition period as the detection target. Execute the pulse detection process,
Pulse measuring device. The pulse measuring device according to claim 8, wherein
The body motion recognition unit calculates an acceleration spectrum classified by a pulse component based on the acceleration signal, and further outputs body motion spectrum data indicating the content of the acceleration spectrum,
The pulse detection unit further receives the body motion spectrum data in the body motion recognition period, and calculates an acceleration pulse component that is a pulse component at a peak position in the acceleration spectrum and a pulse component at a peak position of the Fourier calculation result When the result pulse component matches, the invalidation processing for invalidating the calculation result pulse component without using the pulse detection result is executed.
Pulse measuring device. The pulse measurement device according to claim 8 or 9, wherein
The body movement recognition unit outputs a notification signal instructing notification at a predetermined timing in the body movement recognition period,
The pulse measuring device is
A body movement state notification unit for notifying the measurement subject of the occurrence of a body movement state when the notification signal indicates notification;
The reference pulse width setting unit further receives the body motion recognition signal and the Fourier calculation result, determines that the state in which the body motion recognition signal does not indicate a body motion state continues for a predetermined time as the rest state, Performing the reference pulse width setting process based on the pulse component at the peak position of the Fourier calculation result in a resting state,
Pulse measuring device.

Images