JP2019141300A - Biological information acquisition device - Google Patents

Abstract

To solve a problem that complete cancellation has not been achieved and there is a room for improving accuracy in a scheme that perform canceling using a sensor for body motion noise acquisition in addition to a sensor for pulse wave waveform acquisition during acquisition of a pulse wave waveform with a multiaxial pressure sensor.SOLUTION: In acquisition of a pulse wave waveform with a multiaxial pressure sensor, a pair of a sensor for pulse wave waveform acquisition and a sensor for body motion noise acquisition is formed to perform cancellation. In order to increase cancellation accuracy, another sensor set is put on different places that are in substantially the same environment, for example, left and right wrists, and a correction coefficient is calculated on the basis of a pulse wave waveform acquired there and applied.SELECTED DRAWING: Figure 7

Description

The present invention relates to an apparatus for acquiring biological information using a pressure sensor.

Attempts have been made to acquire biological information such as pulse waves with high frequency using a pressure sensor. The tonometry method is already known for this purpose. However, when performing measurement by the tonometry method, a certain level of skill is required to attach the sensor to the subject, and the measurement site must be stationary during the measurement, so that general users can live daily. There is much room for improvement in order to easily perform high-frequency biological information measurement while sending.
As another directionality, for example, there is a method of acquiring biological information using a pressure sensor having sensitivity in three axial directions, such as a technique disclosed in JP-A-2016-202908 and JP-A-2017-006672. ing.

JP, 2006-202908, A JP 2017-006672 A

(Body motion cancellation accuracy problem)
However, even with the tonometry method or the triaxial pressure sensor method, there is a problem that the acquisition of biological information is hindered by the body movement of the measurement subject, and it is difficult to acquire continuously continuously in daily life as well as sports. In order to perform useful measurements, improvement for body motion cancellation is required.

In the above-described prior art, a technique for canceling a body motion component from biological information using two triaxial pressure sensors is disclosed, but it is still not sufficient, and a pulse wave waveform as biological information is accurately obtained. Further improvements are required to obtain it.

The present invention has been made to solve the above-described problems, and can be worn by a person to be measured by using a plurality of multi-axis pressure sensors that can detect pressures of two or more axial components. Another object of the present invention is to provide a biological information acquisition device that can be used continuously in daily life.

The present invention has been made to solve the above problems.

The actuating device of the present invention comprises:
“A biological information acquisition device for measuring biological information of a person being measured with a pressure sensor,
A plurality of multi-axis pressure sensors that detect pressures in two or more axial directions that intersect at a predetermined angle, a signal detection unit that detects a component signal of each axis of the multi-axis pressure sensor, and a multi-axis detected by the signal detection unit An arithmetic device that calculates the signal of each axis of the pressure sensor,
The plurality of multi-axis pressure sensors are attached to the measurement subject, and two of the plurality of multi-axis pressure sensors are predetermined from the multi-axis pressure sensor A and the multi-axis pressure sensor A attached near the point where the biological information acquisition is performed. A multi-axis pressure sensor B is mounted at a position separated by a distance of
The biological information of the person being measured is input to the multi-axis pressure sensor pair as pressure,
The arithmetic device performs an operation based on the signal of each axis component detected by the multi-axis pressure sensors A and B of the multi-axis pressure sensor pair, and mixes in the biological information acquired by the multi-axis pressure sensor A based on the operation result. Reduce body movement noise,
The two multi-axis pressure sensor pairs are attached to the subject so that the vascular impedances from the heart are equal and the movement of each other is not affected by each other, and the two multi-axis pressure sensor pairs are the pair Q and the pair R. Yes, the side that mainly acquires biometric information is the pair Q, and the other side is the pair R,
The arithmetic device acquires biological information in a situation where body movement occurs on the pair Q side and body movement does not occur on the pair R side, and the body of the pair Q sensor A is obtained based on the pair A sensor A and the pair R sensor A. The motion component is extracted, the correction coefficient is calculated from the relationship between the Q motion sensor A body motion component and the signal detected by the pair Q sensor B, and the motion cancel calculation based on the pair Q sensor A and the pair Q sensor B is performed. Apply correction factors to improve the accuracy of body movement cancellation calculations. ''
It is characterized by that.

With this feature, it is possible to cancel body movements with higher accuracy than when only one multi-axis pressure sensor pair is used to acquire biological information.

As a result, the biological information can be acquired with high accuracy, and the biological information can be acquired by the measurer himself or continuously in daily life.

According to the present invention, it is possible to achieve highly accurate and continuous biometric information acquisition, and the above-described body motion cancellation accuracy problem can be solved.

The present invention will be described below with reference to the drawings.

It is a figure explaining the tonometry method which is the conventional method. It is a figure which shows the influence of the body movement with respect to a sensor. It is a figure which shows the influence of the body motion with respect to a pulse wave waveform. It is a wave form diagram which shows the concept of body movement cancellation. It is a wave form diagram which shows the concept of body movement cancellation. It is a figure (substitute photograph) which shows the outline | summary of this invention apparatus. It is a figure (substitute photograph) which shows the experimental condition of this invention apparatus. It is a figure (substitute photograph) which shows the experimental condition of this invention apparatus. It is a graph which shows a pulse wave waveform about experiment of this invention apparatus. It is a graph shown about the correlation of the right and left wrist which is an experimental result of the device of the present invention. It is a graph shown about the correction coefficient and waveform which are the experimental results of this invention apparatus. It is a graph shown about the waveform after noise removal which is an experimental result of the device of the present invention. It is a graph shown about the effect of amendment which is an experimental result of the device of the present invention. Table 1 is an experimental result of the device of the present invention. Table 2 is an experimental result of the device of the present invention.

First, the tonometry method as a conventional technique will be described with reference to FIG.
The Tonomelli method is a method of measuring blood pressure by pressing a pressure sensor against the skin immediately above the blood vessel on the wrist. As shown in FIG. 1, since the vertical component of the tension acting on the blood vessel wall is zeroed by pressing the blood vessel so that the upper part of the blood vessel is flat, the pressure applied to the pressure sensor is equal to the blood pressure. Can be measured. However, blood pressure measurement by the Tonomé method has a problem that measurement cannot be performed correctly unless the influence of body motion such as the wrist joint movement, the inertial force when the arm is shaken, or the height change from the heart is removed.

Accordingly, the inventors have advanced research aimed at reducing these effects. As shown in FIG. 2, the body motion sensor measures noise of the same magnitude as the noise included in the waveform measured by the blood pressure pulse wave sensor, and the effect of body motion is eliminated by subtracting the two waveform data. Was.
However, since the noise contained in the blood pressure pulse wave does not necessarily match the noise due to body movement, the effect of body movement could not be sufficiently removed even if the two waveform data are simply subtracted. .
The apparatus of the present invention also uses a set of two multi-axis pressure sensors as shown in FIG. The “multi-axis pressure sensor A” described in the claims is a sensor for acquiring pulse waves, and the “multi-axis pressure sensor B” is a sensor for acquiring noise due to body movement. “Pair” means a set of these two sensors.

FIG. 2 shows the principle of body motion noise removal using two three-axis pressure sensors.
The blood pressure pulse wave is measured using two sensors, a sensor for measuring the blood pressure pulse wave and a sensor for measuring noise due to body movement.
The blood pressure pulse wave sensor measures noise caused by body movement in addition to force generated in blood vessels by pulsation. On the other hand, the force generated in the blood vessel by pulsation is not measured by the body motion sensor, and only noise due to body motion is measured. Measurement by these two sensors is started at the same time, and by subtracting noise measured by the body motion sensor from the waveform measured by the blood pressure pulse wave sensor, a blood pressure pulse wave with reduced influence due to body motion can be acquired. However, with this method, since the magnitude of noise included in the waveform measured by the blood pressure pulse wave sensor cannot be confirmed, the effect of body movement cannot be sufficiently removed by simple subtraction alone. Therefore, noise included in the blood pressure pulse wave during body movement is extracted, and a correction coefficient is calculated so that the noise measured by the body movement sensor matches the extracted noise magnitude. After correcting the noise measured by the body motion sensor using the correction coefficient, the principle described above is used.

In order to extract only noise due to body movement from blood pressure pulse wave data during body movement, measurement is performed using both hands. While the left hand performs body movement during measurement, the right hand always measures in a resting state. Then, as shown in FIG. 3, the blood pressure pulse wave sensor and the noise due to body movement are measured from the left hand blood pressure pulse wave sensor, whereas only the blood pressure pulse wave is measured from the right hand blood pressure pulse wave sensor. . Therefore, only the noise due to body movement can be extracted by subtracting the output of the right hand blood pressure pulse wave sensor from the left hand blood pressure pulse wave sensor. By this method, the magnitude and waveform of noise included in the waveform measured by the blood pressure pulse wave sensor can be confirmed.
The “multi-axial pressure sensor pair Q” described in the claims means the sensor pair attached to the left wrist in this embodiment, and the “multi-axial pressure sensor pair R” is attached to the right wrist in this embodiment. Sensor pair.

The extracted noise and the noise measured by the body motion sensor are compared, and the ratio of the magnitude is calculated as a correction coefficient. The extracted noise and the waveform of the noise measured by the body motion sensor are shown in FIG.
The correction coefficient calculation method first obtains the difference between the maximum peak and the minimum peak for each amplitude for two noise waveforms. Next, the ratio of the extracted noise amplitude to the noise amplitude measured by the body motion sensor is calculated for each mountain. Then, the average value of the calculated ratio is calculated as a correction coefficient.
As shown in FIG. 5, after correcting the waveform measured by the body motion sensor by the correction coefficient, the corrected waveform is subtracted from the waveform measured by the blood pressure pulse wave sensor as described above, thereby Noise can be removed.

The apparatus of the present invention is intended to finally remove noise due to body movement with only one hand. In the method disclosed in this specification, an optimum correction coefficient for removing noise can be obtained for each body movement, but it cannot be obtained with only one hand. However, on the other hand, if the correction coefficient is constant, it is possible to remove noise with only one hand only by applying the correction coefficient. Therefore, an experiment was performed by confirming the range of body movement in which the correction coefficient becomes a substantially constant value and the value of the correction coefficient in the range while changing the body movement range. By applying a constant correction coefficient to the output of the body motion sensor, noise due to body motion can be removed with only one hand within the determined body motion range.

FIG. 6 shows an outline of the device of the present invention that was prototyped in the process of the present invention.
The device of the present invention is a band-type device that can be worn on the wrist, jointly developed by Chuo University and Shinano Kenshi Co., Ltd., and has a size of 92 × 76 × 24 mm. Further, the apparatus is composed of a MEMS triaxial pressure sensor, a pressing adjustment mechanism, and a fixing band. The pressing adjustment mechanism is a mechanism that can adjust three axes. The mechanism can adjust the position in a direction perpendicular to the traveling direction of the blood vessel, and can adjust the pressing angle by a screwing mechanism. . Further, the pressing force can be adjusted by adjusting the pressing displacement against the blood vessel by the screw feeding mechanism.

As the blood pressure measurement sensor, a MEMS 3-axis pressure sensor array for blood pressure measurement jointly developed with Chuo University, Touchence Corporation, and Shinano Kenshi Corporation was used. The pressure sensor is in the shape of a truncated cone having a pressure receiving surface diameter of 2.3 mm and a height of 1.0 mm, and five pressure sensors are arranged in a cross shape. The distance between each sensor is 3.4 mm. Further, the MEMS triaxial pressure sensor is connected to a signal processing board, and the output from the sensor is amplified by this board to enable A / D conversion, arithmetic processing, and communication. The signal processing board can be fixed to the arm using a band in the same manner as the device of the present invention. Since both wrists are used in the embodiment of the present invention, two devices of the present invention are used. The device worn on the left hand uses two of the five elements, one measures blood pressure pulse waves and the other measures noise due to body movement. On the other hand, in the device worn on the right hand, the blood pressure pulse wave is measured by one of the five-element sensors.

A setup for performing blood pressure pulse wave measurement during body movement is shown in FIG.
The output measured by the MEMS triaxial pressure sensor is transmitted to the personal computer via the substrate, and the measured data is recorded by software jointly developed by Chuo University and Shinano Kenshi Co., Ltd. Moreover, in order to perform measurement using both hands, the device of the present invention is attached to both wrists.
As shown in FIG. 8, the radial artery was marked so that the measurement position was not changed, the adjustment mechanism of the device of the present invention was adjusted, and the height of the measurement position was not changed.
The body movement is 0 ° when the wrist is not bent, and 0 ° to 10 °, 0 ° to 20 °, 0 ° to 30 °, 0 ° to 40 °, 0 ° to 40 ° before checking the angle with a protractor. There were seven types in the range of ° to -10 °, 0 ° to -20 °, and 0 ° to -30 °. Measurement of both wrists was started simultaneously, and after 10 seconds, the left hand was moved at about 1 Hz for 10 seconds. The right hand was always at rest at the same height as the measurement position of the left hand, and blood pressure pulse waves were measured.

FIG. 9 shows the measurement results with the right hand measured simultaneously with the left hand in the range of body movement of 0 ° to −30 °. From FIG. 9, it can be confirmed that the waveform measured with the left hand is disturbed in the waveform of the blood pressure pulse wave for 10 seconds after 10 seconds from the start of measurement.
First, in order to compare the amplitudes of two pulse waves, a correlation between blood pressure pulse waves in a resting state before body movement is calculated for the two waveforms. FIG. 10 shows a graph in which the vertical axis represents the right hand blood pressure pulse wave and the horizontal axis represents the left hand blood pressure pulse wave. When the regression line was calculated in this figure,
y = 0.998x + 0.0055
It became.
Since the two waveforms before the body movement showed a high correlation, the waveform measured with the right hand was multiplied by the reciprocal of the slope of the calculated formula to match the two pulse wave amplitudes. From this state, noise due to body movement was extracted by the method described above. The extracted noise is as shown in FIG.

Next, a correction coefficient is calculated.
FIG. 11 shows the noise extracted by the above-described method and the output of the body motion sensor measured simultaneously.
A correction coefficient is calculated for this waveform by the method described above. The ratio of the extracted noise to the noise measured by the body motion sensor was 0.499 in the range of 0 ° to −30 °. That is, this value is a correction coefficient for matching the noise measured by the body motion sensor with the extracted noise.
Continue to calibrate noise due to body movement. After correcting using the correction coefficient calculated with respect to the waveform measured by the body motion sensor, noise due to body motion was removed using the calibration method described above.

The waveform after noise removal is shown in FIG. In FIG. 12, it can be confirmed that the waveform after removal can reduce the disturbance of the blood pressure pulse wave due to body movement.
In order to evaluate the noise reduction effect, the correlation coefficient between the waveform before and after noise removal and the waveform with the right hand measured simultaneously in a resting state was calculated. FIG. 13 (a) shows the blood pressure pulse wave of the right hand on the horizontal axis, FIG. 13 (b) shows the thing with the waveform after noise removal on the vertical axis, and FIG. 13 (b) shows the thing with the waveform before removal.
In FIG. 13, the correlation coefficient between the waveform after noise removal and the right hand is 0.89, and the correlation coefficient between the waveform before removal and the right hand is 0.60. Since the correlation after noise removal was higher than that before removal, the waveform after noise removal was able to reduce noise due to body movement by the method disclosed in this example.

Next, a series of processes were also performed for body motions in the other six ranges, and correction coefficients were calculated from noise extraction, and correlation coefficients between the waveform and the right hand waveform before and after noise removal were calculated. The results are shown in FIG. 14 (Table 1). From the results of the correction coefficients in Table 1, the correction coefficients are close to each other when considered in terms of two body movements in front of and behind the wrist, although there are variations in each range. Therefore, it shows that there is a different tendency when the wrist is moved forward and backward. From the correlation coefficient results in Table 1, the waveform after noise removal shows a higher correlation than the waveform before removal in all ranges. For these reasons, noise caused by body movement can be reduced by the method disclosed in this embodiment regardless of the range of body movement.

As described above, in order to remove noise due to body movement with only one hand, a change in correction coefficient due to a change in body movement range was confirmed.
First, the correction coefficient was calculated by measuring five times in each range. FIG. 15 (Table 2) shows the average value and standard deviation of the correction coefficients calculated five times in each range.
As shown in Table 2, the average value of the correction coefficients was close in the range of 0 ° to 30 °. In the range of 0 ° to 40 °, it was shown that the average value was different from that of the other ranges of body movement, and the variation was large. Therefore, the average value of each correction coefficient in the range of 0 ° to 30 ° was calculated. The calculated correction coefficient was 0.696.

In addition, the average value of the correction coefficients was a close value even within the range of 0 ° to −30 °. Therefore, the average value of each correction coefficient in the range of 0 ° to −30 ° was calculated in the same manner. The calculated correction coefficient was 0.534.
Finally, noise was removed from all measurement data using the average value of the correction coefficients calculated. The noise reduction effect was evaluated using the correlation coefficient between the waveform and the right hand waveform before and after noise removal.

As a result, in the body movement in front of the wrist, the average correlation coefficient before noise removal was 0.85, whereas the average correlation coefficient after noise removal was 0.92. After the noise removal, the correlation coefficient is improved by 8.2% than before the removal.
In the body movement behind the wrist, the average correlation coefficient before noise removal was 0.81, and the average correlation coefficient after noise removal was 0.90. After the noise removal, the correlation coefficient is improved by 11% than before the removal.
From these facts, noise due to body movement can be removed with only one hand by using the average value of the calculated correction coefficients.

Up to this point, the results of experiments showing the principle of body motion noise removal and the basis of the present invention have been described.
By using the technique disclosed by the present invention, it is possible to effectively remove noise due to body movement and to measure biological information with high accuracy.
In the embodiment of the present invention, wrist joint movement is given as an example, but the effect of body motion noise when another muscle, tendon, or joint is moved is evaluated by a pressure sensor pair, and the correction coefficient is obtained to obtain higher accuracy. It becomes possible to measure.
In the embodiment of the present invention, the pair of multi-axis pressure sensors Q and R is the left and right wrists, but is not limited thereto. Any place where the blood vessel distance from the heart and the impedance are substantially equal and the pulse waveform that can be acquired is synchronized on the time axis may be used. Furthermore, it is more preferable if the movement of the left wrist does not directly affect the right wrist as in both wrists.

Using pressure sensors that can be mass-produced, biometric information can be obtained with high accuracy, and there are advantages such as high economic efficiency and practicality, and wide applicability.

None

Claims (1)

A biological information acquisition device for measuring biological information of a measurement subject with a pressure sensor,
A plurality of multi-axis pressure sensors for detecting two or more axial pressures that intersect at a predetermined angle;
Signal detection means for detecting a signal of a component of each axis of the multi-axis pressure sensor;
An arithmetic unit that calculates signals of the respective axes of the multi-axis pressure sensor detected by the signal detection means;
With
The plurality of multi-axis pressure sensors are attached to the subject,
Two of the plurality of multi-axis pressure sensors are a multi-axis pressure sensor A attached in the vicinity of a point where biological information acquisition is performed, and a multi-axis pressure sensor attached at a position away from the multi-axis pressure sensor A by a predetermined distance. B and a multi-axis pressure sensor pair,
The biological information of the measurement subject is input to the multi-axis pressure sensor pair as pressure,
The arithmetic unit performs an operation based on signals of the respective axis components detected by the multi-axis pressure sensors A and B of the multi-axis pressure sensor pair, and the biological information acquired by the multi-axis pressure sensor A based on the calculation result Body motion noise mixed in
The multi-axis pressure sensor pair is attached to the subject so that the two have the same vascular impedance from the heart and are not affected by body movements.
The two multi-axis pressure sensor pairs are a pair Q and a pair R, the side mainly acquiring biological information is the pair Q, and the other side is the pair R,
The arithmetic unit acquires biological information in a situation where body movement occurs on the pair Q side and body movement does not occur on the pair R side, and based on the pair A sensor A and the pair R sensor A, A body motion component is extracted, a correction coefficient is calculated from the relationship between the body motion component of the pair Q sensor A and the signal detected by the pair Q sensor B, and body motion cancellation based on the pair Q sensor A and the pair Q sensor B is performed. A biological information acquisition apparatus characterized in that the correction coefficient is applied to the calculation to improve the accuracy of the body motion cancellation calculation.