JP2017144229A - Health management device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a health management device capable of determining a respiratory condition of a subject in a supine posture to perform daily health management.SOLUTION: A health management device includes: a sensor 11 that detects vibration caused by respiration of a subject in a supine posture on a bed 10; a stimulator 13 that gives stimulation to the subject; an abnormal respiration display device 21 that displays or stores respiratory arrest, slow respiration, tachypnea, low respiration, and hyperpnea of the subject; and a processing unit 20 that extracts a signal component of a frequency band of respiration from a sensor signal, obtains a distribution of intensity spectrum of the extracted signal component, determines a frequency of a maximum intensity spectrum, determines respiratory arrest, slow respiration, and tachypnea on the basis of the frequency of the maximum intensity spectrum, determines low respiration and hyperpnea on the basis of a magnitude of the maximum intensity spectrum, causes the stimulator to give electrical stimulation to the subject at a time of respiratory arrest, and causes the abnormal respiration display device to display or store the respiratory arrest, slow respiration, tachypnea, low respiration, and hyperpnea.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a health management device, and more particularly to a device that can perform daily health management by determining the respiratory state of a subject in a supine posture.

  Breathing is closely related to health status, and knowing changes in breathing especially during sleep is important in daily health care.

  In general, polysomnographic examination is known for measuring respiration during sleep, but it is necessary to insert a nasal cannula into the nasal cavity, which is difficult to use on a daily basis.

  On the other hand, when a breathing sound or snoring sound during sleep is recorded and a silent period is continuously detected for a predetermined time, for example, 10 seconds to 120 seconds, between the continuous breathing sound and snoring sound, the silent period is detected. Is known as an apnea state (Patent Document 1, Patent Document 2, Patent Document 3).

By the way, apnea is categorized as an abnormality in the number of breaths, but in addition to the abnormality in the number of breaths, there are other slow breaths where the number of breaths is less than normal breathing and tachycardia where the number of breaths is larger than normal breathing There is. In addition to abnormal breathing, abnormal breathing has normal breathing rates, such as hypopneas that reduce the amount of ventilation per breath and hyperventilation that increases the amount of ventilation per breath. It is also known that there is an abnormality in the amount of ventilation once.
However, although the method of recording the above breathing sound and snoring sound can detect the apnea state, it is difficult to detect other abnormal breathing.

  In contrast, the inventors of the present invention provided an acceleration sensor at a position on the bed corresponding to the chest and abdomen of a person who is supine, extracted a signal component of the respiration frequency band from the signal of the acceleration sensor, and extracted signals The frequency spectrum of the components was analyzed to find the distribution of the intensity spectrum, and the method of classifying the apnea depth into deep breathing, shallow breathing and apnea from the maximum value of the intensity spectrum to determine the apnea state has been proposed. (Non-Patent Document 1).

JP 2009-219713 A JP 2006-167427 A JP 2005-21255 A

Tottori University Graduate School Tsuyoshi Nakagawa, Daisuke Kushida, Akira Kitamura “Measurement and Analysis of Sleep Respiration Using Acceleration Sensors for Simple Diagnosis” The 59th System Control Information Society Conference Presentation Meeting (May 20-22, 2015) Osaka)

  In the method described in Non-Patent Document 1, since deep breathing and shallow breathing / apnea are classified, then shallow breathing and apnea are classified. Respiration can be determined, and slow breathing of 9 times / min or less and tachypnea of 25 times / min or more can be determined from the frequency in the maximum intensity spectrum of the signal component, and it is expected that simple breathing can be diagnosed.

  This invention pays attention to this point, and makes it a subject to determine the respiratory state of the test subject in a supine posture, and to provide the health management apparatus which enabled daily health management.

  Therefore, a health management device according to the present invention includes a bed in which a subject can take a supine posture, a sensor that detects vibrations caused by breathing of the subject in the supine posture on the bed, and a stimulator that stimulates the subject An abnormal breath display device that displays or records the apnea, slow breath, tachypnea, hypopnea, and hyperpnea of the subject, and extracts a signal component of the breathing frequency band from the sensor signal, and the extracted signal The distribution of the intensity spectrum of the component is obtained, the frequency of the maximum intensity spectrum is determined, and apnea, slow breathing and tachypnea are determined from the frequency of the maximum intensity spectrum, while hypopnea and hyperpnea are determined from the size of the maximum intensity spectrum. When apnea, apply the above stimulator to apply electrical stimulation to the subject, and display or record apnea, slow breathing, tachypnea, hypopnea, and hyperpnea using the abnormal breath display device. An arithmetic processing unit that, characterized by comprising a.

One of the features of the present invention is to detect vibration caused by breathing of a subject in a supine posture, extract a signal component of the breathing frequency band from the obtained sensor signal, and perform frequency analysis, for example, FFT (Fast Fourier Transform). The frequency analysis is performed to determine the spectrum distribution of the signal intensity, the frequency of the maximum intensity spectrum of the signal component is determined, and apnea, slow breathing and tachypnea are determined from the frequency of the maximum intensity spectrum, while the maximum intensity It is in the point that hypopnea and hyperpnea are judged from the magnitude of the spectrum and abnormal breathing is displayed.
Thereby, a change in respiration generally called abnormal breathing, in particular, a change in breathing during sleep can be detected, and abnormal breathing can be immediately determined.
Here, FIG. 4 shows waveforms of vibrations caused by normal breathing and abnormal breathing, for example, respiratory stop, sleep apnea, slow breathing, tachypnea, hypopnea, and hyperpnea.

  The vibration of the subject's breathing can be detected by an acceleration sensor or a highly responsive pressure sensor. Since the acceleration sensor captures the inclination, it may be provided on the bed corresponding to the subject's chest or abdomen, or may be provided on the leg of the bed. On the other hand, since the pressure sensor captures pressure, it is preferable to provide it on a bed corresponding to the subject's chest or abdomen.

The second feature of the present invention is that a stimulus is given to the subject when apnea is detected.
As a result, the subject can return to normal breathing from apnea even during sleep by stimulation, and can expect to significantly reduce the burden on the body, especially during sleep.
Examples of the stimulus include an electrical stimulus, a stimulus by vibration, and a stimulus by sound.

  The sensor and the arithmetic processing unit may be connected by a signal line, or may be connected wirelessly. However, if you connect wirelessly, there is a risk of signal loss due to obstacles between the sensor and the processing unit, and it is common to restore signals lost due to communication loss. It is difficult, and it is impossible to predict when the signal loss will appear. Therefore, it is preferable to perform interpolation by moving average processing on a plurality of sensor signals over time.

  Further, the sensor signal has a low S / N ratio, and it is difficult to extract only respiration data. Therefore, a signal component of a respiration frequency band, for example, 0.1 Hz to 0.6 Hz (6 to 36 times per minute) is extracted by BPF (band pass filter) processing.

  The number of sensors may be one, but it is preferable to obtain more information due to respiration by detecting vibration due to respiration of the subject. When using a plurality of sensors, normalization is preferably performed in consideration of matching the signal levels of the sensors and knowing only the signal variables.

The following formula can be used for normalization.

  When a plurality of sensors are used, the frequency spectrum of the respiratory frequency band components obtained from the plurality of sensors (sensor 1, sensor 2, sensor 3) is subjected to frequency analysis, and the intensity spectrum as shown in FIG. When the distribution is obtained and the intensity spectrum is subjected to spectrum synthesis as shown in FIG. 2B, the characteristics of the signal component can be further clarified.

Furthermore, if the maximum intensity spectrum of the signal component and its frequency are acquired while shifting along the time axis, the breathing rhythm and breathing depth can be determined. However, the difference in maximum intensity spectrum between apnea and hypopnea is small as shown in FIG.
Therefore, when the RMS value (root mean square) of the sensor signal having the largest maximum intensity spectrum among the signals of the plurality of sensors is obtained, as shown in FIG. 3B, between apnea and hypopnea. Since there is a difference, apnea and hypopnea can be classified.
Since the maximum intensity spectrum and the RMS value have different dimensions, normalization was performed using Equation 2 below. Here, Xmin and Xmax are the minimum value and the maximum value of the data.

1 is an overall configuration diagram showing a preferred embodiment of a health management device according to the present invention. It is a figure which shows the example of the intensity spectrum of the signal component in this invention, and the example of a spectrum synthesis | combination. It is a figure which shows the example of the intensity spectrum in the apnea and the hypopnea in this invention, and the example of RMS value. It is a figure which shows the waveform of the vibration resulting from normal breathing and abnormal breathing. It is a figure which shows the usage example of 2nd Embodiment. It is a circuit block diagram of the acceleration sensor module in 2nd Embodiment. It is a figure which compares the output waveform (b) of the acceleration sensor module in 2nd Embodiment with the output waveform (a) of one acceleration sensor. It is a figure which shows the characteristic calculated | required by the inclination of the flexible sheet | seat 31 at the time of a test subject's supine posture a and the side posture b in 2nd Embodiment.

  Hereinafter, the present invention will be described in detail based on specific examples shown in the drawings. FIG. 1 shows a preferred embodiment of a health care apparatus according to the present invention. In the figure, an acceleration sensor (for example, a 9-axis wireless motion sensor IMU-Z manufactured by ZMP Co., Ltd.) 11 is mounted on the bed 10 at positions corresponding to the chest and abdomen of the subject. For example, the signal shown in FIG. 1B is output to the arithmetic processing unit 20 as a radio signal.

  The arithmetic processing unit 20 is constituted by, for example, a CPU or a storage medium, and performs the following arithmetic processing. Data interpolation is performed by smoothing a plurality of, for example, 30 signals over time of the acceleration sensor 11 by moving average processing, and BPF (band pass filter) processing is performed from the sensor signal subjected to the data interpolation. As a result, a signal component having a respiration frequency of 0.1 Hz to 0.6 Hz (6 times / min to 36 times / min) is extracted.

  Next, frequency analysis is performed on the signal component in the respiratory frequency band by FFT (Fast Fourier Transform) to obtain the intensity spectrum distribution of the signal component. For example, 4096 signals are used for frequency analysis, but 512 signals (sampling interval: 0.1 second) can also be used, and normalization is performed for each processing data section. For example, the above-described formula 1 is used for normalization.

  Further, spectrum synthesis of intensity spectrum distributions (sensor 1, sensor 2, sensor 3) obtained from the signals of the three acceleration sensors 11 is performed, and the maximum intensity and the frequency of the synthesized intensity spectrum distribution are along the time axis. Is acquired while shifting.

The presence or absence of apnea / hypopnea and hyperventilation is determined from the magnitude of the maximum intensity of the obtained composite intensity spectrum. Further, an RMS value (root mean square) is obtained for the signal of the acceleration sensor 11 from which the maximum intensity spectrum is obtained, and apnea and hypopnea are discriminated.
Since the maximum intensity spectrum and the RMS value have different dimensions, normalization is performed. The above-described formula 2 is used for normalization.

  Further, the frequency of the maximum intensity spectrum is obtained in the obtained composite intensity spectrum distribution, and it is determined whether or not the breath is 9 breaths / min or less from whether the frequency is 0.15 Hz or less, and 0.4 Hz or more. Whether or not the tachypnea is 25 times / minute or more is determined.

Furthermore, the bed 10 and the acceleration sensor 11 are provided with a stimulator 13 that gives vibration to the subject with vibrations and sounds. When an apnea is detected, the subject is stimulated by a signal from the arithmetic processing unit 20. ing. A stimulator 13 that applies electrical stimulation to the subject's fingers may be attached, but the stimulator 13 is provided on the bed 10 or the acceleration sensor 11 in view of the simple examination of breathing without subject restraint. It is good.
Further, when apnea, slow breathing, tachypnea, hypopnea, and hyperpnea are detected, they are displayed on the abnormal breath display device 21 and recorded.

  As described above, a change in breathing called abnormal breathing, particularly a breathing change during sleep can be detected, and abnormal breathing can be easily detected.

  In addition, since a stimulus is given to the subject when apnea is detected, the subject can return to consciousness and return to normal breathing from sleep, especially during sleep. The burden on the body can be greatly reduced.

  In addition, you may add to the arithmetic processing unit 20 the function to detect a change in body temperature during sleep and to apply electrical stimulation to the subject when the body temperature changes abruptly.

  By the way, as shown in FIG. 5, a plurality of acceleration sensor modules 32 are mounted on a flexible sheet 31 to form an acceleration sensor sheet 30, and the acceleration sensor sheet 30 is placed on the comforter of the subject, It is proposed to detect the acceleration caused by the breathing of the subject without restriction. Such a method is easy to install and has the advantage that there is less risk of failure as in the case of installing under the subject and there is less discomfort during sleeping because there is no direct contact with the body.

  Common causes of apnea during sleep include sublingual tongue subtraction and soft palate subsidence. These occur in the supine position, are unlikely to occur in the lateral position, and are less likely to cause apnea in the lateral position. However, in the method of detecting the acceleration related to the respiration of the subject in the supine posture, if the subject is in the lateral posture, the component related to the respiration hardly appears in the detected signal, and there is a possibility that it may be determined as apnea. There is little need to estimate respiration in a lateral posture that is unlikely to become apnea.

  Therefore, by detecting the curvature of the acceleration sensor sheet from the static acceleration by the acceleration sensor, detecting the body movement change, and estimating the posture, it is possible to omit detection of respiratory abnormalities that are likely to be wasted in the lateral posture.

  That is, an acceleration sensor module is configured by a plurality of acceleration sensors and an addition averaging circuit that calculates an average of signals from the plurality of acceleration sensors, and the plurality of acceleration sensor modules are mounted side by side on a flexible sheet. In this configuration, the inclination of the flexible sheet is obtained from the signals of the plurality of acceleration sensor modules, and when the inclination of the flexible sheet exceeds the set inclination, it is determined as a lateral posture and the determination of abnormal breathing is stopped. be able to.

  FIG. 6 shows an example of the acceleration sensor module. As shown in FIG. 6, when the signals of the four acceleration sensors Acc1 to Acc4 are input to the inverting input of the operational amplifier (inverting addition averaging circuit) 40 to obtain the averaging, the white noise can be reduced to ½. it can.

  FIG. 7A shows a signal waveform in the case of one acceleration sensor, and FIG. 7B shows a signal waveform when the outputs of the four acceleration sensors are added and averaged, and has a high S / N ratio. Is confirmed.

  In addition, a low-pass filter 41 having a cutoff frequency, for example, 13 Hz, is configured in the circuit of FIG.

Now, the acceleration value of the X axis of each acceleration sensor module is Sout, the output value of 0 [G] of the X axis of the acceleration sensor module is S0out, and the output value of 1 [G] of the X axis of the acceleration sensor module is Assuming that S1out, the slope can be calculated by Equation 3.

  The inclination value can be drawn on two-dimensional coordinates and the posture of the subject can be estimated from the shape of the curve. An example is shown in FIG. 8, a diagram a shows a supine posture, a diagram b shows a scoliosis posture, and the subject takes a scoliosis posture when the apex of the curve height exceeds a threshold value c, such as apnea Stop detecting abnormal breathing.

10 Bed 11 Acceleration sensor 13 Stimulator 20 Arithmetic processor 21 Respiratory abnormality display device

Claims (2)

A health management device capable of performing a health management by determining a respiratory state of a subject in a supine posture,
A bed (10) in which the subject can take a supine posture;
A sensor (11) for detecting vibration caused by breathing of a subject in a supine posture on the bed (10);
A stimulator (13) for stimulating the subject;
An abnormal breath display device (21) for displaying or recording the subject's apnea, slow breath, tachypnea, hypopnea, hyperpnea;
The signal component of the respiration frequency band is extracted from the signal of the sensor (11), the distribution of the intensity spectrum of the extracted signal component is obtained, the frequency of the maximum intensity spectrum is determined, and the frequency of the maximum intensity spectrum is determined from the frequency of the maximum intensity spectrum. While judging breathing, slow breathing and tachypnea, hypopnea and hyperpnea are judged from the magnitude of the maximum intensity spectrum, the stimulator (13) is used to stimulate the subject during apnea, and an abnormal breath display device (21) an arithmetic processing unit which displays or records apnea, slow breathing, tachypnea, hypopnea, and hyperpnea (20),
A health management device comprising: The acceleration sensor module (30) is configured by a plurality of acceleration sensors (32) and an addition average circuit for calculating an addition average of signals of the plurality of acceleration sensors (32), and the plurality of acceleration sensor modules (30) are flexible. It is mounted side by side on the seat (31),
The arithmetic processing unit (20) obtains the inclination of the flexible sheet (31) from the signals of the plurality of acceleration sensor modules (30), and the inclination of the flexible sheet (31) exceeds a set inclination. The health management device according to claim 1, wherein the health management device is determined to have a scoliosis posture and abnormal breathing determination is stopped.

Images