JP6716888B2 - Respiratory analysis device, respiratory analysis method and program - Google Patents

Description

Embodiments of the present invention relate to a respiratory analysis device, a respiratory analysis method, and a program.

Sleep apnea syndrome awakens the brain when breathing is resumed from an apnea state and makes the depth of sleep shallower, which causes increased fatigue during the day and decreased concentration, which may lead to traffic accidents. A disease that increases the likelihood of causing it. In recent years, sleep apnea syndrome has been known as a disease that causes various lifestyle-related diseases such as cardiovascular disease and diabetes, and is a disease recommended for early treatment and improvement of lifestyle and sleep environment. is there.

At present, for patients who come to the hospital because they feel that they have sleep apnea syndrome, either spontaneously or as pointed out by others, the sleep apnea syndrome tester is used during the examination admission and the biological information during sleep is used. By analyzing the measured biological information and clarifying the onset tendency of respiratory disorders, changes in sleeping posture, pattern of body movements and snoring, etc., medical examination and appropriate treatment are performed. There is.

However, sleep apnea syndrome has the inconvenience that there are few subjective symptoms and the patient is not fully manifested. For this reason, in recent years, a simple sleep apnea syndrome test apparatus has been developed, but there is an inconvenience that the burden on the patient is large, for example, it is necessary to wear a nasal cannula or the like. Therefore, there is a demand for the development of a new technique that can reduce the burden on the patient and can fully realize the patient with sleep apnea syndrome.

JP, 2007-292514, A

The problem to be solved by the present invention is to provide a respiratory analysis device, a respiratory analysis method, and a program that put a small burden on a patient and can realize the manifestation of a patient with sleep apnea syndrome.

According to the embodiment, the respiratory analysis device is wearable on the human body. This respiration analysis device measures a wearer's acceleration and a first frequency band obtained by executing a first filtering process on a waveform showing a temporal change of the measured acceleration. Detection means for detecting movement of the human body based on outliers in the waveform, and discrimination means for discriminating whether the detected movement of the human body is movement caused by respiration or movement caused by body movement. And. In the waveform of the first frequency band, the detection unit detects a time at which the frequency is less than a first threshold value as the outlier, and the determination unit sets a second value with respect to the measured acceleration waveform. In the waveform of the second frequency band obtained by executing the filtering process of step S1, the detection is performed based on the first characteristic amount indicating the characteristic of the waveform of the first section including the time detected as the outlier. It is determined whether the movement of the human body is the movement caused by the breathing or the movement caused by the body movement.

The figure which shows the schematic structure of the respiratory analysis system which concerns commonly on each embodiment. The figure which shows the hardware structural example of the biometric sensor apparatus which concerns commonly on each embodiment. The figure which shows the circuit structural example of the biosensor apparatus which concerns commonly on each embodiment. The block diagram which shows an example of a function structure of the respiratory analysis application program which concerns on 1st Embodiment. The figure which shows an example of the raw waveform of the acceleration signal which concerns on the same embodiment. The figure which shows an example of the waveform of the acceleration signal after the filtering process which concerns on the same embodiment. The figure which shows another example of the waveform of the acceleration signal after the filtering process which concerns on the same embodiment. The figure which shows another example of the waveform of the acceleration signal after the filtering process which concerns on the same embodiment. The figure for demonstrating the amplitude value in the same embodiment. The figure which shows an example of the screen which displayed the analysis result information obtained by running the respiratory analysis application program which concerns on the same embodiment. The flowchart which shows an example of a series of processing procedures implement|achieved by running the respiratory analysis application program which concerns on the same embodiment. The block diagram which shows an example of a function structure of the respiratory analysis application program which concerns on 2nd Embodiment. The flowchart which shows an example of the procedure of the process performed by the sleeping posture determination part which concerns on the same embodiment. The block diagram which shows an example of a function structure of the respiratory analysis application program which concerns on 3rd Embodiment. The flowchart which shows an example of the procedure of the process performed by the sleep time calculation part and the severity calculation part which concern on the same embodiment. The figure which shows another schematic structure of the respiratory analysis system which concerns commonly on each embodiment.

Hereinafter, embodiments will be described with reference to the drawings.
<First Embodiment>
FIG. 1 shows a schematic configuration example of a respiratory analysis system according to each embodiment. As shown in FIG. 1, the respiratory analysis system 1 includes a biometric sensor device 10 and an electronic device 11, and the biometric sensor device 10 and the electronic device 11 are communicably connected to each other. Note that some of the various functions of the biological sensor device 10 may be realized by the electronic device 11. The biometric sensor device 10 is removably attached to (on the sternum of) a human body by, for example, an adhesive tape (adhesive member) or the like in order to constantly measure biometric information. The method of attachment to the human body may be attachment by a band, attachment by a belt, embedding by clothes, etc., as well as attachment by attachment. The biological sensor device 10 simultaneously measures a plurality of biological information such as a pulse wave, an electrocardiogram, temperature, and acceleration, analyzes the measured biological information, and wirelessly analyzes the analysis result information obtained as a result of the analysis. It also has the function of sending to. Furthermore, the biological sensor device 10 also has a function of wirelessly receiving a control signal or the like from the electronic device 11.

The electronic device 11 includes a communication unit 12, an output unit 13, and the like, is a module capable of receiving biometric information and analysis result information transmitted from the biometric sensor device 10, and displaying and outputting the biometric information on a screen, such as a smartphone or a tablet computer. Wearable terminals such as watches and earphones fall under this category. An application for displaying biometric information and analysis result information sent from the biometric sensor device 10 is installed in advance in the electronic device 11.

It should be noted that the biometric sensor device 10 and the electronic device 11 may be connected so as to be able to communicate by wire instead of wirelessly.

Here, the hardware configuration of the biosensor device 10 will be described with reference to FIG. 2.
Although the biometric sensor device 10 has a plurality of sensors, the analog front end of each sensor is required to have both flexibility and high performance because the specifications are different for each sensor, which may result in an increase in size. However, here, a sensor module of several millimeters square is realized by integrating a plurality of analog front ends, a CPU and the like on a single chip by using the pseudo SoC technology. Pseudo SoC technology is a technology that achieves both downsizing equivalent to SoC and design flexibility equivalent to SiP by integrating components on a wafer. By connecting a small peripheral component such as an antenna and a battery to this module, a small-sized, lightweight (about several tens of grams) and thin (about several mm) biosensor device 10 is realized. It should be noted that it is also possible to realize miniaturization by a component built-in technology or a configuration using a dedicated LSI.

The biosensor device 10 has, for example, an elliptical shape having a major axis of about several cm, and as shown in FIG. 2, the electrocardiogram electrode (R) 20a, the electrocardiogram electrode (L) 20b, and the photoelectric unit are provided on the surface to be adhered to the human body. 22, a temperature sensor 24, and a charging terminal 26 are arranged. Since the electrocardiographic electrodes 20a, 20b need to be located on the left and right of the heart, they are spaced along the long axis. The photoelectric unit 22 optically detects a pulse wave, and a window portion made of a transparent material that transmits light is provided on the front surface thereof.

FIG. 3 is a block diagram showing a circuit configuration of the biosensor device 10. The biological sensor device 10 includes an electrocardiograph 30, an acceleration sensor 32, a pulse wave meter 34, and Bluetooth (registered trademark) in addition to the electrocardiographic electrodes 20a and 20b, the photoelectric unit 22, the temperature sensor 24, and the charging terminal 26 described above. A module 36, a system controller 38, an embedded controller (EC) 40, a lithium secondary battery 42, a CPU 44, a main memory 46, a BIOS-ROM 48, a flash memory 50 and the like are included.

The electrocardiogram electrode (R) 20a and the electrocardiogram electrode (L) 20b are connected to an electrocardiograph 30, which is an analog front end for electrocardiogram. The electrocardiograph 30 obtains an electrocardiogram by analyzing a time-series signal obtained by sampling the potential difference between the electrocardiogram electrode (R) 20a and the electrocardiogram electrode (L) 20b.

The photoelectric unit 22 is for detecting a volume pulse wave, and includes a light emitting element (for example, a blue LED) 22a that is a light source and a photodiode (PD) 22b that is a light receiving unit. A transparent window is provided on the front surface of the photoelectric unit 22, light from the blue LED 22a is applied to the skin surface through the window, and reflected light is incident on the PD 22b through the window. The blue LED 22a and the PD 22b are connected to a pulse wave meter 34 which is an analog front end for pulse waves. The sphygmograph 34 detects the fluctuation of the reflected light that changes due to the change of blood flow in the capillaries, and obtains the pulse wave by analyzing the detection signal.

The electrocardiograph 30, the acceleration sensor 32, the pulse wave meter 34, and the temperature sensor 24 are connected to the system controller 38. The temperature sensor 24 measures the temperature of the human body surface, and the acceleration sensor 32 measures the movement of the human body. The acceleration sensor 32 is assumed to be a triaxial acceleration sensor with a sampling frequency of 1 KHz.

The CPU 44 is a processor that controls the operation of each module and each component of the biological sensor device 10. As described above, the biological sensor device 10 analyzes various biological information (eg, body temperature, skin temperature, pulse rate, heart rate, autonomic nerve activity) by analyzing the output of each sensor or the combination of the outputs of a plurality of sensors. Index, blood pressure, sleep time, etc.) can be continuously measured.

The system controller 38 is a bridge device that connects the CPU 44 to each module and each component. The system controller 38 is also connected to the Bluetooth module 36, embedded controller (EC) 40, CPU 44, main memory 46, BIOS-ROM 48, and flash memory 50.

The embedded controller 40 is a power management controller for performing power management of the biometric sensor device 10, and controls charging of a built-in secondary battery, for example, a lithium secondary battery 42. When the biosensor device 10 is attached to the charger 52, the charging terminal 26 comes into contact with the terminal of the charger 52, and the charging current from the charger 52 is supplied to the biosensor device 10 via the charging terminal 26. The secondary battery 42 is charged. The embedded controller 40 supplies operating power to each module and each component based on the electric power from the lithium secondary battery 42.

The CPU 44 is a processor that controls the operation of each module and each component in the biological sensor device 10. The CPU 44 executes various software loaded from the flash memory 50 into the main memory 46. The software includes an operating system (OS) and various application programs. The application program includes a respiratory analysis application program (hereinafter, simply referred to as a respiratory analysis application) 100.

In the present embodiment, the hardware configuration of the biometric sensor device 10 has been described with reference to FIGS. 2 and 3, but the hardware configuration of the biometric sensor device 10 is not limited to this. Various functions realized by causing the CPU 44 to execute the respiratory analysis application 100 described below can operate at least when the biosensor device 10 has the acceleration sensor 32 built therein. For example, from the biosensor device 10. The temperature sensor 24, the electrocardiograph 30, the pulse wave meter 34, etc. may be omitted. The biosensor device 10 may be configured to further include a display capable of displaying analysis result information.

Generally, the waveform of the acceleration signal is various, such as when resuming breathing from apnea or hypopnea symptoms, when snoring occurs, when rolling over, during minute body movements (limb movements and upper body movements) that occur during sleep, etc. It varies in various cases. As a general respiratory analysis method, there is a method of estimating whether or not an apnea symptom occurs by detecting the peak value related to respiratory fluctuation in the waveform of the acceleration signal and capturing the peak intensity difference as the respiratory intensity. Proposed. However, with this respiratory analysis method, only the occurrence of apnea symptoms can be estimated, and there is the inconvenience that it is not possible to estimate the occurrence of hypopnea symptoms, the occurrence of snoring, the rolling motion and the occurrence of minute body movements ( That is, there was a disadvantage that it was not possible to distinguish between the occurrence of apnea symptoms and the occurrence of other events). The respiratory analysis application 100 according to the present embodiment detects the movement of the wearer based on the acceleration signal acquired by the acceleration sensor 32, and determines (identifies) what caused the movement of the wearer. ) ・Has a function to decide. Hereinafter, the breath analysis application 100 will be described in detail.

FIG. 4 is a block diagram showing an example of the functional configuration of the respiratory analysis application 100. As shown in FIG. 4, the respiratory analysis application 100 includes an acceleration event detection unit 101, an event type determination unit 102, and the like. Further, the event type determination unit 102 includes a body movement respiratory event determination unit 103, a body movement event attribute determination unit 104, a respiratory event attribute determination unit 105, and the like. In the present embodiment, the respiratory analysis application 100 is stored in the main memory 46 of the biological sensor device 10 and executed by the CPU 44 of the biological sensor device 10. However, for example, the respiratory analysis application 100 is It may be stored in a memory (not shown) of the electronic device 11 and executed by a CPU (not shown) in the electronic device 11. In this case, the acceleration signal measured by the acceleration sensor 32 is output to the electronic device 11. Further, some of the functional units included in the respiratory analysis application 100 may be mounted on the electronic device 11.

The details of the functions of the respective units 101 to 105 will be described below.
Upon receiving the input of the acceleration signal output from the acceleration sensor 32 (when acquiring the acceleration signal from the acceleration sensor 32), the acceleration event detection unit 101 detects the acceleration event from the acceleration signal. Specifically, the acceleration event detection unit 101 performs a filtering process on a temporal change waveform of the acceleration indicated by the input acceleration signal, and the frequency is preset from the waveform after the filtering process. An outlier time T1 that is less than the first threshold value is detected as a time when an acceleration event occurs. The acceleration event indicates a non-resting state, that is, the wearer of the biometric sensor device 10 has moved. The outlier time T1 indicates the time when the acceleration event occurs, that is, the time when the wearer of the biometric sensor device 10 has moved. Further, the filtering process here is a process of mounting a high-pass filter at the first cutoff frequency to remove noise. The first cutoff frequency is preferably 2 [Hz] or higher, which is higher than the frequency of respiratory motion. In the filtering process here, the bandpass filter may be mounted at the second cutoff frequency higher than the first cutoff frequency instead of mounting the highpass filter. According to this, it is possible to remove the noise that is constantly superimposed regardless of the occurrence of the acceleration event. In this case, the second cutoff frequency is preferably 100 [Hz] to 150 [Hz], which is close to a respiratory-related fundamental frequency such as snoring.

Here, the process of detecting an acceleration event (the process of detecting an outlier T1) will be described in more detail with reference to FIGS. 5 and 6.
FIG. 5 shows an example of the waveform (raw waveform) of the acceleration signal output from the acceleration sensor 32. The waveform of the acceleration signal shown in FIG. 5 shows a temporal change in the acceleration of the axis perpendicular to the chest surface of the wearer of the biometric sensor device 10. FIG. 5A shows the waveform of the acceleration signal when the wearer's leg myoelectricity is active during sleep and the wearer is performing limb movements. FIG. 5B shows the waveform of the acceleration signal when the wearer awakens during sleep and moves the upper body. FIG. 5C shows the waveform of the acceleration signal when the wearer is rolling over. FIG. 5D shows the waveform of the acceleration signal when the wearer is snoring. FIG. 5E shows the waveform of the acceleration signal when the wearer is in an apnea state. FIG. 5F shows the waveform of the acceleration signal when the wearer is in a low breathing state.

In general, a high frequency vibration is superimposed on the acceleration signal as shown in FIGS. 5(a) to 5(c) in association with minute body movements during sleep (limb movements or upper body movements) and turning over movements. Also, when snoring occurs and when resuming breathing from an apnea or hypopnea state, the vocal tract and the vicinity of the oral cavity vibrate, producing a large airway sound with a frequency higher than the respiratory frequency, which is transmitted to the chest surface. Therefore, the waveform of the acceleration signal as shown in FIGS. 5D to 5F is obtained.

FIG. 6 shows an example of the waveform of the acceleration signal after performing the filtering process on the various waveforms shown in FIG. FIGS. 6A to 6F show the waveforms of the acceleration signals shown in FIGS. 5A to 5F after the high-pass filter is mounted at the first cutoff frequency and the filtering process is executed. The respective waveforms of the acceleration signal (of the first frequency band) are shown. The acceleration event detection unit 101 sets the time when the frequency becomes less than the preset first threshold value as the outlier time T1, that is, the acceleration event, as shown by the crosses in FIGS. 6A to 6F. Is detected as the time when the occurrence occurs. In the waveform of the high frequency band extracted by the above-described filtering process, all the times when the frequency becomes less than the first threshold value may be detected as the outlier time T1 or may have a predetermined continuous length. Even if the time when the moving standard deviation value σ1 and the moving average value M1 in the section are calculated and a value statistically outside the range of M1±α1*σ1 (where α1 is a constant) is detected as the outlier time T1 Good. FIG. 6 shows a case where the outlier time T1 is detected by using the latter method. Specifically, the latter method detects three outlier time points T1 in FIG. 6A, four outlier time points T1 in FIG. 6B, and five outlier time points T1 in FIG. 6C. In the case where the outlier time T1 is detected, 24 outlier times T1 are regularly detected in FIG. 6D, and two outlier time T1 are both detected in FIGS. 6E and 6F. Shows.

The information indicating the detected outlier time T1 is sent to the event type determination unit 102.

Again, returning to the description of FIG. The body movement/breathing event determination unit 103 in the event type determination unit 102 determines that the acceleration event detected by the acceleration event detection unit 101 (in other words, the movement of the wearer of the biosensor device 10 that occurred at the outlier time T1) is Processing is performed to determine whether the acceleration event (movement) is caused by body movement or the acceleration event (movement) is caused by respiration.

Generally, when an acceleration event such as a small body movement or a rolling motion occurs due to a body movement, the posture of the biosensor device 10 changes significantly, that is, the posture change amount of the biosensor device 10 increases. Tend. On the other hand, when an acceleration event due to respiration occurs, the posture of the biometric sensor device 10 does not change much (because the movement associated with the acceleration event is a vertical movement of the chest, etc.), that is, the biometric sensor. The posture change amount of the device 10 tends to be small. Using this tendency, the body movement/breathing event determination unit 103 calculates the above-described posture change amount as the first feature amount, and the calculated first feature amount is equal to or greater than the preset second threshold value. By executing a process of determining whether the acceleration event detected by the acceleration event detection unit 101 is an acceleration event caused by body movement or an acceleration event caused by respiration. To do.

Here, the process of calculating the posture change amount (first feature amount) of the biological sensor device 10 will be described in more detail with reference to FIG. 7.
FIG. 7 shows an example of the waveform of the acceleration signal after performing the filtering process on the various waveforms shown in FIG. 7(a) to 7(f) show the waveforms of the acceleration signals shown in FIGS. 5(a) to 5(f) after the low pass filter is mounted at the third cutoff frequency to remove noise. The waveforms of the acceleration signal (that is, after performing the filtering process) (in the second frequency band) are respectively shown. The third cutoff frequency is a value lower than the frequency of respiratory movement, and is preferably 0.02 [Hz] or more and less than 0.1 [Hz].

The body movement/breathing event determination unit 103 calculates the posture change amount in the predetermined section T2 including the outlier time T1 as the first feature amount in the various waveforms y(t) shown in FIGS. 7A to 7F. .. Specifically, the body motion/breathing event determination unit 103 determines the median value in the range of y{T1-(T2/2)} to y(T1) and y(T1) to y{T1+(T2/2)}. The difference from the median in the range is obtained as the first feature amount. The predetermined section T2 is, for example, a section provided at a predetermined interval before and after the outlier time T1 as a center, and in the present embodiment, 120 seconds (that is, a section 60 seconds before and after the outlier time T1) is predetermined. It is assumed to be the section T2. The first feature amount is not limited to the above-described method, and for example, an average value in the range of y{T1-(T2/2)} to y(T1) and y(T1) to y{T1+(T2 /2)} may be obtained as a difference with the average value in the range, or between the maximum value and the minimum value in the range of y{T1-(T2/2)} to y{T1+(T2/2)}. It may be calculated as a difference, or may be calculated as a variance value or a standard deviation value in the range of y{T1-(T2/2)} to y{T1+(T2/2)}.

Again, returning to the description of FIG. If the acceleration event detected by the acceleration event detection unit 101 is determined by the body movement respiratory event determination unit 103 to be an acceleration event caused by the body movement, the body movement event attribute determination unit 104 determines that the body movement has occurred. A process of determining whether the resulting acceleration event is an acceleration event due to a small body movement or an acceleration event due to a rolling motion is executed.

In general, the feature amount (posture change amount) when an acceleration event occurs due to a small body movement (limb movement or upper body movement) and the feature amount (posture change when an acceleration event occurs due to rolling over movement) Amount), there is a tendency that the feature amount when the acceleration event caused by the rolling over motion occurs is larger than the feature amount when the acceleration event caused by the small body motion occurs. Using this tendency, the body movement event attribute determination unit 104 determines whether the first feature amount (posture change amount) calculated by the body movement event determination unit 103 is equal to or greater than a preset third threshold value. By determining whether the acceleration event determined by the body movement respiratory event determination unit 103 to be an acceleration event caused by body movement is an acceleration event caused by minute body movement or caused by a rolling over movement. A process of determining whether the event is an acceleration event is executed.

Note that here, the body movement event attribute determination unit 104 executes the determination process described above using the first feature amount (posture change amount) calculated by the body movement respiratory event determination unit 103. For example, instead of using the first feature amount calculated by the body movement/breathing event determination unit 103, the feature amount may be calculated by itself, and the determination process described above may be executed. In this case, the cutoff frequency when the low-pass filter is mounted on the waveform of the acceleration signal in calculating the feature amount is not limited to the above-mentioned third cutoff frequency, and may be equal to or lower than the frequency of body movement.

When the acceleration event detected by the acceleration event detection unit 101 is determined by the body motion respiratory event determination unit 103 to be an acceleration event caused by respiration, the respiratory event attribute determination unit 105 determines the acceleration caused by the respiration. A process of determining whether the event is an acceleration event caused by a snoring, an acceleration event caused by an apnea condition, or an acceleration event caused by a hypopnea condition is executed.

Here, the difference between the acceleration event caused by the snoring, the acceleration event caused by the apnea, and the acceleration event caused by the hypopnea will be described with reference to FIG. FIG. 8 shows an example of the waveform of the acceleration signal after performing the filtering process on the various waveforms shown in FIG. 8(a) to 8(f) show the waveforms of the acceleration signals shown in FIGS. 5(a) to 5(f) after the noise is removed by mounting a low-pass filter at the fourth cutoff frequency. The waveforms of the acceleration signal (that is, after performing the filtering process) (in the third frequency band) are shown. The fourth cutoff frequency is preferably in the frequency band of respiratory movement, and is preferably 0.1 [Hz] or more and less than 1.0 [Hz], for example.

In general, changes in the waveform of the acceleration signal when a snoring-induced acceleration event occurs, and acceleration due to respiratory return from an apnea condition (apnea condition) or a hypopnea condition (hypopnea condition) The following tendency can be seen when comparing with the change observed in the waveform of the acceleration signal when the event occurs. In other words, when an acceleration event due to snoring occurs, as shown in FIG. 8(d), there is no significant change in the waveform characteristics, but in the respiratory recovery from the apnea or hypopnea state. When the resulting acceleration event occurs, as shown in FIGS. 8E and 8F, there is a tendency that a large change occurs in the waveform characteristics before and after the outlier time T1. This is because in the case of an acceleration event due to snoring, since the respiratory movement itself is stationary, there is no significant change in the waveform characteristics, but the acceleration event due to respiratory return from an apnea or hypopnea state In this case, since the tissue around the vocal tract and the oral cavity vibrates at the time of resuming breathing, and chest motion is performed before and after that, there is a tendency that a large change occurs in the waveform characteristics. The respiratory event attribute determination unit 105 uses this tendency to calculate the second characteristic amount in the predetermined section T3 including the outlier time T1 in the waveform y2(t) of the acceleration signal shown in FIG. By determining whether or not the generated second feature amount is equal to or more than a preset fourth threshold value, the acceleration event determined to be the acceleration event caused by respiration by the body movement respiratory event determination unit 103 , A process of determining whether the acceleration event is caused by a snoring or the acceleration event is caused by an apnea state or a hypopnea state.

Further, as shown in FIGS. 8(e) and (f), a case where an acceleration event occurs due to a return of breath from an apnea state and an acceleration event occurs due to a return of breath from a low breathing state. In the case and, when the acceleration event caused by the respiratory return from the apnea state occurs, the waveform characteristics are significantly changed compared with the case where the acceleration event caused by the respiratory return from the hypopnea state occurs. You can see it. Therefore, the respiratory event attribute determination unit 105 presets the second feature amount calculated as described above with respect to the acceleration event determined to be the acceleration event caused by the apnea state or the low respiratory state. Whether the acceleration event determined to be the acceleration event due to the apnea state or the hypopnea state is the acceleration event due to the apnea state by determining whether it is equal to or more than the fifth threshold value, Alternatively, a process for determining whether the acceleration event is caused by the low respiratory state is executed.

The second feature amount (here, the variation ratio of the chest motion before and after the acceleration event) in the predetermined section T3 including the outlier time T1 is calculated based on the following equation (1). At this time, the predetermined section T3 may be a section set so as to include the outlier time T1. For example, in the present embodiment, the predetermined section T3 is 20 seconds (that is, a section 10 seconds before and after the outlier time T1). Is set to. In addition, in the following formula (1), the case where the standard deviation value is used is illustrated, but the variance value may be used instead of the standard deviation value.

In addition, in the present embodiment, the formula (1) is used to calculate the variation ratio (second feature amount) of the chest motion before and after the acceleration event, but the second feature amount is the acceleration event. It may be calculated from the average value of the amplitude values of the waveform y2(t) in the respiratory frequency band before and after. Specifically, it may be calculated by using the following equation (2).

The amplitude value here is, as shown in FIG. 9, the maximum value (circle in FIG. 9) of the waveform y2(t) of the acceleration signal and the minimum value immediately before (indicated by Δ in FIG. 9). And the difference.

The analysis result information obtained as a result of various processes in the respective units 101 to 105 described above is stored in the flash memory 50 and is output to the electronic device 11 via the Bluetooth module 36. At least the analysis result information is associated with the outlier time T1 detected by the acceleration event detection unit 101 and the information indicating what caused the motion of the wearer at the outlier time T1. Information. When receiving the input of the analysis result information output from the biometric sensor device 10, the electronic device 11 can display a screen as shown in FIG. 10, for example. According to this, the wearer (user) can understand when and what caused the wearer to move.

Next, an example of a series of processing procedures executed by the respiratory analysis application 100 will be described with reference to the flowchart in FIG.
First, when the acceleration event detection unit 101 receives an input of an acceleration signal measured by the acceleration sensor 32, the acceleration event detection unit 101 receives a first waveform with respect to a waveform showing a temporal change in acceleration indicated by the acceleration signal that receives the input. A high-pass filter is mounted at the cut-off frequency to perform a filtering process for removing noise, and the waveform of the acceleration signal in the respiratory frequency band or higher is extracted (step S1). That is, the acceleration event detection unit 101 executes the filtering process on the raw waveform of the acceleration signal shown in FIG. 5 to extract the waveform of the acceleration signal in the respiratory frequency band shown in FIG.

Subsequently, the acceleration event detection unit 101 detects an outlier time T1 at which the frequency becomes less than a preset first threshold value from the waveform of the acceleration signal in the respiratory frequency band or higher extracted by the process of step S1 ( Step S2). That is, the acceleration event detection unit 101 detects, from the waveform of the acceleration signal shown in FIG. 6, the time when the frequency becomes less than the first threshold value as the outlier time T1 as indicated by the cross mark in FIG.

Next, when the body motion/breathing event determination unit 103 acquires the acceleration signal input to the acceleration event detection unit 101 from the acceleration sensor 32, the body motion/breathing event determination unit 103 creates a waveform indicating a temporal change in the acceleration indicated by the acquired acceleration signal. On the other hand, a low-pass filter is mounted at the third cutoff frequency to perform a filtering process to remove noise, and the waveform y(t) of the acceleration signal below the respiratory frequency band is extracted (step S3). That is, the body motion respiratory event determination unit 103 executes the filtering process on the raw waveform of the acceleration signal shown in FIG. 5 to extract the waveform y(t) of the acceleration signal below the respiratory frequency band shown in FIG.

Subsequently, the body movement respiratory event determination unit 103 includes a predetermined value including the outlier time T1 detected by the process of step S2 in the waveform y(t) of the acceleration signal below the respiratory frequency band extracted by the process of step S3. The first feature amount in the section T2 is calculated. Specifically, the body-movement-breathing event determination unit 103 sets the median value in the range of y{T1-(T2/2)} to y(T1) in the waveform y(t) of the acceleration signal below the respiratory frequency band. , Y(T1) to y{T1+(T2/2)} is calculated as the first feature amount of the predetermined section T2 including the outlier time T1 (step S4).

Next, the body movement/breathing event determination unit 103 determines whether or not the first feature amount calculated by the process of step S4 is equal to or greater than a preset second threshold value (step S5).

In the process of step S5, when it is determined that the first feature amount is equal to or more than the second threshold value (YES in step S5), the body movement respiratory event determination unit 103 determines that the acceleration event that occurred at the outlier time T1 is It is determined that the acceleration event is caused by body movement (step S6).

Then, when the body motion/breathing event determination unit 103 determines that the acceleration event that occurred at the outlier T1 is an acceleration event caused by body movement, the body movement event attribute determination unit 104 determines in step S4. It is determined whether or not the first characteristic amount calculated by the process is equal to or larger than a preset third threshold value (step S7).

In the process of step S7, when it is determined that the first feature amount is equal to or larger than the third threshold value (YES in step S7), the body movement event attribute determination unit 104 determines that the acceleration event that occurred at the outlier time T1 is It is determined that the acceleration event is caused by the rolling motion, the analysis result information indicating the result of the series of processes is output to the electronic device 11 (step S8), and the process here is ended.

On the other hand, in the process of step S7, when it is determined that the first feature amount is less than the third threshold value (NO in step S7), the body movement event attribute determination unit 104 determines that the acceleration occurred at the outlier time T1. It is determined that the event is an acceleration event caused by a small body motion, and analysis result information indicating the result of the series of processes is output to the electronic device 11 (step S9), and the process here is ended.

In the process of step S5 described above, when it is determined that the first feature amount is less than the second threshold value (NO in step S5), the body motion respiratory event determination unit 103 determines that the acceleration that occurred at the outlier T1. It is determined that the event is an acceleration event caused by respiration (step S10).

Next, when the respiratory event attribute determination unit 105 acquires the acceleration signal input to the acceleration event detection unit 101 from the acceleration sensor 32, with respect to the temporal change waveform of the acceleration indicated by the acquired acceleration signal, A low-pass filter is mounted at the fourth cutoff frequency to perform filtering processing to remove noise, and the waveform y2(t) of the acceleration signal in the respiratory frequency band is extracted (step S11). That is, the respiratory event attribute determination unit 105 executes the filtering process on the raw waveform of the acceleration signal shown in FIG. 5, and extracts the waveform y2(t) of the acceleration signal in the respiratory frequency band shown in FIG.

Subsequently, the respiratory event attribute determination unit 105, in the waveform y2(t) of the acceleration signal in the respiratory frequency band extracted by the process of step S11, the predetermined section T3 including the outlier time T1 detected by the process of step S2. The second feature amount in is calculated. Specifically, the respiratory event attribute determination unit 105 sets the standard deviation value in the range y2(T1) to y2{T1+(T3/2)} to y2 in the waveform y2(t) of the acceleration signal in the respiratory frequency band. The quotient obtained by dividing by the standard deviation value in the range of {T1-(T3/2)} to y2(T1) is calculated as the second feature amount of the predetermined section T3 including the outlier time T1 (step. S12).

Next, the respiratory event attribute determination unit 105 determines whether or not the second characteristic amount calculated by the process of step S12 is equal to or greater than a preset fourth threshold value (step S13). In the process of step S13, when it is determined that the second feature amount is less than the fourth threshold value (NO in step S13), the respiratory event attribute determination unit 105 determines that the acceleration event occurred at the outlier time T1. Determines that the acceleration event is caused by snoring, outputs analysis result information indicating the result of the series of processes to the electronic device 11 (step S14), and ends the process here.

On the other hand, in the process of step S13, when it is determined that the second feature amount is equal to or larger than the fourth threshold value (YES in step S13), the respiratory event attribute determination unit 105 sets the second feature amount in advance. It is determined whether or not it is equal to or more than the fifth threshold value (step S15).

When it is determined in the process of step S15 that the second feature amount is equal to or larger than the fifth threshold value (YES in step S15), the respiratory event attribute determination unit 105 determines that there is no acceleration event that occurred at the outlier time T1. It is determined that the acceleration event is caused by the breathing state, the analysis result information indicating the result of the series of processes is output to the electronic device 11 (step S16), and the process here is ended.

On the other hand, in the process of step S15, when it is determined that the second feature amount is less than the fifth threshold value (NO in step S15), the respiratory event attribute determination unit 105 determines that the acceleration event that occurred at the outlier time T1. Is determined to be an acceleration event caused by the low breathing state, and analysis result information indicating the result of the series of processes is output to the electronic device 11 (step S17), and the process here is ended.

According to the first embodiment described above, the biosensor device 10 detects the movement of the wearer based on the acceleration measured by the acceleration sensor 32, and the movement of the wearer is a movement caused by what. Since the respiratory analysis application 100 capable of discriminating such details is provided, the wearer (user) can grasp the movement of the wearer during sleep. That is, the user can easily understand whether he/she may develop sleep apnea syndrome.

<Second Embodiment>
Next, a second embodiment will be described. In the present embodiment, a case will be described in which the respiratory analysis application 100 further includes a sleeping posture determination unit (sleeping posture estimation unit) 106 as illustrated in FIG. 12 in addition to the above-described units 101 to 105. In addition, in this embodiment, the same reference numerals are given to the respective units having the same functions as those in the above-described first embodiment, and the detailed description thereof will be omitted.

The sleeping posture determination unit 106 determines when the acceleration event detected by the acceleration event detection unit 101 is determined by the body motion event attribute determination unit 104 in the event type determination unit 102 to be an acceleration event caused by a rolling over movement. A process of estimating (determining) the sleeping posture of the wearer of the biometric sensor device 10 is executed. The information indicating the estimated sleeping posture is information indicating that the outlier value time T1 detected by the acceleration event detection unit 101 and the motion of the wearer at the outlier value time T1 are motions caused by the rolling over motion. And is stored in the flash memory 50 and output to the electronic device 11 via the Bluetooth module 36.

Here, an example of the processing procedure executed by the sleeping posture determination unit 106 will be described with reference to the flowchart in FIG. 13. Note that, here, it is assumed that the body motion event attribute determination unit 104 has determined that the acceleration event generated at the outlier time T1 is the acceleration event caused by the rolling motion, that is, steps S1 to S8 shown in FIG. 11. The processing will be described as if it has already been executed.

First, the sleeping posture determination unit 106 acquires, from the waveform y(t) of the acceleration signal extracted by the body motion event attribute determination unit 104, section data from the outlier time T1 until a predetermined time T4 elapses ( Step S21). In addition, although the case where the predetermined time T4 is 60 seconds will be described here, the predetermined time T4 is not limited to this, and the predetermined time T4 may be an arbitrary time.

Then, the sleeping posture determination unit 106 is an acceleration signal measured during a lapse of a predetermined time T4 from the outlier T1, and is an axis that is horizontal to the chest surface and perpendicular to the body axis. The acceleration signal of the component is acquired from the acceleration sensor 32 (step S22).

Next, the sleeping posture determination unit 106 sets the waveform indicating the temporal change of the acceleration signal acquired by the process of step S22 to the third cutoff frequency in the same manner as the waveform y(t) described above. Then, a low-pass filter is installed to perform filtering processing for removing noise, and the waveform y3(t) of the acceleration signal in the low frequency band is extracted (step S23).

Then, the sleeping posture determination unit 106 obtains section data from the outlier time T1 until a predetermined time T4 elapses from the waveform y3(t) of the acceleration signal in the low frequency band extracted by the process of step S23. Yes (step S24).

After that, the sleeping posture determination unit 106 respectively calculates a third characteristic amount based on the section data acquired by the process of step S21 and a fourth characteristic amount based on the section data acquired by the process of step S24. Specifically, the sleeping posture determination unit 106 calculates the median value or the average value from y(T1) to y(T1+T4) as the third feature amount, and calculates the median from y3(T1) to y3(T1+T4). A value or an average value is calculated as the fourth characteristic amount (step S25).

Then, the sleeping posture determination unit 106 determines which of the supine position, the left lateral position, the right lateral position and the prone position, the sleeping position corresponds to, based on the calculated third and fourth characteristic amounts. Is executed, the sleeping posture is estimated, the analysis result information indicating the result of the series of processes is output to the electronic device 11 (step S26), and the process here is ended.

Note that when determining whether the sleeping posture corresponds to the supine position, the left lateral position, the right lateral position, and the prone position based on the third and fourth characteristic amounts, for example, one of neural networks is used. The error back propagation method or the like may be used. It is assumed that the weight used when using this back propagation method is previously learned and recorded in the flash memory 50 or the like in advance. Further, when determining which of the above-mentioned body postures the sleeping posture corresponds to, for example, another neural network such as a support vector machine may be used. Furthermore, the average value of the two feature amounts for each body position is obtained in advance, and the sleeping position clustering is performed by searching for which body position the calculated third and fourth feature amounts are closest to. May be performed. Further, it may be possible to determine which of the above-mentioned body positions the sleeping posture corresponds to, based on the threshold values respectively corresponding to the third and fourth characteristic amounts.

According to the second embodiment described above, the biological sensor device 10 further includes a function of estimating the sleeping posture when it is determined that the motion of the wearer is the motion caused by the rolling motion. The wearer (user) can use the estimation result for formulating a sleep environment improvement policy.

<Third Embodiment>
Next, a third embodiment will be described. In the present embodiment, a case will be described in which the respiratory analysis application 100 further includes a sleep time calculating unit 107 and a severity calculating unit 108, as shown in FIG. 14, in addition to the above-described units 101 to 105. In addition, in the present embodiment, the same reference numerals are given to the respective units having the same functions as those in the above-described first and second embodiments, and the detailed description thereof will be omitted. 14, the sleeping posture determination unit 106 described in the second embodiment is omitted for convenience of description, but the respiratory analysis application 100 has a sleeping posture determination unit in addition to the configuration shown in FIG. 106 may be further provided.

The sleep time calculation unit 107 calculates the sleep time based on the bedtime and the wakeup time of the wearer of the biometric sensor device 10. The bedtime and the wake-up time may be input by the wearer himself/herself via an input unit (not shown) provided in the biological sensor device 10, or may be acquired from an external device (for example, the electronic device 11 or the like). However, it may be estimated from the operation history of the external device (specifically, the sleep time or the wake-up time is acquired from the alarm function provided in the external device) (specifically, the operation of the external device is interrupted). Time is estimated as the bedtime, and the time when the operation of the external device is restarted is estimated as the wakeup time).

Further, at bedtime and wake-up time, a function capable of estimating (determining) the posture in which the upper body is tilted is added to the sleeping posture determination unit 106 in addition to the supine position, the left lateral position, the right lateral position and the prone position. However, it may be estimated by using the function. Specifically, the time at which the upper body shifts from a tilted posture to another sleeping posture is defined as the bedtime, and the upper body tilts from any of the supine position, the left lateral position, the right lateral position and the prone position. The time at which the posture is changed may be the wake-up time.

The severity calculating unit 108 acquires information indicating the bedtime, the wakeup time, and the sleeping time from the sleeping time calculating unit 107. Further, when the respiratory event attribute determination unit 105 determines that the predetermined acceleration event is an acceleration event caused by an apnea state or a hypopnea state, the severity calculation unit 108 causes the acceleration event detection unit 101 to determine the predetermined acceleration event. The information indicating the outlier time T1 at which the acceleration event occurs is acquired. Further, when the acquired outlier value time T1 is a time between the bedtime and the wake-up time, the severity calculating unit 108 adds 1 to the apnea hypopnea count, and the absent count that has been counted by the wake-up time. An apnea hypopnea index (AHI) is calculated based on the respiratory hypopnea frequency and the sleep time. As shown in the following formula (3), AHI indicates the number of apnea hypopnea per hour during sleep, and is generally used as an index of the severity of sleep apnea syndrome for screening or It is used to formulate treatment policies.

AHI=sum of apnea and low breathing rate/sleep time (3)
In the above description, if the acceleration event occurring at the outlier time T1 is either the acceleration event caused by the apnea state or the acceleration event caused by the hypopnea state, it is added as the number of times of apnea hypopnea. However, for example, the number of times of apnea and the number of times of low breathing may be counted separately. That is, when the acceleration event that occurred at the outlier time T1 is an acceleration event caused by an apnea and the outlier time T1 is included between the bedtime and the wakeup time, the severity calculation unit 108 You may add 1 to the number of apnea. Similarly, when the acceleration event occurring at the outlier time T1 is an acceleration event caused by the low breathing state and the outlier time T1 is included between the bedtime and the wakeup time, the severity calculating unit 108 , 1 may be added to the low respiratory rate.

In this case, the index of the severity of sleep apnea syndrome is, as shown in the following equations (4) and (5), an index of the severity of apnea and an index of the severity of hypopnea. Alternatively, it may be calculated separately.

Severity index for apnea frequency = total apnea frequency / sleep time (4)
Severity index for low respiratory rate = total low respiratory rate / sleep time (5)
The information indicating the calculated AHI is stored in the flash memory 50 and is also output to the electronic device 11 via the Bluetooth module 36.

Here, an example of a processing procedure executed by the sleep time calculating unit 107 and the severity calculating unit 108 will be described with reference to the flowchart shown in FIG. 15. Here, it is assumed that the bedtime and the wakeup time are input by the wearer of the biometric sensor device 10 immediately before going to bed and immediately after getting up.

First, the sleep time calculation unit 107 receives an input of the bedtime of the wearer according to the operation of the wearer (step S31).

After that, the acceleration event detection unit 101, the body motion respiratory event determination unit 103, the body motion event attribute determination unit 104, and the respiratory event attribute determination unit 105 repeatedly execute the above-described processing of steps S1 to S17, and a predetermined acceleration event is generated. If it is determined that the acceleration event is caused by the apnea or hypopnea state, the severity calculating unit 108 adds 1 to the apnea hypopnea count (step S32).

Then, the sleep time calculation unit 107 determines whether or not the input of the wake-up time of the wearer has been received according to the operation of the wearer (step S33). In addition, in the process of step S33, when it is determined that the input of the wake-up time has not been accepted (NO in step S33), the process returns to the process of step S32 described above.

On the other hand, in the process of step S33, when it is determined that the input of the wake-up time has been received (YES in step S33), the sleep time calculation unit 107 receives the bedtime received the input in the process of step S31 and the input. The sleep time is calculated based on the wake-up time (step S34). The information indicating the calculated sleep time is sent to the severity calculating section 108.

Thereafter, the severity calculating unit 108 calculates the AHI based on the sleep time calculated by the process of step S34 and the number of apnea hypopnea counted by the process of step S32 between the bedtime and the wakeup time. The calculated result information indicating the result of the series of processes is output to the electronic device 11 (step S35), and the process here is ended.

According to the third embodiment described above, the biological sensor device 10 is further provided with a configuration for calculating the apnea hypopnea index by counting the number of times of apnea hypopnea while calculating the sleep time. The wearer (user) can more clearly understand whether or not he/she may develop sleep apnea syndrome.

In addition, in the above-mentioned 1st-3rd embodiment, although the respiratory analysis system demonstrated the structure containing the biosensor apparatus 10 and the electronic device 11 as an example, as shown in FIG. 16, for example, a respiratory analysis system. May further include a cloud service 14 including one or more server devices. That is, the respiratory analysis system may be implemented as a system using cloud computing. In this case, some or all of the various functions of the respiratory analysis application 100 can be provided in the server device in the cloud service 14. For example, the server device in the cloud service 14 acquires the acceleration signal measured by the biological sensor device 10 in response to a request from the electronic device 11, and performs the various processes described in the above-described first to third embodiments. Can be executed. According to this, various processes that are supposed to be executed in the biometric sensor device 10 can be executed in the server device in the cloud service 14, so that the processing load can be distributed.

According to at least one embodiment described above, since no nasal cannula or the like is attached, the burden on the patient is small, and the manifestation of a patient with sleep apnea syndrome can be sufficiently realized.

Since the processing of the present embodiment can be realized by a computer program, simply by installing and executing this computer program in a computer through a computer-readable storage medium storing this computer program, The same effect can be easily realized.

Although some embodiments of the present invention have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the spirit of the invention. These embodiments and modifications thereof are included in the scope and the gist of the invention, and are also included in the invention described in the claims and an equivalent range thereof.

DESCRIPTION OF SYMBOLS 1... Respiration analysis system, 10... Biological sensor device, 11... Electronic device, 12... Communication part, 13... Output part, 100... Respiration analysis application program, 101... Acceleration event detection part, 102... Event type determination part, 103... Body movement respiratory event discrimination unit, 104... Body movement event attribute determination unit, 105... Respiration event attribute determination unit, 106... Sleeping posture determination unit, 107... Sleep time calculation unit, 108... Severity degree calculation unit.

Claims (11)

A respiratory analysis device that can be worn on the human body,
Measuring means for measuring the wearer's acceleration,
Detection means for detecting movement of a human body based on an outlier in a waveform of a first frequency band obtained by executing a first filtering process on a waveform showing a temporal change in the measured acceleration. ,
And a determination unit that determines whether the detected movement of the human body is a movement caused by respiration or a movement caused by body movement .
The detection means is
In the waveform of the first frequency band, the time when the frequency becomes less than a first threshold value is detected as the outlier,
The discrimination means is
In the waveform of the second frequency band obtained by executing the second filtering process on the waveform of the measured acceleration, the characteristic of the waveform of the first section including the time detected as the outlier is shown. A respiratory analysis apparatus that determines whether the detected movement of the human body is a movement caused by respiration or a movement caused by body movement based on the first characteristic amount . The discrimination means is
It is determined whether the movement caused by the breathing is a movement associated with snoring, an apnea symptom, or a hypopnea symptom, and the movement caused by the body movement is a movement associated with either rolling over movement or minute body movement. The respiratory analysis apparatus according to claim 1, which determines whether or not there is. The discrimination means is
When the first characteristic amount is less than the second threshold value, it is determined that the detected movement of the human body is a movement caused by respiration, and the first characteristic amount is equal to or more than the second threshold value. If the detected body motion is determined to be a motion due to body motion, breath analyzer of claim 1. The discrimination means is
Regarding the movement of the human body determined to be the movement caused by the body movement, when the first feature amount is equal to or more than a third threshold value, the movement caused by the body movement is a movement accompanying a rolling over movement. The respiratory analysis apparatus according to claim 3 , wherein when the first feature amount is less than the third threshold value, the motion caused by the body motion is determined to be a motion associated with a small body motion. The discrimination means is
Regarding the movement of the human body determined to be the movement caused by the respiration, the outlier in the waveform of the third frequency band obtained by executing the third filtering process on the waveform of the measured acceleration. When the second characteristic amount indicating the characteristic of the waveform of the second section including the time detected as is less than the fourth threshold value, it is determined that the motion caused by the respiration is a motion associated with snoring, and When the second characteristic amount is equal to or more than the fourth threshold value and less than the fifth threshold value, it is determined that the movement caused by the respiration is a movement associated with a hypopnea symptom, and the second characteristic amount is The respiratory analysis apparatus according to claim 3 , wherein when it is equal to or more than the fifth threshold, it is determined that the movement caused by the respiration is a movement associated with an apnea symptom. The storage device further includes a storage unit that stores the time of the detected outlier and the information indicating what caused the motion of the human body detected based on the outlier in association with each other. Respiratory analysis device described in. The display device further comprises a display unit that displays, as an analysis result, the time of the detected outlier and the information indicating what caused the motion of the human body detected based on the outlier. Respiratory analysis device described in. The respiratory analysis apparatus according to claim 2, further comprising: an estimation unit that estimates a sleeping posture before and after the rolling motion when it is determined that the motion caused by the body motion is a motion associated with the rolling motion. First calculating means for calculating sleep time based on the bedtime and the wake-up time of the wearer,
From the bedtime to the wake-up time, the number of times that the movement caused by the breath is determined to be the movement associated with an apnea symptom or a hypopnea symptom, and sleep based on the calculated sleep time. The respiratory analysis apparatus according to claim 2, further comprising: a second calculation unit that calculates the severity of respiratory apnea syndrome. A respiratory analysis method applied to a respiratory analysis device that can be worn on a human body,
Measuring the wearer's acceleration,
Detecting a movement of a human body based on an outlier in a waveform of a first frequency band obtained by executing a first filtering process on a waveform showing a temporal change of the measured acceleration;
Determining whether the detected movement of the human body is movement caused by respiration or movement caused by body movement ,
The detecting is
In the waveform of the first frequency band, including detecting a time when the frequency is less than a first threshold value as the outlier,
The determination is
In the waveform of the second frequency band obtained by executing the second filtering process on the waveform of the measured acceleration, the characteristic of the waveform of the first section including the time detected as the outlier is shown. A respiratory analysis method comprising determining whether the detected movement of the human body is a movement caused by respiration or a movement caused by body movement based on the first characteristic amount . A respiratory analysis device program that can be worn on the human body,
The respiratory analysis device,
Measuring means for measuring the wearer's acceleration,
Detection means for detecting movement of a human body based on an outlier in a waveform of a first frequency band obtained by executing a first filtering process on a waveform showing a temporal change in the measured acceleration. ,
The detected movement of the human body is operated as a determination means for determining whether it is a movement caused by respiration or a movement caused by body movement ,
The detection means is
In the waveform of the first frequency band, the time when the frequency becomes less than a first threshold value is detected as the outlier,
The discrimination means is
In the waveform of the second frequency band obtained by executing the second filtering process on the waveform of the measured acceleration, the characteristic of the waveform of the first section including the time detected as the outlier is shown. A program for determining whether the detected movement of the human body is a movement caused by respiration or a movement caused by body movement based on the first characteristic amount .

Images