JP5583257B2 - Biological state acquisition device, biological state acquisition program, device provided with biological state acquisition device, and air conditioner - Google Patents

Description

  The present invention relates to a biological state acquisition device that acquires biological state information by processing biological signals such as respiration, heartbeat, and body movement, a biological state acquisition program, a device including the biological state acquisition device, and an air conditioner. .

  2. Description of the Related Art Conventionally, an apparatus using a Doppler radar sensor is known as a method for acquiring a biological state such as respiration, heartbeat, and body movement without contact. As a biological state acquisition device using a Doppler radar sensor, for example, a microwave is transmitted toward a human body, a frequency of a Doppler signal that is a change in wavelength of the transmitted wave and a reflected wave from the human body is obtained, and a human being is obtained from the frequency. There is a device that calculates the pulse rate or respiratory rate of the patient (see, for example, Patent Document 1).

  In recent years, using biological state information obtained by this type of biological state acquisition device, technology for measuring the sleep depth of a sleeping human, technology for estimating the state of autonomic nerve function, and biological state information Techniques for controlling various devices based on these have also been proposed (see, for example, Patent Document 2 and Patent Document 3).

JP 2002-71825 A (page 3) JP 2006-263032 A (pages 6-8, FIG. 1) JP 05-92040 A (2nd page, 3rd page, FIG. 1)

  In the above-described conventional technology, the frequency of the Doppler signal is obtained when acquiring the biological state, and complicated processing such as frequency analysis is required for the reflected wave. When the biological state information is used for controlling various devices, it is required to process the biological signal (sensor information) at high speed and with high accuracy to acquire the biological state in real time. For this reason, in the prior art, there is a problem that high processing capability is required for the CPU, and it is difficult to reduce the cost and simplify the device. Also, in each technique for measuring sleep depth using biological state information, estimating an autonomic nerve state, and controlling a device, it is required to acquire biological state information at high speed and with high accuracy.

  The present invention has been made in view of such points, and acquires a biological signal of a living body in a non-contact manner, and acquires information on the living body state without performing complicated processing such as frequency analysis on the living body signal. An object of the present invention is to provide a biological state acquisition device, a biological state acquisition program, a device including the biological state acquisition device, and an air conditioner.

  The biological state acquisition device according to the present invention transmits an electromagnetic wave to the body surface of a biological body, IQ-detects the reflected wave, and outputs an I signal and a Q signal. IQ signal acquisition means for sequentially acquiring the time series, and biological state acquisition means for acquiring the biological state based on the locus on the IQ plane of the acquisition signal acquired by the IQ signal acquisition means. The means has a heartbeat detecting means for detecting a heartbeat of a living body, and the heartbeat detecting means calculates a norm of a position vector of an acquired signal on the IQ plane, and performs a low-pass filter process on time series data of the vector norm. A signal is extracted, a heartbeat signal corresponding to one heartbeat is detected based on periodic fluctuations in the waveform of the extracted heartbeat signal, a heart rate in a unit period is calculated as heart rate information, and a heart rate in the unit period is calculated for a certain period Calculate the heart rate frequency distribution within a certain period, and determine whether the heart rate is normally counted based on whether the frequency distribution has a substantially normal distribution shape To do.

  According to the present invention, since the state of the living body is acquired based on the trajectory on the IQ plane of the acquired signal acquired by the IQ signal acquiring means, processing with high load such as frequency analysis is unnecessary, and low processing capability With this, it is possible to acquire information on the biological state.

It is a block diagram which shows the structure of the biological condition acquisition system provided with the biological condition acquisition apparatus which concerns on Embodiment 1 of this invention. It is a flowchart which shows the flow of the biological condition acquisition process in the arithmetic unit of FIG. It is a figure which shows IQ plane of the reflected wave according to the motion of the body surface at the time of human respiration. It is a figure which shows an example of the time series data of the norm of a velocity vector. It is explanatory drawing (the 1) of the arithmetic processing for detecting respiration from the time series data of FIG. It is explanatory drawing (the 2) of the arithmetic processing for detecting respiration from the time series data of FIG. It is explanatory drawing (the 3) of the arithmetic processing for detecting respiration from the time series data of FIG. It is a figure which shows an example of the time series data of the displacement amount (average value) of the body surface when the motion of the body surface is complicated. It is a figure which shows IQ plane at the time of measuring the motion of the body surface of FIG. It is a figure which shows the time series data of the norm of the velocity vector of IQ plane of FIG. It is a figure which shows the respiration cycle (breathing time) in case the respiration count is performed normally, and its distribution. It is a figure which shows the respiratory cycle (breathing time) in case the respiration count is not performed normally, and its distribution. It is a flowchart which shows the flow of the necessity determination process of correction | amendment of FIG. 2, and a correction | amendment process (when correction is required). It is explanatory drawing (the 1) of the calculation processing at the time of counting respiration from the time series data of the norm of the velocity vector of FIG. It is explanatory drawing (the 2) of the arithmetic processing at the time of counting respiration from the time series data of the norm of the velocity vector of FIG. It is explanatory drawing (the 3) of the arithmetic processing at the time of counting respiration from the time series data of the norm of the velocity vector of FIG. It is a block diagram which shows the structure of the biological condition acquisition system which concerns on Embodiment 2 of this invention. It is a flowchart which shows the flow of the biological condition acquisition process in the arithmetic unit which concerns on Embodiment 2 of this invention. It is a figure which shows IQ signal output of the state in which the heart rate signal and the body motion signal were mixed. FIG. 20 is a diagram illustrating a vector norm waveform after the signal output of FIG. 19 is subjected to envelope processing and low-pass filter processing. It is a figure which shows the time series data of the signal intensity | strength (Amplitude = The square root of each square sum of I and Q) of the I signal and Q signal after a low-pass filter process. It is explanatory drawing (the 1) of the arithmetic processing for detecting a heartbeat from the amplitude time series data of FIG. It is explanatory drawing (the 2) of the arithmetic processing for detecting a heartbeat from the amplitude time series data of FIG. It is explanatory drawing (the 3) of the arithmetic processing for detecting a heartbeat from the amplitude time series data of FIG. It is a figure which shows the time series data of the signal intensity | strength (amplitude = square root of the square sum of Q) of I signal and Q signal after a low-pass filter process in case a body surface pulsates in two steps in 1 beat. It is explanatory drawing (the 1) of the calculation process at the time of counting a heart rate from the time series data of FIG. It is explanatory drawing (the 2) of the arithmetic processing at the time of counting a heart rate from the time series data of FIG. It is explanatory drawing (the 3) of the arithmetic processing at the time of counting a heart rate from the time series data of FIG. It is a flowchart which shows the flow of the necessity determination process of correction | amendment of FIG. 18, and a correction process (when correction is required). It is a figure which shows the heart rate for every unit period when the heart rate count is performed normally, and its distribution. It is a figure which shows the heart rate for every unit period when the heart rate count is not performed normally, and its distribution. It is a block diagram which shows the structure of the biological condition acquisition system which concerns on Embodiment 3 of this invention. It is a flowchart which shows the flow of the biological condition acquisition process in the arithmetic unit which concerns on Embodiment 3 of this invention. It is a block diagram which shows the structure of the biological condition acquisition system which concerns on Embodiment 4 of this invention. It is the figure which showed the characteristic of the body movement, respiration, and heart rate in each case of shallow sleep (including awakening), deep sleep, and REM sleep. It is a flowchart which shows the flow of the biological condition acquisition process in the arithmetic unit which concerns on Embodiment 4 of this invention. It is a block diagram which shows the structure of the biological condition acquisition system which concerns on Embodiment 5 of this invention. It is a figure which shows the stable locus | trajectory (deep sleep) of IQ plane within a certain period. It is a figure which shows the unstable locus | trajectory (except deep sleep) of IQ plane within a certain period. It is a flowchart which shows the flow of the biological condition acquisition process in the arithmetic unit which concerns on Embodiment 5 of this invention. It is a block diagram which shows the structure of the air conditioner provided with the biological condition acquisition system which concerns on Embodiment 1-5 of this invention. It is a flowchart which shows control of the air conditioner which concerns on Embodiment 6 of this invention.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. In Embodiments 1 to 5, the biological state acquisition device will be described in detail. In the biological state acquisition device described below, the activity state of a living body such as breathing, heartbeat, and body movement is acquired as basic data, and further, based on this basic data, the state of the autonomic nerve (sympathetic dominant, parasympathetic nerve) Dominance etc.) and the sleep depth of the living body. In each of the following embodiments, a case where the living state of a living body (for example, a human) during sleep is acquired will be described as an example.

Embodiment 1 FIG. (Respiration detection)
FIG. 1 is a block diagram illustrating a configuration of a biological state acquisition system including a biological state acquisition device according to Embodiment 1 of the present invention. In Embodiment 1, the respiration information regarding respiration is acquired as basic data of a biological state. A case where the state of the autonomic nerve and the human sleep depth are calculated (acquired) based on the acquired respiratory information will be described.
The biological state acquisition system 100 transmits an electromagnetic wave (microwave) to a sleeping human and receives a reflected wave from the human of the transmitted wave, an IQ detector 20, and a bandpass filter. 30 and an AD converter 40. The biological state acquisition system 100 further includes an arithmetic device 50 as a biological state acquisition device and a storage device 60 that stores various data (such as learning data described later).

  The Doppler radar sensor 10 radiates electromagnetic waves toward a sleeping human and is arranged so as to be able to capture reflected waves reflected from the human body surface. The Doppler radar sensor 10 is composed of a module containing an antenna that receives a reflected wave as a received wave, an input / output amplifier, an oscillator, an IQ mixer (detector), a power source, and peripheral components (both shown in FIG. Not shown).

  The IQ detector 20 decomposes the reflected wave received by the antenna of the Doppler radar sensor 10 into an in-phase component (I signal) and a quadrature component (Q signal) with respect to the incident wave, and outputs them to the bandpass filter 30. The band-pass filter 30 has a low-frequency band-pass filter 31 for respiration detection, extracts a target signal and outputs it to the AD converter 40. Since the output of the IQ detector 20 is a signal in which not only respiration but also heartbeat and body movement are all superimposed, a respiration signal is extracted by passing the signal through the band-pass filter 31. Then, the signal after passing through the band pass filter 31 is converted into a digital signal by the AD converter 40 and output to the arithmetic unit 50. Note that the pass frequency band of the bandpass filter 31 for respiration detection is set in advance.

  The arithmetic device 50 is constituted by a microcomputer and includes a CPU, a ROM and a RAM inside, and operates according to various programs stored in the ROM. When the CPU executes the biological state acquisition program stored in the ROM, the IQ signal acquisition unit 51 that acquires the IQ signal from the AD converter 40 and the biological state acquisition unit 52 are functionally configured. .

  The biological state acquisition unit 52 applies the Doppler radar sensor 10 to a sleeping human and uses the Doppler effect caused by the movement of the human body surface by body movement such as breathing, heartbeat, and turning over the biological state (breathing). , Autonomic state, sleep depth, etc.).

  The biological state acquisition means 52 detects the respiration of the living body and calculates respiration information such as the respiration rate, and the autonomy for determining the autonomic nervous state of the living body based on the respiration information calculated by the respiration detection means 53. And a nerve state determination means 54. The biological state acquisition unit 52 further includes a sleep depth determination unit 55 that determines the sleep depth of the biological body based on the respiration information calculated by the respiration detection unit 53.

  Although FIG. 1 shows an example in which the band pass filter 31 and the AD converter 40 are provided in the subsequent stage of the Doppler radar sensor 10, the band pass filter 31 and the AD converter 40 may be incorporated in the module of the Doppler radar sensor 10. good. Further, the band pass filter 31 may be configured by a digital filter and disposed at the subsequent stage of the AD converter 40. Further, when the output of the IQ detector 20 is insufficient, an amplifier may be further arranged in front of the AD converter 40, and the configuration may be any configuration as long as the IQ signal can be accurately filtered and input to the arithmetic unit 50. .

Hereinafter, the operation of the biological state acquisition system 100 will be described.
The Doppler radar sensor 10 emits electromagnetic waves toward a sleeping human and receives a reflected wave from the human with an antenna (not shown). Then, the Doppler radar sensor 10 amplifies the received reflected wave with an amplifier and outputs it to the IQ detector 20. The signal input to the IQ detector 20 is decomposed into an I signal and a Q signal, a respiratory signal is extracted by a band pass filter, converted into a digital signal by the AD converter 40, and then output to the arithmetic unit 50. Is done. Respiration signals (I signal and Q signal) from the AD converter 40 are sequentially input to the arithmetic device 50 in time series.

  The arithmetic device 50 acquires the I signal and the Q signal from the AD converter 40 by the IQ signal acquisition means 51, and performs respiration detection based on the trajectory of the acquired signal on the IQ plane. Hereinafter, the respiration detection method in the respiration detection means 53 of the arithmetic unit 50 will be described in detail.

  First, the measurement principle of the biological state will be briefly described. In humans, the body surface moves due to movements of respiratory muscles and diaphragm due to breathing, pulsation due to heartbeat, and body movement. In the case of breathing, the body surface of the chest is moved by breathing. For this reason, a Doppler shift corresponding to the speed of movement of the body surface due to respiration occurs in the reflected wave from the person received by the antenna of the Doppler radar sensor 10.

  When the behavior of the body surface during breathing is analyzed in detail, the speed of movement of the body surface is substantially zero before the start of breathing (start of inhalation), and after the inhalation starts, the speed gradually increases and reaches a peak. Then, the velocity decreases this time toward the end point of breathing (when switching to sucking and exhaling), and becomes substantially zero at the end point of breathing. After the start of exhalation, the speed of movement of the body surface gradually increases and reaches a peak, and then the speed decreases toward the exhalation end point (at the end of exhalation), and the speed is almost zero at the exhalation end point. It becomes. Respiration is detected by detecting such a change in the velocity of the body surface from the detection result of IQ detection.

FIG. 2 is a flowchart showing the flow of the biological state acquisition process in the arithmetic device of FIG. Hereinafter, the flow of the biological state acquisition process will be described with reference to FIG.
(S1: I signal and Q signal acquisition)
The IQ signal acquisition means 51 of the arithmetic device 50 acquires IQ signals sequentially output in time series from the AD converter 40 in accordance with the movement of the human body surface. The biological state acquisition unit 52 first calculates the norm of the velocity vector based on the acquired I signal and the locus of the Q signal on the IQ plane. Hereinafter, calculation of the norm of the velocity vector will be described.

(S2: Speed vector norm calculation, S2: Respiration detection)
When the acquired signal acquired by the IQ signal acquiring means 51 is plotted on the IQ plane, a locus as shown in FIG. 3 is drawn according to the movement of the human body surface.

FIG. 3 is a diagram illustrating an example of an IQ plane of a reflected wave according to the movement of the body surface during human breathing. 3A shows a case where the body surface approaches the Doppler radar sensor 10, and FIG. 3B shows a case where the body surface moves away from the Doppler radar sensor 10.
The arrow in the counterclockwise direction in FIG. 3A indicates the direction of the locus of coordinates on the IQ plane of the IQ signal when the body surface approaches the Doppler radar sensor 10. The clockwise arrow in FIG. 3B indicates the direction of the locus of coordinates on the IQ plane of the IQ signal when the body surface moves away from the Doppler radar sensor 10. Further, points 1 to 9 in FIG. 3A and points 1 to 9 in FIG. 3B are plots of IQ signal coordinates for each sampling time.
As described above, the movement of the body surface during breathing reaches a peak when the inhalation is started and the speed gradually increases after the inhalation starts. Then, the velocity decreases this time toward the end point of breathing (when switching to sucking and exhaling), and becomes substantially zero at the end point of breathing. When the body surface moves quickly, the phase change of the reflected wave due to the Doppler effect increases. For this reason, on the IQ plane, the distance between the points in FIG. 3A is narrow at the start of the operation, gradually widens, becomes longest at the middle part of the trajectory, and then narrows again. The state of going is shown. An IQ plane having similar characteristics is also obtained during discharge.

  The velocity vector corresponds to a vector difference between points (point 1 to point 9 in the figure) on the IQ plane obtained at each sampling interval. Arrow (a) indicates the velocity vector between points 5 and 6. The length of the vector difference between each point corresponds to the norm of the velocity vector at that timing. When the norm of the velocity vector is illustrated in time series, the following diagram is shown in FIG.

FIG. 4 is a diagram showing time-series data of norms of velocity vectors. In FIG. 4, the horizontal axis represents time, and the vertical axis represents the norm of the velocity vector.
When the body surface moves due to breathing, as described above, the body surface moves at the highest speed approximately halfway between the inhalation and exhalation operations, and at the end point of breathing (when switching between inhalation and exhalation), the body surface The speed of is almost zero. Accordingly, the time series data of the velocity vector norm draws a mountain-shaped curve in each of the suction operation and the discharge operation, as shown in FIG. Therefore, two peaks in the time series data of the norm of the velocity vector indicate one breath. Therefore, it is possible to sequentially calculate the norm of the velocity vector from the IQ signal and to detect respiration from the time series data of the norm.

  In this way, using the feature of the movement of the body surface due to respiration, one respiration is extracted from the respiration signal, with the trajectory speed on the IQ plane being substantially zero at the end point of respiration. That is, a respiratory signal corresponding to one breath is detected based on periodic fluctuations in the waveform of the time series data of the velocity vector norm due to the velocity vector norm being substantially zero when switching between breathing in and breathing out. . Since no frequency analysis is required when calculating the norm, it is possible to detect respiration with a simple and low load.

  The method shown in FIGS. 5 to 7 can be used as a specific calculation process for detecting two mountain curves as one breath from time-series data of the velocity vector norm. In the method of FIG. 5, a constant value is subtracted from the time series data of the velocity vector norm, and the zero cross point of the time series data after the subtraction is detected. One breath is counted for every four zero-cross points. As another method, as shown in FIG. 6, the peak of the velocity norm vector data may be extracted, and the time from the appearance of one peak to the appearance of the second peak may be counted as one breath. As a method opposite to that shown in FIG. 6, the bottom of the valley of the velocity vector norm data is extracted as shown in FIG. 7, and one breath is counted until one valley appears. May be. In the method of FIG. 5, when the amplitude of the norm of the velocity vector varies, a portion that does not zero-cross occurs depending on the setting of a constant numerical value, and there is a possibility that respiration of that portion cannot be counted. On the other hand, in the case of the method of FIGS. 6 and 7, respiration can be counted even when the amplitude of the norm of the velocity vector is not constant.

(S3: Calculation of respiratory rate and fluctuation range (standard deviation) of respiratory cycle)
The respiration detection means 53 detects two mountain curves as one respiration from the time-series norm calculation result as described above, and calculates respiration information. As the respiration information, the respiration rate for a certain period (for example, the past 2 minutes) is calculated. In addition, the fluctuation width of the respiratory cycle for a certain period (the fluctuation width of the time required for one breath (standard deviation)) is calculated.

(S4: Respiratory rate correction necessity determination / correction (if correction is required))
During sleep, when the movement of the body surface due to respiration is stable, respiration can be detected by the above method, but when the movement of the body surface is complicated, the respiration may not be accurately counted. For example, a case where a plurality of muscles are active to cancel the Doppler shift and the movement of the body surface cannot be detected can be considered. In such a case, it is necessary to correct the respiratory rate. Therefore, in step S4, a respiration rate correction necessity determination process is performed to determine whether respiration can be accurately counted. The details of the correction necessity determination process will be described later. Here, the description of the biological state acquisition process when it is determined that the respiration rate can be accurately counted and correction is unnecessary is continued.

(S5: Autonomic nerve state determination)
When the state of the autonomic nerve is the sympathetic nerve dominant state, the respiratory rate is large and the fluctuation range of the respiratory cycle is large. On the other hand, when the parasympathetic nerve is dominant, the respiratory rate is small and the fluctuation range of the respiratory cycle is small. Therefore, the state of the autonomic nerve can be determined by obtaining the fluctuation range of the respiratory rate and the respiratory cycle.

  The autonomic nerve state determination unit 54 calculates an index for determining the state of the autonomic nerve based on the respiration information calculated by the respiration detection unit. This index may be the value of the fluctuation rate itself of the respiratory rate or the respiratory cycle, or may be a value obtained by substituting each into a certain function, and takes a larger value as the sympathetic nerve is dominant. Shall. The state of the autonomic nerve can be determined by this index. For example, the index is compared with a preset threshold. If the index is larger than the threshold, it is determined that the sympathetic nerve is dominant, and if the index is less than the threshold, it is determined that the parasympathetic nerve is dominant. In addition, for example, the degree of sympathetic nerve activity may be determined.

(S6-S10: Sleep depth determination)
Moreover, the sleep depth determination means 55 determines the sleep depth based on respiration information. Next, the operation of the sleep depth determination means 55 will be described.

  Here, before explaining the operation of the sleep depth determination means 55, the sleep depth will be described first. Sleep is generally divided into REM sleep, which is light sleep, and non-REM sleep, which is deep sleep. The sleep depth is further defined in detail, and six states of awakening, REM sleep, and sleep depths 1, 2, 3, and 4 are defined. The sleep depths 1, 2, 3, and 4 are obtained by further dividing the non-REM sleep into four stages. In the non-REM sleep, the sleep depth 1 is the shallowest and the sleep depth 4 is the deepest.

  REM sleep, sleep depth 1, 2, 3, 4 and so on until the next awakening from the start of sleep onset, then sleep depth 3, 2, 1, REM The sleep cycle of shifting to sleep is usually repeated with a period of about 90 minutes. In this example, one of REM sleep, shallow sleep (for example, sleep depths 1 and 2), and deep sleep (for example, sleep depths 3 and 4) is determined based on the respiratory rate and the fluctuation range of the respiratory cycle.

  It is known that the aspect of human respiratory rate during sleep changes depending on the depth of sleep. In general, the respiratory rate during deep sleep is low and stable (the fluctuation range of the respiratory cycle is small), and the respiratory rate during shallow sleep is high and unstable (the fluctuation range of the respiratory cycle is large). Moreover, it is the most unstable during REM sleep, and the fluctuation range of the respiratory cycle is even larger. Therefore, the first respiratory rate threshold value, the second respiratory rate threshold value (<first respiratory rate threshold value), the first variation width threshold value, and the second variation width threshold value (<first variation) for determining REM sleep, shallow sleep, and deep sleep. (Variation width threshold value) is obtained and set in advance by experiments or the like, and the sleep depth is determined by comparison with each threshold value.

  That is, if the respiration rate is equal to or greater than the first respiration rate threshold and the variation width of the respiration cycle is equal to or greater than the first variation width threshold, it is determined that REM sleep (S6, S8). If the respiratory rate is less than the first respiratory rate threshold value and greater than or equal to the second respiratory rate threshold value, and the fluctuation range of the respiratory cycle is less than the first fluctuation range threshold value and greater than or equal to the second fluctuation range threshold value, it is determined to be shallow sleep (S6, S9). . If the respiration rate is less than the second respiration rate threshold value and the fluctuation range of the respiration cycle is less than the second fluctuation range threshold value, deep sleep is determined (S6, S10).

  These threshold values may be set for each individual based on the sleep data obtained by providing a learning period and acquiring sleep data of one sleep cycle or more. Further, sleep data during the learning period may be analyzed based on a predetermined algorithm, and a threshold value may be determined and automatically set.

(Detailed description of determination of necessity of correction of respiration rate in step S4 in FIG. 2)
Hereinafter, details of the determination of necessity of correction of the respiration rate will be described. The case where the correction of the respiration rate is necessary corresponds to the case where the movement of the body surface during respiration is complicated and the Doppler shift is canceled as described above. Next, FIG. 8 shows an example of time-series data of the displacement amount (average value) of the body surface when the movement of the body surface is complicated.

FIG. 8 is a diagram illustrating an example of time-series data of the displacement amount (average value) of the body surface when the movement of the body surface is complicated. The solid line shows the case where the movement of the body surface is complicated. Note that the dotted line extending in the vertical direction in FIG. 8 indicates the case where the body surface moves normally for reference, and the amount of displacement of the body surface gradually increases over time, and then starts to decrease. It shows how the operation is repeated.
In the example of FIG. 8, as an example of a case where the movement of the body surface is complicated, an example in which there is a period in which the average value of the movement of the entire body surface does not fluctuate in each operation of suction and discharge. A portion surrounded by a horizontally long ellipse in FIG. 8 indicates the period. The IQ plane when the movement of the body surface indicated by the solid line in the figure is measured is as shown in FIG.

FIG. 9 is a diagram showing an IQ plane when the movement of the body surface as shown in FIG. 8 is measured. FIG. 9 shows the suction operation and the discharge operation in one breath of FIG.
In the period in which the average value of the movement of the entire body surface does not change, the movement close to the Doppler radar sensor and the movement away from the Doppler radar sensor are mixed on the entire body surface. As a result, the Doppler shift is offset. Therefore, the velocity vector on the IQ plane of the IQ signal in that period becomes zero. In FIG. 9, the part surrounded by the vertically long ellipse shows the part where the velocity vector becomes zero. Accordingly, the time series data of the norm of the velocity vector as shown in FIG. 9 is as shown in FIG.

FIG. 10 is a diagram showing time-series data of norms of velocity vectors on the IQ plane of FIG.
In the IQ plane of FIG. 9, in the time series data of the norm of the velocity vector, each of the suction operation and the discharge operation appears as two peaks as shown in FIG. That is, one breath appears as four mountains. As described above, the breath detection means 53 employs an algorithm that counts two mountains as one breath. Therefore, if one breath appears as four mountains, the portion that is one breath accurately counts as two breaths. It will be done.

11 and 12 are diagrams showing a respiratory cycle (breathing time) and its distribution. FIG. 11 shows a case where the breath count is performed normally, and FIG. 12 shows a case where the breath count is not performed normally. Yes. FIG. 11 and FIG. 12 are diagrams in which the respiratory cycle of each breath is plotted as the vertical axis in the order of each breath taken on the horizontal axis.
In the case of the norm of the velocity vector in FIG. 4, the respiratory cycle is a time indicated by times Ta, Tb, and Tc. If the frequency distribution of the respiratory cycle is taken, a substantially normal distribution shape is obtained as shown in FIG. On the other hand, in the case of the norm of the velocity vector in FIG. 10, the four peaks should be counted as one breath and the breathing cycle should be calculated as time TA + TB. Calculated as TA, TB... Therefore, as shown in FIG. 12, the frequency distribution of the respiratory cycle is clearly a bipolar distribution and has a shape having two peaks. Therefore, the respiration detecting means 53 calculates the frequency distribution of the respiration cycle after counting the respiration rate, and can determine whether or not the respiration count is normally performed depending on whether or not the frequency distribution has a substantially normal distribution shape. .

  By the way, the case where the movement of the body surface due to breathing is complicated specifically corresponds to, for example, the case where the person is sleeping sideways, and the posture at the time of sleep influences. Therefore, it is unlikely that a normal respiratory count and an abnormal respiratory count are mixed in a short period of time. Even if they are mixed, the number is small. Based on this assumption, it is possible to determine whether respiration counting is normally performed or not from the frequency distribution of the respiration cycle.

FIG. 13 is a flowchart showing the flow of the correction necessity determination process and the correction process (when correction is necessary) in FIG. Hereinafter, the flow of the necessity determination process and the correction process (when correction is necessary) will be described with reference to FIG.
The respiration detection means 53 calculates the frequency distribution of the respiration cycle (S21), and checks whether the distribution has a substantially normal distribution shape (S22). If the distribution is a substantially normal distribution shape, it is determined that normal respiration rate counting is being performed, and it is determined that correction is unnecessary (S23). On the other hand, if the respiratory cycle distribution deviates from the substantially normal distribution shape, it is determined that normal respiration rate count is not performed, and it is determined that correction is necessary (S24). Whether or not the frequency distribution of the respiratory cycle has a substantially normal distribution shape is determined by calculating a moment of the frequency distribution of the respiratory cycle and determining the threshold value. Specifically, for example, the degree of kurtosis, which is a third moment, is calculated, and if the degree of kurtosis deviates from a preset threshold value 3 (normal distribution), it is determined that the degree of kurtosis deviates from a substantially normal distribution shape. .

  When it is determined that correction is required by the above algorithm, the respiration detecting means 53 estimates the number N of peaks in the frequency distribution of the respiration cycle (for example, by maximum likelihood estimation) (S25). 8 to 12 exemplify the case where N = 2, the number is not limited to two depending on the movement of the body surface, and there may be a plurality of cases. Then, based on the number N, the respiration rate and respiration cycle in the past fixed period calculated in step S3 (see FIG. 2) are corrected (S26). Hereinafter, this correction will be described with a specific example.

14 to 16 are explanatory diagrams of calculation processing when counting respiration from the time-series data of the norm of the velocity vector in FIG. 14 shows the count by the zero cross point, FIG. 15 shows the count by the peak extraction, and FIG. 16 shows the count by the valley extraction. Here, it is assumed that the number N of peaks in the distribution of the respiratory cycle is N = 2.
In step S3, an algorithm that counts two mountains as one breath is adopted. Therefore, in the case of the measurement data in FIG. 10, as shown in FIGS. Be counted. Therefore, correction is performed to recount 2 breaths as 1 breath and the respiratory cycle is corrected. Specifically, after calculating the respiration rate for a certain period, the respiration rate of the calculation result is divided by 2 to obtain a corrected respiration rate. It should be noted that, from the time-series data of the norm of the velocity vector for a certain period, for example, correction is performed by recounting one breath for 4 (number of zero crosses) × N (number of peaks) by a counting method using zero cross points, for example. It may be. For example, the correction of the breathing cycle is performed by correcting the time TA + TB as the breathing cycle as described in the example of FIG.

  As described above, if it is determined in step S24 that correction is necessary, correction is performed. Then, based on the corrected respiratory rate and respiratory cycle, the determination of the state of the autonomic nerve in step S5 and the determination of the sleep depth in steps S6 to S10 are performed.

  The computing device 50 outputs the determination result of the state of the autonomic nerve and the sleep depth determined as described above to an external device such as an air conditioner. On the device side that has received the determination result, device control according to the determination result is performed. A specific example of device control will be described in a later-described sixth embodiment.

  As described above, in the first embodiment, a difference vector (velocity vector) is calculated from the locus of the IQ plane, a norm of the difference vector is calculated, and respiratory information (respiration rate and respiration) is calculated from the time series data of the norm. Period). Therefore, processing with a high load such as frequency analysis as in the past is not necessary, and respiratory information can be obtained using an inexpensive arithmetic device 50. Moreover, the state of the autonomic nerve and the sleep depth of the living body that are closely correlated with the respiratory information can be determined at high speed. Note that since the state of the autonomic nerve and the sleep depth are related to each other, the sleep depth may be determined based on the index indicating the state of the autonomic nerve. By estimating the activity of the autonomic nerve accompanying breathing, it becomes possible to determine the sleep state according to the physiological model.

  In addition, according to the first embodiment, it is possible to detect a counting error that counts one breath as a plurality of breaths. When a counting error is detected, the respiratory rate is corrected. Calculation accuracy can be improved. As a result, the determination accuracy of the state of the autonomic nerve and the sleep depth can be improved.

  Moreover, since the determination of the state of the autonomic nerve and the determination of the sleep depth are performed in the multidimensional space using both the respiratory rate and the fluctuation range of the respiratory cycle, it is possible to determine with high accuracy. Although some accuracy degradation cannot be denied, the determination of the state of the autonomic nerve and the determination of the sleep depth may be performed based on one of the respiratory rate and the fluctuation range of the respiratory cycle. In the first embodiment, the respiratory information and the fluctuation range of the respiratory cycle are given as examples of the respiratory information. However, the present invention is not limited to this, and the respiratory information fluctuation range may be further included in the respiratory information. Good.

  In addition, when the moving speed of the body surface is fast, the phase of the IQ signal may change by 360 degrees or more. In this case, since the coordinates on the IQ plane are the same, it is difficult to count one breath from simple IQ plane coordinates. However, in the first embodiment, since the norm of the velocity vector of the IQ signal between each sampling is used and the respiration is detected based on an event such as whether the body surface has moved or stopped during each sampling time, Respiration can be counted regardless of the moving speed of the body surface.

  In the first embodiment, the case of detecting respiration has been described. However, if changes such as heartbeats and pulse waves can be acquired as information by IQ detection, they can be detected by the same method as described above. In the case of respiration, since the body surface has a larger movement range than that of heartbeats and pulse waves and the movement is slow, it is easy to find that the speed becomes substantially zero at the turnaround point of respiration. For this reason, the method of Embodiment 1 is particularly preferable in the case of breathing. In the case of breathing, since the body surface has a large movement width, detection is possible even at a position away from the body surface.

  In the first embodiment, an example of detecting respiration mainly during sleep has been described. However, respiration detection can be performed not only during sleep but also in a relaxed state or in another state such as driving by changing a filter. Is possible. If it is a digital filter, it is possible to change the filter automatically, and it is also possible to automatically change the setting according to the scene from which the biological state is acquired.

  As a method for preventing multiple false counts of respiration, for example, the locus itself on the IQ plane is filtered with a moving average to absorb the portion where the velocity becomes zero in the waveform of FIG. Based on this, the respiratory rate may be counted. In this case, it is possible to prevent multiple breaths from being erroneously counted. The filter is a moving average, but any filter can be used as long as it can suppress minute changes and steep changes.

Embodiment 2. FIG. (Heart rate detection)
The second embodiment describes a method suitable for acquiring a biological state in which the movement of the body surface due to body movement is small like a heartbeat and the movement is complicated.

FIG. 17 is a block diagram showing a configuration of the biological state acquisition system according to Embodiment 2 of the present invention. In FIG. 17, the same parts as those in the first embodiment shown in FIG.
In the biological state acquisition system 200 of the second embodiment, the bandpass filter 30 includes a high-frequency bandpass filter 32 for heartbeat and body movement instead of the respiratory bandpass filter 31 of the first embodiment. In addition, the biological state acquisition unit 52 includes a heartbeat detection unit 56 that detects a heartbeat of a living body and calculates heartbeat information such as a heart rate, an autonomic state determination unit 54A, and a sleep depth determination unit 55A. Other configurations are the same as those of the first embodiment.

  Since the output of the IQ detector 20 is a signal in which all of respiration, heartbeat, and body movement are superimposed, the heartbeat and body movement are obtained by passing this signal through a high-frequency bandpass filter 32 for detecting heartbeat and body movement. Signal is extracted. The extracted signal is converted into a digital signal by the AD converter 40 and input to the arithmetic unit 50. The pass frequency band of the band pass filter 32 is set in advance.

  The heartbeat detection unit 56 includes a low-pass filter (not shown) that removes high-frequency components from the acquisition signals (I signal and Q signal) acquired by the IQ signal acquisition unit 51. The heartbeat signal is extracted by passing the measurement data input to the arithmetic unit 50 through a low-pass filter. The heartbeat detecting means 56 detects a heartbeat signal corresponding to one heartbeat based on the periodic fluctuation of the waveform of the heartbeat signal, and calculates the heart rate in the unit period as heartbeat information.

  The autonomic nerve state determination unit 54 </ b> A calculates an index indicating the autonomic nerve state based on the heartbeat information calculated by the heartbeat detection unit 56.

  The sleep depth determination unit 55 </ b> A determines the sleep depth based on the heartbeat information calculated by the heartbeat detection unit 56.

In the following, the second embodiment will be described focusing on the differences from the first embodiment.
FIG. 18 is a flowchart showing the flow of the biological state acquisition process in the arithmetic device according to the second embodiment of the present invention.
(S31: I signal and Q signal acquisition)
The I signal and the Q signal output from the IQ detector 20 in the biological state acquisition system are signals in which respiration, heartbeat, and body motion are all mixed. For this reason, the I signal and Q signal output from the IQ detector 20 are passed through the high-frequency bandpass filter 32 for detecting heartbeat and body motion, thereby extracting the heartbeat signal and body motion signal. This signal is converted into a digital signal by the AD converter 40 and input to the IQ signal acquisition means 51. The pass frequency band of the band pass filter 32 is set in advance. In this way, the IQ signal acquisition means 51 acquires IQ signals (heartbeat and body motion signals) sequentially output in time series from the AD converter 40 according to the movement of the human body surface.

(S32: Calculate time series data of norm of IQ vector)
The heartbeat detecting unit 56 expresses a point on the IQ plane of the acquired signal (a signal in which the heartbeat signal and the body motion signal are mixed) acquired by the IQ signal acquiring unit 51 as a vector from the origin (hereinafter, this vector). (The position vector of the acquired signal on the IQ plane is called the IQ vector). Then, the norm of the IQ vector (amplitude = square root of the sum of squares of I and Q) is calculated.

(S33: Heartbeat signal extraction)
Here, since the acquisition signal acquired by the IQ signal acquisition unit 51 is a signal in which the heartbeat signal and the body motion signal are mixed, the heartbeat detection unit 56 passes the low-pass filter (not shown) as described above. . As a result, a waveform of a heartbeat signal indicating pulsation due to a heartbeat is obtained.

FIG. 19 is a diagram showing the IQ signal output in a state where the heartbeat signal and the body motion signal are mixed, and shows the IQ signal after the band-pass filter. FIG. 20 is a diagram showing a vector norm waveform after the signal output of FIG. 19 is subjected to envelope processing and low-pass filter processing. FIG. 20 also shows a reference (actual pulsation waveform) for reference.
As apparent from comparison between FIG. 19 and FIG. 20, the waveform of the heartbeat signal synchronized with the actual pulsation waveform is obtained by passing through a low-pass filter (not shown).

  The heartbeat detecting means 56 detects a heartbeat signal corresponding to one heartbeat based on the periodic fluctuation of the waveform of the heartbeat signal (I signal and Q signal) after passing through the low-pass filter, and uses the heart rate in the unit period as heartbeat information. calculate. Hereinafter, specific processing of heartbeat detection will be described.

(S34: Calculating fluctuation range of heart rate during unit period and heart rate during certain period)
FIG. 21 is a diagram showing time-series data of the signal strength (amplitude = square root of sum of squares of I and Q) of the I signal and the Q signal after the low-pass filter processing. FIG. 21 corresponds to a partially enlarged view of FIG.
The heart rate detecting means 56 counts the heart rate from time series data (hereinafter referred to as amplitude time series data) of the signal intensity (amplitude = square root of sum of squares of I and Q) of the I signal and the Q signal after low pass filter processing. To do. As shown in FIG. 21, the time from the peak of the amplitude time series data to the next peak is counted as one beat.

  As a specific calculation process for counting the heart rate from the amplitude time series data, the methods shown in FIGS. 22 to 24 can be used. In the method of FIG. 22, a constant value is subtracted from the amplitude time series data, and the zero cross point of the data after subtraction is detected. One beat is counted for every three zero-cross points. As another method, as shown in FIG. 23, the peak of amplitude time series data may be extracted, and the time from the appearance of one peak to the appearance of the next peak may be counted as one beat. As another method, as shown in FIG. 24, the bottom of the valley of the amplitude time series data is extracted, and the time from when one valley bottom appears until the next valley bottom appears is counted as one beat. Also good.

(Heart rate time series data generation)
The heart rate detecting means 56 calculates the heart rate for a unit period (for example, the past one minute) by the above method. The above heart rate is calculated over a certain period (for example, 3 minutes) to generate heart rate time-series data. The heart rate detecting means 56 calculates the fluctuation range (standard deviation) of the heart rate within a certain period. As described above, the heart rate detecting means 56 calculates the heart rate and the fluctuation range of the heart rate as the heart rate information.

(S35: Heart rate correction necessity determination / correction (if correction is required))
During sleep, when the pulsation of the body surface due to a heartbeat is complicated, for example, the body surface may pulsate in two stages in one beat. In this case, in the above method, one beat may be counted as a plurality of beats. In this case, the heart rate needs to be corrected. Therefore, in step S35, it is determined whether or not the heart rate needs to be corrected in order to determine whether the heart rate can be accurately counted. The details of the correction necessity determination process will be described later. Here, the description of the biological state acquisition process when it is determined that the heart rate can be accurately counted and correction is unnecessary is continued.

(S36: Autonomic nerve state determination)
The autonomic nerve state determination unit 54 </ b> A calculates an index for determining the state of the autonomic nerve based on the heartbeat information calculated by the heartbeat detection unit 56. This index may be the value of the heart rate per unit time or the fluctuation range of the heart rate over a certain period, or may be a value obtained by substituting each into a certain function, etc. , Take a large value. The state of the autonomic nerve can be determined by this index. For example, the index is compared with a preset threshold. If the index is larger than the threshold, it is determined that the sympathetic nerve is dominant, and if the index is less than the threshold, it is determined that the parasympathetic nerve is dominant. In addition, for example, the degree of sympathetic nerve activity may be determined.

(S37-S41: Sleep depth determination)
The sleep depth determination unit 55 </ b> A determines the sleep depth based on the heart rate information calculated by the heart rate detection unit 56. Next, the operation of the sleep depth determination means 55A will be described.

  It is known that the human heart rate during sleep changes in appearance depending on the sleep depth, similar to the above-described respiratory rate. In general, the heart rate during deep sleep is low and stable (the heart rate fluctuation range is small), and during shallow sleep, the heart rate is high and unstable (the heart rate fluctuation range is large). Moreover, it is the most unstable during REM sleep, and the heart rate fluctuation range is even larger. Therefore, the first heart rate threshold value, the second heart rate threshold value (<first heart rate threshold value), the first heart rate fluctuation range threshold value, and the second heart rate fluctuation range threshold value for determining REM sleep, shallow sleep, and deep sleep. (<First heart rate fluctuation range threshold) is set in advance, and the sleep depth is determined by comparison with each threshold.

  That is, if the heart rate is equal to or greater than the first heart rate threshold and the heart rate variation is equal to or greater than the first heart rate variation, the REM sleep is determined (S37, S39). If the heart rate is less than the first heart rate threshold value and greater than or equal to the second heart rate threshold value, and the heart rate fluctuation width is less than the first heart rate fluctuation width threshold and greater than or equal to the second heart rate fluctuation width threshold value, it is determined to be shallow sleep (S38). , S40). If the heart rate is less than the second heart rate threshold value and the heart rate fluctuation range is less than the second heart rate fluctuation range threshold value, it is determined to be deep sleep (S38, S41).

  These threshold values may be set for each individual based on the sleep data obtained by providing a learning period and acquiring sleep data of one sleep cycle or more. The sleep data during the learning period may be analyzed based on a predetermined algorithm, and a threshold value may be determined and automatically set.

(Detailed description of the heart rate correction necessity determination in step S35 in FIG. 18)
Hereinafter, details of the necessity determination of heart rate correction will be described. The case where the heart rate needs to be corrected corresponds to, for example, the case where the body surface pulsates in two stages in one beat as described above. FIG. 25 shows an example of time-series data of the displacement amount (average value) of the body surface when the movement of the body surface is complicated.

FIG. 25 is a diagram showing amplitude time-series data, and shows a heartbeat signal when the body surface pulsates in two stages in one beat.
When the heart rate detector 56 counts the heart rate from the amplitude time series data, the measurement data in FIG. 25 is counted as 2 beats as shown in FIGS. 26 to 28 correspond to the counting methods shown in FIGS. 22 to 24, respectively. The heart rate correction necessity determination algorithm is basically the same as the respiration rate correction necessity determination algorithm of the first embodiment, and the heart rate correction necessity determination process will be briefly described below.

FIG. 29 is a flowchart showing a flow of the correction necessity determination process and the correction process (when correction is necessary) in FIG. 30 and 31 are diagrams showing the heart rate for each unit period and the frequency distribution of the heart rate. FIG. 30 shows the case where the heart rate count is normally performed, and FIG. The case where it is not broken is shown.
As can be seen from comparison with FIG. 30 and FIG. 31, when the heart rate count is normally performed, the shape is substantially normal distribution, whereas when it is not normally performed, a plurality of (here 2 Shape).

  Therefore, the heart rate detection means 56 calculates the frequency distribution of the heart rate (S51) and checks whether or not the distribution has a substantially normal distribution shape (S52). If the distribution is a substantially normal distribution shape, it is determined that normal heart rate calculation has been performed, and it is determined that correction is unnecessary (S53). On the other hand, if the distribution deviates from the substantially normal distribution shape, it is determined that normal heart rate calculation has not been performed, and it is determined that correction is necessary (S54). The same method as in the first embodiment can be used to determine whether the heart rate distribution has a substantially normal distribution shape.

  When it is determined that correction is required by the above algorithm (S55), the heart rate detecting means 56 estimates the number N of peaks of the heart rate distribution (for example, by maximum likelihood estimation) (S56). Based on the number N, the heart rate calculated in step S34 (see FIG. 18) is corrected over the past certain period (S57). Specifically, when the number of peaks is N, N beats are counted again as one beat. Thus, accurate heart rate information can be obtained.

  As described above, when it is determined in step S54 that correction is necessary, correction is performed. Then, based on the corrected heartbeat information, the state determination of the autonomic nerve in step S36 and the sleep depth determination in steps S37 to S41 are performed.

  The computing device 50 outputs the determination result of the state of the autonomic nerve and the sleep depth determined as described above to an external device such as an air conditioner. On the device side that has received the determination result, device control according to the determination result is performed. A specific example of device control will be described in a later-described sixth embodiment.

  As described above, in the second embodiment, the norm of the IQ vector is calculated, and the heart rate is detected from the time series data of the norm. Therefore, the conventional high load processing such as frequency analysis is not required, and the heart rate can be calculated at high speed with low load. In addition, since a process with a high load is not required, the heart rate can be obtained using an inexpensive arithmetic device 50. In addition, the state of the autonomic nerve and the sleep depth of the living body that are closely correlated with the heart rate can be determined at high speed. Note that since the state of the autonomic nerve and the sleep depth are related to each other, the sleep depth may be determined based on the index indicating the state of the autonomic nerve. By estimating the activity of the autonomic nerve accompanying the heartbeat, it becomes possible to determine the sleep state according to the physiological model.

  Further, according to the second embodiment, it is possible to detect a count error that counts one heartbeat into a plurality of heartbeats, and when a count error is detected, the heart rate is corrected. Calculation accuracy can be improved. As a result, the determination accuracy of the state of the autonomic nerve and the sleep depth can be improved.

  Note that since the pulsation of the body surface due to the heartbeat is fast and complicated, the heartbeat signal may not be extracted by the method of the first embodiment. In other words, there are many points where the speed becomes zero in addition to the turning point of the heartbeat (the individual difference is also large). For this reason, it is difficult to detect the heartbeat with the method of the first embodiment, which is based on the fact that the speed becomes zero. On the other hand, in the method of the second embodiment, since the heartbeat is detected from the time series change of the signal intensity, the heartbeat can be easily detected.

  In the second embodiment, the case of detecting a heartbeat has been described. However, the present invention is not necessarily limited to a heartbeat, and can be detected by a similar method as long as changes such as respiration and pulse wave can be acquired as information by IQ detection. .

  Moreover, in this Embodiment 2, although the example which performs the state determination of an autonomic nerve and sleep depth determination based on both a heart rate and a heart rate fluctuation range was shown, either one of a heart rate and a heart rate fluctuation range is shown. The determination may be made based on the above. When both the heart rate and the heart rate fluctuation range are used, it is possible to determine with high accuracy by determining in a multidimensional space. Further, since the state of the autonomic nerve and the sleep depth are related to each other, the sleep depth may be determined based on the index indicating the state of the autonomic nerve.

Embodiment 3 FIG. (Body motion detection)
In the third embodiment, a biological state acquisition device that acquires body movements as the biological state will be described.

FIG. 32 is a block diagram showing a configuration of the biological state acquisition system according to Embodiment 3 of the present invention. In FIG. 32, the same parts as those of the second embodiment shown in FIG.
In the biological state acquisition system 300 according to the third embodiment, the biological state acquisition unit 52 includes a body motion detection unit 57, an autonomic nerve state determination unit 54B, and a sleep depth determination unit 55B, and other configurations are implemented. This is the same as the second embodiment. In the following, the third embodiment will be described focusing on the differences from the second embodiment.
The body motion detecting unit 57 detects body motion such as turning over from the locus on the IQ plane of the acquired signals (I signal and Q signal) acquired by the IQ signal acquiring unit 51.

  The body motion detection means 57 expresses a point on the IQ plane of the acquisition signal (I signal and Q signal) acquired by the IQ signal acquisition means 51 as a vector from the origin (hereinafter referred to as this vector (acquisition on the IQ plane). Signal position vector) is called IQ vector). When body movement occurs, the amount of displacement of the body surface movement is larger than that of heartbeat and respiration, and the displacement time is short. Therefore, the norm of the IQ vector at the moment when the body motion occurs becomes a large value. The body motion detection means 57 performs body motion detection using this feature. That is, the body motion detection means 57 calculates the norm of the IQ vector for each sampling or the square of the norm, and integrates these values obtained in a unit period (for example, 5 seconds) (calculates integration or summation). Then, threshold determination is performed on the integrated value, and the presence or absence of body movement in the unit period is determined. In addition, the body motion detecting unit 57 calculates the number of body motions within a certain period (for example, 8 minutes) based on the result of the presence / absence determination of body motion.

  The autonomic nerve state determination unit 54 </ b> B determines the state of the autonomic nerve based on the number of body movements calculated by the body motion detection unit 57.

  The sleep depth determination unit 55 </ b> B determines the sleep depth based on the number of body motions calculated by the body motion detection unit 57.

FIG. 33 is a flowchart showing the flow of the biological state acquisition process in the arithmetic device according to Embodiment 3 of the present invention.
(S61: I signal and Q signal acquisition)
The I signal and the Q signal output from the IQ detector 20 in the biological state acquisition system are signals in which respiration, heartbeat, and body motion are all mixed. For this reason, the I signal and Q signal output from the IQ detector 20 are passed through the high-frequency bandpass filter 32 for detecting heartbeat and body motion, thereby extracting the heartbeat signal and body motion signal. This signal is converted into a digital signal by the AD converter 40 and input to the IQ signal acquisition means 51. The pass frequency band of the band pass filter 32 is set in advance. In this way, the IQ signal acquisition means 51 acquires IQ signals (heartbeat and body motion signals) sequentially output in time series from the AD converter 40 according to the movement of the human body surface.

(S62: Norm integration)
The body motion detection means 57 calculates the norm (= square root of the sum of squares of I and Q) of the IQ vector of the acquired signal (signal in which the heartbeat signal and the body motion signal are mixed) for each sampling time, and the unit period The integrated value of the norm (for example, 5 seconds) is calculated.

(S63-66: calculation of the number of body movements during a certain period)
The body motion detection means 57 compares the norm integrated value of the unit period with a preset threshold value (S63), and determines that there is body motion if the integrated value of the norm of the unit period is equal to or greater than the threshold value. The number of occurrences is counted (S64). If the integrated value of the norm of the unit period is less than the threshold value, it is determined that there is no body movement (S65). The processes from step S31 to S34 are performed for a certain period (for example, 8 minutes) (S66).

(S67: Autonomic nerve state determination)
When the certain period has elapsed, the autonomic nerve state determination unit 54B calculates an index for determining the state of the autonomic nerve based on the number of body movements within the certain period counted in steps S63 to S66. This index may be the value of the number of body movements itself, or may be a value obtained by substituting each into a certain function. Here, the larger the sympathetic nerve is, the larger the value is. With this index, the state of the autonomic nerve can be determined. For example, the index is compared with a preset threshold. If the index is larger than the threshold, it is determined that the sympathetic nerve is dominant, and if the index is less than the threshold, it is determined that the parasympathetic nerve is dominant. In addition, for example, the degree of sympathetic nerve activity may be determined.

(S68-S74: Sleep depth determination)
Moreover, the sleep depth determination means 55B determines the sleep depth based on the number of body movements during a certain period. Next, the operation of the sleep depth determination means 55B will be described.

  It is known that aspects of human body movement during sleep change depending on the depth of sleep. Generally, the number of body movements during deep sleep and REM sleep is small, and the number of body movements increases as sleep becomes shallower, and the number of body movements increases most in the awake state. Therefore, the first body motion threshold, the second body motion threshold (<first body motion threshold), and the third body motion threshold (<first motion) for determining awakening, light sleep, REM sleep, and deep sleep, respectively. (2 body motion number threshold) is obtained and set in advance by experiments or the like, and the sleep depth is determined by comparison with each threshold.

  That is, if the number of body motions during a certain period is equal to or greater than the first body motion number threshold value, it is determined that the user is awake (S68, S69). If the number of body motions in a certain period is less than the first body motion number threshold and greater than or equal to the second body motion number threshold, it is determined that the sleep is shallow (S70, S71). If the number of body motions in a certain period is less than the second body motion number threshold value and greater than or equal to the third body motion number threshold value, it is determined as REM sleep (S72, S73). If the number of body motions during a certain period is less than the third body motion number threshold, it is determined that the sleep is deep (S72, S74).

  These threshold values may be set for each individual based on the sleep data obtained by providing a learning period and acquiring sleep data of one sleep cycle or more. Further, sleep data during the learning period may be analyzed based on a predetermined algorithm, and a threshold value may be determined and automatically set.

  The computing device 50 outputs the determination result of the state of the autonomic nerve and the sleep depth determined as described above to an external device such as an air conditioner. On the device side that has received the determination result, device control according to the determination result is performed. A specific example of device control will be described in a later-described sixth embodiment.

  As described above, in the third embodiment, the presence or absence of body movement is detected from the integrated value of the norm of the IQ vector. That is, the conventional high-load processing such as frequency analysis is not required, and it is possible to detect the presence / absence of body movement and calculate the number of body movements within a certain period of time with a low load. Further, since a process with a high load is not required, body motion information can be obtained using an inexpensive arithmetic device 50. In addition, the state of the autonomic nerve and the sleep depth that are deeply correlated with the body motion information can be determined at high speed. Note that since the state of the autonomic nerve and the sleep depth are related to each other, the sleep depth may be determined based on the index indicating the state of the autonomic nerve. By estimating the activity of the autonomic nerve accompanying the heartbeat, it becomes possible to determine the sleep state according to the physiological model.

Embodiment 4 FIG. (Combination of breathing, heart rate and body movement)
In the first to third embodiments, the example in which the biological state acquisition unit 52 includes the respiration detection unit 53, the heartbeat detection unit 56, and the body motion detection unit 57 has been described. The biological state acquisition unit 52 in the biological state acquisition system according to the fourth embodiment is configured to include all of these detection units.

FIG. 34 is a block diagram showing a configuration of a biological state acquisition system according to Embodiment 4 of the present invention. 34, the same reference numerals are given to the same portions as those in the block diagrams of the first to third embodiments shown in FIGS.
In the biological state acquisition system 400 according to the fourth embodiment, the biological state acquisition unit 52 includes a respiration detection unit 53, a heartbeat detection unit 56, and a body motion detection unit 57 similar to those in the first to third embodiments. Furthermore, the biological state acquisition means 52 obtains the detection results (respiration rate of human (living body), respiratory cycle fluctuation (respiration rate fluctuation), heart rate, heart rate fluctuation, body motion number) of each detection means 53, 56, 57. An autonomic nerve state determination unit 54C that determines an autonomic nerve state in an appropriate combination is provided. In addition, the biological state acquisition unit 52 obtains the detection results (respiration rate of human (living body), respiratory cycle variation (respiration rate variation), heart rate, heart rate variation, body motion number) of the detection units 53, 56, and 57. A sleep depth determination means 55C for determining a human sleep depth in an appropriate combination is provided. By combining these multiple detection results to determine the state of the autonomic nerve and the sleep depth, it is more accurate than the method of determining the state of the autonomic nerve and the sleep depth using only the heartbeat, respiration, and body movement. Judgment is possible. In the following, the fourth embodiment will be described focusing on the differences from the first to third embodiments.

  FIG. 35 is a diagram showing characteristics of body motion, respiration, and heart rate in each case of light sleep (including awakening), deep sleep, and REM sleep. It is the figure plotted to the series. For body motion, the IQ vector norm time-series data is a diagram showing the same characteristics as in FIG. Since light sleep and awakening have the same characteristics in each of body movement, breathing, and heart rate, they are combined into one item. The body movement at the time of awakening has a feature that the movement of the body movement is larger and the number of occurrences is larger than that in the case of shallow sleep, and is indicated by a dotted line in the figure.

As is clear from FIG. 35, the body movement is larger in the number of times of occurrence in the case of shallow sleep (including awakening) than in the case of deep sleep and REM sleep. For this reason, it can distinguish from the 1st state which is a shallow sleep, or the 2nd state which is either deep sleep or REM sleep by body motion information. In addition, as described in the third embodiment, it is possible to distinguish deep sleep and REM sleep only by the number of body movements, but since both the sleeping states have small movements and few occurrences. It is difficult to determine a sleep state with high accuracy. However, focusing on the respiration and heartbeat in FIG. 35, the respiration and heartbeat are characterized by being stable in deep sleep and unstable in REM sleep. Therefore, deep sleep and REM sleep can be distinguished by further using at least one of respiration and heartbeat.
The computing device 50 according to the fourth embodiment determines the sleep depth based on the above characteristics.

FIG. 36 is a flowchart showing the flow of the biological state acquisition process in the arithmetic device according to Embodiment 4 of the present invention.
(S81-S83: Acquisition of information on body movement, breathing, heart rate)
Each of the body motion detection unit 57, the respiration detection unit 53, and the heart rate detection unit 56 of the biological state acquisition unit 52 detects the body motion, respiration, and heart rate within a certain sleep depth determination period (for example, 8 minutes). The operations of the body motion detection unit 57, the respiration detection unit 53, and the heart rate detection unit 56 are the same as those in the above embodiment. The body motion detection unit 57 calculates the number of body motions (S81), and the respiration detection unit 53 The fluctuation range of the respiratory rate and the respiratory cycle is calculated (S82). The heart rate detecting means 56 calculates a heart rate and a heart rate fluctuation range (S83).

(S84: Autonomic state determination)
Then, the autonomic nerve state determination unit 54C determines the state of the autonomic nerve based on various calculation results related to body movement, respiration, and heartbeat. That is, an index for determining the state of the autonomic nerve is calculated using a combination of the number of body movement occurrences, the number of breaths, the fluctuation range of the respiratory cycle, the heart rate, and the fluctuation range of the heart rate. This index may be a value obtained by substituting each into some function, and takes a larger value as the sympathetic nerve is more dominant. With this index, the state of the autonomic nerve can be determined. For example, the index is compared with a preset threshold. If the index is larger than the threshold, it is determined that the sympathetic nerve is dominant, and if the index is less than the threshold, it is determined that the parasympathetic nerve is dominant. In addition, for example, the degree of sympathetic nerve activity may be determined.

(S85-S91: Sleep depth determination)
The sleep depth determination means 55C first determines whether the sleep depth is the first state of wakefulness or shallow sleep based on the number of body movements for a certain period calculated by the body motion detection means 57, whether REM sleep or deep sleep. The second state is distinguished. That is, the number of body motions is greater than or equal to the second body motion number threshold (see FIG. 33 of Embodiment 3), and the integrated value of the norm of the IQ vector of the body motion signal is greater than the preset first body motion integrated value threshold. It is determined whether or not is larger (S85). If this determination is YES, it is determined that the sleep depth is in the first state, and that it is either awakening or shallow sleep.

  In this case, subsequently, a determination is made to distinguish whether the sleep depth is awake or shallow sleep. That is, the second body motion integration in which the number of body motions during a certain period is equal to or greater than the first body motion number threshold (see FIG. 33 of the third embodiment) and the norm integrated value of the IQ vector of the body motion signal is preset. When the value is greater than or equal to the value threshold (> first body motion integrated value threshold) (S86), it is determined that the user is awake (S87). On the other hand, if the determination in step S87 is NO, it is determined to be light sleep (S88).

  If the determination in step S85 is NO and the second state is determined, then the sleep depth determination means 55C determines whether the sleep depth is REM sleep or deep sleep. That is, it is determined whether at least one of respiration and heartbeat is stable or unstable. If it is determined to be unstable, it is determined to be REM sleep (S90), and if it is determined to be stable, it is determined to be deep sleep (S91). The determination of stable or unstable may be made based on the fluctuation rate of the respiratory rate and the respiratory cycle, or the heart rate and the fluctuation rate of the heart rate.

  As described above, in the fourth embodiment, the state of the autonomic nerve and the sleep depth are determined by combining the calculation results of the respiration detection unit 53, the heartbeat detection unit 56, and the body motion detection unit 57. Thereby, it is possible to make a detailed determination with high accuracy as compared with the method of determining the state of the autonomic nerve and the sleep depth using only the heartbeat, respiration, and body movement. That is, it becomes possible to improve the determination accuracy of deep sleep and REM sleep, which is difficult to determine with high accuracy only by the calculation result of the body motion detection means 57 of the third embodiment. In the flowchart of FIG. 36, both respiration and heartbeat are detected, but at least one of them may be detected, and in step S89, it may be determined whether the detected side is stable or unstable.

By the way, the setting of each threshold value (first respiratory rate threshold value, second respiratory rate threshold value, first heart rate threshold value, etc.) used in the first to fourth embodiments described above is combined with measurement data such as an electroencephalogram, for example, It can also be obtained in this way. That is, each threshold value is obtained by the following method based on the determination result of the sleep depth based on the measurement data of another device such as an electroencephalogram.
Learn the sleep depth cycle from sleep onset to wake up with a learning period that measures sleep state for several days, and the learning data in that learning period matches the result of sleep depth discrimination based on the measurement data of other devices such as brain waves Each threshold is set as follows.

  In addition, the sleep depth determination means 55 provides a learning period to learn the sleep depth cycle from sleep onset to wake-up, and changes each threshold so that the sleep depth cycle in the learning period approaches the basic sleep depth cycle. May be. Specifically, for example, each threshold is adjusted so that the deep sleep time is 20 to 30% of the total time. Since the ratios of REM sleep, shallow sleep, and deep sleep change depending on the setting of the threshold, the threshold is determined so that the ratio of deep sleep becomes 20-30% of the whole while changing the threshold value little by little. It may be.

  Moreover, when determining the sleep depth only by body movement, when determining the threshold value for distinguishing between REM sleep and deep sleep which are difficult to distinguish only by body movement, the following may be set. For example, the threshold value is set so that the ratio of deep sleep is 30 to 40% (a value slightly higher than 20 to 30% of the total) with respect to the sum of deep sleep and REM sleep.

  In addition, as an example of changing each threshold so that the sleep depth cycle in the learning period approaches the basic sleep depth cycle, for example, it is also possible to set each threshold from the number of 90-minute sleep cycles repeated during sleep It is. The number of repetitions of the 90-minute sleep cycle is set to several times (for example, four times), and each threshold value is set so as to be the number of times. Moreover, you may determine from the ratio of deep sleep and the repetition frequency of a sleep cycle. Further, for example, when the tendency of the deep sleep ratio or the number of repetitions of the sleep cycle varies depending on the age, each threshold value may be set based on the tendency corresponding to the age.

Embodiment 5 FIG. (Autonomic nerve state and sleep depth determination based on the stability of the trajectory on the IQ plane)
In each of the above embodiments, the state of the autonomic nerve and the sleep depth are determined based on respiration, heartbeat, and body movement. However, in Embodiment 5, the state of the autonomic nerve directly from the locus drawn on the IQ plane and A biological state acquisition system for determining the sleep depth will be described.

FIG. 37 is a block diagram showing a configuration of the biological state acquisition system according to Embodiment 5 of the present invention. In FIG. 37, the same parts as those in the first embodiment shown in FIG.
In the biological state acquisition system 500 of the fifth embodiment, the biological state acquisition unit 52 includes a stability calculation unit 58, an autonomic state determination unit 54D, and a sleep depth determination unit 55D. Further, the pass frequency band of the band pass filter 30 </ b> A is set in advance as a band suitable for the stability calculation processing in the stability calculation means 58. Other configurations are the same as those of the first embodiment. The following description will focus on the differences of the fifth embodiment from the first embodiment.
The stability calculation means 58 calculates the stability of the trajectory on the IQ plane of the acquisition signals (I signal and Q signal) acquired by the IQ signal acquisition means 51. A method for calculating the stability will be described later. When the living body to be measured is in deep sleep, the muscle activity is stabilized, so that the acquired signal repeatedly draws substantially the same locus on the IQ plane as shown in FIG. On the other hand, in cases other than deep sleep, the acquired signal draws an unstable locus as shown in FIG. 39 on the IQ plane. Thus, the stability of the trajectory increases in the order of awakening, REM sleep, shallow sleep, and deep sleep, and is correlated with the sleep depth. Therefore, it is possible to determine the sleep depth from the stability of the trajectory.

Hereinafter, a specific example of the stability calculation method will be described.
(Stability calculation method 1)
The IQ plane is expressed by a quantized finite plane composed of M × M pixels. The number of pixels used in the trajectory drawn by the measurement data on the IQ plane within a certain period is counted, and this reciprocal is used as an index of the stability of the trajectory. The number of pixels used on the quantized IQ plane increases according to the degree to which the locus is unstable. Accordingly, the stability index becomes a smaller value as it becomes unstable, and this is used as the stability index.

(Stability calculation method 2)
The similarity between the trajectory drawn on the IQ plane for a certain period and the trajectory drawn immediately after this is calculated, and this is used as an index of the stability of the trajectory. The degree of similarity uses mutual information, correlation coefficient, and the like.

  The sleep depth determination unit 55D determines the sleep depth by determining the threshold value of the stability index of the trajectory calculated by the stability calculation unit 58. The threshold values (first stability threshold value, second stability threshold value, and third stability threshold value) used for the determination may be stored in advance in the memory in the arithmetic device 50, or may be automatically set for each individual. Good. In the case of automatic setting, sleep data of one sleep cycle or more is acquired by providing a learning period, the distribution of the stability of the trajectory is estimated from the sleep data, and a threshold is appropriately set based on the distribution.

FIG. 40 is a flowchart showing the flow of the biological state acquisition process in the arithmetic device according to Embodiment 5 of the present invention.
(S101: I signal and Q signal acquisition)
Since the I signal and the Q signal output from the IQ detector 20 in the biological state acquisition system 500 are signals in which respiration, heartbeat, and body motion are all mixed, the stability is calculated by passing the signal through the bandpass filter 30A. A signal suitable for is extracted. This signal is converted into a digital signal by the AD converter 40 and input to the IQ signal acquisition means 51. As described above, the IQ signal acquisition unit 51 acquires the IQ signals sequentially output in time series from the AD converter 40 in accordance with the movement of the human body surface.

(S102: Plot measurement data for a certain period on the IQ plane)
The stability calculation means 58 plots the acquired signals (I signal and Q signal) obtained at each sampling time within a certain period on the IQ plane.

(S103: locus stability calculation)
The stability calculation means 58 calculates the stability of the trajectory drawn by the acquired signal for each sampling time plotted on the IQ plane. The stability of the trajectory itself may be used as an index for determining the state of the autonomic nerve. The state of the autonomic nerve can be determined by this index. For example, the index is compared with a preset threshold. If the index is larger than the threshold, it is determined that the sympathetic nerve is dominant, and if the index is less than the threshold, it is determined that the parasympathetic nerve is dominant. In addition, for example, the degree of sympathetic nerve activity may be determined.

(S104-S110: Sleep depth determination)
The sleep depth determination means 55D includes the stability calculated by the stability calculation means 58, a preset first stability threshold, a second stability threshold (<first stability threshold), and a third stability threshold (< 2nd stability threshold) and a sleep depth is determined. That is, if the stability is equal to or higher than the first stability threshold, it is determined that the person is awake (S104, S105). ). Further, if the stability is less than the second stability threshold and not less than the third stability threshold, it is determined as shallow sleep (S108, S109), and if it is less than the third stability threshold, it is determined as deep sleep (S108, S110).

  The computing device 50 outputs the determination result of the sleep depth determined as described above to an external device such as an air conditioner. On the device side that has received the determination result, device control according to the determination result is performed. A specific example of device control will be described in a later-described sixth embodiment.

  As described above, in the fifth embodiment, the state of the autonomic nerve and the sleep depth are determined based on the stability of the trajectory drawn on the IQ plane by the acquired signal within a certain period. For this reason, frequency analysis etc. are unnecessary similarly to the case of the said Embodiment 1-4, and the determination of the state of an autonomic nerve and a sleep state can be performed at low load and high speed. In addition, the state of the autonomic nerve and the sleep depth can be determined directly and simply from the stability of the trajectory without calculating the heart rate or the respiratory rate. Therefore, even when the movement of the body surface is complicated and erroneous calculation of heart rate and respiration rate cannot be avoided, it is possible to determine the state of the autonomic nerve and the sleep depth without being affected by it.

  In addition to the method of determining the threshold value numerically based on the stability of the trajectory, the following method may be used. Of the trajectories drawn on the IQ plane, the trajectory pattern drawn on the IQ plane within a certain period is stored in advance in the storage device 60 for each sleep depth (or for each autonomic nerve state). Then, the sleep depth is determined by comparing the trajectory pattern held in the storage device 60 with the trajectory pattern based on the measurement data (IQ signal) and searching for the most similar one. Even when the trajectories of the same degree as the stability index are drawn numerically, the sleep state can be classified in more detail according to the shape of the trajectory by determining the pattern in this way. Moreover, if the locus pattern for collation is prepared for each age, sex, and further for each user, the sleep depth can be determined with higher accuracy. Similarly, the state of the autonomic nerves can be determined in advance, for example, by storing in advance a trajectory pattern corresponding to each state such as sympathetic nerve dominance and parasympathetic nerve dominance. Become.

  Further, by comparing the trajectory based on the acquired signal with the trajectory data corresponding to the sleep state of each user, it is possible to identify who the user is sleeping. For this reason, as will be described later in Embodiment 6, when the output (sleep depth) of the biological state acquisition device is used for controlling a device (for example, an air conditioner), device control tailored to the user is possible. It becomes.

  By the way, the sleep depth determination means of the living body state acquisition systems of Embodiments 1 to 5 described above can also determine a person's wake-up state (first awakening state). Generally, it is said that it is easy to wake up before and after REM sleep, and can wake up comfortably when getting up after REM sleep. For this reason, if a person is awakened by detecting a wake-up state and driving a device having an alarm function at that timing, for example, a comfortable wake-up can be provided.

Hereinafter, a method for determining the wake-up state will be briefly described.
A state in which at least one of three conditions is satisfied after a certain period of time from the start of falling asleep, after a REM sleep after repeating a predetermined number of sleep cycles, or when the REM sleep time is a certain time or more. Are determined to be in the wake-up state (the state of waking up first). During sleep from sleep onset to wake up, the REM sleep state is entered multiple times, but the duration of each REM sleep gradually increases toward wake up. Therefore, on the condition that the REM sleep time is a certain time or more, it can be determined that the REM sleep condition that satisfies the condition is in the wake-up state. Each condition for determining the wake-up state may be set based on a learning function.

  In addition, in the case of the apparatus that determines the sleep depth using the trajectory pattern on the IQ plane as in the fifth embodiment, it is possible to determine the wake-up state as follows. A learning period is provided, sleep data including at least from bedtime to waking up is collected, and a trajectory on the IQ plane that appears a certain period before waking up is acquired. Then, the trajectory data is held as wake-up sign data, and the timing at which measurement data that matches (similar) the wake-up sign data is obtained is determined as the wake-up state.

  As described above, the wake-up state obtained as described above can be used for controlling a device having a wake-up function, for example. Specifically, control for prompting awakening, such as gradually increasing the illuminance in the room or generating a sound, may be performed so that the user wakes up while in the wake-up state. Note that if control for forcibly waking up at the wake-up time is performed, a comfortable awakening cannot be obtained if the human sleep state is not a wake-up state. Therefore, control for prompting the user to wake up is performed when the user wakes up before the wake-up time. This makes it possible to wake up comfortably and to wake up before the set time, so that time can also be used effectively.

Embodiment 6 FIG.
The sixth embodiment relates to a device including the biological state acquisition system described in the first to fifth embodiments. In particular, an air conditioner will be described here.

FIG. 41 is a block diagram illustrating a configuration of an air conditioner including the biological state acquisition system according to Embodiments 1 to 5 of the present invention. FIG. 41 shows an example including the biological state acquisition system 100 of the first embodiment, but any biological state acquisition system of the second to fourth embodiments may be used. In FIG. 41, only the configuration of the main part related to the present invention is shown, and illustration of various components normally provided in the air conditioner is omitted.
The air conditioner 600 controls the living body state acquisition system 100, the air conditioning unit 610 that air-conditions the indoor space, and the air conditioner 610 based on the acquisition result of the living body state acquisition system and the entire air conditioner 600. 620. The arithmetic device 620 is constituted by a microcomputer, and includes a CPU, a ROM, and a RAM inside, and operates according to various programs stored in the ROM. In addition, you may make it give the function as a biological condition acquisition apparatus to the arithmetic device 620 by memorize | storeing the biological condition acquisition program of the arithmetic device 50 of the biological condition acquisition system 100 in the arithmetic device 620.

  The arithmetic device 620 controls the air conditioning unit 610 based on the sleep depth determination result acquired by the biological state acquisition unit 52, and performs air conditioning control for providing an energy-saving and comfortable sleep environment. It is assumed that a target temperature corresponding to the sleep depth is preset in the arithmetic device 620. This target temperature is set in consideration of energy saving and comfort.

FIG. 42 is a flowchart showing control of the air conditioner according to Embodiment 6 of the present invention.
The computing device 620 acquires the sleep depth determination result from the biological state acquisition system 100 (S121), and performs air conditioning control according to the determination result. That is, when the determination result is REM sleep or shallow sleep, the target temperature corresponding to the sleep depth is set (S122, S123). If the determination result is deep sleep, the target temperature is set for deep sleep (S122, S124), and if the user is getting up, the target temperature is set for getting up (S122, S125). Then, the arithmetic device 620 controls the air conditioning unit 610 so that the temperature of the indoor space becomes the target temperature (S126).

  As described above, according to the sixth embodiment, the air-conditioning means 610 is automatically controlled based on the determination result of the sleep depth acquired by the biological state acquisition system 100, so that an energy-saving and comfortable sleep environment is provided. Is possible.

  In addition to the alarm clock and the air conditioner as described above, the device having the biological state acquisition function is a device having a function of stimulating five senses such as lighting and aroma function, and an AV device such as a TV or a music player. It may be an air conditioner such as a hot water bottle, a humidifier, a dehumidifier, or an air purifier.

  Moreover, you may comprise the apparatus which performs the following controls as an apparatus which provides comfortable sleep environment. When falling asleep, sound and light are gradually lowered over a certain period of time (for example, 30 minutes) to induce a sleep rhythm. During sleep, the ambient noise is reduced by functions such as noise cancellation. Since the illuminance has the effect of controlling the biological rhythm, which is a necessary condition for getting up, the illuminance of light is increased over a certain time (for example, 30 minutes) before the wake-up time. The sound is adjusted to increase during the wake-up time and can be controlled to an unpleasant volume by setting. In addition, a human body detecting means is provided and adjusted so that light does not directly hit a human. The adjusting method may be any method such as lens wrinkles, actuators, and ON / OFF of a plurality of LEDs. In the human body detection, the antenna of the Doppler radar sensor 10 may be directed toward the person, and the LED may be directed to a position different from that direction.

  In the case of hot water bottles, the temperature is gradually lowered and raised again when getting up. It is also possible to warm to a level that does not sweat with a temperature and humidity sensor.

  In each of the above embodiments, the description has been made assuming that the sensing is always performed by the Doppler radar sensor 10. However, the Doppler radar sensor 10 is stopped at regular intervals to save power, and the sleep depth is determined. An interval may be provided. In this case, power consumption can be reduced as compared with the case where sensing is always performed. However, when the stop time is long, the sleep depth determination is hindered. Since there is almost no body movement in the state of deep sleep and it is hard to get up, the probability of shifting to the next sleep depth is low. For this reason, if the stop time is set longer in the deep sleep state and the sensing interval is made longer, it is possible to save power without hindering the determination of the sleep depth. Moreover, when deep sleep time continues more than predetermined time, you may make it the system of stopping a sensing temporarily. By shortening the sensing time, further energy saving can be performed. In shallow sleep, on the contrary, the sensing interval is shortened so that the sleep depth determination is not hindered. When such control is performed, the energization time is shortened, which is effective in extending the life.

  Further, since there is a low probability of waking up by external noise such as sound, light, and temperature change during deep sleep, control that causes noise as a device may be performed during deep sleep. For example, an air conditioner can perform a cleaning function, a wind direction change, a ventilation fan drive, and the like.

  10 Doppler radar sensor, 20 IQ detector, 30 band pass filter, 30A band pass filter, 31 band pass filter, 32 band pass filter, 40 AD converter, 50 arithmetic unit, 51 IQ signal acquisition means, 52 biological state acquisition means 53 Respiratory detection means 54 Autonomic state determination means 54A Autonomic state determination means 54B Autonomic state determination means 54C Autonomic state determination means 54D Autonomic state determination means 55 Sleep depth determination means 55A Sleep depth Determination means, 55B sleep depth determination means, 55C sleep depth determination means, 55D sleep depth determination means, 56 heart rate detection means, 57 body motion detection means, 58 stability calculation means, 60 storage device, 100 biological state acquisition system, 200 biological body Status acquisition system, 300 Biological condition acquisition system, 400 Biological condition acquisition system, 500 Biological condition acquisition system, 600 Air conditioner, 610 Air conditioning means, 620 Arithmetic unit.

Claims (12)

IQ signal acquisition that transmits electromagnetic waves to the body surface of a living body, IQ detection of the reflected waves, and sequentially acquires I and Q signals output from an IQ detector that outputs I and Q signals in time series Means,
Biological state acquisition means for acquiring the state of the biological body based on a locus on the IQ plane of the acquisition signal acquired by the IQ signal acquisition means;
The biological state acquisition means has a heartbeat detection means for detecting a heartbeat of the biological body,
The heartbeat detection means calculates a norm of a position vector of the acquired signal on the IQ plane, extracts a heartbeat signal by low-pass filtering the time series data of the vector norm,
Detecting a heartbeat signal corresponding to one heartbeat based on a periodic fluctuation of the waveform of the extracted heartbeat signal, calculating a heart rate in a unit period as heartbeat information,
The heart rate in the unit period is calculated over a certain period, the frequency distribution of the heart rate within the certain period is calculated, and the heart rate is calculated based on whether the frequency distribution has a substantially normal distribution shape. A biological state acquisition apparatus characterized by determining whether or not the count is normally performed.   2. The biological state acquisition according to claim 1, wherein the heart rate detection means calculates a heart rate in the unit period over a certain period, and calculates a fluctuation range of the heart rate within the certain period as heart rate information. apparatus.   The biological state acquisition unit includes an autonomic state determination unit that determines an autonomic state of the living body based on at least one of the heart rate information calculated by the heart rate detection unit. The biological state acquisition device according to claim 1 or 2.   The said biological state acquisition means is provided with the sleep depth determination means which determines the sleep depth of the said biological body based on at least 1 or more of the said heartbeat information calculated by the said heartbeat detection means. Item 3. The biological state acquisition device according to Item 2.   The heart rate detecting means determines that the normal heart rate is not counted when the frequency distribution of the heart rate is not a substantially normal distribution shape but a shape having two or more peaks. 5. The heart rate within the predetermined period in the past is corrected based on the number N of peaks, so that the N heart rate becomes 1 heart rate. 6. Biological state acquisition device.   The heart rate detecting means calculates a moment of the frequency distribution of the heart rate, and determines whether or not the frequency distribution has a substantially normal distribution shape based on the moment calculation result. The biological state acquisition apparatus as described in any one of Claims 1-5.   The biological state acquisition device according to claim 4, wherein the sleep depth determination means determines whether the sleep depth is REM sleep, shallow sleep, or deep sleep by threshold determination based on a preset threshold. .   The sleep depth determination means learns a sleep depth cycle from falling asleep to waking up by providing a learning period, and changes the threshold so that the sleep depth cycle in the learning period approaches a basic sleep depth cycle. The biological state acquisition device according to claim 7.   The sleep depth determination means includes at least one of three conditions when a predetermined time has elapsed from the start of falling asleep, after a REM sleep after repeating a predetermined number of sleep cycles, or when the REM sleep time is a certain time or more. The biological state acquisition device according to claim 7, wherein a state that satisfies the above conditions is determined to be a wake-up state that is awakening. Computer
The IQ signal acquisition means according to any one of claims 1 to 9,
The biological condition acquisition program for functioning as the said biological condition acquisition means as described in any one of Claims 1-9. The biological state acquisition device according to any one of claims 1 to 9,
An apparatus comprising a biological state acquisition device, comprising: an arithmetic unit that controls operation of the device main body based on the biological state acquired by the biological state acquisition device. The biological state acquisition device according to any one of claims 1 to 9,
Air-conditioning means for air-conditioning the indoor space;
An air conditioner comprising: an arithmetic unit that controls the air conditioning unit based on a biological state acquired by the biological state acquisition device.

Images