JP6011240B2 - Biological information acquisition method and biological information acquisition apparatus - Google Patents

Description

  The present invention relates to a biological information acquisition method and a biological information acquisition apparatus that acquires biological information from a sensor.

  In recent years, there are an increasing number of people who have problems related to sleep, such as insomnia and light sleep. If you are unable to get a good quality sleep, you will suffer from drowsiness in the daytime, which will interfere with your daily life. In view of this, there has been provided a device that provides good quality and good sleep by appropriately controlling the awakening light and the vibration pillow according to the sleep state of the person. In such an apparatus, biological information such as heart rate, respiratory rate, and body movement is detected by a sensor, and a sleep state is determined based on the detection result.

  Usually, a plurality of sensors are attached to bedding or the like so as not to disturb human sleep. A sensor (optimum sensor) that detects human biological information most accurately is selected from a plurality of sensors, and the signal of the sensor is acquired as the biological information at that time. Optimal sensors change during sleep due to body movement.

  Patent Document 1 discloses the following invention. A respiratory sound of the human body is detected by a plurality of sensors, and a frequency spectrum is obtained by frequency-converting the respiratory sound signal for each sensor. Extract the respiratory synchronization component that changes in synchronization with the breath from the spectrum, compare the frequency distribution of the respiratory synchronization component of each sensor, select the optimal sensor, and estimate the respiratory state of the human body from the respiratory synchronization component of the optimal sensor .

  Patent Document 2 discloses the following invention. A human body's respiration and heart rate are detected by a plurality of pressure sensors, and a signal of a pressure change from each sensor is evaluated by SQI (Signal Quality Index), and a sensor having a signal with the highest SQI is selected as an optimum sensor. After the sensor signal is wavelet transformed, it is separated into a respiratory rate detection signal and a heart rate detection signal to estimate the respiratory rate and heart rate.

JP 2007-14501 A JP 2007-61587 A

  However, any of the above-described conventional techniques calculates the frequency spectrum from the detection signal from the sensor in the process of selecting the optimum sensor. In order to calculate a frequency spectrum, it is necessary to convert detection signals in a predetermined time domain at a time. Therefore, before calculation, time for collecting signals corresponding to the number of time series points in the predetermined time domain (for example, 10 Second), a large-capacity memory for storing the signal is required. For this reason, there is a problem that time delay (time lag) occurs between detection signal output by a plurality of sensors and selection of the optimum sensor, and the optimum sensor cannot be selected in real time.

  In view of the above problems, an object of the present invention is to provide a method and apparatus for selecting a sensor capable of acquiring optimal biological information from a plurality of sensors in real time.

  In order to achieve the above object, the biological information acquisition method according to the present invention is characterized by the steps of detecting vibrations generated by a biological body with a plurality of sensors, and signals of frequency components of biological information to be acquired from output signals of the sensors. Only, extracting a signal whose total amplitude value is outside a predetermined range in the extracted signal as noise, generating a pulse signal corresponding to the noise, and a new evaluation value of the sensor, Calculating by adding the pulse signal to a value obtained by multiplying the previous evaluation value of the sensor by a predetermined constant; and selecting a sensor that acquires the biological information based on the new evaluation value. It is in point to include.

  In the above method, a frequency spectrum is not used to select an optimum sensor from a plurality of sensors. For this reason, the calculation of the frequency spectrum becomes unnecessary, and the memory for storing the sensor detection signal necessary for the calculation becomes unnecessary. Furthermore, there is no time lag from detection signal output by a plurality of sensors to selection of the optimum sensor, and it becomes possible to select the optimum sensor in real time and acquire biological information.

  In the biological information acquisition method according to the present invention, it is preferable that the sensor having the smallest new evaluation value is selected in the step of selecting a sensor that acquires the biological information. According to this method, a sensor having a small noise component is always selected as the optimum sensor from a plurality of sensors, so that accurate biological information can be acquired.

  In the biometric information acquisition method according to the present invention, it is preferable that the time interval for calculating the new evaluation value is longer than the time interval for detecting vibration by the sensor. There is usually a few seconds or more until the next body motion occurs after the body motion is settled during sleep, and the possibility that the optimum sensor is frequently changed is low. Therefore, by increasing the time interval for executing the process of calculating the evaluation value and selecting the optimum sensor with respect to the time interval (about 0.01 to 0.1 seconds) for detecting vibration with the sensor, Power consumption can be suppressed and an inexpensive microcomputer can be used.

  In the biological information acquisition method according to the present invention, it is preferable that the time interval for detecting vibration by the sensor and the time interval for calculating the new evaluation value can be changed. According to this method, it is possible to set a time interval for selecting an optimum sensor according to the sleep state. For example, the optimal sensor is unlikely to change for a short time when body movements do not occur frequently, such as during deep sleep, so the calculation load is increased by increasing the time interval for calculating the evaluation value and selecting the optimal sensor. Can be reduced.

  The characteristic configuration of the biological information acquisition device according to the present invention includes a plurality of sensors that detect vibrations generated by a living body and output signals corresponding to the vibrations, and extract only frequency components of biological information to be acquired from the signals. A filter processing unit that outputs an extraction signal, a signal having a total amplitude value outside the predetermined range in the extraction signal is determined as noise, a detection state determination unit that outputs a pulse signal corresponding to the noise, and a sensor A sensor that outputs a new evaluation value by adding an integrated value of the pulse signal to a value obtained by multiplying the previous evaluation value of the sensor by a predetermined constant, and acquires the biological information based on the new evaluation value. And a sensor evaluation / selection unit to be selected.

  In the above configuration, there is no means for calculating the frequency spectrum, and the optimum sensor can be selected from a plurality of sensors without calculating the frequency spectrum. Further, a memory for storing the sensor detection signal necessary for calculating the frequency spectrum is not necessary, and can be mounted on an inexpensive microcomputer. Furthermore, there is no time lag from detection signal output by a plurality of sensors to selection of the optimum sensor, and it becomes possible to select the optimum sensor in real time and acquire biological information.

In the biological information acquisition apparatus according to the present invention, it is preferable that the filter processing unit includes a plurality of low-pass filter processing units having different cutoff frequencies in parallel.
According to this configuration, the filter processing unit outputs a plurality of extraction signals obtained by extracting the frequency components of the biological information at different cutoff frequencies for each low-pass filter processing unit for each sensor. It is possible to select an optimum sensor according to the selection.

In the biological information acquisition apparatus according to the present invention, the filter processing unit includes in parallel a plurality of combinations of a bandpass filter processing unit and a lowpass filter processing unit that processes a signal output from the bandpass filter processing unit, It is preferable that the cut-off frequency of each of the band-pass filter processing units included in each combination is made different and the cut-off frequency of each of the low-pass filter processing units is made different.
According to this configuration, the filter processing unit extracts, for each sensor, the frequency component of the biological information with a different cutoff frequency for each bandpass filter processing unit and a different cutoff frequency for each lowpass filter processing unit. By outputting the extraction signal, it is possible to select an optimum sensor according to the physiological characteristics of the living body.

  In the biological information acquiring apparatus according to the present invention, it is preferable that the sensor evaluation / selection unit selects a sensor having the smallest new evaluation value. In this configuration, a sensor having a small noise component is always selected as the optimum sensor from a plurality of sensors, so that accurate biological information can be acquired.

  In the biological information acquisition apparatus according to the present invention, it is preferable that the time interval at which the sensor evaluation / selection unit outputs the new evaluation value is longer than the time interval at which the signal is output by the sensor. There is usually a few seconds or more until the next body motion occurs after the body motion is settled during sleep, and the possibility that the optimum sensor is frequently changed is low. Therefore, the time interval for executing the process of calculating the new evaluation value and selecting the optimum sensor is increased with respect to the time interval (about 0.01 to 0.1 second) for detecting vibration by the sensor. Thus, power consumption can be suppressed.

  In the biological information acquisition apparatus according to the present invention, it is preferable that the time interval for outputting a signal by the sensor and the time interval for outputting the new evaluation value by the sensor evaluation / selection unit can be changed. In this configuration, it is possible to set a time interval for selecting an optimum sensor according to the sleep state. For example, when there is almost no body movement such as during a deep sleep, the optimal sensor is unlikely to be changed many times in a short time, so the time interval for executing the process of calculating the new evaluation value and selecting the optimal sensor The calculation load can be reduced by lengthening.

  The biometric information acquisition apparatus according to the present invention is preferably provided in a bedding. By providing the bedding with the biometric information acquisition device, it becomes possible to accurately estimate the sleep state of a person based on the biometric information. As a result, it is possible to control the bedroom environment such as lighting, temperature, and humidity based on the estimation, and provide a good quality and good sleep to a person.

  The biometric information acquisition apparatus according to the present invention is preferably provided in a vehicle seat. For example, by providing the biometric information acquisition device in the backrest portion of the vehicle seat, it is possible to accurately estimate the state of the driver's body based on the biometric information. As a result, for example, it is possible to notify the driver of arousal, for example, by giving the driver a stimulus based on the assumption that the driver is experiencing strong sleepiness.

It is a perspective view which shows the bed with a biometric information acquisition function which concerns on the 1st Embodiment of this invention. It is a figure which shows the structure of a biosensor assembly. It is a block diagram of the whole system containing a biometric information acquisition apparatus. It is a figure which shows the processing flow for selecting the optimal piezoelectric sensor. It is explanatory drawing of the cutoff frequency corresponding to a high heart rate test subject. It is explanatory drawing of the cutoff frequency corresponding to a low heart rate test subject. It is a block diagram of a filter process part. It is a figure which shows the output signal waveform of a piezoelectric sensor. It is a figure which shows the waveform of an envelope signal. It is the graph which represented the RR space | interval measured about each of the envelope signal in time series. It is a graph which shows the time transition of the pulse output by an abnormality flag. It is the figure which expanded the envelope signal of the heartbeat waveform. It is a figure which shows the output signal waveform of the piezoelectric sensor corresponding to a high heart rate test subject. It is a figure which shows the waveform of the envelope signal corresponding to a high heart rate test subject. It is a figure which shows the output signal waveform of the piezoelectric sensor corresponding to a low heart rate test subject. It is a figure which shows the waveform of the envelope signal corresponding to a low heart rate test subject. It is a figure which shows the flow of a detection state determination. It is a figure which shows operation | movement of an incomplete integrator. It is a figure which shows the flow of evaluation of a piezoelectric sensor, and selection of an optimal piezoelectric sensor. It is a figure which shows the time transition of the evaluation value of each piezoelectric sensor, and the selected piezoelectric sensor. It is a figure which shows the detection rate according to test subject. It is a block diagram of the filter process part which concerns on the 2nd Embodiment of this invention. It is a perspective view which shows the vehicle seat which concerns on the 3rd Embodiment of this invention.

[First Embodiment]
Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a perspective view showing a bed 1 with a biological information acquisition device according to the present embodiment. The bed 1 with a biological information acquisition device includes a biological information acquisition device 10, a mattress 28, and a headboard 29 with illumination. The mattress 28 is placed on the bed frame of the bed 1 with the biological information acquisition device, and the biological information acquisition device 10 is installed between the bottom surface of the mattress 28 and the upper surface of the bed frame. And a signal processing unit 11 for processing a signal output from the biosensor assembly 18. The biosensor assembly 18 includes two sensor assembly sheets 19 and 20. The sensor assembly sheets 19 and 20 are installed so that the longitudinal direction thereof is parallel to the lateral direction of the mattress 28 and around the scapula and buttocks of a person (subject) lying on the bed 1 with the biological information acquisition device. Yes.

  FIG. 2 is an enlarged view of the biosensor assembly 18. As shown in FIG. 2, the sensor assembly sheets 19 and 20 have the same size, and have a rectangular shape with a longitudinal length of about 600 millimeters and a short side length of about 150 millimeters. Three piezoelectric sensors 19 a, 19 b, 19 c are built in the sensor assembly sheet 19, and three piezoelectric sensors 20 a, 20 b, 20 c are built in the sensor assembly sheet 20. These piezoelectric sensors 19a, 19b, 19c, 20a, 20b, and 20c (hereinafter, when six piezoelectric sensors are collectively expressed, they may be described as piezoelectric sensors 19a to 19c and 20a to 20c) Vibration is detected and a detection signal corresponding to the magnitude of the vibration is output. The piezoelectric sensors 19a to 19c and 20a to 20c are examples of sensors in the present invention.

  Processing of detection signals output from the piezoelectric sensors 19a to 19c and 20a to 20c will be described with reference to FIGS. FIG. 3 is a block diagram of the entire system including the biological information acquisition apparatus 10 according to the present embodiment. The biological information acquisition device 10 is a device that detects vibrations generated by a subject with the piezoelectric sensors 19a to 19c and 20a to 20c, and measures biological information such as a heartbeat of the subject therefrom to estimate a sleep depth. The signal processing unit 11 of the biological information acquisition device 10 includes a filter processing unit 12, a biological information measurement unit 13, a detection state determination unit 14, a sensor evaluation / selection unit 15, a biological information output unit 16, and a sleep depth estimation unit 17. ing. Hereinafter, a detailed description will be given of a processing flow for selecting a piezoelectric sensor (hereinafter referred to as an optimum piezoelectric sensor) that most accurately detects the heartbeat of a subject in the system and each block constituting the system.

  FIG. 4 is a processing flow for selecting the optimum piezoelectric sensor in the biological information acquiring apparatus 10 according to the present embodiment. As shown in this flow, the piezoelectric sensors 19a to 19c and 20a to 20c detect vibrations and output detection signals (S41), and extract frequency components corresponding to biological information such as the heartbeat of the subject from the detection signals. Filter processing is performed (S42). Thereafter, the heartbeat interval of the subject is measured by biological information measurement (S43), and at the same time, detection state determination is performed to determine whether or not a signal other than heartbeat is included in the detection signal (S44). Next, the piezoelectric sensors 19a to 19c and 20a to 20c are evaluated based on the detection state determination result, and the optimum piezoelectric sensor is selected (S45). Finally, the heartbeat interval signal (biological information) of the selected optimum piezoelectric sensor is output (S46), and the sleep depth of the subject is estimated (S47). Although not shown in FIG. 4, the external controller 23 controls the illumination 24, the air conditioner 25, the curtain 26, the vibration pillow 27, and the like based on the estimation result, and provides good quality and good sleep to the subject.

The filter processing unit 12 shown in FIG. 3 acquires the detection signals of the piezoelectric sensors 19a to 19c and 20a to 20c at an operation period of 0.01 seconds, performs the filtering process, and converts the detection information into biological information such as the heartbeat of the subject. The corresponding frequency component is extracted and an envelope signal of the heartbeat waveform is output.
In the filter processing unit 12, a pulsation frequency of a heartbeat is extracted by a bandpass filter processing unit (hereinafter referred to as a BPF processing unit) 50 that performs bandpass filter processing, and a rectification processing unit 51 that performs rectification processing and lowpass filter processing are performed. A low-pass filter processing unit (hereinafter referred to as an LPF processing unit) 52 acquires an envelope signal of a heartbeat waveform. One period of the envelope signal corresponds to one beat of the heartbeat.

As illustrated in FIG. 5, for example, when a low-pass filter process is performed at a low cutoff frequency fL for a high heart rate subject with a high (many) heart rate during sleep (an example of biological information), the basic component is blocked. Therefore, detection omission is likely to occur.
Conversely, as illustrated in FIG. 6, if the low-pass filter process is performed at a high cutoff frequency fH for a low heart rate subject with a low (less) heart rate during sleep, the influence of harmonics cannot be canceled out. Also in this case, detection omission is likely to occur.
Therefore, for heart rate detection, it is necessary to set an appropriate cutoff frequency according to the physiological characteristics of the subject (heart rate), but since the heart rate varies greatly depending on individual differences, a single low-pass filter is used. There are limits to dealing with it.

  Therefore, as shown in FIG. 7, the filter processing unit 12 in the present embodiment includes a plurality of (two in the present embodiment) LPF processing units 52 having different cutoff frequencies, that is, a high frequency range with a high heart rate. High-frequency LPF processing unit 52H with a high cut-off frequency specialized for detection of a low frequency, and a low-frequency LPF processing unit specialized for detection of a low-frequency with a low heart rate lower than the high-frequency LPF processing unit 52H 52L are provided in parallel with each other.

  And after performing the band pass filter process and rectification | straightening process by the same cutoff frequency about each of the detection signal of each piezoelectric sensor 19a-19c, 20a-20c, the low frequency LPF process part 52H from which a cutoff frequency differs mutually, and low A low-pass filter process is performed by the area LPF processing unit 52L.

FIG. 8A is a graph showing the time transition of the detection signals output from the piezoelectric sensors 19a to 19c and 20a to 20c. The horizontal axis represents time and the vertical axis represents the magnitude of vibration.
FIG. 8b shows the envelope signal of the heartbeat waveform after the detection signal of FIG. 8a is processed by the filter processing unit 12, and for each piezoelectric sensor, a high-frequency LPF processing unit 52H and a low-frequency LPF processing unit. Two envelope signals 19aH, 19aL to 19cH, 19cL, 20aH, 20aL to 20cH, and 20cL that are low-pass filtered with 52L are obtained.

  The signals indicated by 19aH to 19cH and 20aH to 20cH correspond to the envelope signal obtained by processing the detection signal output from the corresponding piezoelectric sensor by the high frequency LPF processing unit 52H, and are indicated by 19aL to 19cL and 20aL to 20cL. The signal corresponds to an envelope signal obtained by processing the detection signal output from the corresponding piezoelectric sensor by the low-frequency LPF processing unit 52L.

  FIG. 10a illustrates a detection signal (sensor waveform) output from one of the piezoelectric sensors corresponding to a high heart rate subject. FIG. 10b is obtained by processing the detection signal output corresponding to the high heart rate test subject by the high frequency LPF processing unit 52H and the low frequency LPF processing unit 52L. The envelope signal L is illustrated.

  As shown in FIGS. 10a and 10b, in the high heart rate test subject, the envelope signal L obtained by processing by the low-frequency LPF processing unit 52L is greatly attenuated, so that it is processed by the high-frequency LPF processing unit 52H. By using the obtained envelope signal H, it is possible to acquire an appropriate heart rate (biological information) according to the physiological characteristics of the subject with a high heart rate.

  FIG. 11a illustrates a detection signal (sensor waveform) output from one of the piezoelectric sensors corresponding to a low heart rate subject. FIG. 11 b is obtained by processing the detection signal output corresponding to the low heart rate test subject by the high-frequency LPF processing unit 52H and the low-frequency LPF processing unit 52L. The envelope signal L is illustrated.

  As shown in FIGS. 11a and 11b, in the low heart rate test subject, since the harmonic component remains in the envelope signal H obtained by processing by the high frequency LPF processing unit 52H, the low frequency LPF processing unit 52L By using the envelope signal L obtained by processing, it is possible to acquire an appropriate heart rate (biological information) according to the physiological characteristics of the subject having a low heart rate.

In the biological information measurement unit 13, the envelope signals 19aH, 19aL to 19cH output from the high-frequency LPF processing unit 52H and the low-frequency LPF processing unit 52L for the heartbeat waveforms of the piezoelectric sensors 19a to 19c and 20a to 20c, 19cL, 20aH, 20aL to 20cH, 20cL are detected to create a heartbeat pulse wave, and the time of the pulse interval is measured and output.
This pulse interval time is called RR interval, which is the heartbeat interval of the subject. Heart rate is the reciprocal of the RR interval. FIG. 8c is a graph showing the RR intervals measured for the envelope signals 19aH, 19aL to 19cH, 19cL, 20aH, 20aL to 20cH, and 20cL in time series.
The horizontal axis is time, and the vertical axis is the RR interval. The more stable the value on the vertical axis is, the more constant the heart rate is. The value on the vertical axis fluctuates indicates that the heart rate is changing. For example, the heart rate detected from the piezoelectric sensor 19c is stable, and the heart rate detected from the piezoelectric sensor 20c is fluctuating.

  The detection state determination unit 14 obtains a peak value that is a total amplitude value (Peak-to-Peak value) for each period of the envelope signal of the heartbeat waveform in FIG. 8B, and the signal of the period is obtained from the magnitude of the peak value. It is determined whether or not it is a heartbeat. If it is not a heartbeat, an abnormal flag is set and a pulse is output. FIG. 8d is a graph showing the time transition of the pulse output by the abnormality flag for each detection signal (envelope signal). In this graph, the location of the HI pulse is a location determined not to be a heartbeat.

これより、波高値が所定の範囲内の大きさであれば心拍であると判断され異常フラグのパルスは出力されず、その範囲以外の大きさであれば心拍ではないと判断され異常フラグのパルスが出力される。図９に図８ｂの一部を拡大したグラフを示す。これより、波高値の大きさによって検出状態の判断が異なることがわかる。 Table 1 shows the relationship between the magnitude of the peak value, the detection state, and the on / off state of the abnormality flag.

From this, if the peak value is a size within a predetermined range, it is determined that the pulse is a heartbeat, and an abnormal flag pulse is not output. Is output. FIG. 9 shows an enlarged graph of a part of FIG. 8b. From this, it can be seen that the determination of the detection state differs depending on the magnitude of the peak value.

FIG. 12 shows a flow of detection state determination in the detection state determination unit 14. This is a detailed description of S44 in FIG. The determination is made for each of the six piezoelectric sensors 19a to 19c and 20a to 20c. First, the piezoelectric sensor (CH = 1) for determining the state is selected (S71), the timing of the vertex is detected for each of the two envelope signals of the piezoelectric sensor (S72), and the peak value is calculated (S73).
That is, for each piezoelectric sensor (CH = 1 to 6), after the operation of S72 to S76 is performed on the envelope signal H (S71a) obtained by processing by the high-frequency LPF processing unit 52H, the low-frequency LPF processing unit The operations of S72 to S76 are executed for the envelope signal L (S71b) obtained by processing at 52L.

  Next, for each of the two envelope signals, it is determined whether or not the peak value is within a predetermined range (S74), and if it is within the predetermined range, an abnormality flag is not raised (S75), and the piezoelectric sensor is detected. The state determination ends. If the peak value is not within the predetermined range, an abnormality flag is set (S76), and the detection state determination of the piezoelectric sensor ends.

  Next, it is confirmed whether or not the detection state determination has been performed for all of the six piezoelectric sensors 19a to 19c and 20a to 20c (S77). If there is a piezoelectric sensor that has not performed the detection state determination, 1 is added to CHNo. Then (S78), the detection state determination of the next piezoelectric sensor is started. If the detection state determination of all six piezoelectric sensors has been performed, the detection state determination flow ends.

  The sensor evaluation / selection unit 15 calculates the evaluation value of each of the piezoelectric sensors 19a to 19c and 20a to 20c using the pulse output based on the determination made by the detection state determination unit 14, and detects the heartbeat of the subject most accurately. Select the selected piezoelectric sensor. First, an evaluation value calculation method for the piezoelectric sensors 19a to 19c and 20a to 20c will be described.

  Evaluation values of the piezoelectric sensors 19a to 19c and 20a to 20c are calculated by the incomplete integrator 21 shown in FIG. The incomplete integrator 21 is an integrator that operates so as to forget past information at a constant rate. This is because the detection state changes depending on the sleeping position or sleeping posture of the subject on the bed 1 with the biological information acquisition device, and this is to cope with this change.

FIG. 13 is a schematic diagram showing the operation of the incomplete integrator 21. The graph on the left side is the same as the graph showing the time transition of the pulse output by the abnormality flag for each detection signal (envelope signal) shown in FIG. When this pulse signal is input to the incomplete integrator 21, it is converted by a transfer function, and an output signal shown in the graph on the right side is obtained. This graph on the right side shows the time transition of the evaluation value of the piezoelectric sensor.
The graph of the evaluation value of the piezoelectric sensor in FIG. 13 and FIG. 15 described later illustrates six detection signals in consideration of the visibility of the figure, but in reality, it has 12 envelope signals. Evaluate the piezoelectric sensor.
At each time, the piezoelectric sensor having the smallest output signal, that is, the smallest evaluation value is selected as the optimum piezoelectric sensor. The piezoelectric sensor having the smallest value of the output signal indicates that the HI pulse is generated at the shortest time in the input signal of the calculation target time, and this is the sensor having the longest time to detect the subject's heartbeat. is there. It is clear from the applicant's knowledge that the longer the time at which the subject's heartbeat is detected, the more accurately the subject's heartbeat is detected, which will be described later.

で表される。αは帰還率であり、不完全積分器２１の設定パラメータである。αの値によって、過去情報を忘却する割合が定まる。帰還率１００％すなわちα＝１のときが通常の積分器である。 The transfer function of the incomplete integrator 21 is U (z) as input and Y (z) as output.

It is represented by α is a feedback rate and is a setting parameter of the incomplete integrator 21. The ratio of forgetting past information is determined by the value of α. The normal integrator is when the feedback rate is 100%, that is, α = 1.

となる。これより今回の評価値ｙ（ｋ）は、前回の評価値ｙ（ｋ−１）に帰還率αを乗じて一定割合忘却させて今回の入力値ｕ（ｋ）を加算した値になることがわかる。このように帰還率αと前回の評価値ｙ（ｋ−１）から今回の評価値ｙ（ｋ）が求められるので、周波数スペクトルを算出するときのような大容量のメモリは不要となる。 When the above equation (1) is expressed by a difference equation,

It becomes. As a result, the current evaluation value y (k) may be a value obtained by multiplying the previous evaluation value y (k−1) by the feedback rate α and forgetting a certain percentage and adding the current input value u (k). Recognize. As described above, since the current evaluation value y (k) is obtained from the feedback rate α and the previous evaluation value y (k−1), a large-capacity memory as in the case of calculating the frequency spectrum becomes unnecessary.

これを解くと、

となる。本発明の実施の形態でＴｃ＝１０秒、Ｔｐ＝０．０１秒と設定すると、（４）式よりα＝０．９９９を得る。 Next, the feedback rate α is obtained as follows. In the incomplete integrator 21, the time required for the influence of the current input to be reduced to 1 / e (= 36.8%) is equal to the time constant Tc of the incomplete integrator 21. Further, when the sampling period of the input signal of the incomplete integrator 21 is Tp, the following relationship is established.

Solving this,

It becomes. When Tc = 10 seconds and Tp = 0.01 seconds are set in the embodiment of the present invention, α = 0.999 is obtained from the equation (4).

FIG. 14 shows a flow of selecting an optimum piezoelectric sensor in the sensor evaluation / selection unit 15. This is a detailed description of S45 in FIG. The determination is performed for each of the two envelope signals obtained by processing the detection signals output from the six piezoelectric sensors 19a to 19c and 20a to 20c by the high frequency LPF processing unit 52H and the low frequency LPF processing unit 52L. First, the piezoelectric sensor for determining the state is selected (CH = 1), CHNo (BestCh) of the optimum piezoelectric sensor is set to 0, and the evaluation value (J_min) of the optimum piezoelectric sensor is set to the lower limit value (inf) (S91). .
That is, for each piezoelectric sensor (CH = 1 to 6), after performing the operation of S92 to S94 on the envelope signal H (S91a) obtained by processing by the high-frequency LPF processing unit 52H, the low-frequency LPF processing unit The operation of S92 to S94 is executed for the envelope signal L (S91b) obtained by processing at 52L.

  Next, the evaluation value (J) of the selected piezoelectric sensor is calculated as described above (S92), and the evaluation value (J) is compared with the evaluation value (J_min) of the optimum piezoelectric sensor (S93). If J ≧ J_min, BestCh and J_min are not changed. If J <J_min, BestCh is changed to CHNo of the piezoelectric sensor, and J_min is changed to the evaluation value (S94). Next, it is confirmed whether or not evaluation values have been calculated and evaluated for all six piezoelectric sensors (S95). If there is a piezoelectric sensor that has not yet been evaluated and evaluated, 1 is added to CHNo (S96). ), The evaluation value calculation and evaluation of the next piezoelectric sensor are started. If the evaluation values of all six piezoelectric sensors have been calculated and evaluated, the optimum piezoelectric sensor selection flow ends, and CHNo of the optimum piezoelectric sensor is output.

  The time required from the output of the detection signals of the piezoelectric sensors 19a to 19c, 20a to 20c to the output of CHNo of the optimum piezoelectric sensor in the embodiment of the present invention is about 0.2 seconds. From this, it can be seen that the embodiment of the present invention is a method that can select the optimum piezoelectric sensor in real time without a time lag as compared with the method of calculating the frequency spectrum and selecting the optimum piezoelectric sensor.

  FIG. 15 shows the time transition of the evaluation values of the piezoelectric sensors 19a to 19c and 20a to 20c. This is the same as the graph shown as a function of Y (Z) in FIG. The selected piezoelectric sensor is indicated by a bold line in FIG. As described above, the piezoelectric sensor having the lowest evaluation value is always selected.

  In the biological information output unit 16, information on the optimum piezoelectric sensor selected by the sensor evaluation / selection unit 15 and signals of RR intervals of the piezoelectric sensors 19 a to 19 c and 20 a to 20 c measured by the biological information measurement unit 13. Is inputted, and only the signal of the R-R interval of the optimum piezoelectric sensor is outputted. As described above, since the RR interval signals of all six piezoelectric sensors 19a to 19c and 20a to 20c are input to the biological information output unit 16, the optimum piezoelectric sensor selected by the sensor evaluation / selection unit 15 is changed. Even immediately, it becomes possible to output an RR interval signal of a new optimum piezoelectric sensor, and it is possible to immediately detect a change in the state of the subject (such as turning over) and acquire the biological information thereof.

  As described above, the heartbeat detected by the piezoelectric sensor having the smallest evaluation value accurately detects the heartbeat of the subject. In the following, the results proved by experiments will be described. In FIG. 16, the alphabet on the horizontal axis represents the subject, and the vertical axis represents the heart rate detection rate acquired by the biological information acquisition device 10. The detection rate refers to the measurement of the actual heart rate by an electrocardiograph at the same time as the acquisition of the heart rate by the biometric information acquisition device 10 for each subject, and the biometric information acquisition device 10 is correct when the difference between the two heart rates is within one beat. Based on the definition that heartbeats could be detected, this is the rate at which heartbeats were detected correctly.

  It shows that the higher the detection rate, the more accurately the heartbeat can be detected by the biological information acquisition apparatus 10. A subject with the same alphabet indicates that the same subject was tested on different days. FIG. 16 shows that the detection rate of each subject is high. Therefore, it can be said that the acquisition of the heartbeat by the biological information acquisition apparatus 10 reflects the actual heart rate.

  The biological information such as the subject's heartbeat is input by the sleep depth estimation unit 17 and the sleep depth of the subject is estimated based on the biological information. The sleep depth can be estimated in four stages, such as awakening, REM sleep, shallow non-REM sleep, and deep non-REM sleep. And the signal according to the estimation result of sleep depth is output to the external controller 23 or the estimation result display part 22, for example.

  The external controller 23 is in an optimal state for the lighting 24, the air conditioner 25, the curtain 26, the vibration pillow 27, etc. according to the sleep depth in order to provide a more comfortable and high quality sleep environment according to the sleep depth estimation result. Control to become.

  As described above, in the bed 1 with the biological information acquisition device according to the embodiment of the present invention, the biological information acquisition device 10 quickly outputs the accurate biological information of the subject in response to the change in the state of the subject. It becomes possible to provide the subject with a more comfortable and high quality sleep environment.

  In the present embodiment, the sampling period of the piezoelectric sensors 19a to 19c and 20a to 20c and the evaluation value calculation period in the sensor evaluation / selection unit 15 are both 0.01 seconds, but the movement of the subject's body during sleep is not much. Since it is not quick, the sampling period may be 0.1 seconds, for example. Further, the sampling period of the piezoelectric sensor and the evaluation value calculation period are not necessarily the same period, and the evaluation value calculation period may be set to a period of 0.1 seconds. Since it is considered that the movement of the subject's body is further reduced when deep non-REM sleep is estimated, the sampling period and the evaluation value calculation period may be changed according to the sleep depth, such as further increasing the evaluation value calculation period.

The sensor evaluation / selection unit 15 selects the piezoelectric sensor having the smallest evaluation value, but outputs information of five piezoelectric sensors other than the piezoelectric sensor having the largest evaluation value, and the RR interval of these piezoelectric sensors. You may make it output the signal which averaged the signal from the biometric information output part 16 as a test subject's heartbeat.
In addition, the filter process part 12 of the biometric information acquisition apparatus 10 in this embodiment may be provided with the 3 or more LPF process part from which a detection area differs or a part overlaps.

[Second Embodiment]
FIG. 17 shows a second embodiment of the present invention.
In the present embodiment, the filter processing unit 12 includes a plurality of combinations (two in the present embodiment) of the BPF processing unit 50 and the rectification processing unit 51 and the LPF processing unit 52 that process the signal output from the BPF processing unit 50. Are provided in parallel with each other.
The cut-off frequencies of the BPF processing units 50 included in each combination are made different from each other, and the cut-off frequencies of the LPF processing units 52 are made different from each other.

That is, the combination of the high frequency BPF processing unit 50H having a high cutoff frequency, the high frequency rectification processing unit 51H and the high frequency LPF processing unit 52H for processing the signal output from the high frequency BPF processing unit 50H, and the cutoff frequency A low-frequency BPF processing unit 50L, and a combination of a low-frequency rectification processing unit 51L and a low-frequency LPF processing unit 52L for processing a signal output from the low-frequency BPF processing unit 50L are provided in parallel.
Other configurations are the same as those of the first embodiment.
Note that the filter processing unit 12 of the biological information acquisition apparatus 10 according to the present embodiment may include a combination of three or more BPF processing units and LPF processing units having different detection regions or overlapping a part thereof.

[Third Embodiment]
Next, a third embodiment of the present invention will be described. FIG. 18 is a schematic view of the vehicle seat 2 showing the third embodiment. This vehicle is equipped with a biological information acquisition device 30. The signal processing unit 31 is mounted on the ECU 39, and the biosensor assembly 38 is built in the backrest portion of the vehicle seat 2.

  Since the configuration and operation of the biological information acquisition apparatus 30 are the same as those in the first or second embodiment, detailed description thereof is omitted. However, in the present embodiment, the biological information acquisition device 30 is mounted on a vehicle, and the purpose is to prevent a drowsy driving. Therefore, the final output signal is not a signal corresponding to the estimation result of the sleep depth, but, for example, a signal for estimating the arousal state, weak sleepiness, strong sleepiness, and sleep state (hereinafter referred to as an awakening degree estimation signal) is output. .

  Upon receiving the estimated signal of the awakening level output from the signal processing unit 31 of the biological information acquisition device 30, the controller 40 on the ECU 39 controls various devices so as to prompt the driver to awaken. For example, the sleep / wake device 41 that gives some stimulus to the driver is operated. In addition, the driver may be awakened by other methods such as blowing air from the air conditioner on the driver's face or increasing the volume of the in-vehicle acoustic device.

  According to the present invention, since the sensor that can acquire the most accurate biological information among the biological information detected from a plurality of sensors can be selected in real time, the sensor can be used for a device that performs various controls using the biological information that can be acquired correctly. be able to.

1: Bed with biological information acquisition device 2: Seat for vehicle 10, 30: Biological information acquisition device 11, 31: Signal processing unit 12: Filter processing unit 13: Biometric information measurement unit 14: Detection information determination unit 15: Sensor evaluation / Determination unit 16: Biological information output unit 19a, 19b, 19c: Piezoelectric sensor 20a, 20b, 20c: Piezoelectric sensor 50H, 50L: Band pass filter processing unit 52H, 52L: Low pass filter processing unit

Claims (12)

Detecting a vibration generated by a living body with a plurality of sensors;
Extracting only the signal of the frequency component of biological information to be acquired from the output signal of the sensor;
Determining a signal whose total amplitude value is outside a predetermined range in the extracted signal as noise, and generating a pulse signal corresponding to the noise;
Calculating a new evaluation value of the sensor by adding the pulse signal to a value obtained by multiplying the previous evaluation value of the sensor by a predetermined constant;
Selecting a sensor that acquires the biological information based on the new evaluation value.   The biological information acquisition method according to claim 1, wherein in the step of selecting a sensor that acquires the biological information, the sensor having the smallest new evaluation value is selected.   The biological information acquisition method according to claim 1, wherein a time interval for calculating the new evaluation value is longer than a time interval for detecting vibration by the sensor.   The biological information acquisition method according to claim 1, wherein a time interval for detecting vibration by the sensor and a time interval for calculating the new evaluation value can be changed. A plurality of sensors for detecting vibrations emitted by the living body and outputting signals corresponding to the vibrations;
A filter processing unit that extracts only the frequency component of biological information to be acquired from the signal and outputs an extracted signal;
A detection state determination unit that determines a signal whose total amplitude value is outside a predetermined range in the extracted signal as noise, and outputs a pulse signal corresponding to the noise;
A sensor that outputs the new evaluation value of the sensor by adding the pulse signal to a value obtained by multiplying the previous evaluation value of the sensor by a predetermined constant, and acquires the biological information based on the new evaluation value. A biological information acquisition apparatus comprising: a sensor evaluation / selection unit to select.   The biological information acquisition apparatus according to claim 5, wherein the filter processing unit includes a plurality of low-pass filter processing units having different cut-off frequencies in parallel.   The filter processing unit includes a plurality of combinations of a band-pass filter processing unit and a low-pass filter processing unit that processes a signal output from the band-pass filter processing unit in parallel, and the band-pass filter processing included in each combination 6. The biological information acquisition apparatus according to claim 5, wherein the cut-off frequency of each of the units is made different and the cut-off frequency of each of the low-pass filter processing units is made different.   The biometric information acquisition apparatus according to claim 5, wherein the sensor evaluation / selection unit selects a sensor having the smallest new evaluation value.   9. The time interval for outputting the new evaluation value by the sensor evaluation / selection unit is longer than the time interval for outputting a signal by the sensor. 9. Biological information acquisition device.   10. The time interval for outputting a signal by the sensor and the time interval for outputting the new evaluation value by the sensor evaluation / selection unit can be changed. 10. Biological information acquisition device.   A bedding comprising the biological information acquisition device according to any one of claims 5 to 10.   The vehicle seat provided with the said biometric information acquisition apparatus as described in any one of Claim 5 to 10.

Images