JP2013123524A - Wakefullness determining device, wakefullness determining method and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce decline of the accuracy in determining the wakefulness, caused by a noise component included in a detected heartbeat signal.SOLUTION: An acquisition part 11 detects a heartbeat signal representing a signal of the heartbeat of a subject from the subject, and acquires the cycle of the heartbeat signal as a heartbeat interval. An estimation part 12 estimates a model of the heartbeat interval of the subject. The estimation is executed by updating the model of the heartbeat interval, which is estimated based on the past acquired value of the heartbeat interval, based on an error between a new value of the heartbeat interval acquired by the acquisition part 11 and an estimated value of the newly acquired value of the model to reduce the error. A determination part 13 calculates a spectrum to the variation in the heartbeat interval based on the model of the heartbeat interval estimated by the estimation part 12, and determines the degree of wakefulness of the subject based on the spectrum.

Description

  The embodiments discussed herein relate to techniques for determining a subject's arousal level.

The technique using the heartbeat signal (pulse) of the subject, which is known as one of the techniques for measuring the degree of arousal and sleepiness of the subject, has less burden on the subject for acquiring the heartbeat signal.
For example, a technique of detecting a peak interval (beat interval) of a heartbeat signal of a subject, calculating a spectrum with respect to the fluctuation of the heartbeat interval, and determining the arousal level of the subject based on a frequency at which the calculated spectrum is maximized It has been known. This technique uses the fact that the heartbeat interval fluctuates due to the influence of the cardiovascular system, and the spectrum of the heartbeat interval changes depending on the state of the living body.

  Sympathetic nerves predominate during biological activities (at wakefulness), increase heart rate and increase myocardial contractility, but parasympathetic nerves predominate when the body is resting (such as when going to sleep), and heart rate Decrease myocardial contractility. The technique described above uses this fact to determine the degree of arousal of a subject by inferring a change in dominance between the sympathetic nerve and the parasympathetic nerve from the analysis result of fluctuations in the heartbeat interval of the subject. It is.

  Further, as a similar technique, a feature amount is extracted from the heartbeat of the subject, and the subject's sleepiness is calculated from the statistical distribution of the feature amount within a time width set based on the frequency of the rising peak of the frequency spectrum distribution of the feature amount. A technique for determining the degree is known.

  As another background art, in a biological information monitor that monitors changes in biological information, a technique is known in which a measurement operation mode being used on the monitor is switched to another measurement operation mode according to the measured biological information. ing.

  In addition, as another background art, when measurement of a body motion trajectory representing a temporal change in the body motion of a human body is selected, a period of taking the body motion detection data into the storage unit is set as the detection data. A technique is known in which the measurement of the exercise intensity calculated from the above is made shorter than when the measurement is selected.

International Publication No. 2008/065724 JP 2010-131061 A JP 2010-104430 A International Publication No. 2009/040949

In the above-described wakefulness determination technique using the analysis result of the fluctuation of the heartbeat interval, if the detected heartbeat signal includes a noise component, there is a concern that the determination accuracy of the wakefulness may be lowered.
In view of the above-described problems, the wakefulness determination device described later in this specification reduces a decrease in the determination accuracy of the wakefulness caused by the noise component included in the detected heartbeat signal.

  One of the wakefulness determination devices for determining the wakefulness of a subject, which will be described later in this specification, includes an acquisition unit, an estimation unit, and a determination unit. Here, the acquisition unit detects a heartbeat signal representing a heartbeat signal of the subject from the subject, and acquires a cycle of the heartbeat signal as a heartbeat interval. The estimation unit estimates a model of the heartbeat interval of the subject. This estimation is based on a model of a heartbeat interval estimated based on a past acquisition value of the heartbeat interval, and a new acquisition value of the heartbeat interval by the acquisition unit and a predicted value of the new acquisition value in the model. Based on the error, the update is performed so that the error decreases. Then, the determination unit calculates a spectrum for fluctuations in the heartbeat interval based on the model of the heartbeat interval estimated by the estimation unit, and determines the arousal level of the subject based on the spectrum.

  In addition, one of the arousal level determination methods for determining the arousal level of a subject, which will be described later in this specification, is to first detect a heart rate signal representing a subject's heart rate signal from the subject and determine the cycle of the heart rate signal as a heart rate. Get as interval. Next, the model of the subject's heart rate interval is estimated. This estimation is based on the error between the new acquired value of the heart rate interval and the predicted value of the new acquired value in the model, based on the model of the heart rate interval estimated based on the past acquired value of the heart rate interval. This is done by updating so that the error is reduced. Then, a spectrum for heartbeat interval fluctuation is calculated based on the model of the heartbeat interval estimated by this estimation, and the arousal level of the subject is determined based on the spectrum.

  One of the programs described later in this specification is a program for causing a computer to determine the degree of arousal of a subject. This program causes a computer to perform the following processing. In this process, first, a heartbeat signal representing a heartbeat signal of the subject is detected from the subject, and the period of the heartbeat signal is acquired as a heartbeat interval. Next, the model of the subject's heart rate interval is estimated. This estimation is based on the error between the new acquired value of the heart rate interval and the predicted value of the new acquired value in the model, based on the model of the heart rate interval estimated based on the past acquired value of the heart rate interval. This is done by updating so that the error is reduced. Then, a spectrum for heartbeat interval fluctuation is calculated based on the model of the heartbeat interval estimated by this estimation, and the arousal level of the subject is determined based on the spectrum.

  According to the arousal level determination device described later in this specification, there is an effect that a decrease in the accuracy of determination of the arousal level due to a noise component included in the detected heartbeat signal is reduced.

It is a functional block diagram illustrating the configuration of the arousal level determination device according to one embodiment. It is a hardware structural example of the arousal level determination apparatus of FIG. It is the flowchart which illustrated the processing content of the arousal degree determination process. It is the flowchart which illustrated the processing content of the heartbeat interval model estimation process. It is the flowchart which illustrated the processing content of the 1st example of a Kalman gain control process. It is an example of a control amount table. It is the flowchart which illustrated the processing content of the 2nd example of a Kalman gain control process.

First, FIG. 1 will be described. FIG. 1 is a functional block diagram illustrating the configuration of a wakefulness determination device according to one embodiment.
The arousal level determination device 10 in FIG. 1 is a device that determines the awakening level of a subject, and includes an acquisition unit 11, an estimation unit 12, and a determination unit 13.

  The acquisition unit 11 detects a heartbeat signal representing a heartbeat signal of the subject from the subject, and acquires a cycle of the heartbeat signal as a heartbeat interval. The heartbeat signal is detected from, for example, a potential difference between the electrodes when a voltage is applied to the electrodes brought into contact with the subject. Further, the heartbeat interval is acquired by, for example, calculating a cycle of the maximum value of the waveform of the heartbeat signal corresponding to one beat of the heartbeat (interval of the maximum value of the R wave in the electrocardiogram waveform).

  The estimation unit 12 estimates a model for the heartbeat interval of the subject using the heartbeat interval acquired by the acquisition unit 11. This estimation is performed by updating an estimation model of a heartbeat interval that has been estimated based on a past acquired value of the heartbeat interval. This update is performed so that the error is reduced based on the error between the new acquired value of the heartbeat interval by the acquiring unit 11 and the predicted value of the new acquired value in the estimated model.

The estimation unit 12 may estimate the model of the heartbeat interval by using, for example, a Kalman filter.
The determination unit 13 calculates a spectrum for fluctuations in the heartbeat interval of the subject based on the model of the heartbeat interval estimated by the estimation unit 12, and determines the arousal level of the subject based on the spectrum.

  According to the configuration described above, the estimation unit 12 performs estimation of the heartbeat interval of the subject as described above, thereby reducing the influence of noise components included in the heartbeat signal on the spectrum with respect to fluctuations in the heartbeat interval. Therefore, by determining the arousal level based on the spectrum with respect to the fluctuation of the heartbeat interval estimated by the estimation unit 12, a decrease in the accuracy of the arousal level determination due to the noise component is reduced.

In addition, as represented using a broken line in FIG. 1, the arousal level determination device 10 may further include a detection unit 14 and a control unit 15.
The detection unit 14 detects the amount of time fluctuation of the physiological state of the subject. The detection unit 14 detects, for example, a time fluctuation amount of the average heart rate of the subject as the time fluctuation amount of the physiological state of the subject. Or the detection part 14 detects the amount of time fluctuations of the acceleration of a test subject's body part, for example as the amount of time fluctuations of the physiological state of this test subject.

  The control part 15 controls the estimation part 12 when the detection result by the detection part 14 about the amount of time fluctuation of the subject's physiological state becomes a predetermined amount or more. This control is a control in which the amount of decrease in the error per update in updating the estimation model of the heartbeat interval performed by the estimation unit 12 is larger than that before the control.

  When the estimation unit 12 estimates a heart rate interval model by using a Kalman filter, the control unit 15 controls the Kalman gain in the Kalman filter to perform the above-described control. May be.

  For example, consider using this arousal level determination device 10 to detect that sleepiness has occurred in a driver who is driving a car. The physiological state of the driver who is the subject at this time may fluctuate rapidly depending on the actual driving situation. For example, when an obstacle suddenly appears on the road or when the road surface condition of the road suddenly deteriorates, the driver's workload increases and the physiological state fluctuates rapidly. On the other hand, the estimation unit 12 estimates the heart rate interval model by updating the heart rate interval estimation model based on the past acquired values of the heart rate interval. For this reason, immediately after such a sudden change occurs, the difference between the estimation result of the heart rate interval model by the estimation unit 12 and the actual arousal state model of the driver after the occurrence of the change becomes large. It is conceivable that the accuracy of the driver's sleepiness detection temporarily decreases. On the other hand, when the control unit 15 performs the above-described control, the responsiveness of the estimation of the heart rate interval model by the estimation unit 12 with respect to the rapid change in the physiological state of the driver as described above is improved. A reduction in the accuracy of the driver's sleepiness detection is suppressed.

Next, a hardware configuration example of the arousal level determination device 10 will be described with reference to FIG.
In the configuration example of FIG. 2, the arousal level determination device 10 includes a computer 20, a heart rate sensor 41, and an acceleration sensor 42.

  The computer 20 includes an MPU 21, ROM 22, RAM 23, hard disk device 24, input device 25, output device 26, interface device 27, and recording medium drive device 28. Note that these components are connected via a bus line 29, and various data can be exchanged under the management of the MPU 21.

An MPU (Micro Processing Unit) 21 is an arithmetic processing unit that controls the operation of the entire computer 20.
A ROM (Read Only Memory) 22 is a read-only semiconductor memory in which a predetermined basic control program is recorded in advance. The MPU 21 reads out and executes this basic control program when the computer 20 is activated, thereby enabling operation control of each component of the computer 20.

  A RAM (Random Access Memory) 23 is a semiconductor memory that can be written and read at any time and used as a working storage area as needed when the MPU 21 executes various control programs.

  The hard disk device 24 is a storage device that stores various control programs executed by the MPU 21 and various data. The MPU 21 can perform various control processes by reading and executing a predetermined control program stored in the hard disk device 24.

  The input device 25 is, for example, a touch panel switch or a keyboard device. When the input device 25 is operated by a subject who is awakened by the awakening level determination device 10 or the like, the input device 25 inputs various information from the user associated with the operation content. The acquired input information is sent to the MPU 21.

  The output device 26 is, for example, a liquid crystal display or a speaker, and displays various texts and images according to display data sent from the MPU 21 and various sounds according to sound generation data sent from the MPU 21.

  The interface device 27 is connected with a heart rate sensor 41 and an acceleration sensor 42 connected to the computer 20. The interface device 27 manages the exchange of various information with these devices.

  The recording medium driving device 28 is a device that reads various control programs and data recorded on the portable recording medium 30. The MPU 21 can read out and execute a predetermined control program recorded on the portable recording medium 30 via the recording medium driving device 28 to perform various control processes described later. As the portable recording medium 30, for example, a flash memory equipped with a CD-ROM (Compact Disc Read Only Memory), a DVD-ROM (Digital Versatile Disc Read Only Memory), or a USB (Universal Serial Bus) standard connector. and so on.

  In order to configure the arousal level determination device 10 using such a computer 20, for example, a control program for causing the MPU 21 to perform the awakening level determination process described below is created. The created control program is stored in advance in the hard disk device 24 or the portable recording medium 30. Then, a predetermined instruction is given to the MPU 21 to read and execute this control program. By doing so, it is possible to provide the functions of the acquisition unit 11, the estimation unit 12, the determination unit 13, the detection unit 14, and the control unit 15 of FIG.

  The heartbeat sensor 41 detects a heartbeat signal representing a subject's heartbeat signal from the subject. More specifically, for example, as described above, the heart rate sensor 41 detects a heart rate signal from a potential difference between the electrodes when a voltage is applied to the electrodes brought into contact with the subject. Instead of this, as the heart rate sensor 41, one that captures the heartbeat of the subject and outputs it as an electrical signal, such as an electrocardiograph, a pulse wave meter, or a heart sound sensor, may be used. The functions as the acquisition unit 11 and the detection unit 14 in FIG. 1 are provided by a combination of the heart rate sensor 41 and the control process performed by the computer 20.

The acceleration sensor 42 is mounted on the subject and detects acceleration caused by the body movement of the subject. A function as the detection unit 14 in FIG. 1 is provided by a combination of the acceleration sensor 42 and the control process performed by the computer 20.
The arousal level determination apparatus 10 of FIG. 2 is configured as described above.

Next, processing performed by the arousal level determination device 10 and its processing procedure will be described.
FIG. 3 is a flowchart illustrating the contents of the arousal level determination process performed by the arousal level determination device 10.

  The processing from S101 to S103 in FIG. 3 is processing performed by the acquisition unit 11, and detects a heartbeat signal representing a heartbeat signal of the subject from the subject and acquires the cycle of the heartbeat signal as a heartbeat interval. It is processing.

First, in S101, the acquisition unit 11 performs a process of detecting a heartbeat signal of a subject.
Next, in S102, the acquisition unit 11 performs a process of acquiring a heartbeat interval from the heartbeat signal acquired by the process of S101 and storing the obtained heartbeat interval. In this process, for example, the acquisition unit 11 detects a range in which the magnitude of the signal is equal to or greater than a predetermined threshold from the heartbeat signal acquired in the process of S101, and then the magnitude of the signal within the range. The peak position is detected as the position of the maximum value of the R wave in the electrocardiogram waveform. Then, the acquisition unit 11 measures the time between the detected peaks with a timer (for example, a timer provided in the MPU 21), and uses the measured result as observation value data of the heartbeat interval as a storage unit (for example, the RAM 23 in FIG. 2). ) In a predetermined storage area.

  Next, in S103, the acquisition unit 11 performs a process of determining whether or not new observation value data of the heartbeat interval has been acquired by the process of S102. Here, when the acquisition unit 11 determines that new observation value data of the heartbeat interval has been acquired (when the determination result is Yes), the processing proceeds to S104. On the other hand, when the acquisition unit 11 determines that new observation value data of the heartbeat interval cannot be acquired (the magnitude of the heartbeat signal detected in the latest processing of S101 was not the peak described above) (determination If the result is No), the process returns to S101 and the above-described process is repeated.

  Next, in S104, the estimation unit 12 performs the heartbeat interval model estimation process. The heartbeat interval model estimation process is a process of estimating a heartbeat interval model (in this embodiment, a heartbeat interval estimation data group) of the subject using the heartbeat interval observation value data acquired by the acquisition unit 11. Details of the processing will be described later.

  The processing from S105 to S109 is processing performed by the determination unit 13, and calculates a spectrum for heartbeat interval variation based on the heartbeat interval model estimated by the estimation unit 12, and the subject is based on the spectrum. This is a process for determining the degree of awakening. In addition, the determination of the arousal level is such that the frequency of the maximum (maximum within a predetermined frequency range) spectrum decreases and the maximum spectrum increases (the spectrum is concentrated) as the subject shifts from the awake state to the sleep state. To do). This determination method is also used in the technique described in Patent Document 1 described above, for example.

  First, in S105, a fast Fourier transform (FFT) is performed using a predetermined number of estimated values obtained most recently among the estimated values of the heartbeat interval obtained by the heartbeat interval model estimating process in S104. The determination unit 13 performs a process of calculating a spectrum for fluctuations in the heartbeat interval.

In S106, the determination part 13 performs the process which specifies the largest spectrum within a predetermined frequency range from the spectrum obtained by the process of S105.
In S107, the determination unit 13 performs a process of determining whether or not the frequency of the spectrum specified by the process of S106 is equal to or lower than a predetermined reference frequency. Here, when the determination unit 13 determines that the frequency of the identified spectrum is equal to or lower than the reference frequency (when the determination result is Yes), the determination unit 13 proceeds to S108. On the other hand, when the determination unit 13 determines that the frequency of the specified spectrum is higher than the reference frequency (when the determination result is No), the test subject determines that the subject is in an awake state at this time, The processing is returned to S101 and the above-described processing is repeated.

  In S108, the determination part 13 performs the process which determines whether the magnitude | size of the spectrum specified by the process of S106 is more than a predetermined reference value. Here, when the determination unit 13 determines that the specified spectrum size is greater than or equal to the reference value (when the determination result is Yes), the determination unit 13 proceeds with the process to S109. On the other hand, when the determination unit 13 determines that the size of the identified spectrum is smaller than the reference value (when the determination result is No), the determination unit 13 determines that the subject is in an awake state at this time. , The process is returned to S101 and the above-described process is repeated.

  When the determination results of the determination processes of S107 and S108 described above are both Yes, the determination unit 13 determines that the user is not in an awake state (ie, feels sleepy) and executes the process of S110. To do. In S109, for example, the determination unit 13 performs a process of controlling the output device 26 of FIG. 2 to output an alarm indicating that sleepiness has been detected (for example, causing the speaker to emit an alarm sound). Thereafter, the arousal level determination process ends.

Next, details of the heartbeat interval model estimation process, which is the process of S104 in FIG. 3, will be described. FIG. 4 is a flowchart illustrating the processing contents of the heartbeat interval model estimation processing.
As described above, the heartbeat interval model estimation process estimates a heartbeat interval model (in this embodiment, a heartbeat interval estimation data group) of the subject using the heartbeat interval observation value data acquired by the acquisition unit 11. It is processing to do. This estimation is based on an error between an estimated value of a heartbeat interval estimated based on a past acquired value of the heartbeat interval and a predicted value of the new acquired value in the estimated model. This is done by updating so that the error is reduced. More specifically, in this heartbeat interval model estimation process, a heartbeat interval model is estimated by using a Kalman filter.

  When the process of FIG. 4 is started, first, in S201, the estimation unit 12 performs a process of determining whether or not the heartbeat interval model estimation process is executed for the first time. Here, when the estimation unit 12 determines that the execution of this estimation process is the first time (when the determination result is Yes), the process proceeds to S202, and when it is determined that the execution of this estimation process is not the first time If the determination result is No, the process proceeds to S203.

  In S202, the estimation unit 12 sets the initial value of the estimation model of the heartbeat interval and the initial value of the prior estimation error covariance in the Kalman filter and stores them in a predetermined storage area of the storage unit (for example, the RAM 23 in FIG. 2). Do. Here, the prior estimation error covariance is the difference between the observation value data group of the heartbeat interval actually acquired by the acquisition unit 11 and the heartbeat interval estimation model (preliminary estimation value) based on the past acquisition value of the heartbeat interval. This is the error covariance. Note that, as these initial values, for example, an average of these values obtained in advance from measured values of heartbeat intervals for a large number of subjects is set in advance.

  In S203, the estimation unit 12 performs a process of calculating a Kalman gain in the Kalman filter using the prior estimation error covariance stored in the storage unit at the time of this process. In the calculation of the Kalman gain, the covariance of the observation error of the heartbeat interval by the acquisition unit 11 is assumed to be a normal distribution with an average of “0” in the present embodiment.

  Next, in S204, the Kalman gain control process is performed by the control unit 15 or the like. The Kalman gain control process is a process of detecting the time variation amount of the physiological state of the subject and controlling the Kalman gain value calculated by the process of S203 when the time variation amount is equal to or greater than a predetermined amount. Details of the processing will be described later.

  Next, in S205, the estimation unit 12 performs a process of updating the heartbeat interval estimation model and storing it in a predetermined storage area of the storage unit (for example, the RAM 23 in FIG. 2). In this process, the latest observed value data of the heartbeat interval acquired most recently by the process of S102 of FIG. 3, the estimation model of the heartbeat interval stored by the process of S202 or S207 described later, and the processes of S203 and S204 The later Kalman gain is used. In the process of S205, the estimation unit 12 first subtracts the prediction value for the latest observation value data in the estimation model (preliminary estimation value) of the heartbeat interval from the latest observation value data, thereby calculating a prediction error. Ask. Next, the estimation unit 12 multiplies the prediction error by the Kalman gain, and adds the multiplication result to the estimation model (preliminary estimation value) of the heartbeat interval. ). By this calculation performed using the Kalman gain, the estimation model of the heartbeat interval is updated so as to reduce the prediction error. The new estimation model of the heartbeat interval obtained by the processing of S205 is used as the estimation data group of the heartbeat interval in the processing after S105 in FIG.

  Next, in S206, the estimation unit 12 performs a process of updating the pre-estimation error covariance stored in the storage unit at the time of this process to the post-estimation error covariance. Here, the posterior estimation error covariance is an error between the observation value data group of the heartbeat interval actually acquired by the acquisition unit 11 and the new estimation model (post posterior estimation value) of the heartbeat interval acquired by the processing of S205. Is the covariance of.

  Next, in S207, the next observed value data of the heartbeat interval is predicted from the new estimated model (post-mortem estimated value) of the heartbeat interval obtained by the processing of S204, and the heartbeat when the observed value data is obtained. The estimation unit 12 performs a process of calculating a new estimation model (preliminary estimation value) of the interval. The estimation unit 12 also performs a process of storing the calculated estimation model in a predetermined storage area of a storage unit (for example, the RAM 23 in FIG. 2).

  Next, in S208, a process of calculating the prior estimation error covariance from the posterior estimation error covariance stored in the storage unit at the time of this processing and storing it in a predetermined storage area of the storage unit (eg, RAM 23 in FIG. 2). Is performed by the estimation unit 12. This pre-estimation error covariance is a covariance of errors between the observation value data group of the heartbeat interval actually acquired by the acquisition unit 11 and the new estimation model of the heartbeat interval calculated by the process of S207. This prior estimation error covariance is used when the Kalman gain calculation process (S203) is subsequently executed.

  After completion of the process of S208 described above, the heartbeat interval model estimation process of FIG. 4 ends, and thereafter the process returns to FIG.

  Next, the Kalman gain control process that is the process of S204 of FIG. 4 will be described. As described above, the Kalman gain control process detects the amount of time variation of the subject's physiological state, and when the amount of time variation exceeds a predetermined amount, the value of the Kalman gain calculated by the process of S203. It is a process to control.

  First, a first example of the Kalman gain control process will be described. In this first example, the time fluctuation amount of the average heart rate of the subject is detected as the time fluctuation amount of the physiological state of the subject, and when the time fluctuation amount exceeds a predetermined amount, the value of the Kalman gain It is a process to control.

FIG. 5 will be described. FIG. 5 is a flowchart illustrating the contents of the first example of the Kalman gain control process.
First, in S301, a predetermined number (for example, data for the last two minutes) of the observed value data of the heartbeat interval stored by repeatedly executing the processing of S101 and S102 in FIG. The detection unit 14 performs a process of calculating an average value of the intervals.

  Next, in S302, whether or not the average heart rate calculated in the process of S301 has changed by a predetermined ratio (for example, 10%) or more from the calculated value of the average heart rate in the previously executed Kalman gain control process. The control part 15 performs the process which determines. Here, when it is determined that the average heart rate has changed by a predetermined rate or more (when the determination result is Yes), the control unit 15 advances the process to S303. On the other hand, when the control unit 15 determines that the change in the average heart rate is less than the predetermined ratio (when the determination result is No), the control unit 15 advances the process to S305. When the Kalman gain control process is executed for the first time, the result of the determination process in S302 may be Yes.

  In S303, the control unit 15 performs a process of reading out and acquiring the control amount associated with the above-described change rate of the average heart rate from the control amount table.

FIG. 6 shows an example of a control amount table in which the rate of change in average heart rate is associated with the control amount. Note that the value of the control amount is a value larger than 1.
Note that the control unit 15 may calculate the control amount from the rate of change in the average heart rate using a predetermined calculation formula instead of acquiring the control amount from the control amount table.

  Next, in S304, the control unit 15 performs a process of increasing the value of the Kalman gain by multiplying the value of the Kalman gain calculated by the process of S203 by the control amount acquired by the process of S303. When the value of the Kalman gain is increased in this way, the amount of decrease in the prediction error per update in the heartbeat interval estimation model update process of S205 in FIG. 4 performed after this process increases. As a result, the responsiveness of estimation of the model of the heart rate interval by the estimation unit 12 with respect to a sudden change in the average heart rate of the subject is improved.

  Next, in S305, the detection unit 14 performs a process of storing the calculated value of the average heart rate calculated in the process of S301 in a predetermined storage area of a storage unit (for example, the RAM 23 in FIG. 2). The storage of the calculated value of the average heart rate is used in the determination process of S302 when the Kalman gain control process is executed again thereafter.

  After completion of the process of S305 described above, the Kalman gain control process of FIG. 5 ends, and thereafter the process returns to FIG.

  Next, a second example of the Kalman gain control process will be described. In this second example, the time fluctuation amount of the acceleration of the subject's body part is detected as the time fluctuation amount of the physiological state of the subject, and when the time fluctuation amount exceeds a predetermined amount, the Kalman gain This is a process for controlling the value.

FIG. 7 will be described. FIG. 7 is a flowchart illustrating the contents of a second example of the Kalman gain control process.
First, in S351, acceleration generated by the body movement of the subject is detected using the acceleration sensor 42 of FIG. 2, and an average value (for example, data for the latest two minutes) of acceleration data representing the acceleration is detected ( The detection unit 14 performs a process of calculating an average speed.

  Next, in S352, it is determined whether or not the average acceleration calculated in the process of S351 has changed by a predetermined ratio (for example, 10%) or more from the calculated value of the average acceleration in the previously executed Kalman gain control process. The control part 15 performs the process to perform. Here, when it is determined that the average acceleration has changed by a predetermined ratio or more (when the determination result is Yes), the control unit 15 advances the process to S353. On the other hand, when the control unit 15 determines that the change in the average acceleration is less than the predetermined ratio (when the determination result is No), the control unit 15 advances the process to S355. When the Kalman gain control process is executed for the first time, the result of the determination process in S352 may be Yes.

  In S353, the control unit 15 performs a process of reading out and acquiring the control amount associated with the above-described change rate of the average acceleration from the control amount table. This control amount table may be the same as that illustrated in FIG.

  Note that the control unit 15 may calculate the control amount from the rate of change in the average acceleration using a predetermined calculation formula instead of acquiring the control amount from the control amount table.

  Next, in S354, the control unit 15 performs a process of increasing the value of the Kalman gain by multiplying the value of the Kalman gain calculated by the process of S203 by the control amount acquired by the process of S353. When the value of the Kalman gain is increased in this way, the amount of decrease in the prediction error per update in the heartbeat interval estimation model update process of S205 in FIG. 4 performed after this process increases. As a result, the responsiveness of estimation of the model of the heart rate interval by the estimation unit 12 with respect to a sudden change in the average heart rate of the subject is improved.

  Next, in S355, the detection unit 14 performs a process of storing the calculated value of the average acceleration calculated in the process of S351 in a predetermined storage area of the storage unit (for example, the RAM 23 in FIG. 2). The storage of the calculated value of the average acceleration is used in the determination process in S352 when the Kalman gain control process is executed again thereafter.

  After completion of the process of S355 described above, the Kalman gain control process of FIG. 7 ends, and thereafter the process returns to FIG.

  By performing each of the above-described processes in the above procedure in the awakening level determination device 10, the determination of the awakening level of the subject is performed, and a predetermined warning is issued when it is detected that sleepiness has occurred.

  In addition, the procedure of the process described above performed in the arousal level determination device 10 is not limited to the above-described procedure.

DESCRIPTION OF SYMBOLS 10 Arousal degree determination apparatus 11 Acquisition part 12 Estimation part 13 Determination part 14 Detection part 15 Control part 20 Computer 21 MPU
22 ROM
23 RAM
24 hard disk device 25 input device 26 output device 27 interface device 28 recording medium drive device 29 bus line 30 portable recording medium 41 heart rate sensor 42 acceleration sensor

Claims (18)

A wakefulness determination device for determining a subject's wakefulness,
An acquisition unit for detecting a heartbeat signal representing a heartbeat signal of the subject from the subject and acquiring a cycle of the heartbeat signal as a heartbeat interval;
A model for the heartbeat interval estimated based on a past acquisition value of the heartbeat interval is obtained by calculating a new acquisition value of the heartbeat interval by the acquisition unit and a prediction value for the new acquisition value in the model. An estimation unit that estimates the heartbeat interval model by updating the error based on the error, and a spectrum for the fluctuation of the heartbeat interval based on the heartbeat interval model estimated by the estimation unit And a determination unit that determines the arousal level based on the spectrum,
A wakefulness determination device comprising: A detection unit that detects a temporal variation amount of the physiological state of the subject, and an update in updating the heart rate interval model by performing control of the estimation unit when the temporal variation amount of the physiological state becomes a predetermined amount or more. A control unit for increasing the amount of decrease in the error per operation larger than before the control;
The wakefulness determination device according to claim 1, further comprising: The estimation unit performs estimation of the model of the heartbeat interval by using a Kalman filter,
The control unit controls a Kalman gain in the Kalman filter to increase the amount of decrease in the error per update in updating the heart rate interval model more than before the control.
The arousal level determination apparatus according to claim 2.   The wakefulness determination device according to claim 2, wherein the detection unit detects a time fluctuation amount of the average heart rate of the subject as a time fluctuation amount of the physiological state of the subject.   The wakefulness determination device according to claim 2, wherein the detection unit detects a time variation amount of acceleration of the body part of the subject as a time variation amount of the physiological state of the subject.   The wakefulness determination apparatus according to claim 1, wherein the estimation unit estimates the model of the heartbeat interval by using a Kalman filter. A method for determining a degree of arousal for determining a degree of arousal of a subject,
Detecting a heartbeat signal representing a signal of the heartbeat of the subject from the subject, and obtaining a cycle of the heartbeat signal as a heartbeat interval;
Based on the error between the new acquired value of the heartbeat interval and the predicted value of the new acquired value in the model, the model for the heartbeat interval estimated based on the past acquired value of the heartbeat interval, By updating the error to reduce the model of the heartbeat interval,
Calculating a spectrum for fluctuation of the heartbeat interval based on a model of the heartbeat interval estimated by the estimation, and determining the arousal level based on the spectrum;
A method of determining arousal level, characterized by: The wakefulness determination method according to claim 7, further comprising:
Detecting the amount of time variation of the physiological state of the subject,
The estimation control is performed when the amount of time fluctuation of the physiological state exceeds a predetermined amount, and the amount of decrease in the error per update in updating the heart rate interval model is larger than before the control. To
A method of determining arousal level, characterized by: The estimation of the model of the heartbeat interval is performed using a Kalman filter,
In the estimation control, the Kalman gain in the Kalman filter is controlled so that the amount of decrease in the error per update in updating the heart rate interval model is larger than before the control.
The arousal level determination method according to claim 8.   The arousal level determination method according to claim 8 or 9, wherein a time fluctuation amount of the average heart rate of the subject is detected as the time fluctuation quantity of the physiological state of the subject.   The awakening degree determination method according to claim 8 or 9, wherein a temporal fluctuation amount of acceleration of the body part of the subject is detected as a temporal fluctuation quantity of the physiological state of the subject.   The wakefulness determination method according to claim 7, wherein the estimation of the model of the heartbeat interval is performed using a Kalman filter. A program for causing a computer to determine the degree of arousal of a subject,
Detecting a heartbeat signal representing a signal of the heartbeat of the subject from the subject, and obtaining a cycle of the heartbeat signal as a heartbeat interval;
Based on the error between the new acquired value of the heartbeat interval and the predicted value of the new acquired value in the model, the model for the heartbeat interval estimated based on the past acquired value of the heartbeat interval, By updating the error to reduce the model of the heartbeat interval,
Calculating a spectrum for fluctuation of the heartbeat interval based on a model of the heartbeat interval estimated by the estimation, and determining the arousal level based on the spectrum;
A program that causes a computer to execute processing. The program according to claim 13, further comprising:
Detecting the amount of time variation of the physiological state of the subject,
The estimation control is performed when the amount of time fluctuation of the physiological state exceeds a predetermined amount, and the amount of decrease in the error per update in updating the heart rate interval model is larger than before the control. To
A program that causes a computer to execute processing. The estimation of the model of the heartbeat interval is performed using a Kalman filter,
In the estimation control, the Kalman gain in the Kalman filter is controlled so that the amount of decrease in the error per update in updating the heart rate interval model is larger than before the control.
The program according to claim 14.   The program according to claim 14 or 15, wherein the time fluctuation amount of the average heart rate of the subject is detected as the time fluctuation amount of the physiological state of the subject.   The program according to claim 14 or 15, wherein a time fluctuation amount of acceleration of the body part of the subject is detected as a time fluctuation amount of the physiological state of the subject.   The program according to claim 13, wherein the estimation of the model of the heartbeat interval is performed using a Kalman filter.

Images