JP2010142456A - Heartbeat detecting apparatus - Google Patents

Abstract

An apparatus for detecting a heartbeat of an occupant of a vehicle detects RRI with high accuracy.
A heart rate detecting device for detecting a heart rate of an occupant in a vehicle, wherein a sensor unit for detecting an occupant's body surface motion and a heart rate source signal, which is an output of the sensor unit, are detected by a band pass filter to obtain a peak wave of the heart rate. A heartbeat peak signal calculation unit that calculates an enhanced heartbeat peak signal, an R wave calculation unit that calculates a peak wave as an R wave from the heartbeat peak signal based on amplitude information, and two temporally adjacent ones calculated by the R wave calculation unit An RRI calculating unit that calculates a time interval between two R waves as an RRI and determines whether the RRI calculated from comparison with the adjacent RRI is an abnormal value, and the RRI calculated by the RRI calculating unit is an abnormal value If it is determined, the R wave calculation unit performs an appropriate R wave calculation again by comparing with the adjacent RRI, and then the RRI calculation unit is recalculated by the R wave calculation unit. Again it calculates the RRI based on R waves.
[Selection] Figure 1

Description

  The present invention relates to a heartbeat detection device and a heartbeat detection method, and more particularly, to a heartbeat detection of an automobile occupant.

  In recent years, traffic accidents caused by low-concentration due to drowsy driving and sleepiness of vehicle occupants have become social problems. It is desired to realize a drowsy driving prevention system that objectively detects that the wakefulness of the occupant has decreased to a state that hinders driving, and awakens the occupant or encourages a break.

  Thus, the heartbeat is attracting attention as a physiological index for easily determining the degree of wakefulness of an occupant at a relatively low cost. In this method, the degree of arousal is determined by measuring the activity level of the positive and parasympathetic nerves by performing fluctuation analysis of the heartbeat interval. In general, fluctuations in the heartbeat interval change under the influence of both sympathetic and parasympathetic nerves (Respiratory Sinus Arrhythmia: RSA), the amplitude of which decreases when the effect of parasympathetic nerves is greatly tensed, and the amplitude increases when relaxed. Blood pressure fluctuations or Mayer rhythm fluctuations (Mayer Wave Relative Sinus Arrhythmia: MWSA) and fluctuations related to body temperature regulation are known.

  In order to apply such heartbeat interval fluctuation analysis to a drowsiness driving prevention system that requires real-time evaluation, RRI (abbreviation of R-R Interval) must be accurately calculated in real time. RRI is a time interval between two adjacent R waves (peak waves having the largest amplitude in one pulsation). Therefore, it is important to accurately calculate the timing of the R wave generated by each pulsation.

  On the other hand, various sensors have been studied as means for measuring heartbeats. However, a non-contact sensor that does not require the occupant to wear the sensor is desired for installation in the passenger compartment. Examples of such non-contact sensors include radio wave radar, ultrasonic radar, piezoelectric sensor, and pneumatic sensor. While these non-contact sensors have a merit on the product that does not restrain the occupant, there is a problem that vibrations during travel and movements of the body other than the heartbeat are detected as noise. As described above, in determining the degree of wakefulness of an occupant using a heartbeat, it is required to accurately calculate the RRI necessary for heartbeat interval fluctuation analysis from reception signals with a lot of noise components acquired by a non-contact sensor.

  Patent Document 1 discloses a heartbeat measuring device that measures an occupant's RRI using a body surface motion detection sensor such as a radio wave Doppler sensor.

  Specifically, it is described that the peak value of a signal obtained by band-limiting the output signal of the body surface motion detection sensor with a bandpass filter is calculated as an R wave, and the time interval between adjacent R waves is calculated as an RRI. . The peak value is calculated using a maximum value in time that is the maximum value within a predetermined calculation time. Here, the predetermined calculation time is determined based on the maximum value of the heartbeat to be detected. For example, when the maximum value of the heart rate is 300 beats / minute, the predetermined calculation time is 0.2 seconds.

Furthermore, it is described that the peak value calculation is limited to 50% or more and 100% or less of the maximum amplitude, thereby avoiding erroneous calculation of the peak value corresponding to noise as an R wave.
JP 2006-55406 A

  In the prior art, R wave calculation processing is performed based on amplitude information, such as calculating the peak value of the maximum amplitude within a predetermined calculation time as an R wave. However, when the radio wave type Doppler sensor is irradiated to the subject and the actually acquired received signal is observed, the peak wave corresponding to the R wave in the received signal is a plurality of peaks observed in one pulsation section. Most of the waves have the maximum amplitude, but the amplitude of the T wave, which is the second peak wave generated after the R wave in one pulsation, and other peak waves caused by the movement of the subject are larger. There is also a case of taking Therefore, even if the peak value calculation is limited to 50% or more and 100% or less of the maximum amplitude, the conventional technique for calculating the peak value corresponding to the R wave based on the amplitude information causes the peak wave due to noise or the like. There are many cases where R is erroneously calculated as an R wave.

  Further, in the conventional technique, a histogram is generated from the time interval of the peak value corresponding to the R wave calculated based on the amplitude information, and the peak value shorter than the maximum number of time intervals of the histogram is deleted to calculate the RRI. Is described. However, the peak value corresponding to the R wave calculated based on the amplitude information includes a peak wave due to an erroneously calculated noise or the like, and a lot of data and time are required to improve accuracy. Therefore, there is a problem that accuracy and real-time performance are lacking in applying to a doze driving prevention system that objectively detects the wakefulness of the occupant and wakes up the occupant or encourages a break.

  The present invention has been made to solve such a problem, and is useful for heart rate interval fluctuation analysis used when evaluating a comprehensive change in consciousness including driver's tension and drowsiness when driving a vehicle. An object of the present invention is to provide a heartbeat detecting device and a heartbeat detecting method capable of accurately calculating a real RRI in real time.

A first aspect for achieving the above object is as follows:
A heart rate detecting device for detecting a heart rate of an occupant in a vehicle, wherein a heart rate peak signal is emphasized by a sensor unit detecting a body surface motion of the occupant and a heart rate source signal output from the sensor unit by a band pass filter. A heartbeat peak signal calculation unit that calculates a peak signal, an R wave calculation unit that calculates a peak wave as an R wave based on amplitude information from the heartbeat peak signal, and two temporally adjacent R waves calculated by the R wave calculation unit An RRI calculating unit that calculates a time interval as RRI and determines whether the RRI calculated by comparing with the adjacent RRI is an abnormal value, and the RRI calculated by the RRI calculating unit is determined to be an abnormal value. Then, the R wave calculation unit again calculates an appropriate R wave from the comparison with the adjacent RRI, and then the RRI calculation unit performs RR based on the R wave recalculated by the R wave calculation unit. And calculates again.

  In addition, the RRI calculation unit determines that the difference is an abnormal value when the difference between the RRI to be determined and the RRI that is immediately before is equal to or greater than a predetermined value.

  In addition, the R wave calculation unit is characterized in that the R wave that is closest to the previous RRI in time is recalculated as a new R wave.

  In addition, a heartbeat cycle signal calculation unit that calculates a sinusoidal heartbeat cycle signal that indicates the periodicity of an occupant's heartbeat by using a bandpass filter from the heartbeat source signal, and a period in which an R wave exists based on the heartbeat cycle signal And an R wave existence section setting unit for setting an R wave existence section to be limited to the R wave existence section, wherein the R wave calculation section calculates one peak wave as an R wave in each of the R wave existence sections, To do.

  In this way, the R wave can be calculated more accurately.

  Further, the R wave existence section setting unit calculates all the maximum values of the heartbeat period signal, sets the time at which the maximum value is obtained as the heartbeat period signal maximum time, and sets the Rwave presence to the section of the heartbeat period signal maximum time ± time offset Tofs. It is characterized by setting a section.

  Further, the R wave existence section setting unit may set the time offset Tofs to at least 1/4 or less of an average period of the heartbeat period signal.

  In the second aspect, the RRI calculation unit obtains a difference value from the previous RRI in terms of time for each calculated RRI, and whether the RRI calculated using the average value of the difference values is an abnormal value. It is characterized by judging.

The third aspect is
A heart rate detection method for detecting the heart rate of an occupant in a vehicle, and calculating a heart rate peak signal in which a heart rate source signal that is an output of a sensor unit that detects an occupant's body surface motion is emphasized by a band pass filter. A heartbeat peak signal calculating step, an R wave calculating step for calculating a peak wave as an R wave from the heartbeat peak signal based on amplitude information, and a time interval between two temporally adjacent R waves calculated by the R wave calculating step. RRI calculating step for determining whether or not the RRI calculated by comparing with the adjacent RRI is an abnormal value, and comparing the RRI calculated in the RRI calculating step with the adjacent RRI Recalculating an appropriate R wave from and recalculating the RRI based on the recalculated R wave. Characterized in that it.

  As described above, according to the heartbeat detection device of the present invention, since the RRI is calculated in consideration not only of the amplitude characteristic of the heartbeat but also the periodic characteristic, the non-contact sensor in which a lot of noise components are included in the output signal is provided. Even when used, the RRI can be accurately calculated in real time. Therefore, it is possible to provide a heart rate detection device and a heart rate detection method that can be applied to a drowsy driving prevention system for a vehicle occupant or the like using heart rate interval fluctuation analysis.

(First embodiment)
Hereinafter, a heartbeat detection device according to a first embodiment of the present invention will be described. FIG. 1 is a block diagram illustrating a configuration of the heartbeat detection device 100. The heartbeat detection device 100 includes a sensor unit 101 and a calculation unit 102. As shown in FIG. 2, the heart rate detection device 100 is typically installed inside the backrest of a driver seat 202 on which an occupant 201 is seated. When the heart rate detection device 100 is installed, it is sufficient that at least the sensor unit 101 is on the backrest unit, and the calculation unit 102 may be installed in another place.

  The sensor unit 101 is, for example, a radio wave type Doppler sensor, and outputs a reception signal whose amplitude changes according to the heartbeat, respiration, or other object movement of the occupant 201 (hereinafter, this reception signal is referred to as “heartbeat source”). Signal "). In addition, a body surface motion detection sensor such as an ultrasonic Doppler sensor, a piezoelectric sensor, or a pneumatic sensor can be used.

  The computing unit 102 calculates an RRI (abbreviation of RR interval: time interval between two R waves adjacent in time) in the heartbeat of the occupant 201 by various computations from a heartbeat source signal that is an output signal of the sensor unit 101. . The calculation unit 102 includes a heartbeat cycle signal calculation unit 103, an R wave existence section setting unit 104, a heartbeat peak signal calculation unit 105, an R wave calculation unit 106, and an RRI calculation unit 107.

  Here, a human electrocardiogram will be described. FIG. 3 shows a time waveform for the first 20 seconds of an electrocardiogram of a human in a seated position measured for 3 minutes using a chest electrode. FIG. 4 is a 3-minute RRI trend graph obtained by calculating the RRI in the electrocardiogram. From the above graph, the following three things can be said about the human heartbeat. The R wave has the maximum amplitude in one pulsation. 2. The RRI changes in a range of approximately 800 to 1000 ms (60 to 75 bpm when converted to an average heart rate), and the RRI does not become extremely small or large beyond the range. 3. The absolute value of the difference between two adjacent RRIs in time is about several tens of ms, and is 60 ms at the maximum, and the difference value does not become extremely large even if it becomes small or becomes zero. Thus, the heartbeat signal has clear amplitude characteristics and periodic characteristics.

  Next, a heartbeat source signal that is output from the reception signal of the sensor unit 101 will be described with reference to FIG. A dotted line shows a time waveform between 20 seconds and 30 seconds of the electrocardiogram in FIG. The solid line shows the time waveform between 20 seconds and 30 seconds in the heart rate source signal measured simultaneously with the electrocardiogram measurement of FIG. From these waveforms, the following three things can be said about the heartbeat source signal. 1. The global cycle of the heartbeat source signal is generally consistent with the electrocardiogram. 2. At the time when the R wave of the electrocardiogram exists, a peak wave having a large amplitude also exists in the heartbeat source signal, and in most cases, the amplitude is maximum in the pulsation interval. 3. Since a non-contact sensor is used, a peak wave caused by a motion other than a heartbeat including an R wave is generated, and some of them have a larger amplitude than the peak wave corresponding to the R wave.

  Therefore, in the present embodiment, when an R wave is calculated based on amplitude information for such a heartbeat source signal as in the prior art, a peak wave caused by noise is erroneously calculated as an R wave. Process to avoid that. As described above, RRI can be detected with high accuracy in real time by utilizing not only the amplitude characteristic of the heartbeat signal but also the periodicity. Details will be described later.

  Hereinafter, specific processing of the calculation unit 102 will be described.

  First, in step S601 to step S603 in FIG. 6, in order to calculate the R wave generated with a period of about 1 second with high accuracy and efficiency, the R wave existence section is set as a section having a high probability that the R wave exists. First, the processing is based on the idea of calculating R waves mainly using amplitude information.

  In step S601, a heartbeat source signal is passed through a narrow band-pass filter to calculate a signal (heartbeat cycle signal) in which the heartbeat periodicity is emphasized. That is, it is said that the heart rate of a human is 70 bpm at rest, about 60 bpm for a slow person, and about 90 bpm for a fast person, so the pass band of the bandpass filter is, for example, 0.8 to 1.5 Hz. And Thereby, the heartbeat cycle signal becomes a signal close to a sine wave.

  In step S602, all local maximum values in the sinusoidal heartbeat cycle signal obtained in step S601 are calculated. The time at which the local maximum value is obtained with respect to the i-th local maximum value (i = 0, 1, 2,...) From the start of measurement is the cardiac cycle signal maximum time Tlm [i].

  In step S603, a certain section centered on each cardiac cycle signal maximum time calculated in step S602 is set as a section having a high probability that an R wave exists (an R wave existence section). For example, the start point Zst [i] and the end point Zed [i] of the R wave existence section for the i-th maximum value (i = 0, 1, 2,...) From the start of measurement are as follows with the time offset as Tofs. The following equation is set: Zst [i] = Tlm [i] −Tofs, Zed [i] = Tlm [i] + Tofs. If the time offset Tofs is too large, the number of peak waves existing in one R wave existence section increases, and the possibility of erroneous detection of R waves increases. On the other hand, if the time offset Tofs is too small, when the R wave exists outside the R wave existence section, the accuracy of the RRI may be significantly lowered. For example, Tofs may be set to 350 ms.

  The time offset Tofs may be set to at least 1/4 of the average period of the heartbeat period signal. When there are n local maximum values of the heartbeat cycle signal, the average cycle Cav of the heartbeat cycle signal is the difference between the first heartbeat cycle signal maximum time and the nth heartbeat cycle signal maximum time by the maximum number n−1. Since it is divided, it is calculated by the following formula: Cav = (Tlm [n] −Tlm [1]) / (n−1), Tofs ≦ Cav / 4.

  By setting Tofs as described above, adjacent R wave existence sections hardly overlap in time, so that the possibility of erroneously calculating R waves in each R wave existence section is reduced. be able to. FIG. 7 shows a time waveform of 45 to 50 seconds after the start of measurement, among the 3-minute heartbeat source signals acquired previously. A maximum value was further calculated from a heartbeat cycle signal (thin solid line) calculated from a heartbeat source signal (thick solid line), and an R wave existence section (section where the value is 1 with a thin one-dot broken line) was set.

  On the other hand, from step S604 to step S609, the R wave is calculated using the R wave existence section set in step S603 to calculate the RRI. This will be described in detail below.

  In step S604, the heartbeat source signal is passed through a broadband bandpass filter to calculate a heartbeat peak signal in which the peak wave of the heartbeat is emphasized. The pass band of the band pass filter is, for example, 0.8 to 10 Hz. If the pass band is too large, the number of peak waves calculated increases, and there is a risk of calculating the R wave incorrectly. On the other hand, if it is too small, the peak wave is dull in the amplitude direction and the time direction, so that the temporal position accuracy of the R wave may be deteriorated. In calculating the heart rate peak signal, it is also possible to use a difference signal of the heart rate source signal in addition to using the heart rate source signal as described above. The difference signal is a signal obtained by taking a difference between a heartbeat source signal at a certain time and a heartbeat source signal at the previous time in time, and obtaining a signal with enhanced peak characteristics. it can.

  In step S605, all peak waves in the heartbeat peak signal obtained in step S604 are calculated.

  In step S606, among all the peak waves calculated in step S605 in all the R wave existence sections obtained in step S603, the amplitude is the largest among all the peak waves existing in each R wave existence section. The thing is calculated as an R wave. As described above, by setting the R wave existence section as a section having a high probability that the R wave exists, the R wave can be calculated with high accuracy and efficiency. In the present embodiment, a method for calculating RRI using an R wave existence section will be described, but RRI can be calculated without using an R wave existence section.

  FIG. 8 shows the result of executing Steps S604 to S606 using the heartbeat source signal for 45 seconds to several seconds after the start of the measurement described above. A graph having a peak wave flag value of 1 indicated by a solid line with square marks is obtained by calculating a peak wave having a high possibility of being an R wave from a heartbeat peak signal indicated by a thin broken line. For example, a peak wave having a high possibility of being an R wave is calculated by a process such as omitting a continuous peak wave and omitting a peak wave having an amplitude value smaller than a predetermined value. A graph in which the value of the R-wave flag indicated by a thin solid line with a triangle mark is 1, all peaks existing in each R-wave existence section (a section with a thin solid line with a value of 2). The wave having the maximum amplitude is calculated as an R wave.

  Returning to FIG. 6, the description will be continued. In step S607, RRI is calculated from the R wave calculated in step S606. The time at which the R wave exists in the i th (i = 0, 1, 2,...) R wave existence section from the start of measurement is Tr [i], and the i th (i = 0, 1, 2,. ..) Is RRI [i], RRI [i] is obtained by the following formula: RRI [i] = Tr [i] -Tr [i-1].

  FIG. 9 shows an RRI (hereinafter referred to as “sensor RRI”) calculated in step S607 from the heartbeat source signal measured for 3 minutes previously and an RRI calculated from an electrocardiogram simultaneously measured by the chest electrode (hereinafter referred to as “Refer RRI”). ")" Is a trend graph. The sensor RRI coincides with the referrer RRI as a global trend, but it can be confirmed that an abnormal value is taken in some places. In most cases, this is because the R wave in the R wave existence section is not the maximum amplitude in the section. That is, it occurs when the amplitude of the peak wave corresponding to the T wave, the peak wave generated by the movement of the entire body of the subject, or the like becomes larger than the amplitude of the R wave.

  FIG. 10 shows the data of 22 to 24.5 seconds after the start of measurement in the same experimental data as FIG. In the R wave existence section (one-dot broken line) of 23.0 to 23.6 seconds, the peak wave of the maximum amplitude of the heartbeat peak signal (broken line) (around 23.4 seconds, indicated by a square mark) is calculated as an R wave. Yes. However, this peak wave corresponds to a T wave, and a true R wave is a peak wave around 23.1 seconds. This is also clear from the comparison with the rifa heartbeat (thick solid line).

  Therefore, it is determined whether the RRI has an abnormal value using the periodicity of the heartbeat signal (step S608), and the R wave is recalculated (step S609). In step S610, RRI is calculated again with the recalculation of the R wave performed in step S609. This completes a series of RRI calculation processing.

  FIG. 12 shows a trend graph in which RRI is calculated again after performing R-wave recalculation processing on the RRI calculation result of FIG. It can be seen that the abnormal values seen in FIG. 9 almost disappear and the sensor RRI follows the referrer RRI.

  Here, determination of whether RRI is an abnormal value and recalculation of the R wave will be described with reference to FIG. For the determination of the abnormal value of RRI, the above-described periodic characteristics of the human heartbeat are used. That is, the RRI does not deviate greatly from the range of approximately 800 to 1000 ms, and the absolute value of the difference between two RRIs that are adjacent in time is about several tens of ms, and thus does not take an extremely large value. do. Specifically, when RRI [i] is abnormally larger or smaller than RRI [i-1] for a certain integer i, it is determined that calculation of the R wave in the i-th R wave existence section has failed. To do. That is, it is determined that RRI is an abnormal value. Then, a peak wave that matches RRI [i−1] is calculated as a new R wave from the remaining peak waves in the R wave existence section. For example, a peak wave that minimizes the difference between RRI [i] and RRI [i−1] is calculated as a new R wave.

  Returning to FIG. In the R wave existence section of 23.0 to 23.6 seconds, the RRI that had been changing at around 900 ms suddenly became 1100 ms, so it was determined that the calculation of the R wave failed. Then, among the remaining peak waves in the R wave existence section, a peak wave (star mark in FIG. 10) around 23.1 seconds in which the difference between the RRI and the RRI immediately before it becomes the smallest is newly added. Calculated as a simple R wave.

Thus, calculating the R wave accurately from the peak wave in the R wave existence section in real time is when evaluating the comprehensive change in consciousness including the driver's tension and sleepiness when driving the vehicle. It is very effective for heartbeat interval fluctuation analysis. That is, in order to use the evaluation result of the driver's consciousness change in a drowsy driving prevention system or the like, it is required to measure the consciousness change in a short time in real time with a simple calculation. When calculating the heart rate using RRI as in the prior art, an average value can be obtained from many samples even if an R wave that is erroneously calculated from a peak wave due to noise or the like is included. However, it is not suitable for systems that require results in real time, such as evaluating changes in driver awareness while driving.
(Second Embodiment)
Hereinafter, a heartbeat detection device according to a second embodiment of the present invention will be described.

  The processing of the RRI calculating unit 107 in the second embodiment is the same as the flowchart of the first embodiment shown in FIG. 6 except for the R wave recalculation (step S608). Therefore, only the R wave recalculation process will be described here, and the description of other parts will be omitted.

  In the first embodiment, if RRI [i] is abnormally larger or smaller than RRI [i−1] for a certain integer i in step S608, the R wave is calculated in the i-th R wave existence section. Judged that it failed. The second embodiment is characterized in that a difference value of RRI is held, and an abnormal value determination of RRI is performed using a deviation of the difference value of RRI. That is, for each RRI calculated in step S607, a difference value (RRI difference value) from the previous RRI in time is calculated, and an average value (RRI difference average value) is calculated from all the RRI difference values. Further, a deviation from the RRI difference average value (RRI difference deviation) is calculated for each RRI difference value. If the absolute value of the i-th RRI difference deviation for a certain integer i exceeds a predetermined value, it is determined that the R-wave calculation has failed in the i-th R-wave calculation section. Then, a peak wave that matches RRI [i−1] is calculated as a new R wave from the remaining peak waves in the R wave existence section. For example, a peak wave that minimizes the difference between RRI [i] and RRI [i−1] is calculated as a new R wave. As described above, the RRI changes from moment to moment, but the difference between adjacent RRIs is about 50 ms at most, which is based on the heartbeat period characteristic.

  As described above, in the heartbeat detection device according to the embodiment of the present invention, the peak wave corresponding to the R wave of the heartbeat is maximized in one pulsation section using the non-contact sensor that detects the body surface motion of the human body. Even when the amplitude is not taken, it is possible to accurately calculate the RRI by accurately detecting the R wave by utilizing not only the heartbeat amplitude characteristic but also the period characteristic. Therefore, the heartbeat interval fluctuation analysis using the RRI obtained from the heartbeat detection device can improve the determination accuracy of the wakefulness of the occupant, and is a non-contact drowsiness detection system such as an occupant of a moving body such as an automobile. It is suitable for realizing the above.

  In the first and second embodiments, an example of detecting the heartbeat of an occupant of an automobile has been shown. However, the present invention can be applied not only to an occupant but also to an occupant. Also widely applicable. Further, various parameters in the embodiment can be changed as appropriate.

  The heartbeat detection device of the present invention is useful as a dozing driving prevention system using a non-contact sensor mounted in a vehicle.

The block diagram which shows the structure of the heart-rate detection apparatus in the 1st Embodiment of this invention. The figure which shows a mode that the heartbeat detection apparatus in the 1st Embodiment of this invention was installed in the vehicle interior. The figure which shows the time waveform of the electrocardiogram acquired with the chest electrode The figure which shows the RRI trend graph of the electrocardiogram acquired by the chest electrode The figure which shows the time waveform of the electrocardiogram acquired by the heart-rate source signal acquired by the heart-rate detection apparatus in the 1st Embodiment of this invention, and the chest electrode The flowchart which shows operation | movement of the RRI calculation part of the heart rate detection apparatus in the 1st Embodiment of this invention. The figure which shows the time waveform and R wave presence area of the heartbeat source signal and heartbeat period signal which were acquired by the heartbeat detection apparatus in the 1st Embodiment of this invention. The figure which shows a mode that R wave is calculated from the heartbeat source signal acquired by the heartbeat detection apparatus in the 1st Embodiment of this invention. The figure which shows the trend graph of sensor RRI computed by the referrer RRI and the heart rate detection apparatus in the first embodiment of the present invention. The figure which shows the miscalculation of the typical R wave of the heart rate detection apparatus in the 1st Embodiment of this invention. The figure which showed typically R wave recalculation operation | movement of the heart-rate detection apparatus in the 1st Embodiment of this invention. The figure which shows the trend graph of referrer RRI and the sensor RRI re-calculated by the heart rate detection apparatus in the 1st Embodiment of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Heartbeat detection apparatus 101 Sensor part 102 Calculation part 103 Heartbeat period signal calculation part 104 R wave presence area setting part 105 Peak wave calculation part 106 R wave calculation part 107 RRI calculation part 201 Crew 202 Driver's seat

Claims (8)

A heart rate detection device for detecting a heart rate of an occupant in a vehicle,
A sensor unit for detecting the body surface movement of the occupant;
A heart rate peak signal calculation unit that calculates a heart rate peak signal in which a heart rate source signal that is an output of the sensor unit is emphasized by a band pass filter.
An R wave calculating unit that calculates a peak wave as an R wave based on amplitude information from the heartbeat peak signal;
An RRI calculation unit that calculates a time interval between two R waves that are temporally adjacent calculated by the R wave calculation unit as an RRI and determines whether the RRI calculated by comparison with the adjacent RRI is an abnormal value; Prepared,
If it is determined that the RRI calculated by the RRI calculation unit is an abnormal value, the R wave calculation unit calculates an appropriate R wave again by comparing with the adjacent RRI, and then the RRI calculation unit A heart rate detection device that calculates RRI again based on the R wave recalculated by the wave calculation unit. The RRI calculation unit determines that the difference is an abnormal value when a difference between an RRI to be determined and an RRI that is immediately before is equal to or greater than or equal to a predetermined value. Heart rate detector. The heartbeat detection device according to claim 1, wherein the R wave calculation unit recalculates an R wave that is closest to the previous RRI in terms of time as a new R wave. A heartbeat cycle signal calculation unit that calculates a sinusoidal heartbeat cycle signal indicating the periodicity of the occupant's heartbeat by a bandpass filter from the heartbeat source signal;
An R-wave presence section setting unit that sets an R-wave presence section that temporally limits a section in which an R wave exists based on the heartbeat cycle signal;
The heartbeat detection device according to claim 1, wherein the R wave calculation unit calculates one peak wave as an R wave in each of the R wave existence sections. The R wave existence section setting unit calculates all the maximum values of the heartbeat period signal, sets the time when the maximum value is obtained as the heartbeat period signal maximum time, and sets the section of the heartbeat period signal maximum time ± time offset Tofs as the R wave The heartbeat detection device according to claim 4, wherein the heartbeat detection device is set as an existing section. 6. The heartbeat detection device according to claim 5, wherein the R wave existence section setting unit sets the time offset Tofs to at least 1/4 or less of an average period of the heartbeat period signal. The RRI calculation unit obtains a difference value from the previous RRI in terms of time for each calculated RRI, and determines whether the calculated RRI is an abnormal value using the average value of the difference values. The heartbeat detecting device according to claim 1, wherein A heart rate detection method for detecting the heart rate of an occupant in a vehicle,
A heart rate peak signal calculation step of calculating a heart rate peak signal in which a heart rate peak signal is emphasized by a band pass filter from a heart rate source signal that is an output of a sensor unit that detects the body surface motion of the occupant;
An R wave calculating step of calculating a peak wave as an R wave based on amplitude information from the heartbeat peak signal;
An RRI calculating step of calculating, as an RRI, a time interval between two temporally adjacent R waves calculated by the R wave calculating step, and determining whether or not the RRI calculated by comparison with the adjacent RRI is an abnormal value;
If it is determined that the RRI calculated in the RRI calculation step is an abnormal value, an appropriate R wave is calculated again by comparing with the adjacent RRI, and the RRI is calculated again based on the recalculated R wave. And a heart rate detection method.

Images