JP2016202603A - Biological information processing system, program and control method for biological information processing system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a biological information processing system having a small calculation load and performing highly accurate processing when pulse wave information is obtained on the basis of pulse wave sensor information, and a program, a control method of the biological information processing system and the like.SOLUTION: A biological information processing system includes: an information acquisition section 110 acquiring pulse wave sensor information from a pulse wave sensor 42; and a processing section 130 obtaining pulse wave information on the basis of the pulse wave sensor information. The processing section 130 obtains estimated pulse information, sets a parameter of an approximate expression approximating a wave form of a pulse wave identified from the pulse wave sensor information on the basis of the estimated pulse information obtained and obtains the pulse wave information on the basis of the approximate expression having the parameter set.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a biological information processing system, a program, a control method for the biological information processing system, and the like.

  An electrocardiogram that is often used for examining the function of the heart is a graph that records minute potential fluctuations that occur as the heart beats. Among the P, QRS, and T waves of the electrocardiogram, the R wave, which is the sharpest peak, is an electrical signal that is generated when the ventricle contracts rapidly and pumps blood out of the heart. The time difference between the previous R wave generation times is the heartbeat interval (RR interval).

  The heartbeat interval is not always constant and varies with periodicity. This fluctuation of the heartbeat interval is greatly related to the function of the autonomic nerve, and has been studied for many years in the medical and physiological fields. Also, in the field of biomedical engineering, by measuring the heart rate interval and constantly monitoring the autonomic nervous function, it is possible to grasp the mental state of a person, predict the risk of heart disease, and examine various applications. Has been.

  In addition, the pulse wave is measured using a pulse wave sensor that can be operated more easily than electrocardiogram measurement, and the peak interval of the pulse wave is information corresponding to the RR interval of the electrocardiogram (hereinafter referred to as pulse wave RR interval). It is also considered to be used for health management and medical diagnosis.

  For example, in Patent Document 1, a pulse wave measured continuously for a certain time is passed through a plurality of digital smoothing polynomial filters, and each differential signal and a time when the differential signal peaks are calculated. A technique for obtaining the RR interval is known.

JP 2012-95795 A JP 2009-72417 A JP 2014-212915 A

  In Patent Literature 1, the frame size m of the digital smoothing polynomial filter (approximate range of the approximate expression, a parameter for determining the approximate expression in a broad sense) is 10, 12, 14, 16, 18, 20,. . . As described above, a plurality of settings are set, differential calculation is performed, and an appropriate frame size m is determined from the calculation result. Therefore, since it is necessary to repeatedly perform differential calculation using a plurality of frame sizes, there is a problem that the calculation load is heavy and the efficiency is deteriorated.

  According to some aspects of the present invention, when obtaining pulse wave information based on pulse wave sensor information, a biological information processing system, a program, and a biological information processing system that perform processing with high computational load and light accuracy A method or the like can be provided.

  One aspect of the present invention includes an information acquisition unit that acquires pulse wave sensor information from a pulse wave sensor, and a processing unit that obtains pulse wave information based on the pulse wave sensor information. Based on the obtained estimated pulse information, pulse information is determined, an approximate expression parameter that approximates the waveform of the pulse wave specified from the pulse wave sensor information is set, and the parameter is set based on the approximate expression. The present invention relates to a biological information processing system for obtaining the pulse wave information.

  In one aspect of the present invention, estimated pulse information is obtained, and parameters of an approximate expression are set based on the estimated pulse information. As a result, parameters that can be appropriately approximated can be set, so that pulse wave information can be obtained with high accuracy, and the processing load is reduced compared to the case where calculation is performed with multiple parameters as in the conventional method. It becomes possible.

  In one aspect of the present invention, the information acquisition unit acquires body motion sensor information from a body motion sensor, and the processing unit obtains a user's behavior pattern based on the body motion sensor information, and When it is determined that the behavior pattern has changed, the parameter resetting process may be performed.

  This makes it possible to reset the parameters at appropriate timing and conditions based on the behavior pattern change determination.

  In one aspect of the present invention, the processing unit obtains an error between a pulse wave identified from the pulse wave sensor information and a pulse wave identified from the approximate expression obtained based on the estimated pulse information, When the error is larger than a given threshold, the obtained pulse wave information may be deleted.

  Thereby, it becomes possible to suppress the output of inappropriate pulse wave information based on the comparison determination between the original waveform and the approximate waveform.

  In the aspect of the invention, the processing unit may perform the parameter resetting process when the error is larger than the given threshold.

  This makes it possible to reset parameters based on the comparison determination between the original waveform and the approximate waveform.

  In the aspect of the invention, the parameter may be a parameter that sets an application range of the approximate expression for the pulse wave sensor information.

  This makes it possible to appropriately set the application range of the approximate expression.

  In the aspect of the invention, the processing unit may obtain pulse wave RR interval information as the pulse wave information.

  Thereby, it becomes possible to obtain the pulse wave RR interval information with high accuracy.

  In the aspect of the invention, the processing unit may be configured to detect the pulse wave based on a peak detection process of the waveform represented by the approximate expression or a process using a differential signal of the waveform represented by the approximate expression. RR interval information may be obtained.

  Thereby, it is possible to obtain the pulse wave RR interval information with high accuracy by processing using peak detection or differential signals.

  In the aspect of the invention, the processing unit may estimate autonomic nerve state information based on the pulse wave information obtained based on the approximate expression.

  Thereby, it becomes possible to estimate the autonomic nerve state information using the obtained pulse wave information.

  In one aspect of the present invention, the processing unit sets a value of a frame size m as the parameter based on the estimated pulse information, and the pulse in the frame corresponding to the set frame size m is set. For the wave sensor information, a smoothing polynomial filter may be applied to obtain the approximate expression, and the pulse wave information may be obtained based on the approximate expression.

  Thereby, it is possible to perform processing for obtaining pulse wave information using a smoothing polynomial filter in which the frame size m is set.

  In one aspect of the present invention, the frame corresponding to the frame size m includes the pulse wave sensor information at the processing target timing and the pulse wave sensor information for m samples before and after the processing target timing. 2m + 1 sample frames, and the processing unit obtains an estimated pulse wave RR interval that is an estimated value of the pulse wave RR interval as the estimated pulse information, and as a frame of 2m + 1 samples represented by the frame size m. A frame corresponding to a length within a given numerical range including the length of the estimated pulse wave RR interval may be set.

  Accordingly, it is possible to set the frame size m in accordance with the length of the estimated pulse wave RR interval.

  In the aspect of the invention, the processing unit may obtain a differential signal of the pulse wave sensor information after applying the smoothing polynomial filter, and obtain the pulse wave information based on the differential signal.

  Thereby, after applying the smoothing polynomial filter, it is possible to obtain pulse wave information by performing a process for obtaining a differential signal.

  In the aspect of the invention, the information acquisition unit may acquire body motion sensor information from a body motion sensor, and the processing unit may obtain the estimated pulse information based on the body motion sensor information. Good.

  This makes it possible to obtain estimated pulse information from body motion sensor information.

  In one embodiment of the present invention, the processing unit obtains exercise intensity information of a user based on the body motion sensor information, and the estimation is performed based on the exercise intensity information, maximum pulse information, and resting pulse information. You may ask for pulse information.

  This makes it possible to obtain estimated pulse information based on exercise intensity information obtained from body motion sensor information.

  In the aspect of the invention, the processing unit may perform a frequency conversion process on the pulse wave sensor information, and obtain the estimated pulse information based on a result of the frequency conversion process.

  Thereby, it is possible to obtain estimated pulse information by frequency conversion processing on pulse wave sensor information.

  In the aspect of the invention, the processing unit may obtain the estimated pulse information based on the pulse wave information obtained before the processing target timing.

  Thereby, it becomes possible to obtain the estimated pulse information using the pulse wave information already obtained.

  According to another aspect of the present invention, a computer functions as an information acquisition unit that acquires pulse wave sensor information from a pulse wave sensor, and a processing unit that obtains pulse wave information based on the pulse wave sensor information. The processing unit obtains estimated pulse information, sets parameters of an approximate expression that approximates a waveform of a pulse wave specified from the pulse wave sensor information based on the obtained estimated pulse information, and sets the approximate parameter set The present invention relates to a program for obtaining the pulse wave information based on an equation.

  In another aspect of the present invention, pulse wave sensor information from a pulse wave sensor is acquired, estimated pulse information is obtained, and a pulse wave identified from the pulse wave sensor information based on the obtained estimated pulse information. This relates to a control method of a biological information processing system that sets parameters of an approximate expression that approximates the waveform of and calculates pulse wave information based on the approximate expression that is parameter-set.

The structural example of the biological information processing system which concerns on this embodiment. 2A to 2C are configuration examples of a system including the biological information processing system according to the present embodiment. 3A and 3B are external views of the biological information detection apparatus. The other external view of a biological information detection apparatus. FIG. 5A shows an example of peak detection when an appropriate parameter is set, and FIG. 5B shows an example of an error between the electrocardiogram RR interval and the pulse wave RR interval. 6A shows an example of peak detection when a parameter that is too small for the RR interval is set, and FIG. 6B shows an example of an error between the electrocardiogram RR interval and the pulse wave RR interval. FIG. 7A shows an example of peak detection when a parameter that is too large for the RR interval is set, and FIG. 7B shows an example of an error between the electrocardiogram RR interval and the pulse wave RR interval. The flowchart explaining the process which calculates | requires the pulse wave information of this embodiment. The flowchart explaining an estimated pulse information calculation process. The other flowchart explaining an estimated pulse information calculation process. The flowchart explaining a pulse wave information calculation process. The other flowchart explaining a pulse wave information calculation process. FIG. 13A is a diagram for explaining the time change of the pulse wave RR interval, and FIG. 13B is a diagram for explaining processing for obtaining LF and HF from the pulse wave RR interval.

  Hereinafter, this embodiment will be described. In addition, this embodiment demonstrated below does not unduly limit the content of this invention described in the claim. In addition, all the configurations described in the present embodiment are not necessarily essential configuration requirements of the present invention.

1. First, the method of this embodiment will be described. As described above, it is known that the heartbeat interval (RR interval) obtained from the electrocardiogram is useful information. For example, the user's autonomic nerve activity is estimated based on fluctuation (temporal fluctuation) of the RR interval. There is a known technique to do this. However, in order to measure an electrocardiogram, it is necessary to attach electrodes to two or more points that sandwich a heart at a given site of the user. There are small and lightweight devices such as Holter ECG, but the current ECG measurement is not easy to do in daily life.

  On the other hand, the pulse wave is measured using a pulse wave sensor, and the peak interval of the pulse wave is used as information corresponding to the RR interval of the electrocardiogram (hereinafter referred to as pulse wave RR interval) for health management and medical diagnosis. It is also considered to use it. The pulse wave sensor can be realized by, for example, a photoelectric sensor including a light emitting unit and a light receiving unit, and the device including the pulse wave sensor is also a wearable device such as a wristwatch type device described later with reference to FIG. realizable. Therefore, the pulse wave RR interval can be measured more easily than the RR interval using an electrocardiogram.

  The RR interval is information indicating the interval between heartbeats. The pulse wave sensor information (AC component in the narrow sense) is information corresponding to the pulse, and the period of the waveform is information indicating the pulse interval. As long as there is no disease, the heartbeat and the pulse correspond to each other. Therefore, if information corresponding to one cycle of the waveform of the pulse wave sensor information, specifically, the peak interval is obtained, the peak interval corresponds to the RR interval. This represents the pulse wave RR interval to be performed.

  However, it is known that the pulse wave sensor information includes various noises such as body motion noise caused by the user's body motion. For this reason, even if the peak interval is simply obtained, the error becomes large. For example, when a differential signal of pulse wave sensor information is obtained and an attempt is made to detect the zero point as a peak, the relative error increases due to the calculation for obtaining the differential (difference).

  On the other hand, a method of smoothing the pulse wave sensor information can be considered. Specifically, the pulse wave sensor information is approximated by a given approximate expression, and peak detection is performed using the approximated information. In this way, the variation in value due to the error is smoothed, so that the peak can be detected with high accuracy and the accuracy of the required pulse wave RR interval is also improved.

  It is necessary to set various parameters in order to determine the approximate expression. For example, an approximate expression cannot be obtained unless a setting is made such that approximation is performed for a range of a given frame size m using a p-order polynomial. In the following, approximation is performed using a smoothing polynomial filter, and the order p of the approximate polynomial and the frame size m that is an approximation range are used as parameters (control parameters) of the smoothing polynomial filter. However, the method of approximation and the parameters of the approximation formula are not limited to this.

  In order to perform appropriate approximation, it is necessary to set appropriate parameters according to the waveform to be approximated. For example, when the parameters are the above-described order p and frame size m, the higher the order p and the smaller the frame size m, the easier the approximation formula follows the original waveform (the waveform of pulse wave sensor information). This is because if p is large, the approximate expression can be a complex waveform, and the smaller m is, the smaller the number of samples of pulse wave sensor information contained in the frame corresponding to the frame size m is, and This is because the waveform becomes a simple waveform. Conversely, as the order p is lower and the frame size m is larger, the approximate expression becomes less likely to follow the original waveform.

  The approximation in this embodiment is for suppressing the influence of noise. Therefore, there is a problem that the approximate expression cannot follow the original waveform at all because p is extremely low or m is extremely large. On the other hand, it is also a problem that the approximate expression follows the original waveform too much. This is because the approximate expression follows up to fluctuations caused by noise as shown in A4 to A6 in FIG. 5A (peaks such as B4 to B6 in FIG. 6A are detected). Therefore, the effect of reducing noise cannot be expected. Specifically, when the waveform of the pulse wave sensor information is considered to be a waveform in which fluctuation due to noise is superimposed on an approximately sine wave, the approximate expression follows the approximate sine wave, and fluctuation due to noise It is preferable to set parameters that do not follow.

  For example, when a value of about p = 3 is used, the approximate expression is a cubic polynomial that can have two extreme values, and is therefore suitable for approximation of a waveform of about one sine wave. On the other hand, if a waveform such as two wavelengths or three wavelengths is included in the frame, only the element of the substantially sine wave in the original waveform has many extreme values such as four or six, and the order is Insufficiently, the error between the approximate expression and the original waveform becomes excessively large. Further, if only a waveform such as a half wavelength or a third wavelength is included in the frame, the approximate expression follows a fluctuation that is finer than that of a substantially sine wave, and the influence of noise cannot be suppressed.

  As described above, the parameters of the approximate expression need to be set appropriately. In the method of Patent Document 1, p is fixed to a value of about 3, and m is variable. However, an appropriate value of m is not set in advance, 10, 12, 14, 16, 18, 20,. . . Thus, approximation is performed using various frame sizes. In Patent Document 1, the pulse wave RR interval is obtained on that basis, and the value of the appropriate m is estimated from the variation of the value. In other words, in Patent Document 1, a procedure is adopted in which the pulse wave RR interval obtained from the frame size determined to be appropriate is adopted after trying a frame size within a certain range with a brute force. Is very big.

  Therefore, the present applicant proposes a method of estimating an appropriate value of the frame size m, instead of performing calculation using various frame sizes as in Patent Document 1. Specifically, as illustrated in FIG. 1, the biological information processing system 100 according to the present embodiment is based on the information acquisition unit 110 that acquires pulse wave sensor information from the pulse wave sensor 42 and the pulse wave sensor information. A processing unit 130 for obtaining pulse wave information. Then, the processing unit 130 obtains estimated pulse information, sets parameters of an approximate expression that approximates the waveform of the pulse wave specified from the pulse wave sensor information based on the obtained estimated pulse information, and sets the parameter-set approximation. Based on the equation, pulse wave information is obtained.

  Here, the estimated pulse information is pulse information estimated by some technique, and an example is an estimated value of the pulse wave RR interval. The pulse information here is information representing a pulse, and is not limited to the pulse wave RR interval, and may be other information such as information representing a pulse frequency.

  The pulse wave RR interval is a time representing the length of one wavelength of the pulse wave waveform. That is, if the pulse wave RR interval is estimated, it is possible to estimate the relationship between the frame corresponding to the frame size m and how many wavelengths of the pulse wave sensor information waveform are included in the frame. Therefore, for example, the length of the frame corresponding to the frame size m is set to the length of one wavelength of the substantially sine wave of the pulse wave waveform as described above, that is, the estimated pulse wave RR interval (estimated pulse wave RR interval). It is possible to set a length corresponding to the length.

  In this way, it is possible to set appropriate parameters without trying various parameters (frame sizes) as in the case of Patent Document 1, and the calculation load can be reduced. At this time, it is not necessary to obtain the estimated pulse information with high accuracy. This is because the estimated pulse information here is used only for setting parameters, and then pulse wave information with high accuracy is obtained using the set parameters.

  For example, as will be described later using the following equation (2), the approximation range (frame length) does not need to be exactly one wavelength of a sine wave, and can be sufficiently approximated even if there is some variation. It is. That is, it is sufficient that the estimated pulse information is accurate enough to set parameters.

  A specific method for obtaining the estimated pulse information will be described later, but may be obtained by performing frequency conversion processing such as FFT on pulse wave sensor information of about 16 seconds, for example. In this case, one pulse information is estimated from the pulse wave sensor information for a relatively long period due to the characteristics of the frequency conversion process, so information such as fluctuations in the pulse wave RR interval is buried. Therefore, the estimated pulse information obtained from the FFT obtains a low frequency component LF of 0.04 Hz to 0.15 Hz and a high frequency component HF of 0.15 Hz to 0.4 Hz as will be described later with reference to FIG. In that case, it is not suitable and its accuracy is low. However, it has sufficient accuracy in terms of the approximate value of the pulse wave RR interval, and it is possible to use the FFT result for parameter setting.

  If the above explanation is taken from another viewpoint, it can be said that the method of the present embodiment assumes that pulse wave information is obtained with relatively high accuracy. This is because if the pulse wave information with low accuracy is sufficient, the estimated pulse information described above may be output as it is, and there is no need to recalculate the pulse wave information using the estimated pulse information. Various advantages of obtaining pulse wave information with relatively high accuracy are conceivable. One example is that autonomic nerve activity can be estimated from the fluctuation of the above-described pulse wave RR interval. The method for estimating the autonomic nerve activity will be described later with reference to FIGS. 13 (A) and 13 (B).

  Hereinafter, after describing a detailed configuration example of the biological information processing system according to the present embodiment, details of processing for obtaining pulse wave information will be described. Finally, a method for estimating the state of autonomic nerve activity based on the obtained pulse wave information will be described.

2. System Configuration Example The biological information processing system 100 according to the present embodiment includes the information acquisition unit 110, the storage unit 120, and the processing unit 130 as described above with reference to FIG. However, the biological information processing system 100 is not limited to the configuration shown in FIG. 1, and various modifications such as addition of other components are possible.

  The information acquisition unit 110 acquires pulse wave sensor information from the pulse wave sensor 42. The information acquisition unit 110 may be realized as an interface for acquiring information from other devices, for example, and acquires information from other devices via the network NE as will be described later with reference to FIG. If it is, it is implement | achieved as a communication part (reception process part) which has a communication function. The information acquisition unit 110 may be directly connected to the pulse wave sensor 42 or may be connected via another device. When the biological information processing system 100 is realized as the biological information detection apparatus 400 shown in FIG. 3A or the like, the pulse wave sensor 42 and the information acquisition unit 110 are assumed to be mounted on the same device and directly connected. The Further, when the biological information processing system 100 is realized as the terminal device 300, the information acquisition unit 110 is assumed to acquire pulse wave sensor information via the communication unit or the like of the biological information detection device 400. When the biological information processing system 100 is realized as the server system 200, the information acquisition unit 110 is assumed to acquire pulse wave sensor information via the biological information detection device 400, the terminal device 300, or the like. .

  The storage unit 120 serves as a work area for the processing unit 130 and the like, and its function can be realized by a memory such as a RAM, an HDD (hard disk drive), or the like. The storage unit 120 stores at least the pulse wave sensor information acquired by the information acquisition unit 110. Further, the storage unit 120 may store a processing result (pulse wave information or autonomic nerve state information) in the processing unit 130.

  The processing unit 130 performs processing for obtaining pulse wave information based on the pulse wave sensor information. Specifically, the estimated pulse information is obtained, the parameters of the approximate expression are set, the pulse wave sensor information is approximated by the approximate expression set with the parameters, and the pulse wave information is obtained based on the result. Further, the processing unit 130 may obtain the autonomic nerve state information based on the pulse wave information. Note that the processing unit 130 may perform processing for displaying the obtained pulse wave information and the like. The process to be displayed here represents a process necessary for displaying on the display unit, and includes, for example, generation of display screen information, display control information, and transmission processing. Of course, in the case where the biological information processing system 100 includes a display unit, a process of displaying on the display unit may be included. Specific processing contents in the processing unit 130 will be described later.

  A specific implementation example of a system including the biological information processing system 100 according to the present embodiment will be described with reference to FIG. The system in FIG. 2A includes a server system 200, a terminal device 300, and a biological information detection device (biological information measurement device) 400, which are connected via a network NE. The network NE can include a public network such as a mobile communication network and an Internet network, or a fixed telephone network. The network NE can be realized by a WAN (Wide Area Network), a LAN (Local Area Network), or the like, regardless of wired or wireless.

  The biological information processing system 100 according to the present embodiment can be realized in various forms. For example, the terminal device 300 in FIG. 2A corresponds to the biological information processing system 100, and the biological information detection device 400 is a processing target. It may be a device used by a user. In this case, the server system 200 may be, for example, an apparatus that performs accumulation of the obtained pulse wave information or the like and processing with a relatively high load. As an example, the server system 200 is configured to be connectable to a plurality of biological information processing systems 100 (terminal devices 300), and is a device that acquires, accumulates, and analyzes a large amount of data from the plurality of terminal devices 300. Further, as shown in FIG. 2B, the server system 200 may be omitted.

  FIGS. 3A to 4 show examples of external views of a biological information detection apparatus 400 that collects biological information (pulse wave information in a narrow sense). The biological information detection apparatus 400 according to this embodiment includes a band unit 10, a case unit 30, and a sensor unit 40. The case part 30 is attached to the band part 10. The sensor unit 40 is provided in the case unit 30.

  The band unit 10 is for wrapping around the wrist of the user and mounting the biological information detection device 400. The band part 10 has a band hole 12 and a buckle part 14. The buckle portion 14 has a band insertion portion 15 and a projection portion 16. The user inserts one end side of the band unit 10 into the band insertion unit 15 of the buckle unit 14, and inserts the protrusion 16 of the buckle unit 14 into the band hole 12 of the band unit 10, whereby the biological information detection device 400 is configured. Wear on your wrist. In addition, the band part 10 is good also as a structure which has a buckle instead of the buckle part 14. FIG.

  The case part 30 corresponds to the main body part of the biological information detecting device 400. Various components of the biological information detection apparatus 400 such as the sensor unit 40 and a circuit board (not shown) are provided inside the case unit 30. That is, the case part 30 is a housing for housing these components.

  The case part 30 is provided with a light emitting window part 32. The light emitting window 32 is formed of a light transmissive member. The case portion 30 is provided with a light emitting portion as an interface mounted on a flexible substrate, and light from the light emitting portion is emitted to the outside of the case portion 30 through the light emitting window portion 32.

  The biological information detection apparatus 400 is worn on the user's wrist as shown in FIG. 2A and the like, and pulse wave information (biological information in a broad sense) is measured in the worn state. Specifically, the sensor unit 40 includes a photoelectric sensor having a light emitting unit and a light receiving unit, and measures pulse wave information using the photoelectric sensor. Note that measurement of pulse wave information using a photoelectric sensor is a well-known technique, and thus detailed description thereof is omitted. Further, the wearing part of the biological information detection apparatus 400 may be an ankle, a finger, an upper arm, or the like.

  The biological information detection device 400 is not limited to the band type (watch type) shown in FIGS. 3A to 4, but is a band type device worn on the chest, a device attached to the neck, etc., and a glasses type (face) Various devices such as a wearable device can be used. Furthermore, the sensor to be mounted is not limited to the photoelectric sensor, and various sensors such as an ultrasonic sensor, a body motion sensor (for example, an acceleration sensor), and a position sensor such as a GPS receiver can be used.

  The pulse wave sensor information detected by the biological information detection device 400 is transmitted to the terminal device 300 via the network NE and acquired by the information acquisition unit 110 of the terminal device 300 (biological information processing system 100). Alternatively, the biological information detection device 400 and the terminal device 300 may be connected by wire using a USB (Universal Serial Bus), a microUSB, or the like.

  Although the terminal device 300 is a small portable terminal such as a smartphone in FIG. 2A, the terminal device 300 is not limited to this and may be a stationary device such as a PC. The terminal device 300 includes a display unit in a narrow sense. Then, the pulse wave information and the autonomic nerve state information generated by the processing unit 130 may be displayed on the display unit of the terminal device 300.

  3A to 4 illustrate an example of a device that does not have a display unit as the biological information detection device 400. However, the present invention is not limited to this, and the biological information detection device 400 has a display unit. May be. As an example, it is a wristwatch type device that is worn on the wrist of the user in the same manner as in FIGS. 3A to 4, and a display unit such as an LCD or an organic EL display is arranged at a position corresponding to the dial of the watch. The device which was made can be considered.

  When the biological information detection device 400 includes a display unit, the pulse wave information obtained by the processing unit 130 may be displayed on the display unit of the biological information detection device 400. However, since the biological information detection device 400 is a device that is worn on the user's wrist as shown in FIG. 2A and the like, it is assumed that the area and resolution of the display unit are inferior to those of the terminal device 300. The For this reason, the processing unit 130 reduces the amount of information to be displayed compared to the case where it is displayed on the display unit of the terminal device 300, or a process in which display is performed in a form that is visually easy to understand using an icon (stamp) or the like. It is good to do.

  Moreover, both the terminal device 300 and the biological information detection device 400 may have a display unit, and pulse wave information or the like may be displayed on both. Moreover, the terminal device 300 is not limited to one, For example, the terminal device 300 can be modified using both a smartphone and a PC. Further, the server system 200 may include a display unit, and pulse wave information and the like may be displayed on the display unit of the server system 200.

  In the above description, the biological information processing system 100 according to the present embodiment is realized by the terminal device 300, but is not limited thereto. For example, the biological information processing system 100 according to the present embodiment may be realized by the server system 200 instead of the terminal device 300. A configuration example in this case is FIG. 2 (A) or FIG. 2 (B). The server system 200 often has a margin in processing performance, storage area, and battery capacity as compared with the terminal device 300 such as a PC, particularly the terminal device 300 realized as a mobile terminal device such as a smartphone. Therefore, by realizing the biological information processing system 100 with the server system 200, it is possible to perform processing for obtaining pulse wave information at high speed (or for more pulse wave sensor information).

  The network NE may include short-range wireless communication (NFC) such as Bluetooth (registered trademark). Therefore, an embodiment in which the biological information detection device 400 and the terminal device 300 are connected by short-range wireless communication and the terminal device 300 is connected to the server system 200 by the Internet or the like is also conceivable. That is, when the pulse wave sensor information from the biological information detection device 400 is acquired by the biological information processing system 100 that is the server system 200, communication via another device such as the terminal device 300 may be performed. Specifically, the pulse wave sensor information is transmitted from the biological information detection device 400 to the terminal device 300 by short-range wireless communication, and the pulse wave sensor information is transmitted to the server system 200 via the Internet or the like. Is possible. FIG. 2C illustrates such a specific example.

  The server system 200 may be configured by one server device (processing device), but is not limited thereto, and may be realized as a set of a plurality of server devices. For example, the server system 200 may be a set of database servers, application servers, and the like. In that case, since each server apparatus should just be able to operate | move in relation with each other, it is not necessary to be provided in the physically close position, The form arrange | positioned in the distant position and connected by network NE may be sufficient. .

  Further, the biological information processing system 100 according to the present embodiment may be realized by a device other than the terminal device 300 and the server system 200. For example, when improvement of terminal performance or usage pattern is taken into consideration, an embodiment in which the biological information processing system 100 according to the present embodiment is realized by the biological information detection device 400 is not denied. In this case, the information acquisition unit 110 acquires pulse wave sensor information from the pulse wave sensor 42 (for example, the sensor unit 40 in FIG. 4) in the same apparatus. When the biological information processing system 100 is realized by the biological information detection device 400, the biological information processing system 100 uses the biological information detection device 400 because it is less necessary to calculate pulse wave information for a large number of users. One or a small number of users may be targeted. In other words, the possibility of satisfying the user's request is also conceivable in the processing performance of the biological information detecting apparatus 400.

  In the above description, the biological information processing system 100 is realized by any one of the server system 200, the terminal device 300, and the biological information detection device 400. However, the present invention is not limited to this. For example, acquisition of pulse wave sensor information and calculation of pulse wave information based on the acquired pulse wave sensor information may be realized by distributed processing of a plurality of devices. Specifically, the biological information processing system 100 may be realized by at least two of the server system 200, the terminal device 300, and the biological information detection device 400. Alternatively, the processing according to the present embodiment may be performed in another device, and the biological information processing system 100 according to the present embodiment can be realized by various devices (or combinations of devices).

  Further, the method of the present embodiment is not limited to that applied to the biological information processing system 100. For example, the method of the present embodiment causes the computer to function as an information acquisition unit 110 that acquires pulse wave sensor information from the pulse wave sensor 42 and a processing unit 130 that obtains pulse wave information based on the pulse wave sensor information. The processing unit 130 obtains estimated pulse information, sets parameters of an approximate expression that approximates the waveform of the pulse wave specified from the pulse wave sensor information based on the obtained estimated pulse information, and sets the approximate expression in which the parameters are set. Based on this, it can be applied to a program for obtaining pulse wave information.

  For example, this program is installed in the terminal device 300 of FIG. 2A and can be used when pulse wave information is obtained using the terminal device 300 (when the terminal device 300 is operated as the biological information processing system 100). is there. As an example, the program according to the present embodiment may be an application that operates on the terminal device 300 that is a smartphone. Further, this program may be one that is installed in the server system 200 and operates the server system 200, or one that is installed in the biological information detection device 400 and operates the biological information detection device 400 (firmware in a narrow sense). There may be one installed in a plurality of devices.

3. Pulse Wave Information Calculation Processing Details of a method for obtaining pulse wave information according to this embodiment will be described next. Specifically, the basic flow of the processing of the present embodiment will be described by explaining processing for obtaining estimated pulse information and setting parameters of an approximate expression, and peak detection processing based on the approximate expression. In addition, as ancillary processing according to the present embodiment, parameter resetting processing and detected peak discard determination processing will be described, and finally the flow of processing according to the present embodiment using the flowcharts of FIGS. 8 to 12. Will be explained.

3.1 Calculation of Estimated Pulse Information and Parameter Setting First, details of processing for obtaining estimated pulse information and processing for setting parameters of an approximate expression based on the obtained estimated pulse information will be described. As described above, the estimated pulse information is information (estimated value) obtained by estimating the pulse information. The pulse information here is information related to the pulse, such as the pulse wave RR interval, the pulse frequency, and the pulse rate.

  The estimated pulse information does not need to be highly accurate as described above. Therefore, for the estimated pulse information, it is possible to widely apply a technique already known as a technique for obtaining the pulse information.

  For example, the information acquisition unit 110 may acquire body motion sensor information from a body motion sensor, and the processing unit 130 may obtain estimated pulse information based on the body motion sensor information. Here, the body motion sensor may be included in, for example, the biological information detection apparatus 400, and specifically, can be realized by various sensors such as an acceleration sensor, a gyro sensor, a geomagnetic sensor, and an atmospheric pressure sensor. Here, an example in which an acceleration sensor is used as a body motion sensor will be described.

  Specifically, the processing unit 130 may obtain the user's exercise intensity information based on the body motion sensor information, and may obtain the estimated pulse information based on the exercise intensity information, the maximum pulse information, and the resting pulse information. .

Conventionally, it is known that there is a correlation between the oxygen intake VO2 and the heart rate HR. For example, when the maximum VO2 is VO2m, the resting VO2 is VO2r, the resting pulse information (resting heart rate) is HRr, and the maximum pulse information (maximum heart rate) is HRm, the following equation (1) The relationship is known. However, the minute oxygen intake is used in the left side of the following formula (1).

  Also, since it is known that the exercise intensity is obtained as a ratio to the maximum oxygen intake VO2m, the exercise intensity information is obtained from the body motion sensor information, and the heart rate HR is calculated by applying the above equation (1) or the like. It is possible to estimate. Various methods such as the method disclosed in Patent Document 2 are known as a method for obtaining exercise intensity from an acceleration signal and obtaining an approximate heart rate, and these methods can be widely applied in the present embodiment. . If the estimated heart rate is obtained as the estimated pulse information, the conversion to the estimated pulse wave RR interval or the like is easy.

  However, the method for obtaining the estimated pulse information is not limited to this. For example, the processing unit 130 may perform frequency conversion processing on the pulse wave sensor information and obtain estimated pulse information based on the result of the frequency conversion processing.

  When frequency conversion processing is performed on pulse wave sensor information, a peak appears ideally only at the frequency corresponding to the frequency of the pulse, so it is only necessary to estimate the frequency of the peak as the pulse frequency. The period corresponding to the frequency may be the estimated pulse wave RR interval. However, since various noises are superimposed on the pulse wave sensor information as described above, a plurality of peaks may be detected after frequency conversion. In that case, the estimated pulse information may be obtained by performing a known noise reduction process. Specifically, body motion information is acquired, and the frequency spectrum of the body motion information is obtained by frequency-converting the body motion information, and the difference between the frequency spectrum of the pulse wave sensor information and the frequency spectrum of the body motion information is obtained. Processing may be performed. If it does in this way, it will become possible to reduce the noise resulting from a body motion from pulse wave sensor information, and to obtain suitable presumed pulse information. Note that various methods are known for noise reduction processing when performing frequency conversion processing, such as using an adaptive filter instead of simply taking a difference, and can be widely applied in the present embodiment.

  In addition, considering that it is possible to detect that inappropriate parameter settings have been made, such as peak discard determination processing described later, emphasis is placed on reducing processing load rather than accuracy when obtaining estimated pulse information. The processing may be performed. For example, the processing unit 130 may obtain the estimated pulse information based on the pulse wave information obtained before the processing target timing.

  At the timing except immediately after the start of the process according to the present embodiment, some pulse wave information has already been obtained by the method according to the present embodiment (peak detection based on the approximate expression). Since the pulse wave information is considered to be a relatively reliable value, the estimated pulse information may be obtained from the pulse wave information. As an example, an average value of n pieces of pulse wave information obtained in the past from the processing target timing may be used as estimated pulse information. Specifically, a process for obtaining the estimated pulse wave RR interval from each pulse wave information and setting the average value as the estimated pulse wave RR interval can be considered.

  The estimated pulse information is not limited to that obtained by one method, and a plurality of methods may be combined. For example, an embodiment in which the estimated pulse information is obtained by frequency conversion processing at a certain timing and the estimated pulse information is obtained from an average of past pulse wave information at other timings is possible.

  When the estimated pulse information is obtained, the processing unit 130 sets parameters of an approximate expression. The parameter here is a parameter for setting an application range of the approximate expression for the pulse wave sensor information. As described above, in an ideal state in which noise is not superimposed, the pulse wave sensor information has a waveform close to a sine wave, although the fluctuation of the pulse rate and the fluctuation of the DC component itself occur. That is, in the approximation of the pulse wave sensor information, an order that can approximate the sine wave may be used, and a value such as 2 or 3 is sufficient for the order p. On the contrary, if the order p is increased, the processing load for obtaining the approximate expression increases. This is also because the approximation formula becomes higher order and the coefficient to be determined (p + 1 to pth order 0 + 1) increases.

  In any case, if the pulse wave sensor information is approximated as in the present embodiment, the necessity to set the order p variably is not so high. Therefore, although setting the order p as a parameter of the approximate expression is not hindered, the parameter having a higher importance is the application range of the approximate expression.

  More specifically, the processing unit 130 sets the value of the frame size m as a parameter based on the estimated pulse information, and targets the pulse wave sensor information in the frame corresponding to the set frame size m, An approximate expression is obtained by applying a smoothing polynomial filter, and pulse wave information is obtained based on the obtained approximate expression.

  Here, the smoothing polynomial filter may be, for example, a Savitzky-Golay smoothing filter (SG filter). In the Savitzky-Golay smoothing filter, a set (p, m) of the frame size m and the order p is used as a control parameter of the filter. In the present embodiment, the frame size m may be set based on the estimated pulse information. Since the technique using the Savitzky-Golay smoothing filter (Savitzky-Golay method, Savitzky-Golay algorithm) is a widely known technique, a detailed description thereof will be omitted. In this embodiment, a general least square method or the like may be used without using the Savitzky-Golay smoothing filter, and various modifications can be made to the method for obtaining the approximate expression.

  Here, the frame corresponding to the frame size m may be a frame of 2m + 1 samples including pulse wave sensor information at the processing target timing and pulse wave sensor information of m samples before and after the processing target timing. . In that case, the processing unit 130 obtains an estimated pulse wave RR interval that is an estimated value of the pulse wave RR interval as the estimated pulse information, and sets the estimated pulse wave RR interval as a frame of 2m + 1 samples represented by the frame size m. Set the frame corresponding to the length within the given numerical range including the length.

As described above, since p = 3 (or p = 2) is assumed in the present embodiment, it is desirable that the signal for one cycle of the waveform of the pulse wave be an application range of the approximate expression. Therefore, in the case of the frame size m, if the applicable range of the approximate expression is 2m + 1 samples, the length corresponding to 2m + 1 samples may correspond to the length of the estimated pulse wave RR interval. As an example, the frame size m may be set so as to satisfy the following expression (2). In the following equation (2), RRI represents an estimated pulse wave RR interval (msec), and Fs represents a sampling rate for acquiring pulse wave sensor information.

In the above equation (2), (RRI × FS) / 1000 represents the number of samples included in the time of the length of the estimated pulse wave RR interval. That is, when the above equation (2) is used, the frame size m is set so that the application range of the approximate equation is 2/3 times or more and 5/4 times or less the estimated pulse wave RR interval. Become. However, the setting of the frame size m is not limited to the above equation (2), and the following equation (3) may be used. The coefficients α and β in the following expression (3) are values that satisfy 0 <α <1 <β.

Note that, when the parameter m is set, the application range (frame length) of the approximate expression is a range of 2m + 1 samples, which is a general technique in the Savitzky-Golay smoothing filter. Since this is based on the definition of the parameter m, if the total number of samples included in the application range is defined as the parameter M of the approximate expression, M may be set so as to satisfy the following expression (4). In addition, since digital filter processing (FIR smoothing filter processing) is assumed here, determination is performed in units of samples. However, if the sampling rate Fs is known, conversion between time and number of samples is easy. It is. Therefore, it is possible to perform modifications such as using a parameter representing time as a parameter for setting the application range of the approximate expression.

  Specific numerical examples will be described with reference to FIGS. 5 (A) to 7 (B). In the following example, the sampling rate Fs is 100 Hz. A1 in FIG. 5A represents the acquired pulse wave sensor information, and the vertical line of A2 represents the detection timing of the R wave obtained from the electrocardiogram. In FIG. 5A, the horizontal axis represents time (sampling timing), and the vertical axis represents a signal value. Since the accuracy of the R wave detection timing of the electrocardiogram of A2 is considered to be high, it may be considered that appropriate processing can be performed if peak detection is performed at the timing corresponding to A2 based on the pulse wave sensor information. it can. Further, as can be seen from FIG. 5A, the pulse wave sensor information has false peaks such as A4 to A6, and it is necessary not to detect these as peaks.

  In the example of FIG. 5A, it is known from the result of the electrocardiogram that the average pulse rate in the range shown in the figure is 82 (bpm). When this is converted to the RR interval, RRI = 60 × 1000/82 = 731.7 (msec). If this RRI is substituted into the above equation (2), the frame size m becomes 24 ≦ m ≦ 45. Hereinafter, it will be described using a specific example that appropriate peak detection can be performed when m within this range is used, and proper peak detection cannot be performed when m outside the range is used.

  FIG. 5A shows an example in which m = 25. In this case, the peak (upper peak) obtained by the peak detection process described later is indicated by A3. As can be seen from FIG. 5A, the R wave timing A2 obtained from the electrocardiogram and the upper peak A3 obtained from the pulse wave sensor information by applying a smoothing polynomial filter are in good agreement. Moreover, there is no case where A4 to A6 etc. are erroneously detected as peaks. As a result, as shown in FIG. 5 (B), the obtained pulse wave RR interval is also close to the RR interval obtained from the electrocardiogram, and in the examples of FIGS. 5 (A) and 5 (B), its absolute error is The average can be suppressed to about 30 msec. That is, when the frame size m within the range obtained from the above equation (2) is used, appropriate processing is possible.

  On the other hand, FIG. 6A shows an example in which m = 15 is set. As in FIG. 5A, B1 is pulse wave sensor information, B2 is R wave detection timing, and B3 is peak detection timing based on pulse wave sensor information. As can be seen from FIG. 6 (A), it is possible to detect the detection timing of the R wave as a peak, but a false peak caused by noise has been detected (B4 to B6). As described above, this is caused by the fact that the approximate expression follows the signal change caused by noise. That is, since an inappropriate point has been detected as a peak, the difference between the pulse wave RR interval obtained as shown in FIG. 6B and the RR interval obtained from the electrocardiogram is large. I can't say that.

  FIG. 7A shows an example in which m = 50 is set. Similarly to FIG. 5A, C1 is pulse wave sensor information, C2 is R wave detection timing, and C3 is peak detection timing based on pulse wave sensor information. As can be seen from FIG. 7A, a false peak due to noise is not detected, but there is a location where the detection timing of the R wave cannot be detected as a peak (C4). As described above, this is due to the fact that the approximation error is large because the application range of the approximate expression is wider than one period of the pulse wave and one or more pulse wave peaks are included. As a result, the difference between the pulse wave RR interval obtained as shown in FIG. 7B and the RR interval obtained from the electrocardiogram is large, which is not an appropriate process.

  As described above, it is possible to perform appropriate processing by setting the frame size m using the above equation (2) or the like.

3.2 Peak Detection Processing With the processing described above, it is possible to set an appropriate frame size m based on the estimated pulse information and obtain an approximate expression. If an approximate expression is obtained, pulse wave information may be obtained based on the approximate expression. Various types of pulse wave information can be considered here. Simply, it may be a pulse wave waveform itself with reduced noise, or other information such as a heart rate (bpm) or a pulse signal frequency (Hz) obtained based on the waveform. Also good.

  For example, the processing unit 130 may obtain pulse wave RR interval information as the pulse wave information. By obtaining the pulse wave RR interval, it becomes possible to obtain autonomic nerve state information and the like as will be described later. The pulse wave RR interval is information corresponding to the period of the pulse wave waveform, and various methods for obtaining the period are conceivable. The pulse wave RR interval information in the present embodiment is not limited to the pulse wave RR interval itself, and may be other information obtained from the pulse wave RR interval. For example, the pulse wave RR interval information may be distribution information of the pulse wave RR interval. The distribution information of the pulse wave RR interval may be, for example, information representing time-series fluctuation of the pulse wave RR interval, or information representing variation in the value of the pulse wave RR interval within a predetermined period. Also good.

  The processing unit 130 may obtain the pulse wave RR interval based on a peak detection process of the waveform represented by the approximate expression or a process using a differential signal of the waveform represented by the approximate expression. Specifically, if the peak detection processing of the waveform represented by the approximate expression is performed, the peak interval becomes a time representing the period, and thus the peak interval can be set as the pulse wave RR interval. However, there is no problem as long as the sine wave does not superimpose noise at all, but in the case of a complex waveform, the search for the peak may not be easy, and general peak detection is performed using a differential signal. In this embodiment, the influence of noise is suppressed by obtaining an approximate expression, but even in that case, the change in the value at the peak position becomes clear by using the differential signal, so that more accurate peak detection is possible. It becomes. Note that the pulse wave information according to the present embodiment includes a waveform peak represented by an approximate expression in a broad sense.

  That is, the processing unit 130 may obtain a differential signal of the pulse wave sensor information (approximate expression) after applying the smoothing polynomial filter, and may obtain the pulse wave information based on the differential signal. As an example, when an approximate expression is obtained with the pulse wave sensor information for 2m + 1 samples centered on a given processing target timing as an application range, the above-mentioned given processing target is obtained by differentiating the approximate expression. What is necessary is just to obtain | require the differential signal value in a timing. In this case, since one differential signal value is obtained from one approximate expression, if the differential signal value is obtained at each sampling timing, the differential signal is acquired as a set of the differential signal values.

  When performing peak detection based on the differential signal, a point where the value of the differential signal is 0 (zero cross point) may be detected. The zero cross point at which the differential signal changes from positive to negative corresponds to the upper peak in the original signal, and the zero cross point at which the differential signal changes from negative to positive corresponds to the lower peak in the original signal. Therefore, for example, if the upper peak is detected as shown in FIG. 5A, a zero cross point where the differential signal changes from positive to negative may be searched.

  Alternatively, since it is only necessary to obtain the period of the waveform represented by the approximate expression, the interval between the zero cross points in the approximate expression may be set as the pulse wave RR interval. In that case, the zero cross point of the approximate expression may be searched directly, or the peak of the differential signal may be detected.

  In summary, the pulse wave RR interval is obtained by searching for the peak or zero cross point of the waveform represented by the approximate expression, or by searching for the zero cross point or peak of the differential waveform of the waveform represented by the approximate expression. be able to.

  In the above description, a primary differential signal is assumed as the differential signal. However, the present invention is not limited to this, and modifications using secondary differentiation, tertiary differentiation, and the like are possible. In this case, if the upper peak is to be obtained, the determination may be made under the condition that it is a zero cross point of the third derivative and the second derivative becomes a negative minimum value.

3.3 Parameter Reset As described above, in the present embodiment, an appropriate parameter (frame size m in a narrow sense) is set based on the estimated pulse information. Therefore, it is considered that the calculation load can be reduced as compared with the method disclosed in Patent Document 1. However, as described above, the appropriate frame size m is a value corresponding to the RR interval at the processing target timing, and it is known that the human RR interval varies greatly depending on the state. For example, it is known that the RR interval varies in the range of about 300 msec to 1500 msec.

  In other words, if the technique as in Patent Document 1 in which omnidirectional processing is performed using a plurality of frame sizes m is not used, processing for changing the frame size m in accordance with the change in the RR interval is necessary. become. If the estimated pulse information is obtained at all processing target timings (at all timings for obtaining the approximate expression) and the frame size m is obtained again, it is possible to cope with the fluctuation of the RR interval. However, such processing is not preferable in consideration of reduction in processing load. That is, it is preferable to change the frame size m at a predetermined condition and timing, but disclosure of the timing and condition for changing the frame size has not been found in the conventional method.

  Therefore, the present embodiment proposes a method of changing the approximate expression parameter (frame size m) under appropriate conditions and timing. Specifically, the information acquisition unit 110 acquires body motion sensor information from the body motion sensor, and the processing unit 130 obtains the user's behavior pattern based on the body motion sensor information, and the behavior pattern has changed. If it is determined, parameter resetting processing may be performed.

  If the user's behavior pattern is known, the heart rate (pulse rate) can be estimated to some extent, so the value of the RR interval can also be estimated to some extent. For example, the RR interval becomes a very large value in the sleep state, whereas the RR interval becomes a very small value when exercising vigorously. Furthermore, if you consider a resting state that is awakening but very little movement, a state of life movement that performs activities in daily life, a state of relatively light exercise such as relaxing walking, etc., For example, the RR interval is assumed to be intense exercise <light exercise <daily motion <rest state <sleep state. In any case, if the user's behavior pattern is analyzed and the result changes, it can be considered that the RR interval also changes.

  Therefore, when it is determined that the behavior pattern has changed, the parameter resetting process may be performed because there is a possibility that the behavior cannot be handled by the frame size m until then. Conversely, if the behavior pattern has not changed, it is considered that the previous frame size m can be used continuously. That is, by setting whether or not the behavior pattern has changed, it is possible to reset (change) the parameters with appropriate conditions and timing. Specifically, since it is possible to prevent the use of an inappropriate parameter as it is, it is possible to ensure the processing accuracy and to prevent the resetting process from being performed unnecessarily, and thus it is possible to suppress an increase in processing load.

  As a behavior pattern analysis method based on body motion sensor information, it is possible to widely apply a known method. For example, as disclosed in Patent Document 3, an action pattern may be analyzed based on acceleration sensor information. In Patent Document 3, in an acceleration sensor having two or more axes (three axes in a broad sense), one of the axes is determined as a main axis, and rest, life movement, running, walking, muscle strength are determined based on information on the main axis. Analyzes behavior patterns such as training, cycling, gymnastics and other exercises.

  The action pattern here is information that can estimate the degree of the RR interval. In other words, the motion state may be analyzed finely as in Patent Document 3, but the present invention is not limited to this, and a simplified behavior pattern analysis may be performed. For example, if it is known that the exercise intensity classified into the exercise classified into the first behavior pattern and the exercise classified into the second behavior pattern are about the same, any of the first and second behavior patterns However, the RR interval is estimated to be similar. In that case, instead of dividing the first and second behavior patterns into two, the processing may be performed in one.

3.4 Deletion (Destroy) Detection of Detected Peak In addition, the peak obtained based on the approximate expression (or pulse wave information such as the pulse wave RR interval obtained from the peak or zero cross point) is adopted as it is. It is not limited to a form, You may perform the determination process of employ | adopting or discarding.

  Specifically, the processing unit 130 determines the pulse wave (original waveform) specified from the pulse wave sensor information and the pulse wave specified from the approximate expression obtained based on the estimated pulse information (the waveform represented by the approximate expression). ), And when the error is larger than a given threshold, the obtained pulse wave information (peak in a narrow sense) is deleted.

  As described above with reference to FIGS. 7A and 7B, when the frame size m is excessively large, the accuracy of the detected peak position is low, and the required pulse wave RR interval also has a large error. . This is an error that occurs because the application range of the approximate expression is wider than one period of the pulse wave, and a waveform including one or more pulse wave peaks is forcibly approximated by a p-order polynomial. In such a case, the pulse wave specified from the obtained approximate expression has a large error from the waveform of the original pulse wave sensor information. That is, when this error is excessively large, it can be estimated that the set parameter is inappropriate, or in a narrow sense, the set frame size m is excessively large.

  In that case, since the reliability of the obtained peak is low, if the peak is adopted as it is, the reliability of the pulse wave information obtained from the peak is also lowered as described above. Therefore, in this case, the obtained peak may be deleted (discarded) without being used for the subsequent processing. In this way, it is possible to suppress outputting information with low accuracy. As the error between the two waveforms, the square sum (or the square root) of the difference between the values at each sampling timing may be used. In addition, since the peak with low reliability is considered not to cause a problem unless it is used for the subsequent processing, the data itself may not be completely deleted and may be held in the storage area. However, it can be said that it is desirable to delete the above-mentioned low-reliability peak in consideration of the possibility that data with low reliability may be erroneously used for processing, or considering the effective use of the storage area.

  As described above with reference to FIGS. 6A and 6B, when the frame size m is excessively small, a false peak is detected because the difference between the original waveform and the approximate expression becomes too small. Resulting in. That is, the determination of whether or not the error is large is basically to detect that the frame size m is too large with respect to the RR interval, and that the frame size m is too small with respect to the RR interval. It is not assumed to be detected.

  Further, the detected peak may be simply discarded, and the processing at the processing target timing may be terminated (shift to the processing at the next processing target timing), but the processing of the present embodiment is not limited to this. For example, the processing unit 130 may perform the parameter resetting process when the error is larger than a given threshold value. And it becomes possible to obtain | require the peak and pulse wave information in a process target timing accurately by calculating | requiring a peak again by the reset parameter.

  As described above, when the error is large, it is considered that the frame size m is too large with respect to the RR interval. Therefore, the resetting process here does not need to obtain the above-described estimated pulse information again, and may be a process of reducing the already obtained frame size m.

  Moreover, even if the same process is performed again as the process for obtaining the estimated pulse information, the obtained estimated pulse information is the same, and the range of the frame size m obtained from the above equation (2) is also the same. If the estimated pulse information itself is largely incorrect, there is a possibility that an appropriate value will not be obtained even if m is switched within the range obtained from the above equation (2).

  Therefore, when the parameter is reset by recalculating the estimated pulse information, the estimated pulse information may be obtained by a method different from the previous method. As an example, if a parameter is set from estimated pulse information obtained based on body motion sensor information and the error is larger than a given threshold, the reset processing performs estimation by performing frequency conversion processing of the pulse wave sensor information Processing such as obtaining pulse information and setting parameters may be performed.

3.5 Flowchart The processing flow of this embodiment will be described with reference to FIGS. FIG. 8 is a flowchart for explaining processing for obtaining pulse wave information (pulse wave RR interval in a narrow sense) according to the present embodiment. When this process is started, the information acquisition unit 110 first acquires pulse wave sensor information (S101). Then, the estimated pulse information is obtained by some method (S102).

  FIG. 9 shows an example of a flowchart for explaining the processing for obtaining the estimated pulse information in S102. FIG. 9 corresponds to a method of obtaining exercise intensity from body motion sensor information of a body motion sensor (acceleration sensor) and obtaining estimated pulse information from the exercise intensity. Specifically, first, acceleration information is acquired as body motion sensor information (S201), and a body motion amount is calculated based on the acceleration information (S202). The amount of body movement is a value such as the sum of absolute values of acceleration, the sum of squares, and the like. Then, the exercise intensity is obtained based on the obtained amount of body movement (S203). It is known that there is a correlation between the amount of body movement and the value of exercise intensity. For example, the method disclosed in Patent Document 2 may be used.

  Further, the heart rate HR is estimated using the exercise intensity, the maximum heart rate (HRm), and the resting heart rate (HRr) (S204). For example, the above equation (1) may be used for S204, and this point is also described in detail in Patent Document 2. Here, since it is assumed that the estimated value of the RR interval is used as the estimated pulse information, the heart rate (estimated heart rate) obtained in S204 is converted into the RR interval to be estimated pulse information (S205).

  Further, as described above, the estimated pulse information may be obtained by other methods. FIG. 10 shows another example of the flowchart for explaining the process for obtaining the estimated pulse information. FIG. 10 corresponds to a technique for obtaining estimated pulse information by performing frequency conversion processing on pulse information for a predetermined period. Specifically, first, pulse wave sensor information for a predetermined period is acquired (S301). The process of S301 may be a process of reading out the pulse wave sensor information acquired before the processing target timing and stored in the storage unit 120 or the like, for example. The predetermined period here is, for example, 16 seconds as described above. Then, frequency conversion processing (FFT in a narrow sense) is performed on the pulse wave sensor information obtained in S301 for the predetermined period, and the average pulse rate in the predetermined period is estimated. Specifically, the peak of information after frequency conversion may be detected, and the processing corresponding to the frequency corresponding to the peak may be performed. Further, the average pulse rate obtained in S302 is converted into an RR interval to obtain estimated pulse information (S303).

  The process for obtaining the estimated pulse information shown in S102 of FIG. 8 is executed by the processes of FIGS. Returning to FIG. After the process of S102, parameters of the approximate expression (parameters of the smoothing polynomial filter) are set using the obtained estimated pulse information (S103). Specifically, the estimated pulse information may be converted into the RR interval fluctuation range using the above equation (2) and the like, and a value within the range may be set as a parameter.

  Next, pulse wave information, here, the pulse wave RR interval is calculated using the set parameters (S104). FIG. 11 is a flowchart for explaining specific processing of S104. As shown in FIG. 11, first, a differential signal of pulse wave sensor information is obtained using the smoothing polynomial filter whose parameters are set in S103, and peak detection is performed using the differential signal (S401). Then, the time corresponding to the peak interval is obtained as the pulse wave RR interval (S402).

  After the process of S104, the pulse wave sensor information is acquired again (S105). Note that S105 is acquisition of pulse wave sensor information that is assumed to be used for processing after the next timing, and the same processing as S101 may be performed.

  Then, a behavior pattern is detected based on information from the acceleration sensor (S106). As for the specific detection process, as described above, the technique disclosed in Patent Document 3 may be used. It is determined whether or not the detected behavior pattern has changed from the previous (immediately previous) behavior pattern (S107). In the case of No in S107, that is, when the behavior pattern has not changed, it is assumed that the fluctuation of the RR interval is small, so even if the approximate expression parameter obtained in the process of S103 is continuously used. Good. Therefore, without resetting the parameters, the process returns to S104 and the calculation process of the pulse wave RR interval is executed.

  On the other hand, in the case of Yes in S107, that is, when the behavior pattern is changed, it is assumed that the RR interval varies greatly, and therefore the parameters of the approximate expression obtained in the process of S103 cannot be continuously used. Therefore, after returning to S102, calculation of the estimated pulse information (S102) and parameter reset (S103) are performed, and then the calculation process of the pulse wave RR interval is executed (S104).

  In the present embodiment, as described above, it may be determined whether or not to discard the peak obtained based on the differential signal. FIG. 12 is a flowchart for explaining the process. FIG. 12 shows a process corresponding to S104 of FIG. 8, and FIG. 12 may be performed instead of FIG. 11 as the process of S104.

  When this processing is started, first, a differential signal of pulse wave sensor information is obtained using the smoothing polynomial filter set in S103 as in S401, and peak detection is performed using the differential signal (S501). . Thereafter, an error between the original waveform of the pulse wave sensor information and the waveform of the approximate expression obtained in S501 is obtained (S502). Then, it is determined whether or not the obtained error is larger than a given threshold value (S503).

  In the case of Yes in S503, it is determined that the parameter set in S103 is not appropriate (specifically, too large with respect to the RR interval), and the peak detected in S501 is not adopted and discarded ( S504). On the other hand, in the case of No in S503, the peak detected in S501 is adopted, and the RR interval is calculated using the peak (S505).

4). Determination of Autonomic Nerve Activity As described above, the processing unit 130 of the present embodiment may estimate the autonomic nerve state information based on the pulse wave information obtained using the approximate expression. The pulse wave information here is a pulse wave RR interval in a narrow sense.

  Autonomic nerves include sympathetic nerves and parasympathetic nerves, and there are fluctuations in the activity state in each day and season. In general, sympathetic nerves dominate during daytime activities, and parasympathetic nerves dominate during nighttime activities. In the season, the sympathetic nerve is dominant from autumn to winter, and the parasympathetic nerve is dominant from spring to summer.

  In order to determine the autonomic nervous activity state from the pulse wave information, first, the pulse wave RR interval (pulse interval) is measured over a certain period of time, so that the time series of the pulse wave RR interval as shown in FIG. Specific data. The pulse wave RR interval is not always constant and varies (fluctuates). It is known that the fluctuation is caused by the activity of the sympathetic nerve and the activity of the parasympathetic nerve, and it is also known that the degree of fluctuation due to the activity of the sympathetic nerve is different from the degree of fluctuation due to the activity of the parasympathetic nerve.

  Therefore, frequency conversion is performed on time-series data of the pulse wave RR interval. An example of data after frequency conversion is shown in FIG. As can be seen from FIG. 13B, a relatively low frequency peak LF and a relatively high frequency peak HF are obtained from the data after frequency conversion.

  LF represents a slow change in the pulse wave RR interval, and mainly reflects both sympathetic and parasympathetic activities. On the other hand, HF represents a quick change of the pulse wave RR interval, and mainly reflects the activity of the parasympathetic nerve.

  If such characteristics are taken into account, the ratio of LF to HF (for example, the ratio of the signal intensity at each peak) is obtained, so that the sympathetic nerve is dominant or the parasympathetic nerve is dominant in the measurement period of the pulse wave information. It is possible to determine whether there is.

  Although the state of the autonomic nerve can be estimated in this way, various usage scenes of information on the estimated autonomic nerve state are conceivable. As an example, the sleep state may be determined from LF and HF, and specifically, the value of LF / HF may be used. LF / HF is larger as the sympathetic nerve is dominant, and becomes smaller as the parasympathetic nerve is dominant. Therefore, the first threshold value Th1 is set, and if LF / HF> Th1, the sympathetic nerve is dominant and thus determined to be awake, and if LF / HF ≦ Th1, the parasympathetic nerve is dominant and determined to be the sleeping state. The method of doing is conceivable.

  In addition, the second threshold Th2 (<Th1) is set for the sleep state, and if TH1 ≧ LF / HF> Th2, the sympathetic nerve is relatively dominant in the sleep state, so that the sleep is shallow (REM sleep) State), and if LF / HF ≦ Th2, the parasympathetic nerve is dominant and the sleep state may be determined to be deep (non-REM sleep state). Moreover, if the determination of the stage in a non-REM sleep state is performed, the determination may be performed by providing a further threshold value in the non-REM sleep state.

  Alternatively, in a disease such as an atrial cell, the response to the disease may differ depending on whether the disease is caused by a sympathetic nerve or a parasympathetic nerve (vagus nerve as an autonomic nervous system). . In that case, it can be said that it is very useful in treatment to identify which nerve the disease is caused by. That is, it is considered possible to facilitate classification of sympathetic nerve-dependent atrial fibrillation and vagus nerve-dependent atrial fibrillation by acquiring information on the activity state of the user's autonomic nerve.

  Although the present embodiment has been described in detail as described above, it will be easily understood by those skilled in the art that many modifications can be made without departing from the novel matters and effects of the present invention. Accordingly, all such modifications are intended to be included in the scope of the present invention. For example, a term described at least once together with a different term having a broader meaning or the same meaning in the specification or the drawings can be replaced with the different term in any part of the specification or the drawings. Further, the configuration and operation of the biological information processing system and the like are not limited to those described in the present embodiment, and various modifications can be made.

10 band part, 12 band hole, 14 buckle part, 15 band insertion part,
16 Projection part, 30 Case part, 32 Light emission window part, 40 Sensor part,
42 pulse wave sensor, 100 biological information processing system, 110 information acquisition unit,
120 storage units, 130 processing units, 200 server systems, 300 terminal devices,
400 Biological information detection device, NE network

Claims (17)

An information acquisition unit for acquiring pulse wave sensor information from the pulse wave sensor;
A processing unit for obtaining pulse wave information based on the pulse wave sensor information;
Including
The processor is
Estimated pulse information is obtained, and based on the obtained estimated pulse information, an approximate expression parameter that approximates the waveform of the pulse wave specified from the pulse wave sensor information is set, and the parameter is set based on the approximate expression. A biological information processing system for obtaining the pulse wave information. In claim 1,
The information acquisition unit
Get body motion sensor information from the body motion sensor,
The processor is
A biological information processing system, wherein a user's behavior pattern is obtained based on the body motion sensor information, and the parameter resetting process is performed when it is determined that the behavior pattern has changed. In claim 1 or 2,
The processor is
An error of a pulse wave specified from the pulse wave sensor information and an error of the pulse wave specified from the approximate expression obtained based on the estimated pulse information is obtained, and obtained when the error is larger than a given threshold value. A biological information processing system, wherein the received pulse wave information is deleted. In claim 3,
The processor is
The biological information processing system, wherein the parameter resetting process is performed when the error is larger than the given threshold. In any one of Claims 1 thru | or 4,
The biological information processing system, wherein the parameter is a parameter for setting an application range of the approximate expression for the pulse wave sensor information. In any one of Claims 1 thru | or 5,
The processor is
A biological information processing system characterized by obtaining pulse wave RR interval information as the pulse wave information. In claim 6,
The processor is
A biological information processing system for obtaining the pulse wave RR interval information based on a peak detection process of a waveform represented by the approximate expression or a process using a differential signal of a waveform represented by the approximate expression . In any one of Claims 1 thru | or 7,
The processor is
A biological information processing system that estimates autonomic nerve state information based on the pulse wave information obtained based on the approximate expression. In any one of Claims 1 thru | or 8.
The processor is
Based on the estimated pulse information, a value of the frame size m is set as the parameter, and a smoothing polynomial filter is applied to the pulse wave sensor information in the frame corresponding to the set frame size m. The biological information processing system is characterized in that the approximate expression is obtained and the pulse wave information is obtained based on the approximate expression. In claim 9,
The frame corresponding to the frame size m is a frame for 2m + 1 samples consisting of the pulse wave sensor information at the processing target timing and the pulse wave sensor information for each m samples before and after the processing target timing,
The processor is
As the estimated pulse information, an estimated pulse wave RR interval that is an estimated value of the pulse wave RR interval is obtained, and the length of the estimated pulse wave RR interval is included as a frame of 2m + 1 samples represented by the frame size m. A biological information processing system characterized by setting a frame corresponding to a length within a given numerical range. In claim 9 or 10,
The processor is
A biological information processing system, wherein a differential signal of the pulse wave sensor information after application of the smoothing polynomial filter is obtained, and the pulse wave information is obtained based on the differential signal. In any one of Claims 1 thru | or 11,
The information acquisition unit
Get body motion sensor information from the body motion sensor,
The processor is
The biological information processing system, wherein the estimated pulse information is obtained based on the body motion sensor information. In claim 12,
The processor is
A biological information processing system that obtains exercise intensity information of a user based on the body motion sensor information and obtains the estimated pulse information based on the exercise intensity information, maximum pulse information, and resting pulse information. In any one of Claims 1 thru | or 11,
The processor is
A biological information processing system, wherein a frequency conversion process is performed on the pulse wave sensor information, and the estimated pulse information is obtained based on a result of the frequency conversion process. In any one of Claims 1 thru | or 11,
The processor is
A biological information processing system characterized in that the estimated pulse information is obtained based on the pulse wave information obtained before the processing target timing. An information acquisition unit for acquiring pulse wave sensor information from the pulse wave sensor;
As a processing unit for obtaining pulse wave information based on the pulse wave sensor information,
Make the computer work,
The processor is
Estimated pulse information is obtained, and based on the obtained estimated pulse information, an approximate expression parameter that approximates the waveform of the pulse wave specified from the pulse wave sensor information is set, and the parameter is set based on the approximate expression. A program for obtaining the pulse wave information. Obtain pulse wave sensor information from the pulse wave sensor,
Find estimated pulse information,
Based on the obtained estimated pulse information, set an approximate expression parameter that approximates the waveform of the pulse wave specified from the pulse wave sensor information,
Find pulse wave information based on the parameterized approximate expression.
A control method for a biological information processing system.