Blood pressure measurement device, blood pressure measurement method, and blood pressure measurement program

Abstract

The present invention stops pressurization of a cuff at a desired timing. This blood pressure measurement device comprises: a cuff wound around a part being measured; a pressure detection unit for detecting a cuff pressure in the cuff; a pressure control unit for controlling the cuff pressure; a pulse wave acquisition unit for acquiring, from the cuff pressure, a pulse wave of a person being measured; a blood pressure calculation unit for calculating the blood pressure of the person being measured on the basis of the pulse wave acquired until the cuff pressure reaches a stop cuff pressure at which pressurization of the cuff is stopped; a reference cuff pressure estimation unit for estimating, on the basis of a feature of the pulse wave acquired at the cuff pressure lower than the stop cuff pressure, a reference cuff pressure that is the value of the cuff pressure at which the amplitude of the pulse wave becomes maximized; and a stop cuff pressure determination unit for determining the value of the stop cuff pressure on the basis of the reference cuff pressure.

Description

Blood pressure measurement device, blood pressure measurement method, and blood pressure measurement program

[0001] The present invention relates to a blood pressure measurement device, a blood pressure measurement method, and a blood pressure measurement program.

[0002] In recent years, health management has become commonplace by measuring information about an individual's physical and health, such as blood pressure values, using a measuring device and recording and analyzing the measurement results. One example of such a measuring device is a sphygmomanometer, which measures blood pressure, including systolic blood pressure, based on a pressure pulse wave acquired during the process of inflating a cuff attached to a measurement part, such as the upper arm or wrist, of the person being measured (see, for example, Patent Document 1).

[0003] The blood pressure monitor described in Patent Document 1 allows the user to select between a low inflation mode and a full inflation mode. In the low inflation mode, inflation is stopped when the pulse wave amplitude starts to decrease from its maximum value, and the systolic blood pressure is calculated from the relationship between the diastolic blood pressure and the mean blood pressure.

[0004] However, in order to actually measure the peak of the envelope of the pulse wave as in the technology described in Patent Document 1, it is necessary to acquire the pulse waveform up to the point where the envelope exceeds the peak, which delays the timing at which the cuff inflation is stopped.

[0005] US Patent Application Publication No. 2010 / 0298726 JP 03-280932

[0006] In view of the above-described conventional techniques, an object of the present invention is to provide a technique that can stop pressurizing the cuff at a desired timing.

[0007] In order to solve the above-mentioned problems, the present invention provides a blood pressure measurement device comprising: a cuff to be wrapped around a part to be measured; a pressure detection unit to detect cuff pressure in the cuff; a pressure control unit to control the cuff pressure; a pulse wave acquisition unit to acquire a pulse wave of the person to be measured from the cuff pressure; a blood pressure calculation unit to calculate the blood pressure of the person to be measured based on the pulse waves acquired before the cuff pressure reaches a stop cuff pressure at which inflation of the cuff is stopped; a reference cuff pressure estimation unit to estimate a reference cuff pressure, which is the cuff pressure value at which the amplitude of the pulse wave is maximum, based on characteristics of the pulse wave acquired at a cuff pressure lower than the stop cuff pressure; and a stop cuff pressure determination unit to determine the value of the stop cuff pressure based on the reference cuff pressure.

[0008] This estimates the reference cuff pressure, which is the cuff pressure at which the amplitude of the pulse wave is maximum, i.e., the cuff pressure at which the envelope curve peaks, and determines the value of the stop cuff pressure based on the estimated reference cuff pressure.Therefore, the stop cuff pressure can be determined so that the application of cuff pressure can be stopped at the desired timing without relying on actual measurement of the peak of the envelope curve.

[0009] In the present invention, the reference cuff pressure estimation unit may estimate the reference cuff pressure based on the cuff pressure corresponding to the diastolic blood pressure of the subject.

[0010] This makes it possible to determine the stop cuff pressure so that the application of cuff pressure can be stopped at a desired timing without relying on the actual measurement of the peak of the envelope.

[0011] Furthermore, in the present invention, an envelope calculation unit may be provided that calculates an envelope of the pulse wave, and the reference cuff pressure estimation unit may have a mathematical model that mathematically expresses the envelope of the pulse wave, and may estimate the reference cuff pressure based on the mathematical model that expresses an estimated envelope that includes the calculated envelope.

[0012] This allows the peak of the envelope to be estimated based on a mathematical model that represents an estimated envelope including the calculated envelope, and therefore allows the stop cuff pressure to be determined so that the cuff pressure can be stopped at the desired timing without relying on actual measurement of the peak of the envelope.

[0013] Furthermore, in the present invention, the reference cuff pressure estimation unit may include: a learning model generated by machine learning the pulse wave or the envelope of the pulse wave relative to the cuff pressure and an index representing a relative position on the envelope as learning data; an index estimation unit that estimates the index at the time of estimation by inputting the pulse wave or the envelope of the pulse wave acquired up to the time of estimation into the learning model; and a determination unit that determines whether the index at the time of estimation has reached the position corresponding to the reference cuff pressure, and the stop cuff pressure determination unit may determine the stop cuff pressure based on the index at the time of estimation.

[0014] This allows the peak of the envelope to be estimated using an index that represents the relative position of the envelope, so that the stop cuff pressure can be determined so that the cuff pressure can be stopped at the desired timing without relying on actual measurement of the peak of the envelope.

[0015] Furthermore, in the present invention, the reference cuff pressure estimation unit may include: a learning model generated by machine learning the pulse wave or the envelope of the pulse wave relative to the cuff pressure and an index indicating a relative position on the envelope as learning data; an index estimation unit that estimates the index at the time of estimation by inputting the pulse wave or the envelope of the pulse wave acquired up to the time of estimation into the learning model; a recording unit that records the cuff pressure corresponding to the index at the time of estimation; and a regression line estimation unit that estimates a regression line indicating the relationship between the recorded index and the cuff pressure, and the reference cuff pressure may be estimated based on the regression line.

[0016] This allows the peak of the envelope to be estimated using a regression line estimated using an index representing the relative position of the envelope, so that the stop cuff pressure can be determined so that the cuff pressure can be stopped at the desired timing without relying on actual measurement of the peak of the envelope.

[0017] The present invention also provides a blood pressure measurement method comprising the steps of: detecting a cuff pressure in a cuff wrapped around a part to be measured; controlling the cuff pressure; acquiring a pulse wave of the person to be measured from the cuff pressure; calculating a blood pressure of the person to be measured based on the pulse wave acquired until the cuff pressure reaches a stop cuff pressure at which inflation of the cuff is stopped; estimating a reference cuff pressure, which is the cuff pressure value at which the amplitude of the pulse wave is maximum, based on characteristics of the pulse wave acquired at a cuff pressure lower than the stop cuff pressure; and determining a value of the stop cuff pressure based on the reference cuff pressure.

[0018] This estimates the reference cuff pressure, which is the cuff pressure at which the amplitude of the pulse wave is maximum, i.e., the cuff pressure at which the envelope curve peaks, and determines the value of the stop cuff pressure based on the estimated reference cuff pressure.Therefore, the stop cuff pressure can be determined so that the application of cuff pressure can be stopped at the desired timing without relying on actual measurement of the peak of the envelope curve.

[0019] In the present invention, the reference cuff pressure may be estimated based on the cuff pressure corresponding to the diastolic blood pressure of the subject.

[0020] This makes it possible to determine the stop cuff pressure so that the application of cuff pressure can be stopped at a desired timing without relying on the actual measurement of the peak of the envelope.

[0021] Furthermore, in the present invention, the method may include a step of calculating an envelope of the pulse wave, and the step of estimating the reference cuff pressure may estimate the reference cuff pressure by expressing an estimated envelope including the calculated envelope using a mathematical model that mathematically expresses the envelope of the pulse wave.

[0022] This allows the peak of the envelope to be estimated based on a mathematical model that represents an estimated envelope including the calculated envelope, and therefore allows the stop cuff pressure to be determined so that the cuff pressure can be stopped at the desired timing without relying on actual measurement of the peak of the envelope.

[0023] Furthermore, the present invention includes the steps of: estimating the index at the time of estimation by inputting the pulse wave or the envelope of the pulse wave acquired up to the time of estimation into a learning model generated by machine learning the pulse wave or the envelope of the pulse wave with respect to the cuff pressure and an index representing a relative position on the envelope as learning data; and determining whether the index at the time of estimation has reached the position corresponding to the reference cuff pressure, wherein the step of determining the stop cuff pressure may determine the stop cuff pressure based on the index at the time of estimation.

[0024] This allows the peak of the envelope to be estimated using an index that represents the relative position of the envelope, so that the stop cuff pressure can be determined so that the cuff pressure can be stopped at the desired timing without relying on actual measurement of the peak of the envelope.

[0025] Furthermore, in the present invention, the step of estimating the reference cuff pressure may include the steps of: estimating the index at the time of estimation by inputting the pulse wave or the envelope of the pulse wave acquired up to the time of estimation into a learning model generated by machine learning the pulse wave or the envelope of the pulse wave relative to the cuff pressure and an index indicating a relative position on the envelope as learning data; recording the cuff pressure corresponding to the index at the time of estimation; and estimating a regression line indicating the relationship between the recorded index and the cuff pressure, and the reference cuff pressure may be estimated based on the regression line.

[0026] This allows the peak of the envelope to be estimated using a regression line estimated using an index representing the relative position of the envelope, so that the stop cuff pressure can be determined so that the cuff pressure can be stopped at the desired timing without relying on actual measurement of the peak of the envelope.

[0027] The present invention also provides a blood pressure measurement program that causes a computer to execute the following steps: detecting the cuff pressure in a cuff wrapped around a part to be measured; controlling the cuff pressure; acquiring the pulse wave of the person to be measured from the cuff pressure; calculating the blood pressure of the person to be measured based on the pulse wave acquired until the cuff pressure reaches a stop cuff pressure at which inflation of the cuff is stopped; estimating a reference cuff pressure, which is the cuff pressure value at which the amplitude of the pulse wave is maximum, based on characteristics of the pulse wave acquired at a cuff pressure lower than the stop cuff pressure; and determining the value of the stop cuff pressure based on the reference cuff pressure.

[0028] This estimates the reference cuff pressure, which is the cuff pressure at which the amplitude of the pulse wave is maximum, i.e., the cuff pressure at which the envelope curve peaks, and determines the value of the stop cuff pressure based on the estimated reference cuff pressure.Therefore, the stop cuff pressure can be determined so that the application of cuff pressure can be stopped at the desired timing without relying on actual measurement of the peak of the envelope curve.

[0029] According to the present invention, it is possible to stop pressurizing the cuff at a desired timing.

[0030] FIG. 1 is a diagram illustrating an outline of the hardware configuration of a blood pressure measurement device according to a first embodiment. FIG. 2 is a functional block diagram of the blood pressure measurement device according to the first embodiment. FIG. 3 is a schematic diagram illustrating the relationship between cuff pressure and a pressure pulse wave in the blood pressure measurement device according to the first embodiment. FIG. 4 is a flowchart illustrating the overall processing procedure of the blood pressure measurement device according to the first embodiment. FIG. 5 is a flowchart illustrating the peak position estimation processing procedure of the blood pressure measurement device according to the first embodiment. FIGS. 6A, 6B, and 6C are diagrams illustrating indices associated with one pulse wave in the blood pressure measurement device according to the first embodiment. FIG. 7 is a functional block diagram of a blood pressure measurement device according to the second embodiment. FIG. 8 is a flowchart illustrating the peak position estimation processing procedure of the blood pressure measurement device according to the second embodiment. FIG. 9 is a diagram illustrating the peak position estimation processing in the blood pressure measurement device according to the second embodiment. FIG. 10 is a diagram illustrating the peak position estimation processing in the blood pressure measurement device according to the third embodiment. FIG. 11 is a diagram illustrating a learning model of the blood pressure measurement device according to the third embodiment. FIG. 12 is a functional block diagram of the blood pressure measurement device according to the third embodiment. Fig. 13 is a flowchart illustrating the procedure of peak position estimation processing in the blood pressure measurement device according to Example 3. Fig. 14 is a diagram illustrating the peak estimation processing in the blood pressure measurement device according to Example 4. Fig. 15 is a functional block diagram of the blood pressure measurement device according to Example 4. Fig. 16 is a flowchart illustrating the procedure of peak position estimation processing in the blood pressure measurement device according to Example 4.

[0031] Hereinafter, specific embodiments of the present invention will be described with reference to the drawings.

[0032] Example 1 An example of an embodiment of the present invention will be described below. However, unless otherwise specified, the dimensions, materials, shapes, relative positions, etc. of the components described in this example are not intended to limit the scope of the present invention to those only.

[0033] (Device Configuration) FIG. 1 is a schematic diagram of the hardware configuration of a blood pressure measurement device 1 according to this embodiment.

[0034] The blood pressure measurement device 1 includes a cuff 11, a pressure sensor 12, a pressure pump 13, an exhaust valve 14, an air tube 15, an oscillator circuit 21, a pump drive circuit 22, a valve drive circuit 23, a display unit 25, a memory 24, an operation switch 26, a power supply 27, and a CPU 100. The blood pressure measurement device 1 corresponds to the blood pressure measurement device of the present invention. The cuff 11 corresponds to the cuff of the present invention. The blood pressure measurement device 1 having the CPU 100 and the memory 24 corresponds to the computer of the present invention.

[0035] The cuff 11 includes an air bag 11a containing air. The cuff 11 is provided with a pressure sensor 12 for detecting the pressure in the air bag 11a of the cuff 11 (hereinafter referred to as "cuff pressure") via an air tube 15, a pressure pump 13 for supplying air to the air bag 11a, and an exhaust valve 14 that opens and closes to maintain the pressure in the air bag 11a or to exhaust the air from the air bag 11a.

[0036] Furthermore, the blood pressure measuring device 1 includes a CPU (Central Processing Unit) 100 for controlling each part of the device, a memory 24 for storing a program executed for determining an inflation stop position (to be described later), a program executed for blood pressure measurement processing, and data such as cuff pressure, pressure pulse wave, and blood pressure measurement results, a display unit 25 for displaying various information such as blood pressure measurement results, an operation switch 26 for inputting various instructions for measurement, and a power supply 27 for supplying power to each part of the device, such as the CPU.

[0037] Furthermore, the oscillator circuit 21 outputs a signal having an oscillation frequency corresponding to the output value of the pressure sensor 12 to the CPU 100. The pump drive circuit 22 controls the drive of the pressure pump 13 based on a control signal output from the CPU 100. The valve drive circuit 23 controls the opening and closing of the exhaust valve 14 based on the control signal output from the CPU 100.

[0038] 2 is a functional block diagram of the CPU 100 of the blood pressure measurement device 1. The CPU 100 includes a pressure detection unit 110, a pressure control unit 120, an SBP calculation unit 130, a DBP calculation unit 140, a peak position estimation unit 150, and a stop cuff pressure calculation unit 160. The pressure detection unit 110, the pressure control unit 120, the SBP calculation unit 130, and the DBP calculation unit 140 correspond to the pressure detection unit, the pressure control unit, and the blood pressure calculation unit, respectively, of the present invention. The peak position estimation unit 150 corresponds to the reference cuff pressure estimation unit, and the stop cuff pressure calculation unit 160 corresponds to the stop cuff pressure determination unit of the present invention.

[0039] The output signal of the oscillator circuit 21 is input to the pressure detection unit 110. The pressure detection unit 110 detects the oscillation frequency of the input signal and converts the detected oscillation frequency into a pressure value signal. The pressure detection unit 110 includes an HPF (High Pass Filter) unit 111 that applies HPF (High Pass Filter) processing to the pressure value signal to extract and output a pressure pulse wave signal, and an LPF (Low Pass Filter) unit 112 that applies LPF (Low Pass Filter) processing to the pressure value signal to extract and output a cuff pressure signal. The pressure pulse wave signal detected in time series by the HPF unit 111 of the pressure detection unit 110 and the cuff pressure signal indicating the cuff pressure detected in time series by the LPF unit 112 are stored in a predetermined area of ​​the memory 24. Here, the HPF unit 111 corresponds to a pulse wave acquisition unit of the present invention.

[0040] 3 is a graph schematically illustrating the relationship between the cuff pressure and the pressure pulse wave detected by the pressure detection unit 110. Generally, as shown in FIG. 3, a diastolic blood pressure (DBP) and a systolic blood pressure (SBP) are calculated as cuff pressures corresponding to specific fluctuation patterns of the pressure pulse wave, and when the cuff 11 is inflated, the diastolic blood pressure is calculated, and after a peak occurs in the envelope of the pressure pulse wave, the systolic blood pressure is calculated at a higher cuff pressure.

[0041] The pressure control unit 120 controls the operation of the pump drive circuit 22 and the valve drive circuit 23 to control the cuff pressure of the cuff 11 .

[0042] The DBP calculation unit 140 and the SBP calculation unit 130 receive the pressure pulse wave signal extracted by the HPF unit 111 of the pressure detection unit 110 and process the received pressure pulse wave signal to calculate the diastolic blood pressure (minimum blood pressure, DBP) and the systolic blood pressure (maximum blood pressure, SBP). The DBP and SBP calculation process will be described later.

[0043] The peak position estimation unit 150 estimates the peak position where the envelope of the acquired pressure pulse wave is at its maximum based on the calculated DBP. Here, the peak position is the cuff pressure corresponding to the peak of the envelope when inflation control of the cuff 11 is performed according to a predetermined algorithm. For example, when the cuff pressure is plotted on the horizontal axis and the amplitude of the pressure pulse wave is plotted on the vertical axis, the peak position is represented by the position on the horizontal axis where the envelope of the pressure pulse wave reaches its peak. The peak position estimation process will be described later.

[0044] The stop cuff pressure calculation unit 160 determines the value of the stop cuff pressure, which is the cuff pressure at which inflation of the cuff 11 is stopped, from the peak position estimated by the peak position estimation unit 150. The stop cuff pressure determination process will be described later.

[0045] (Blood pressure measurement method) Fig. 4 is a flowchart showing the overall processing procedure of the blood pressure measurement method by the blood pressure measurement device 1. The overall processing of blood pressure measurement shown in Fig. 4 is stored in advance in a predetermined area of ​​the memory 24 as a blood pressure measurement program, and is realized by the CPU 100 reading and executing the program from the memory 24. The program may be stored in a computer-readable storage medium and read into the blood pressure measurement device 1 from the storage medium.

[0046] When measuring blood pressure, the subject wraps the cuff 11 around the part to be measured in advance. In the following description, an example is given in which the part to be measured is the upper arm, but the part to be measured is not limited to this and may be the wrist, etc. The description also assumes that the subject performs predetermined settings using the operation switch 26 and issues a command to start blood pressure measurement. Upon receiving the command to start blood pressure measurement, the blood pressure measurement device 1 performs predetermined initialization, such as opening the exhaust valve 14 and setting the cuff pressure to atmospheric pressure (initial pressure).

[0047] When blood pressure measurement is started, the pressure control unit 120 starts pressure control to pressurize the cuff 11 (step S1).

[0048] During the process of pressure control, the HPF unit 111 of the pressure detection unit 110 acquires a pressure pulse wave (step S2).

[0049] Next, the peak position estimation unit 150 estimates the peak position (step S3). Details of estimating the peak position of the envelope of the pressure pulse wave will be described later.

[0050] In step S3, when the peak position of the envelope of the pressure pulse wave is estimated, the stop cuff pressure calculation unit 160 determines the value of the stop cuff pressure based on this peak position (step S4). When the cuff pressure corresponding to the peak of the envelope of the pressure pulse wave (also referred to as the "peak cuff pressure") is Pp and the stop cuff pressure is P1, for example, P1=f １ The relationship between the rest cuff pressure value and the peak cuff pressure can be expressed by a predetermined relational expression, such as (Pp).

[0051] The stop cuff pressure P1 can be set to any appropriate value. However, as described below, when the cuff pressure reaches the stop cuff pressure P1, compression of the measurement site by the cuff 11 is stopped. From the perspective of reducing the burden on the subject, it is desirable to terminate blood pressure measurement earlier after inflation has begun, i.e., by inflating to a lower cuff pressure. On the other hand, from the perspective of ensuring blood pressure measurement accuracy, it is desirable to terminate blood pressure measurement later after inflation has begun, i.e., after inflation to a higher cuff pressure, since this allows for more cuff pressure (pressure pulse wave) data to be acquired. As shown in FIG. 3 , DBP can be detected before the peak in the envelope of the pressure pulse wave, i.e., at a cuff pressure lower than the peak cuff pressure. Furthermore, the peak in the envelope of the pressure pulse wave occurs before the SBP, i.e., the peak cuff pressure is lower than the cuff pressure corresponding to the SBP. Therefore, an appropriate stop cuff pressure P1 is set taking these factors into consideration.

[0052] When the value of the stop cuff pressure is determined in step S4, the pressure control unit 120 continues to control the increase of the cuff pressure (step S5) and determines whether the cuff pressure detected by the pressure detection unit 110 is greater than the stop cuff pressure P1 (step S6).

[0053] When the pressure control unit 120 determines that the cuff pressure has become greater than the stop cuff pressure P1, it controls the pump drive circuit 22 to stop pressurizing the cuff 11 (step S7), and controls the valve drive circuit 23 to open the exhaust valve 14 and discharge the air from the air bag 11a.

[0054] The SBP calculation unit 130 calculates blood pressure, such as SBP, based on the pressure pulse wave acquired before the cuff pressure reaches the stop cuff pressure P1 (step S8). Here, the SBP is estimated based on the pressure pulse wave acquired before the cuff pressure reaches the stop cuff pressure P1 using an estimation model included in the blood pressure measurement program. This estimation model uses, for example, pressure pulse waves detected from the subject while the cuff 11 is inflated until the amplitude of the pressure pulse wave reaches its maximum as training samples, and is machine-learned so that the estimated values ​​obtained by the estimation model match the true blood pressure values ​​(here, SBP) when these training samples are detected. The pressure pulses detected from the subject while the cuff 11 is inflated until the amplitude of the pressure pulse wave reaches its maximum may also be used as training samples. The training samples may include at least one of envelope data and amplitude data of the pressure pulse wave, or may include cuff pressure data and beat-by-beat features of the pressure pulse wave. The beat-by-beat feature of the pressure pulse wave may be, for example, a statistical quantity such as a maximum value, a minimum value, a variance, a standard deviation, an n% tile value, a skewness, or a kurtosis, or may be the RAV, WID, or DFN described above. The machine learning model constituting such an estimation model may be, for example, a neural network, a regression model, a decision tree model, a support vector machine, or other functional formulas (computational models). The machine learning method may be appropriately selected depending on the employed machine learning model (e.g., backpropagation). The estimation model may be configured to directly derive an estimated value of SBP. Alternatively, the estimation model may be configured to indirectly derive an estimated value of SBP by predicting the time point of SBP on the assumption that the increase in cuff pressure follows a predetermined condition (e.g., is constant). In this case, the estimated value of SBP can be derived by calculating the value of cuff pressure at the time point predicted by the estimation model.

[0055] The CPU 100 displays the measurement results, such as the SBP calculated in step S8, on the display unit 25 of the blood pressure measurement device 1 (step S9). The CPU 100 records the measurement results, such as the calculated blood pressure value, in a predetermined area of ​​the memory 24 of the blood pressure measurement device 1, and ends the blood pressure measurement process.

[0056] (Peak Position Estimation Method) Fig. 5 shows a flowchart illustrating the processing procedure of the peak position estimation method according to Example 1. In the flowchart of Fig. 5, the same processes as those in the flowchart of Fig. 4 are denoted by the same reference numerals.

[0057] Here, the DBP calculation unit 140 calculates the DBP from information on the detected pressure pulse wave (step S301). Specifically, the DBP calculation unit 140 determines whether the cuff pressure being inflated has reached the cuff pressure corresponding to the DBP, using an index obtained based on the pressure pulse wave information. The DBP calculation unit 140 continues inflation and further acquires the cuff pressure and pressure pulse wave until the cuff pressure reaches the cuff pressure corresponding to the DBP (step S2), and repeats the process of step S301. The DBP calculation unit 140 then calculates the cuff pressure value corresponding to the DBP as the DBP and stores it in a predetermined area of ​​the memory 24.

[0058] As shown in Patent Document 2, RAV, WID, and DFN are known as indices obtained based on pressure pulse wave information. These indices are calculated for each beat of the pressure pulse wave, and RAV, WID, and DFN respectively represent the area, width, and slope of the pulse wave for one beat. FIG. 6(A) illustrates RAV, FIG. 6(B) illustrates WID, and FIG. 6(C) illustrates DFN. The waveform in FIG. 6(A) shows a pulse wave for one beat. RAV is the pulse wave area for each beat indicated by the diagonal lines in FIG. 6(A) normalized by amplitude, and is expressed as (pulse wave area / pulse wave amplitude within one beat)×100. WID is the waveform width, which is the time width from the maximum amplitude (Maximum) to a threshold (Threshold), normalized by the pulse wave period, and is expressed as (waveform width / pulse wave period)×100, as shown in FIG. 6(B). As shown in FIG. 6C, DFN is the minimum value of the first derivative of the pressure pulse wave normalized by the pulse wave amplitude, and is expressed as (minimum value from 0 of the first derivative of the pulse wave / pulse wave amplitude of the first derivative of the pulse wave).

[0059] When the cuff pressure at which DBP occurs is Pd, Pd = f ２It is expressed as a function of RAV, WID, and DFN, such as (RAV, WID, DFN). Here, Pd is expressed as a function of three indices, RAV, WID, and DFN, but it may be expressed as any one or any two indices. ２ Although the form of these indices is not particularly limited, the DBP calculation unit 140 determines whether the cuff pressure has reached the DBP by calculating these indices for each beat of the acquired pressure pulse wave. For example, it can be determined that the cuff pressure has reached the DBP when the RAV or WID reaches a minimum value or the DFN reaches a maximum value.

[0060] Next, the predicted peak position is estimated based on the DBP (step S302). When the peak cuff pressure is Pp, Pp=f ３ The relationship between the cuff pressure value at which DBP occurs and the peak cuff pressure is expressed by a predetermined relational expression, such as (Pd). ３ The form of Pp is not particularly limited, but for example, Pp=αPd may be expressed as Pp=αPd, where α is a constant. In this way, the peak cuff pressure can be estimated in advance based on the cuff pressure corresponding to the DBP calculated in step S301.

[0061] Based on the peak cuff pressure thus estimated, in step S4, the stop cuff pressure calculation section 160 determines the value of the stop cuff pressure as described above.

[0062] Here, the processing in steps S301 and S302 corresponds to the peak position estimation processing in step S3 in the flowchart of FIG.

[0063] In this way, it is possible to stop pressurizing the cuff 11 at a desired timing based on the peak cuff pressure.

[0064] Example 2 The overall configuration of a blood pressure measurement device 2 according to Example 2 is similar to that of the blood pressure measurement device 1 according to Example 1 shown in Fig. 1. The overall processing procedure of the blood pressure measurement method according to Example 2 is also the same as that described in the flowchart shown in Fig. 4, but the details of the peak position estimation processing in step S3 are different. Differences from Example 1 are described below.

[0065] 7 is a functional block diagram of the CPU 100A of the blood pressure measurement device 2. The blood pressure measurement device 2 having the CPU 100A and the memory 24 corresponds to the computer of the present invention.

[0066] The CPU 100A includes a pressure detection unit 110, a pressure control unit 120, a blood pressure calculation unit 130A, an envelope calculation unit 170, a fitting unit 180, a peak position estimation unit 150A, and a stop cuff pressure calculation unit 160. A mathematical model 242 is stored in the memory 24. The functions of the pressure detection unit 110, the pressure control unit 120, and the stop cuff pressure calculation unit 160 are the same as those in the first embodiment, and therefore description thereof will be omitted. In the second embodiment, the CPU 100A includes a blood pressure calculation unit 130A instead of the SBP calculation unit 130 and the DBP calculation unit 140 shown in FIG. 2 . Here, the envelope calculation unit 170 corresponds to the envelope calculation unit of the present invention, and the fitting unit 180 and the peak position estimation unit 150A correspond to the reference cuff pressure estimation unit of the present invention. Furthermore, the blood pressure calculation unit 130A corresponds to the blood pressure calculation unit of the present invention.

[0067] The envelope calculation unit 170 receives the pressure pulse wave signal extracted by the HPF unit 111 of the pressure detection unit 110 and calculates the envelope of the received pressure pulse wave.

[0068] The fitting section 180 reads out the mathematical model stored in the memory 24, performs fitting with the envelope calculated by the envelope calculation section 170, and outputs a fitted waveform that fits the envelope. The fitting process will be described later.

[0069] The peak position estimation unit 150A estimates the peak position from the fitted waveform obtained by the fitting unit 180. The peak position estimation process will be described later.

[0070] The stop cuff pressure calculation unit 160 determines the value of the stop cuff pressure, which is the cuff pressure at which inflation of the cuff 11 is stopped, based on the peak position estimated by the peak position estimation unit 150A. The stop cuff pressure determination process will be described later.

[0071] (Peak Position Estimation Method) Fig. 8 is a flowchart illustrating the processing steps of a peak position estimation method according to Example 2. In the flowchart of Fig. 8, processes common to the overall processing steps of blood pressure measurement shown in Fig. 4 are denoted by the same reference numerals. Fig. 9 is a diagram illustrating the peak position estimation method according to Example 2, and is a graph with the cuff pressure on the horizontal axis and the amplitude of the pressure pulse wave on the vertical axis. Note that while Fig. 3 shows time on the horizontal axis, because cuff pressure and time are correlated, the envelope may also be expressed on a Cartesian coordinate plane with the cuff pressure on the horizontal axis and the amplitude of the pressure pulse wave on the vertical axis, as shown in Fig. 9.

[0072] In this step, the envelope calculation unit 170 calculates the envelope of the pressure pulse wave from the pressure pulse wave signal extracted by the HPF unit 111 of the pressure detection unit 110 (step S311). As shown in Fig. 9 , the envelope Ec is calculated based on the pressure pulse wave Pw extracted by the HPF unit 111 of the pressure detection unit 110.

[0073] Next, the fitting unit 180 reads out the mathematical model 242 stored in the memory 24 and performs fitting with the envelope calculated by the envelope calculation unit 170 (step S312). The mathematical model 242 used for fitting is not limited, but for example, a Gaussian including fit parameters A, B, and C shown below can be used. Such f ４ The fit parameters A, B, and C of the mathematical model 242 expressed as follows are adjusted to fit the envelope Ec and the fitted waveform Fc of the mathematical model.

[0074] Then, a predicted peak position is calculated based on the fitted waveform Fc, which is the fitted mathematical model 242 (step S313). Here, the fitted waveform Fc corresponds to the estimated envelope of the present invention. In this case, the mathematical model equation f ４ Differentiate and However, as described above, when a Gaussian is used as the mathematical model, the fit parameter B corresponds to the peak cuff pressure, and therefore, this value can be used as the predicted peak position Fp.

[0075] Furthermore, it is determined whether the reliability of the predicted peak position calculated in step S313 is OK, i.e., the reliability of the predicted peak position is determined (step S314). Although the method for determining the reliability is not limited, for example, the degree of fit of the mathematical model 242 used for fitting can be calculated using RMSE shown below, and the reciprocal of this calculation can be defined as the reliability. Here, for a certain cuff pressure, the pressure pulse wave amplitude value calculated by the mathematical model 242 is denoted as y, and the detected value of the pressure pulse wave amplitude is denoted as y ｍ The RMSE is calculated by squaring the differences between the cuff pressures for multiple sampling points, summing them, dividing by the number of samplings, and then taking the square root. A threshold is set for this RMSE, and the reliability is determined by determining whether the reciprocal of the RMSE of the fitted waveform Fc is equal to or greater than the predetermined threshold and whether the reliability is acceptable if it is less than the predetermined threshold. If the reliability of the predicted peak position is determined to be acceptable in step S314, the predicted peak position calculated in step S313 is used as the estimated value of the peak position, and the process proceeds to step S4. If the reliability of the predicted peak position is determined to be unacceptable in step S314, the process returns to step S2, where inflation control of the cuff 11 is continued, and the processes of steps S311 to S314 are repeated.

[0076] Here, the processes of steps S311 to S314 correspond to the peak position estimation process of step S3 in the flowchart of FIG. 4. The processes from determining the stop cuff pressure (step S4) to stopping inflation of the cuff 11 (step S7) are the same as those described in the first embodiment. In the second embodiment, in step S8, the blood pressure calculation unit 130A calculates blood pressure values ​​(DBP and SBP) based on the pressure pulse wave acquired before the cuff pressure reaches the stop cuff pressure. The blood pressure calculation unit 130A acquires the DBP and SBP based on the pressure pulse wave acquired before the cuff pressure reaches the stop cuff pressure, using the estimation model described for the SBP calculation unit 130 in the first embodiment. The processes from step S9 onwards are the same as those described in the first embodiment.

[0077] In this way, it is possible to stop pressurizing the cuff 11 at a desired timing based on the peak cuff pressure.

[0078] Example 3 The overall configuration of a blood pressure measurement device 3 according to Example 3 is similar to that of the blood pressure measurement device 1 according to Example 1 shown in Fig. 1. The overall processing procedure of the blood pressure measurement method according to Example 3 is also the same as that described in the flowchart shown in Fig. 4, but the details of the peak position estimation processing in step S3 and the blood pressure calculation processing in step S8 are different from those in Example 1. The differences from Example 1 are described below.

[0079] In the third embodiment, a learning model that outputs the relative position of the envelope of the pressure pulse wave based on the acquired cuff pressure and pressure pulse wave is used to estimate the relative position of the envelope of the pressure pulse wave and determine whether the estimated position exceeds the peak of the envelope. Then, the inflation stop position of the cuff 11 is determined based on the peak position thus determined.

[0080] Here, the relative position of the envelope (referred to as the "envelope relative position") is a relative position defined with respect to the entire envelope corresponding to the process in which the amplitude of the pressure pulse wave gradually increases after inflation of the cuff 11 starts, passes through a peak where the amplitude is maximum, and then gradually decreases, as shown in Figure 10(A). Figures 10(A) and 10(B) are diagrams showing the relationship between the pressure pulse wave Pw, its envelope Ec, and the envelope relative position, with the horizontal axis representing the envelope relative position and the vertical axis representing the amplitude of the pressure pulse wave.

[0081] The index representing this envelope relative position is not limited. For example, as shown in FIG. 10A, if the envelope relative position corresponding to the diastolic blood pressure on the envelope curve Ec is defined as −1 and the envelope relative position corresponding to the systolic blood pressure is defined as 1, the envelope relative position corresponding to the peak can be approximately 0. By defining the envelope relative position in this manner, any position on the envelope curve Ec can be represented by the envelope relative position. As shown in FIG. 10B, the unshaded portion represents the pressure pulse wave and its envelope acquired after inflation of the cuff 11 began. If the current time is the position indicated by a circle on the envelope curve Ec, the feature values ​​extracted from the acquired pressure pulse wave and its envelope are input into a learning model (described later) to output an estimate of the current envelope relative position.

[0082] 11 is a diagram illustrating learning of a learning model used in Example 3 and estimation using a trained learning model. During machine learning of the learning model, the entire data of the pressure pulse wave, i.e., the data of the entire process from the start of inflation of the cuff 11, when the amplitude of the pressure pulse wave gradually increases, passes through a peak where the amplitude is maximum, and then gradually decreases, is defined, along with the relative position of the envelope calculated for this pressure pulse wave. The relative position defined in this way is the true value of the relative position, and the pressure pulse wave data for which the relative position of the envelope is defined is used as training data when training the learning model.

[0083] At this time, the input pressure pulse wave data is subjected to preprocessing such as noise removal and envelope calculation. Then, feature quantities related to the envelope and feature quantities related to the pressure pulse wave are extracted from the preprocessed data. At this time, the feature quantities related to the pressure pulse wave include feature quantities for each beat and feature quantities of multiple pressure pulse waves. By inputting the extracted feature quantities into a machine learning model, an estimated value of the relative position of the envelope is output. The machine learning model is trained based on the estimated value of the relative position of the envelope thus obtained and the true value. The machine learning model is not limited, and examples that can be used include multiple regression, random forest, neural network, support vector regression, Lasso regression, Ridge regression, and naive Bayes.

[0084] When estimating the relative position of the envelope using the machine learning model generated in this way, the pressure pulse wave data acquired up to the time of estimation (corresponding to a portion of the pressure pulse wave data as training data) indicated by the arrow as a waveform cut in Fig. 11 is subjected to preprocessing and feature quantification, and then input to the trained machine learning model. The machine learning model then outputs an estimated value of the relative position of the envelope at the time of estimation indicated by the arrow.

[0085] FIG. 12 is a functional block diagram of a CPU 100B of a blood pressure measurement device 3 according to the third embodiment. The blood pressure measurement device 3 having the CPU 100B and the memory 24 corresponds to the computer of the present invention. FIG. 13 is a flowchart illustrating the processing procedure of a peak position estimation method according to the third embodiment. Processes common to the processing procedure shown in FIG. 4 are denoted by the same reference numerals.

[0086] The CPU 100B includes a pressure detection unit 110, a pressure control unit 120, a blood pressure calculation unit 130A, a preprocessing unit 190, a feature extraction unit 200, a relative position estimation unit 210, a peak position determination unit 220, and a stop cuff pressure calculation unit 160. A trained machine learning model 243 is stored in the memory 24 along with set parameters, etc. The functions of the pressure detection unit 110, the pressure control unit 120, the blood pressure calculation unit 130A, and the stop cuff pressure calculation unit 160 are the same as those in the second embodiment, and therefore will not be described here. The functions of each block will be described together with the processing procedure of the peak position estimation method. Here, the machine learning model 243 corresponds to the learning model of the present invention. The preprocessing unit 190, the feature extraction unit 200, the relative position estimation unit 210, and the peak position determination unit 220 correspond to the reference cuff pressure estimation unit of the present invention. The relative position estimation unit 210, or the preprocessing unit 190, the feature extraction unit 200, and the relative position estimation unit 210 correspond to the index estimation unit of the present invention. The peak position determining section 220 corresponds to the determining section of the present invention.

[0087] (Peak Position Estimation Method) First, the preprocessing unit 190 performs preprocessing such as envelope calculation and noise removal on the pressure pulse wave signal extracted by the HPF unit 111 of the pressure detection unit 110 (step S321).

[0088] Next, the feature extracting unit 200 extracts feature amounts from the envelope and pressure pulse wave that have been preprocessed by the preprocessing unit 190 (step S322).

[0089] Next, the relative position estimation unit 210 inputs the feature amounts extracted by the feature amount extraction unit 200 into the trained machine learning model 243, and estimates the envelope relative position (step S323).

[0090] The peak position determination unit 220 determines whether the envelope relative position estimated in step S323 exceeds the peak position (in the example shown in Figure 10 (A) , whether the index of the envelope relative position is 0 or greater) (step S324).

[0091] In step S324, if it is determined that the estimated value of the envelope relative index exceeds the peak position, the envelope relative position estimated by the relative position estimation unit 210 is the peak position, and the stop cuff pressure calculation unit 160 determines the value of the stop cuff pressure based on this peak position (step S4).

[0092] Here, the processes of steps S321 to S324 correspond to the peak position estimation process of step S3 in the flowchart of Fig. 4. The processes from determining the stop cuff pressure (step S4) to stopping inflation of the cuff 11 (step S7) and the blood pressure value display process (step S9) are the same as those described in Example 1. The blood pressure calculation process of step S8 is the same as that described in Example 2.

[0093] In this way, it is possible to stop pressurizing the cuff 11 at a desired timing based on the peak cuff pressure.

[0094] <Example 4> The overall configuration of a blood pressure measurement device 4 according to Example 4 is similar to that of the blood pressure measurement device 1 according to Example 1 shown in Fig. 1. The overall processing procedure of the blood pressure measurement method according to Example 4 is also the same as that described in the flowchart shown in Fig. 4, but the details of the peak position estimation processing in step S3 are different. Differences from Example 1 are described below.

[0095] 14 is a graph illustrating a peak position estimation process according to Example 4, with the horizontal axis representing the cuff pressure and the vertical axis representing an index representing the envelope relative position. In Example 4, a learning model that outputs the relative position of the envelope of the pressure pulse wave based on the acquired cuff pressure and pressure pulse wave described in Example 3 is used. A regression line L showing the relationship between the envelope relative position and the cuff pressure is estimated from multiple estimated envelope relative position values ​​and their corresponding cuff pressures (shown as points Pi in FIG. 14 ). This regression line L is used to estimate a peak position Pp (cuff pressure corresponding to the envelope peak), which is the cuff pressure corresponding to the envelope relative position index Ip corresponding to the envelope peak. The inflation stop position of the cuff 11 is then determined based on the estimated peak position, i.e., the peak predicted cuff pressure Pp.

[0096] FIG. 15 is a functional block diagram of a CPU 100C of a blood pressure measurement device 4 according to the fourth embodiment. The blood pressure measurement device 4 having the CPU 100C and the memory 24 corresponds to the computer of the present invention. FIG. 16 is a flowchart illustrating the processing procedure of a peak position estimation method according to the fourth embodiment. Processes common to the processing procedure shown in FIG. 4 are denoted by the same reference numerals.

[0097] The CPU 100C includes a pressure detection unit 110, a pressure control unit 120, a blood pressure calculation unit 130A, a preprocessing unit 190, a feature extraction unit 200, a relative position estimation unit 210, a cuff pressure / relative position recording unit 230, a regression line estimation unit 240, a peak position estimation unit 150B, and a stop cuff pressure calculation unit 160. A trained machine learning model 243 is stored in the memory 24 along with set parameters, etc. The functions of the pressure detection unit 110, the pressure control unit 120, and the stop cuff pressure calculation unit 160 are the same as those in the first embodiment, and the functions of the blood pressure calculation unit 130A are the same as those in the second embodiment, so their description will be omitted. The preprocessing unit 190, the feature extraction unit 200, and the relative position estimation unit 210 are the same as those in the third embodiment, so their description will be omitted. The functions of each block will be described together with the processing procedure of the peak position estimation method. Here, the machine learning model 243 corresponds to the learning model of the present invention. The preprocessing unit 190, the feature extraction unit 200, the relative position estimation unit 210, the cuff pressure / relative position recording unit 230, the regression line estimation unit 240, and the peak position estimation unit 150B correspond to a reference cuff pressure estimation unit of the present invention. The relative position estimation unit 210, the cuff pressure / relative position recording unit 230, and the regression line estimation unit 240 correspond to a usage estimation unit, a recording unit, and a regression line estimation unit of the present invention, respectively.

[0098] (Peak Position Estimation Method) First, the preprocessing unit 190 performs preprocessing such as envelope calculation and noise removal on the pressure pulse wave signal extracted by the HPF unit 111 of the pressure detection unit 110 (step S331).

[0099] Next, the feature extracting unit 200 extracts feature amounts from the envelope and pressure pulse wave that have been preprocessed by the preprocessing unit 190 (step S332).

[0100] Next, the relative position estimation unit 210 inputs the feature extracted by the feature extraction unit 200 into the trained machine learning model 243 stored in the memory 24, and estimates the envelope relative position (step S333).

[0101] The cuff pressure and relative position recording unit 230 records the envelope relative position estimated in step S333 and the corresponding cuff pressure in the memory 24 (step S334). In this manner, the estimated envelope relative position and the corresponding cuff pressure are sequentially recorded in the memory 24, and a plurality of cuff pressures 244 corresponding to the estimated envelope relative position are accumulated in the memory 24.

[0102] Next, the regression line estimation unit 240 reads out the cuff pressures 244 corresponding to the multiple envelope relative positions recorded in the memory 24, and determines whether a regression line L representing the relationship that must be satisfied between the indices representing these envelope relative positions and the corresponding cuff pressures can be defined (step S335).

[0103] If the regression line estimation unit 240 determines that the regression line L cannot be defined from the cuff pressure corresponding to the index representing the relative position of the envelope recorded in the memory 24, the process returns to step S2, and the cuff pressure and pressure pulse wave are acquired while continuing inflation control.

[0104] If the regression line estimation unit 240 determines that a regression line L can be defined from the index representing the relative position of the envelope recorded in the memory 24 and the corresponding cuff pressure, the defined regression line L is used to calculate the peak predicted cuff pressure Pp for the index Ip of the relative position to the peak of the envelope (step S336).

[0105] Based on the predicted peak cuff pressure Pp calculated in step S336, that is, the peak position, the stop cuff pressure calculation section 160 determines the value of the stop cuff pressure.

[0106] Here, the processes of steps S331 to S336 correspond to the peak position estimation process of step S3 in the flowchart of Fig. 4. The processes from determining the stop cuff pressure (step S4) to stopping inflation of the cuff 11 (step S7) and the blood pressure value display process (step S9) are the same as those described in Example 1. The blood pressure calculation process of step S8 is the same as that described in Example 2.

[0107] In this way, it is possible to stop pressurizing the cuff 11 at a desired timing based on the peak cuff pressure.

[0108] DESCRIPTION OF SYMBOLS 1, 2, 3, 4... Blood pressure measurement device 11... Cuff 110... Pressure detection unit 120... Pressure control unit 130... SBP calculation unit 130A... Blood pressure calculation unit 140... DBP calculation unit 150... Stop cuff pressure calculation unit

Claims

1. A blood pressure measuring device comprising: a cuff to be wrapped around a part to be measured; a pressure detection unit to detect the cuff pressure in the cuff; a pressure control unit to control the cuff pressure; a pulse wave acquisition unit to acquire the pulse wave of the person to be measured from the cuff pressure; a blood pressure calculation unit to calculate the blood pressure of the person to be measured based on the pulse waves acquired before the cuff pressure reaches a stop cuff pressure at which inflation of the cuff is stopped; a reference cuff pressure estimation unit to estimate a reference cuff pressure, which is the cuff pressure value at which the amplitude of the pulse wave is maximum, based on characteristics of the pulse wave acquired at a cuff pressure lower than the stop cuff pressure; and a stop cuff pressure determination unit to determine the value of the stop cuff pressure based on the reference cuff pressure.

2. The blood pressure measurement device according to claim 1, wherein the reference cuff pressure estimation unit estimates the reference cuff pressure based on the cuff pressure corresponding to the diastolic blood pressure of the subject.

3. A blood pressure measurement device as described in claim 1, characterized in that it comprises an envelope calculation unit that calculates the envelope of the pulse wave, and the reference cuff pressure estimation unit has a mathematical model that mathematically expresses the envelope of the pulse wave, and estimates the reference cuff pressure based on the mathematical model that expresses an estimated envelope that includes the calculated envelope.

4. The blood pressure measurement device according to claim 1, wherein the reference cuff pressure estimation unit comprises: a learning model generated by machine learning the pulse wave or the envelope of the pulse wave relative to the cuff pressure and an index representing the relative position on the envelope as learning data; an index estimation unit that estimates the index at the time of estimation by inputting the pulse wave or the envelope of the pulse wave obtained up to the time of estimation into the learning model; and a determination unit that determines whether the index at the time of estimation has reached the position corresponding to the reference cuff pressure.

5. The blood pressure measurement device according to claim 1, wherein the reference cuff pressure estimation unit comprises: a learning model generated by machine learning the pulse wave or the envelope of the pulse wave relative to the cuff pressure and an index representing the relative position on the envelope as learning data; an index estimation unit that estimates the index at the time of estimation by inputting the pulse wave or the envelope of the pulse wave obtained up to the time of estimation into the learning model; a recording unit that records the cuff pressure corresponding to the index at the time of estimation; and a regression line estimation unit that estimates a regression line representing the relationship between the recorded index and the cuff pressure, and wherein the reference cuff pressure is estimated based on the regression line.

6. A blood pressure measurement method comprising the steps of: detecting the cuff pressure in a cuff wrapped around a part to be measured; controlling the cuff pressure; acquiring the pulse wave of the person to be measured from the cuff pressure; calculating the blood pressure of the person to be measured based on the pulse wave acquired until the cuff pressure reaches a stop cuff pressure at which inflation of the cuff is stopped; estimating a reference cuff pressure, which is the cuff pressure value at which the amplitude of the pulse wave is maximum, based on characteristics of the pulse wave acquired at a cuff pressure lower than the stop cuff pressure; and determining the value of the stop cuff pressure based on the reference cuff pressure.

7. The blood pressure measurement method according to claim 6, wherein the reference cuff pressure is estimated based on the cuff pressure corresponding to the diastolic blood pressure of the subject.

8. A blood pressure measurement method as described in claim 6, characterized in that it includes a step of calculating an envelope of the pulse wave, and the step of estimating the reference cuff pressure estimates the reference cuff pressure by expressing an estimated envelope that includes the calculated envelope using a mathematical model that mathematically expresses the envelope of the pulse wave.

9. A blood pressure measurement method according to claim 6, comprising: a step of estimating the index at the time of estimation by inputting the pulse wave or the envelope of the pulse wave acquired up to the time of estimation into a learning model generated by machine learning the pulse wave or the envelope of the pulse wave relative to the cuff pressure and an index representing the relative position on the envelope as learning data; and a step of determining whether the index at the time of estimation has reached the position corresponding to the reference cuff pressure, wherein the step of determining the stop cuff pressure determines the stop cuff pressure based on the index at the time of estimation.

10. The blood pressure measurement method according to claim 6, wherein the step of estimating the reference cuff pressure includes the steps of: estimating the index at the time of estimation by inputting the pulse wave or the envelope of the pulse wave acquired up to the time of estimation into a learning model generated by machine learning the pulse wave or the envelope of the pulse wave relative to the cuff pressure and an index representing the relative position on the envelope as learning data; recording the cuff pressure corresponding to the index at the time of estimation; and estimating a regression line representing the relationship between the recorded index and the cuff pressure; and estimating the reference cuff pressure based on the regression line.

11. A blood pressure measurement program that causes a computer to execute the following steps: detecting the cuff pressure in a cuff wrapped around a part to be measured; controlling the cuff pressure; acquiring the pulse wave of the person to be measured from the cuff pressure; calculating the blood pressure of the person to be measured based on the pulse wave acquired until the cuff pressure reaches a stop cuff pressure at which inflation of the cuff is stopped; estimating a reference cuff pressure, which is the cuff pressure value at which the amplitude of the pulse wave is maximum, based on characteristics of the pulse wave acquired at a cuff pressure lower than the stop cuff pressure; and determining the value of the stop cuff pressure based on the reference cuff pressure.

Images