CN104644151A - Photoelectric volume pulse signal-based propagation prediction method for pulse pressure waveform - Google Patents

Abstract

The invention discloses a photoelectric volume pulse signal-based propagation prediction device for a pulse pressure waveform. The photoelectric volume pulse signal-based propagation prediction device for the pulse pressure waveform is characterized by comprising a waveform input module, a waveform conditioning module, a waveform fitting module, a waveform conversion module and a waveform output module, wherein the waveform conditioning module comprises a preprocessing circuit, a single shot separator and a normalization circuit; the waveform fitting module comprises a fitting function setting device, a waveform fitter and a waveform quality discriminator; the waveform conversion module comprises a part setting device, an objective function setting device, a feature population setting device, a parameter converter and a waveform synthesizer. According to the device, pulse pressure waveform signals of all parts of a human body can be predicted by utilizing photoelectric volume pulse signals of all the parts of the human body according to the physiological statistical law; the application range, the prediction effect and the stability of the device are improved to a certain extent in comparison with the existing device.

Description

A kind of pressure pulse wave wave travel Forecasting Methodology based on photoelectricity volume pulse signal

Technical field

The present invention relates to technical field of medical equipment, particularly a kind of method predicting each site pressure pulse waveform based on photoplethysmographic.

Background technology:

Abundant hemodynamics information is contained in pulse wave.Always as the foundation of clinical diagnosis and treatment, a large amount of clinical measured results confirms, feature and the cardiovascular physiology state of pulse wave have close relationship.The integrated information of the aspects such as the form (shape of ripple) that pulse wave shows, intensity (amplitude of ripple), speed (wave propagation velocity) and the rhythm and pace of moving things (wave period) reflects much physiology and the pathological characters of cardiovascular system of human body to a great extent.

In acquisition principle, the acquisition mode of current pulse wave mainly contains the pressure pulse wave collection using pressure transducer or the photoplethysmographic collection using photoelectric sensor.Intraarterial pressure pulse wave has obtained more comprehensive analysis and research, and fluid mechanic model is comparatively clear and definite, and corresponding Hydrodynamic character and cardiovascular physiology meaning are applied also comparatively extensive.But the collection of human pressure's pulse wave is very easily subject to many-sided interference such as collection position, and action required skill set requirements is higher, and lack repeatable, be not easy to continuous detecting.The collection of photoplethysmographic finger tip has good stability and repeatability, but it is to the accuracy that cardiovascular function judges, and more to have studied comparatively perfect pressure pulse wave not good enough.

Summary of the invention:

The non-physiological parameter models such as the transfer function that existing technical scheme mainly utilizes the power spectrumanalysis gathering position photoelectricity plethysmogram signal and corresponding target site pressure pulse wave signal to set up, and mainly pay close attention to finger volume pulse signal and radial artery pressure pulse signal.Because the individual difference of pulse signal is comparatively large, said apparatus is less stable when large-scale application, and function is comparatively single, and institute's Modling model itself is without physiological significance, is unfavorable for the correction to model and improvement.

For solving the problem, the present invention proposes the waveform of a kind of volume pulsation signal based on pulse wave physiological feature and pressure pulse signal respectively containing ginseng expression formula, and utilizes priori statistical law to set up collection position photoplethysmographic expression argument and the regression equation group between corresponding target site pressure pulse waveform expression argument.Thus realize the pressure pulse wave signal photoelectricity volume pulse signal of input being converted to target site.The method of waveform fitting has good stability, and regression equation group has clearer and more definite physiological significance, conveniently carries out finely tuning for different physiological status and improves.Can solve in prior art thus and predict that stability of waveform is poor, and the problem that model cannot be revised.

For achieving the above object, the technical solution used in the present invention is: a kind of pressure pulse wave propagation forecast device based on photoelectricity volume pulse signal, is characterized in that: comprise signal input conditioning module, waveform fitting module, waveform transformation module, Waveform composition module and output module.

Described Waveform Input module is received from the time domain photoelectricity volume pulse signal of human body a certain position actual measurement.

Described Signal-regulated kinase, carries out pretreatment to the time domain photoelectricity volume pulse signal of input, and signal clapped by the list utilizing prior art to be decomposed into corresponding single cardiac cycle, to each single clap signal amplitude and wavelength be normalized.

Described waveform fitting module receives the list after normalization and claps pulse signal, and utilizes given waveform containing ginseng expression formula f _iapplication curves fitting algorithm carries out matching to it, waveform containing ginseng expression formula by represent respectively the main ripple of pulse waveform, dicrotic wave, echo containing ginseng expression formula be added determine, matching gained analytical expression parameters vector is I _i, as waveform feature parameter vector.

The ginseng expression formula that contains of photoelectricity volume pulse signal is:

$f_I=\underset{n=1}{\overset{3}{\mathrm{\Σ}}}{H}_{\mathrm{in}}\×e^{\frac{-{(t-b_{\mathrm{in}})}^2}{W_{\mathrm{in}}}}$

Wherein, H _in, b _in, W _infor parameter, t is independent variable, represents sampling number.N=1 in expression formula, the part of 2,3 distinguishes main ripple, echo, the dicrotic wave waveform of corresponding pulse waveform.Wherein, H _inrepresent the amplitude of fluctuation, b _inrepresent the center of fluctuation, W _inthe width of fluctuation.

Curve fitting process algorithm adopts least-squares algorithm, and limits parameters scope according to each fluctuation physiological significance, sets matching initial condition H afterwards _i1>H _i2>H _i3, b _i1<b _i2<b _i3, W _in>0.The characteristic parameter vector I that matching is determined _ifor:

I _I＝[H _i1,H _i2,H _i3,b _i1,b _i2,b _i3,W _i1,W _i2,W _i3]

At above-mentioned waveform containing under ginseng expression formula condition, fitting effect depends primarily on the annoyance level that pulse wave collection is subject to, therefore with matching determination coefficients R ²as the quantification standard that waveform quality differentiates.Matching determination coefficients R ²the computational methods of conventional judgement two Similar Broken Line degree.R in the present invention ²same as acquisition quality discriminant parameter, corresponding R ²be less than certain value list clap waveform think acquisition quality difference and given up.

R ²computing formula is as follows:

$R^2=1-\frac{\underset{i=1}{\overset{\mathrm{pl}}{\mathrm{\&Sigma;}}}{(f_i-{\hat{f}}_i)}^2}{\underset{i=1}{\overset{\mathrm{pl}}{\mathrm{\&Sigma;}}}{(f_i-\stackrel{\&OverBar;}{f_i})}^2}$

Wherein, f _i, represent measured light Power Capacity Pulse Rate strong point, measured light Power Capacity pulse data meansigma methods and data point matching expected value respectively, pl is single wave number strong point number.

Described waveform transformation module, divide into groups according to measured's sex, age, Mean Arterial forcing up the targets, and according to the priori statistical law between the lower photoplethysmographic actual measurement position of correspondence grouping and required forecast pressure pulse wave position waveform feature parameter, utilize measured light Power Capacity pulse wave characteristic parameters computational prediction corresponding position pressure pulse wave waveform feature parameter.

Priori statistical law method for building up between the lower photoplethysmographic actual measurement position of above-mentioned correspondence classification and required forecast pressure pulse wave position waveform feature parameter is as follows:

(1) first, the crowd participating in experiment divided into groups according to sex, age, mean arterial pressure (MAP), wherein the age was initial with 20 years old, within 5 years old, was interval.Mean arterial pressure is initial with 70mmHg, and 10mmHg is interval.The crowd participating in experiment is divided into groups.Respectively ear, finger end, toe end place photoplethysmographic are detected to above-mentioned each group of Subject Population simultaneously, and utilize pressure transducer to detect radial artery, brachial artery, carotid artery place pressure pulse wave signal.Thus the observed pressure pulse obtaining different crowd feature involves photoplethysmographic signal.

(2) the statistics relation between actual measurement photoplethysmographic and actual measurement target site pressure pulse wave waveform feature parameter is set up after.Similar to above-mentioned photoplethysmographic feature extraction mode, for extracting pressure pulse wave waveform feature parameter, utilize pressure pulse wave expression formula to carry out matching to observed pressure pulse waveform, pressure pulse wave containing ginseng expression formula is:

$f_O=\underset{n=1}{\overset{3}{\mathrm{\&Sigma;}}}{H}_{\mathrm{on}}\&times;e^{-\frac{-{(t-b_{\mathrm{on}})}^2}{W_{\mathrm{on}}}}$

Expression argument and the domain of definition are all and f _iidentical.Fit procedure algorithm adopts least-squares algorithm, and limits parameters scope according to each fluctuation physiological significance, sets matching initial condition H afterwards _o1>H _o2>H _o3, b _o1<b _o2<b _o3, W _on>0.

Its characteristic parameter vector is I _o=[H _o1, H _o2, H _o3, b _o1, b _o2, b _o3, W _o1, W _o2, W _o3]

(3) respectively f is utilized to the measured waveform of different grouping different parts _iand f _oadopt waveform fitting module to carry out matching to the photoplethysmographic waveform of surveying and pressure pulse waveform, obtain the I of each photoelectricity volume pulse signal of corresponding actual measurement _iand the I of pressure pulse signal _ovector.For each site pressure pulse signal I _ovector every parameter, the multiple linear regression equations of the photoelectricity volume pulse signal characteristic parameter vector gathered while setting up corresponding different grouping, that is:

H _o1＝TM ₁₁×H _i1+TM ₁₂×H _i2+......+TM ₁₉×W _i3+CM ₁

H _o2＝TM ₂₁×H _i1+TM ₂₂×H _i2+......+TM ₂₉×W _i3+CM ₂

H _o3＝TM ₃₁×H _i1+TM ₃₂×H _i2+......+TM ₃₉×W _i3+CM ₃

b _o1＝TM ₄₁×H _i1+TM ₄₂×H _i2+......+TM ₄₉×W _i3+CM ₄

b _o2＝TM ₅₁×H _i1+TM ₅₂×H _i2+......+TM ₅₉×W _i3+CM ₅

b _o3＝TM ₆₁×H _i1+TM ₆₂×H _i2+......+TM ₆₉×W _i3+CM ₆

W _o1＝TM ₇₁×H _i1+TM ₇₂×H _i2+......+TM ₇₉×W _i3+CM ₇

W _o2＝TM ₈₁×H _i1+TM ₈₂×H _i2+......+TM ₈₉×W _i3+CM ₈

W _o3＝TM ₉₁×H _i1+TM ₉₂×H _i2+......+TM ₉₉×W _i3+CM ₉

The each term coefficient of TM and CM is determined by multiple linear regression equations.Arrange I _oparameters sets up equation group, and arrange coefficient matrix and constant matrices, can obtain TM and CM, wherein, TM is 9 rank square formations, and CM is 9 element column vectors.

Utilize prior probability gained TM and CM matrix according to above-mentioned, in application process, utilize the waveform feature parameter vector I gathering position _icalculate the characteristic parameter vector I of target site _o.According to surveyed photoplethysmographic, extract its feature I _i, then target site characteristic vector I can be obtained _ofor:

I _O＝TM×I _I+CM

The target site characteristic parameter vector I will obtained afterwards _osubstitute into the pressure pulse wave of corresponding position containing ginseng expression formula f _o, obtain corresponding pressure pulse wave analytical expression.Complete waveform transformation.

Described wave form output module, by the result of above-mentioned Waveform composition module and target site characteristic parameter vector I _oform exports on request.

Beneficial effect of the present invention is:

(1) apparatus of the present invention only need human body single position photoelectricity volume pulse signal, can obtain the pressure pulse wave signal of different parts under certain precision.Simple to operation, and waveform quality is stablized.Overcome testing process in normal pressures pulse wave gatherer process complicated, use inconvenience, and be difficult to the problem obtaining stable pulse wave.Effectively can reduce the direct times of collection of pressure pulse wave in actual application, reduce the demand to operant skill, improve measured's level of comfort.Through a large amount of confirmatory experiments, respond well.

(2) first by the waveform fitting application of installation based on pulse wave physiological fluctuating characteristic in multiple location pulse wave, thus analyze in its communication process physiological change.Facilitate adjustment and the correction of model, for certain basis has been established in larger scale clinical application.

Accompanying drawing illustrates:

Fig. 1 is structured flowchart of the present invention

Fig. 2 is operational flowchart of the present invention

Fig. 3 is waveform fitting and characteristic parameter schematic diagram

Fig. 4 is an actual measurement example, and gives the contrast with observed pressure waveform

Fig. 5 is the experiment effect figure that the present invention is utilizing finger tip photoplethysmographic to predict radial artery pressure pulse wave.R ²data are that observed pressure pulse list claps waveform and the single cross validation effect of clapping between waveform of prediction, test 426 examples altogether.

Detailed description of the invention:

Below in conjunction with accompanying drawing, the comparatively typical detailed description of the invention of one of the present invention is described in detail.

A kind of typical apply scene of the present invention is to utilize human body finger tip photoplethysmographic signal to predict radial artery pressure pulse wave signal.The mature technology that finger tip can be utilized thus to gather and high-quality waveform obtain physiological significance radial artery pressure pulse wave data definitely.

As shown in Figure 1, after forecasting process starts, first according to the Sex, Age of measured, blood pressure selects corresponding characterizing population group, and selects corresponding prediction position as required.With 56 years old age women, pressure value is the measured of 90/130mmHg is example, sets sex in step s 201, age and pressure value, and collection position is finger tip, and target prediction position is radial artery.According to setting in S202, system, namely according to the grouping situation of priori statistical law, selects TM and the CM matrix at corresponding crowd and position automatically.

In step S203, system starts to receive measured waveform, step S204 nurses one's health input waveform, filtering baseline and Hz noise, by continuous wave according to cardiac cycle separated component vertical wave shape, and in units of waveform, carry out the normalization of wave-shape amplitude and wavelength, wherein, the normalization of wavelength is realized by the method for interpolation.

In S205 step, according to collection waveform fitting expression formula given in S203, least square method is utilized to carry out curve fitting to each single waveform of clapping.

Fitting expression is:

$f_I=\underset{n=1}{\overset{3}{\mathrm{\&Sigma;}}}{H}_{\mathrm{in}}\&times;e^{\frac{-{(t-b_{\mathrm{in}})}^2}{W_{\mathrm{in}}}}$

Setting initial condition.Thus obtain each single I clapping waveform corresponding _iand calculate R ².

In this example, waveforms amplitude and wavelength are defined as 100 units respectively, then

I _I＝[H _i1,H _i2,H _i3,b _i1,b _i2,b _i3,W _i1,W _i2,W _i3]＝[43,69,49,15,27,48,14,25,52]

R ²>0.99

According to R in step S206 ²numerical value carries out waveform quality judgement, to R ²the waveform of <0.95 is given up, and ratio given up in record.If occur, waveform is given up, and again extracts next bat waveform and analyzes.R under normal circumstances ²the waveform ratio <3% of <0.95, gives up ratio when being greater than 5%, should consider to adjust acquisition mode.

Step S207 is to the I of waveform up-to-standard in S206 _ichange, calculate the characteristic vector I of target site waveform _o, according to formula:

I _O＝TM×I _I+CM

In this example, TM and CM by the radial artery pressure pulse wave recorded before this patient and finger tip photoplethysmographic signal simultaneously respectively according to f _iwith f _omatching, and set up corresponding regression equation acquisition respectively by the parameter obtained.

Can obtain as calculated:

I _O＝[H _o1,H _o2,H _o3,b _o1,b _o2,b _o3,W _o1,W _o2,W _o3]＝[64,71,35,14,25,49,13,22,48]

By I in step S208 _omiddle parameters substitutes into given target site pressure pulse wave expression formula, namely

$f_O=\underset{n=1}{\overset{3}{\mathrm{\&Sigma;}}}{H}_{\mathrm{on}}\&times;e^{-\frac{-{(t-b_{\mathrm{on}})}^2}{W_{\mathrm{on}}}}$

Obtain target site prediction waveform expression formula and main ripple corresponding respectively, reflection involves dicrotic wave prediction waveform.

Above-mentioned waveform and parameter export according to specified format by step S209.

Claims (1)

1., based on a pressure pulse wave wave travel Forecasting Methodology for photoelectricity volume pulse signal, it is characterized in that: comprise Waveform Input module, Signal-regulated kinase, waveform fitting module, waveform transformation module, and wave form output module;

Described Waveform Input module is received from the time domain photoelectricity volume pulse signal of human body a certain position actual measurement;

Described Signal-regulated kinase, carries out pretreatment to the time domain photoelectricity volume pulse signal of input, and signal clapped by the list being decomposed into corresponding single cardiac cycle, to each single clap signal amplitude and wavelength be normalized;

Described waveform fitting module receives the list after normalization and claps pulse signal, and utilizes given waveform containing ginseng expression formula f _iapplication curves fitting algorithm carries out matching to it, waveform containing ginseng expression formula by represent respectively the main ripple of pulse waveform, dicrotic wave, echo containing ginseng expression formula be added determine, matching gained analytical expression parameters vector is I _i, as waveform feature parameter vector;

The ginseng expression formula that contains of photoelectricity volume pulse signal is:

$f_I=\underset{n=1}{\overset{3}{\mathrm{\&Sigma;}}}{H}_{\mathrm{in}}\&times;e^{\frac{-{(t-b_{\mathrm{in}})}^2}{W_{\mathrm{in}}}}$

Wherein, H _in, b _in, W _infor parameter, t is independent variable, represents sampling number; N=1 in expression formula, the part of 2,3 distinguishes main ripple, echo, the dicrotic wave waveform of corresponding pulse waveform; Wherein, H _inrepresent the amplitude of fluctuation, b _inrepresent the center of fluctuation, W _inthe width of fluctuation;

Curve fitting process algorithm adopts least-squares algorithm, and limits parameters scope according to each fluctuation physiological significance, sets matching initial condition H afterwards _i1>H _i2>H _i3, b _i1<b _i2<b _i3, W _in>0; The characteristic parameter vector I that matching is determined _ifor:

I _I＝[H _i1,H _i2,H _i3,b _i1,b _i2,b _i3,W _i1,W _i2,W _i3]

At above-mentioned waveform containing under ginseng expression formula condition, fitting effect depends primarily on the annoyance level that pulse wave collection is subject to, therefore with matching determination coefficients R ²as the quantification standard that waveform quality differentiates; Matching determination coefficients R ²the computational methods of conventional judgement two Similar Broken Line degree; R ²as acquisition quality discriminant parameter, corresponding R ²be less than certain value list clap waveform think acquisition quality difference and given up;

R ²computing formula is as follows:

$R^2=1-\frac{\underset{i=1}{\overset{\mathrm{pl}}{\mathrm{\&Sigma;}}}{(f_i-{\hat{f}}_i)}^2}{\underset{i=1}{\overset{\mathrm{pl}}{\mathrm{\&Sigma;}}}{(f_i-{\stackrel{\&OverBar;}{f}}_i)}^2}$

Wherein, represent measured light Power Capacity Pulse Rate strong point, measured light Power Capacity pulse data meansigma methods and data point matching expected value respectively, pl is single wave number strong point number;

Described waveform transformation module, divide into groups according to measured's sex, age, Mean Arterial forcing up the targets, and according to the priori statistical law between the lower photoplethysmographic actual measurement position of correspondence grouping and required forecast pressure pulse wave position waveform feature parameter, utilize measured light Power Capacity pulse wave characteristic parameters computational prediction corresponding position pressure pulse wave waveform feature parameter;

Priori statistical law method for building up between the lower photoplethysmographic actual measurement position of above-mentioned correspondence classification and required forecast pressure pulse wave position waveform feature parameter is as follows:

(1) first, the crowd participating in experiment divided into groups according to sex, age, mean arterial pressure, wherein the age was initial with 20 years old, within 5 years old, was interval; Mean arterial pressure is initial with 70mmHg, and 10mmHg is interval; The crowd participating in experiment is divided into groups; Respectively ear, finger end, toe end place photoplethysmographic are detected to above-mentioned each group of Subject Population simultaneously, and utilize pressure transducer to detect radial artery, brachial artery, carotid artery place pressure pulse wave signal; Thus the observed pressure pulse obtaining different crowd feature involves photoplethysmographic signal;

(2) the statistics relation between actual measurement photoplethysmographic and actual measurement target site pressure pulse wave waveform feature parameter is set up after; Similar to above-mentioned photoplethysmographic feature extraction mode, for extracting pressure pulse wave waveform feature parameter, utilize pressure pulse wave expression formula to carry out matching to observed pressure pulse waveform, pressure pulse wave containing ginseng expression formula is:

$f_O=\underset{n=1}{\overset{3}{\mathrm{\&Sigma;}}}{H}_{\mathrm{on}}\&times;e^{\frac{-{(t-b_{\mathrm{on}})}^2}{W_{\mathrm{on}}}}$

Expression argument and the domain of definition are all and f _iidentical; Fit procedure algorithm adopts least-squares algorithm, and limits parameters scope according to each fluctuation physiological significance, sets matching initial condition H afterwards _o1>H _o2>H _o3, b _o1<b _o2<b _o3, W _on>0;

Its characteristic parameter vector is I _o=[H _o1, H _o2, H _o3, b _o1, b _o2, b _o3, W _o1, W _o2, W _o3]

(3) respectively f is utilized to the measured waveform of different grouping different parts _iand f _oadopt waveform fitting module to carry out matching to the photoplethysmographic waveform of surveying and pressure pulse waveform, obtain the I of each photoelectricity volume pulse signal of corresponding actual measurement _iand the I of pressure pulse signal _ovector; For each site pressure pulse signal I _ovector every parameter, the multiple linear regression equations of the photoelectricity volume pulse signal characteristic parameter vector gathered while setting up corresponding different grouping, that is:

H _o1＝TM ₁₁×H _i1+TM ₁₂×H _i2+......+TM ₁₉×W _i3+CM ₁

H _o2＝TM ₂₁×H _i1+TM ₂₂×H _i2+......+TM ₂₉×W _i3+CM ₂

H _o3＝TM ₃₁×H _i1+TM ₃₂×H _i2+......+TM ₃₉×W _i3+CM ₃

b _o1＝TM ₄₁×H _i1+TM ₄₂×H _i2+......+TM ₄₉×W _i3+CM ₄

b _o2＝TM ₅₁×H _i1+TM ₅₂×H _i2+......+TM ₅₉×W _i3+CM ₅

b _o3＝TM ₆₁×H _i1+TM ₆₂×H _i2+......+TM ₆₉×W _i3+CM ₆

W _o1＝TM ₇₁×H _i1+TM ₇₂×H _i2+......+TM ₇₉×W _i3+CM ₇

W _o2＝TM ₈₁×H _i1+TM ₈₂×H _i2+......+TM ₈₉×W _i3+CM ₈

W _o3＝TM ₉₁×H _i1+TM ₉₂×H _i2+......+TM ₉₉×W _i3+CM ₉

The each term coefficient of TM and CM is determined by multiple linear regression equations; Arrange I _oparameters sets up equation group, and arrange coefficient matrix and constant matrices, obtain TM and CM, wherein, TM is 9 rank square formations, and CM is 9 element column vectors;

Utilize prior probability gained TM and CM matrix according to above-mentioned, in application process, utilize the waveform feature parameter vector I gathering position _icalculate the characteristic parameter vector I of target site _o; According to surveyed photoplethysmographic, extract its feature I _i, then target site characteristic vector I is obtained _ofor:

I _O＝TM×I _I+CM

The target site characteristic parameter vector I will obtained afterwards _osubstitute into the pressure pulse wave of corresponding position containing ginseng expression formula f _o, obtain corresponding pressure pulse wave analytical expression; Complete waveform transformation;

Described wave form output module, by the result of above-mentioned Waveform composition module and target site characteristic parameter vector I _oform exports on request.