CN111540436B - Self-adaptive glucose and insulin concentration prediction system and method - Google Patents

Self-adaptive glucose and insulin concentration prediction system and method Download PDF

Info

Publication number
CN111540436B
CN111540436B CN202010304292.3A CN202010304292A CN111540436B CN 111540436 B CN111540436 B CN 111540436B CN 202010304292 A CN202010304292 A CN 202010304292A CN 111540436 B CN111540436 B CN 111540436B
Authority
CN
China
Prior art keywords
insulin
glucose
concentration
plasma
model
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010304292.3A
Other languages
Chinese (zh)
Other versions
CN111540436A (en
Inventor
王少萍
王伟杰
耿艺璇
王兴坚
张超
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beihang University
Original Assignee
Beihang University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beihang University filed Critical Beihang University
Priority to CN202010304292.3A priority Critical patent/CN111540436B/en
Publication of CN111540436A publication Critical patent/CN111540436A/en
Application granted granted Critical
Publication of CN111540436B publication Critical patent/CN111540436B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
    • G16H20/10—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to drugs or medications, e.g. for ensuring correct administration to patients
    • G16H20/17—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to drugs or medications, e.g. for ensuring correct administration to patients delivered via infusion or injection
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/145—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue
    • A61B5/14532—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue for measuring glucose, e.g. by tissue impedance measurement
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/145—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue
    • A61B5/14546—Measuring characteristics of blood in vivo, e.g. gas concentration, pH value; Measuring characteristics of body fluids or tissues, e.g. interstitial fluid, cerebral tissue for measuring analytes not otherwise provided for, e.g. ions, cytochromes
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H70/00—ICT specially adapted for the handling or processing of medical references
    • G16H70/40—ICT specially adapted for the handling or processing of medical references relating to drugs, e.g. their side effects or intended usage

Abstract

The invention relates to a self-adaptive glucose and insulin concentration prediction system and a method, wherein the system acquires monitoring data of a person to be monitored through a data acquisition module; the physiological model set construction module constructs a blood sugar-insulin model according to the collected monitoring data of the person to be monitored; the state space conversion module converts the blood sugar-insulin model into a state space model; the discrete state space model building module converts the state space model into a discrete state space model; the particle wave observer module predicts and updates state variables in the discrete state space model; the automatic eating detection module determines the intake condition of the carbohydrate according to the updated carbohydrate intake factor; the blood glucose prediction module determines a concentration of insulin and a concentration of glucose in the plasma at a future fixed time interval based on the updated state variables. The invention can predict the concentration of glucose and insulin in blood plasma and improve the accuracy of predicting the concentration of glucose in blood plasma.

Description

Self-adaptive glucose and insulin concentration prediction system and method
Technical Field
The invention relates to the field of blood sugar prediction of diabetics, in particular to a self-adaptive glucose and insulin concentration prediction system and a self-adaptive glucose and insulin concentration prediction method.
Background
Diabetes is a metabolic disorder syndrome mainly manifested by high glucose concentration in a fasting state or a postprandial state due to absolute or relative insufficiency of insulin secretion, and is very likely to cause various acute and chronic complications of the whole body. Insulin is the only hypoglycemic hormone in vivo, and for patients with type 1 diabetes and severe type 2 diabetes, which rely on exogenous insulin infusion as the main treatment means, the control of insulin injection amount directly determines the control level of the concentration of glucose in vivo. Timely and appropriate insulin infusion to control blood glucose within a reasonable range without severe high or low blood glucose events is a goal pursued by doctors and patients.
Blood glucose includes the concentration of glucose in the subcutaneous intercellular matrix and the concentration of glucose in plasma. Initially, blood glucose monitoring equipment can realize the real-time monitoring of human subcutaneous glucose concentration, and the insulin pump can realize the subcutaneous continuous microinjection of exogenous insulin. Based on the data, doctors and patients can easily evaluate the blood sugar dynamic change. However, daily blood glucose monitoring equipment collects the glucose concentration in subcutaneous intercellular spaces, the influence of the intake of insulin and carbohydrate on the blood glucose directly acts on the glucose concentration in the blood plasma, and the change of the glucose concentration in the blood plasma influences the glucose concentration transmitted to subcutaneous intercellular spaces to be measured, so that the blood glucose value measured by a blood glucose meter has a large error with the glucose concentration in the blood plasma. Similarly, the process of insulin infusion to take part in the blood sugar reduction process from the subcutaneous transfer plasma has time lag and loss, and the exogenous insulin infusion amount has larger error with the insulin concentration in the plasma. Therefore, establishing accurate hemodynamic models to estimate glucose and insulin concentrations in plasma is a prerequisite for accurate prediction and control of blood glucose.
The individual difference of the human blood sugar regulating system is large, the individual physiological state is changeable, and the mathematical model of the blood sugar regulating system with fixed parameters is difficult to accurately track and predict the dynamic change rule of blood sugar. Therefore, the adaptive adjustment of all blood glucose system parameters of a patient based on real-time acquired data is an important method for improving the prediction accuracy. In addition, external disturbances such as carbohydrate intake, exercise, etc. may cause drastic fluctuations in vivo glucose concentration, and manually inputting the intake time and intake amount of carbohydrates may improve model prediction accuracy, but may cause inconvenience to the patient in life and easily increase the risk of abnormal fluctuations in blood glucose due to personal negligence. The carbohydrate intake state of the patient is automatically recognized, and the prediction model is adaptively adjusted, so that more intelligent and accurate prediction of blood sugar and insulin concentration can be provided for the patient.
Disclosure of Invention
The invention aims to provide a self-adaptive glucose and insulin concentration prediction system and a method, which can realize intelligent and accurate prediction of the concentrations of glucose and insulin in blood plasma.
In order to achieve the purpose, the invention provides the following scheme:
an adaptive glucose and insulin concentration prediction system comprising:
the data acquisition module is used for acquiring monitoring data of a person to be monitored; the monitoring data includes an initial subcutaneous interstitial glucose concentration, an exogenous insulin infusion rate, and a body weight of the subject to be monitored;
the physiological model set construction module is used for constructing a blood sugar-insulin model according to the monitoring data of the person to be monitored; the blood glucose-insulin model comprises: an insulin transmission sub-model, a glycemic insulin dynamics sub-model, and a blood glucose transmission sub-model;
the state space conversion module is used for converting the blood sugar-insulin model into a state space model; the state variables in the state space model comprise state variables converted by the system variables of the blood glucose insulin model and state variables converted by system parameter extension in the blood glucose insulin model;
the discrete state space model building module is used for converting the state space model into a discrete state space model;
the particle wave observer module is used for predicting and updating the state variables in the discrete state space model; the updated state variables include an effective insulin amount in one insulin chamber, an effective insulin amount in two insulin chambers, a concentration of glucose in plasma, a concentration of glucose in subcutaneous intercellular substance, an insulin sensitivity coefficient, a basal effective insulin concentration, a basal blood glucose level, a glycemic autoregulation rate, and a carbohydrate intake factor;
the automatic eating detection module is used for determining the carbohydrate intake condition of the person to be monitored according to the updated carbohydrate intake factor; intake profiles include intake with and without carbohydrate;
and the blood glucose prediction module is used for determining the insulin concentration in the blood plasma, the glucose concentration in the blood plasma and the glucose concentration in the subcutaneous intercellular substance within a future fixed time interval according to the updated state variable.
Optionally, the insulin delivery submodel specifically adopts the following formula:
Figure BDA0002455169760000031
wherein the content of the first and second substances,
Figure BDA0002455169760000032
wherein x is1(t) is the effective insulin amount in an insulin chamber, x2(t) the effective insulin amount in the diloinsulin chamber; u (t) exogenous insulin infusion rate, tI(ii) is the time at which the concentration of available insulin reaches a maximum, x (t) is the concentration of insulin in plasma, W is the body weight of the subject to be monitored, M is the rate of insulin clearance from the human body, and M is 0.017 (l/kg/min);
the blood glucose insulin dynamics submodel specifically adopts the following formula:
Figure BDA0002455169760000033
wherein G (t) is the concentration of glucose in plasma, SISensitivity coefficient of insulin action, Gb(ii) is the concentration level of glucose in the base plasma, K is the rate of concentration self-regulation of glucose in the plasma, u (t) is the carbohydrate uptake factor that causes the concentration change of glucose in the plasma;
the blood glucose transmission sub-model specifically adopts the following formula:
Figure BDA0002455169760000034
wherein IG (t) is the concentration of glucose in the subcutaneous intercellular matrix, and τ is a time-lag factor.
Optionally, the state space conversion module specifically adopts the following formula:
Figure BDA0002455169760000035
wherein X is a state variable, X ═ X1,x2,G,IG,Si,tI,K,Gb,τ,U]，H＝[0,0,0,1,0,0,0,0,0,0]ω (t) is process noise, υ (t) is measurement noise, and z (t) is measurement state.
Optionally, the discrete state space model building module specifically adopts the following formula:
Figure BDA0002455169760000036
wherein Xk-1Is the state variable at time k-1, uk-1Exogenous insulin infusion Rate at time k-1, omegakFor process noise and obedience to be 0, the variance is σwkGaussian noise of (v)kTo measure noise and obey a desired 0, the variance σvkGaussian noise.
Optionally, the particle wave observer module specifically predicts and updates by using the following formula:
p(Xk|z1:k-1)＝∫p(Xk|Xk-1)p(Xk-1|z1:k-1)dXk-1
p(Xk|z1:k)∝p(zk|Xk-1)p(Xk|z1:k-1) Wherein X iskIs the state variable at time k, Xk-1Is a state variable at time k-1, z1:k-1For all blood glucose data sequences acquired before the time point k-1, z1:kAll blood glucose data sequences acquired before the k-th time are acquired.
Optionally, the automatic food intake detection module specifically includes:
a forward difference obtaining unit for obtaining the forward difference of the external interference factor at the time k
Figure BDA0002455169760000041
Wherein the content of the first and second substances,
Figure BDA0002455169760000042
Uk-1carbohydrate uptake factor, U, being the induced change in plasma glucose concentration at time k-1kA carbohydrate uptake factor that is the induced change in concentration of glucose in plasma at time k;
a judging unit for judging whether the forward difference is larger than the predetermined threshold value
Figure BDA0002455169760000043
And (4) judging whether the patient takes the carbohydrate or not.
Optionally, the determining unit specifically includes:
judgment of
Figure BDA0002455169760000044
And
Figure BDA0002455169760000045
whether or not to simultaneously satisfy
Figure BDA0002455169760000046
If at the same time satisfy
Figure BDA0002455169760000047
Then there is carbohydrate intake, otherwise there is no carbohydrate intake; wherein the content of the first and second substances,
Figure BDA0002455169760000048
is the forward difference of the external interference factor at time k,
Figure BDA0002455169760000049
is the forward difference of the external interference factor at the moment of k-1, Threshold is a set Threshold value, Flag k-i0 is an indication of no carbohydrate intake.
Optionally, the blood glucose prediction module specifically adopts the following formula:
Figure BDA00024551697600000410
Figure BDA00024551697600000411
Figure BDA00024551697600000412
Figure BDA00024551697600000413
Figure BDA00024551697600000414
Figure BDA0002455169760000051
wherein the content of the first and second substances,
Figure BDA0002455169760000052
indicating the predicted value of the insulin concentration in the plasma after one period in the future,
Figure BDA0002455169760000053
represents the predicted value of the insulin concentration in plasma after two future sampling periods,
Figure BDA0002455169760000054
represents the predicted value of the glucose concentration in the plasma after one period in the future,
Figure BDA0002455169760000055
represents the predicted value of the glucose concentration in plasma after the next two cycles,
Figure BDA0002455169760000056
represents the predicted value of the subcutaneous interstitial glucose concentration after one period in the future,
Figure BDA0002455169760000057
the subcutaneous interstitial glucose concentration after two future cycles is shown.
An adaptive glucose and insulin concentration prediction method, the prediction method comprising:
collecting monitoring data of a person to be monitored; the monitoring data includes an initial subcutaneous interstitial glucose concentration, an exogenous insulin infusion rate, and a body weight of the subject to be monitored;
constructing a blood sugar-insulin model according to the monitoring data of the person to be monitored; the blood glucose-insulin model comprises: an insulin transmission sub-model, a glycemic insulin dynamics sub-model, and a blood glucose transmission sub-model;
converting the blood glucose-insulin model to a state space model;
converting the state space model into a discrete state space model;
predicting and updating the state variables in the discrete state space model; the updated state variables include an effective insulin amount in one insulin chamber, an effective insulin amount in two insulin chambers, a concentration of glucose in plasma, a concentration of glucose in subcutaneous intercellular substance, an insulin sensitivity coefficient, a basal effective insulin concentration, a basal blood glucose level, a glycemic autoregulation rate, and a carbohydrate intake factor;
determining the carbohydrate intake condition of the person to be monitored according to the updated carbohydrate intake factor; intake profiles include intake with and without carbohydrate;
the insulin concentration in plasma, the glucose concentration in plasma and the glucose concentration in the subcutaneous intercellular substance are determined for a future fixed time interval based on the updated state variables.
Optionally, the insulin delivery submodel specifically adopts the following formula:
Figure BDA0002455169760000061
wherein x is1(t) is the effective insulin amount in an insulin chamber, x2(t) the effective insulin amount in the diloinsulin chamber; u (t) exogenous insulin infusion rate, tIThe time at which the concentration of insulin effective reaches a maximum, x (t) is the plasmaThe concentration of insulin, W is the weight of the person to be monitored, M is the human insulin clearance rate, and M is 0.017 (l/kg/min);
the blood glucose insulin dynamics submodel specifically adopts the following formula:
Figure BDA0002455169760000062
wherein G (t) is the concentration of glucose in plasma, SIIs the sensitivity coefficient of insulin action, Gb(ii) is the concentration level of glucose in the base plasma, K is the rate of concentration self-regulation of glucose in the plasma, u (t) is the carbohydrate uptake factor that causes the concentration change of glucose in the plasma;
the blood glucose transmission sub-model specifically adopts the following formula:
Figure BDA0002455169760000063
wherein IG (t) is the concentration of glucose in the subcutaneous intercellular matrix, and τ is a time-lag factor.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
according to the self-adaptive glucose and insulin concentration prediction system and method provided by the invention, each parameter and system state variable are dynamically updated through the particle observer module, so that the prediction of the concentrations of insulin and glucose in plasma is dynamically updated in real time, and the open-loop estimation of the concentrations of glucose and insulin in plasma is not directly carried out according to a fixed parameter model and collected data. According to the invention, the concentration of insulin and glucose in blood plasma can be predicted by determining the concentration of insulin and glucose in blood plasma, the accuracy of predicting the concentration of glucose in blood plasma can be improved, and the accurate control of the concentration of glucose in blood plasma can be realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a block diagram of an adaptive glucose and insulin concentration prediction system according to the present invention;
FIG. 2 is a graph showing the predicted results of 7 days of plasma glucose concentration in a diabetic patient according to the present invention;
FIG. 3 is a graph showing the predicted results of 7 days of plasma insulin concentration in a diabetic patient according to the present invention;
FIG. 4 is a 7-day prediction of subcutaneous interstitial glucose concentration in a type of diabetic patient according to the present invention;
FIG. 5 is a graph of the automatic detection of 5 days of food intake in a type of diabetic patient according to the present invention;
FIG. 6 is a flow chart of a method for adaptively predicting glucose and insulin concentrations according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a self-adaptive glucose and insulin concentration prediction system and a method, which can realize intelligent and accurate prediction of the concentrations of glucose and insulin.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Fig. 1 is a structural diagram of an adaptive glucose and insulin concentration prediction system provided by the present invention, and as shown in fig. 1, the adaptive glucose and insulin concentration prediction system provided by the present invention includes: the system comprises a data acquisition module 101, a physiological model set construction module 102, a state space conversion module 103, a discrete state space model construction module 104, a particle wave observer module 105, a feeding automatic detection module 106 and a blood sugar prediction module 107. The sampling period of the data acquisition system is T.
The data acquisition module 101 is used for acquiring monitoring data of a person to be monitored; the monitoring data includes initial subcutaneous interstitial glucose concentration, exogenous insulin infusion rate, and body weight of the subject to be monitored.
The physiological model set construction module 102 is configured to construct a blood glucose-insulin model according to the monitoring data of the person to be monitored; the blood glucose-insulin model comprises: an insulin transmission sub-model, a glycemic insulin dynamics sub-model, and a blood glucose transmission sub-model.
The state space conversion module 103 is configured to convert the blood glucose-insulin model into a state space model; the state variables in the state space model comprise state variables converted by the system variables of the blood glucose insulin model and state variables converted by system parameter extension in the blood glucose insulin model.
Discrete state space model construction module 104 is used to convert the state space model into a discrete state space model. The particle wave observer module 105 is configured to predict and update the state variables in the discrete state space model; the updated state variables include the effective insulin amount in one insulin chamber, the effective insulin amount in two insulin chambers, the concentration of glucose in plasma, the concentration of glucose in the subcutaneous intercellular substance, the insulin sensitivity coefficient, the basal effective insulin concentration, the basal blood glucose level, the rate of self-regulation of blood glucose, and the carbohydrate intake factor.
The automatic food intake detection module 106 is used for determining the carbohydrate intake condition of the person to be monitored according to the updated carbohydrate intake factor; the intake profile includes intake with and without carbohydrate.
The blood glucose prediction module 107 is configured to determine a concentration of insulin in plasma, a concentration of glucose in plasma, and a concentration of glucose in subcutaneous intercellular substance at a future fixed time interval based on the updated state variables.
The insulin transmission submodel specifically adopts the following formula:
Figure BDA0002455169760000081
wherein x is1(t) is the effective insulin amount in an insulin chamber, x2(t) the effective insulin amount in the diloinsulin chamber, u (t) the exogenous insulin infusion rate, tIIs the time at which the concentration of active insulin reaches a maximum, x (t) is the concentration of insulin in plasma, W is the body weight of the subject to be monitored, M is the rate of insulin clearance from the human body, and M is 0.017 (l/kg/min).
The blood glucose insulin dynamics submodel specifically adopts the following formula:
Figure BDA0002455169760000091
wherein G (t) is the concentration of glucose in plasma, SIIs the sensitivity coefficient of insulin action, GbK is the rate of concentration self-regulation of glucose in plasma, and u (t) is the carbohydrate uptake factor that causes the concentration change of glucose in plasma.
The blood glucose transmission sub-model specifically adopts the following formula:
Figure BDA0002455169760000092
wherein IG (t) is the concentration of glucose in the subcutaneous intercellular matrix, and τ is a time-lag factor.
The state space conversion module specifically adopts the following formula:
Figure BDA0002455169760000093
wherein X is a state variable, X ═ X1,x2,G,IG,Si,tI,K,Gb,τ,U]，H＝[0,0,0,1,0,0,0,0,0,0]ω (t) is process noise, θ (t) is measurement noise, and z (t) is measurement state.
The discrete state space model construction module specifically adopts the following formula:
Figure BDA0002455169760000094
wherein, Xk-1Is a state variable at time k-1, uk-1Exogenous insulin infusion Rate at time k-1, omegakFor process noise and obedience to be 0, the variance is σwkGaussian noise of (v)kTo measure noise and obey a desired 0, the variance σvkGaussian noise.
In a specific embodiment, the variance vector of each state variance component is represented as σw＝[10-4,10-4,0.01,1,10-6,10-6,10-6,10-6,10-6,1]Variance of observed noise is σvk＝1。
According to the above formula, when the data acquisition module samples 4 times per hour, the continuous state variable of the state space conversion module is converted into a continuous discrete state variable for discrete data.
The particle wave observer module specifically predicts and updates by using the following formula:
p(Xk|z1:k-1)＝∫p(Xk|Xk-1)p(Xk-1|z1:k-1)dXk-1
p(Xk|z1:k)∝p(zk|Xk-1)p(Xk|z1:k-1)
wherein, XkIs a state variable at time k, Xk-1Is the state variable at time k-1, z1:k-1For all blood glucose data sequences acquired before the time k-1, z1:kAll blood glucose data sequences acquired before the k-th time are acquired.
The optimization model in the particle filter observer module is solved based on the Monte Carlo method of Sequential Importance Sampling (SIS), i.e. the probability density function p (X)k|z1:k) By oneSome random particles
Figure BDA0002455169760000102
And their combined approximation, the weights and the positions of the particles can be adjusted according to the measurement results after some weighted random samples are obtained.
The pseudo code of the particle filter observer module is:
Figure BDA0002455169760000101
the automatic food intake detection module specifically comprises: the device comprises a forward difference acquisition unit and a judgment unit.
The forward difference acquisition unit is used for acquiring the forward difference of the external interference factor at the time k
Figure BDA0002455169760000111
Wherein the content of the first and second substances,
Figure BDA0002455169760000112
Uk-1carbohydrate uptake factor, U, being the induced change in plasma glucose concentration at time k-1kCarbohydrate uptake factor which is the induced change in concentration of glucose in plasma at time k.
A judging unit for judging whether the forward difference is larger than the predetermined threshold value
Figure BDA0002455169760000113
And (4) judging whether the patient takes the carbohydrate or not.
The judging unit specifically includes:
judgment of
Figure BDA0002455169760000114
And
Figure BDA0002455169760000115
whether or not to simultaneously satisfy
Figure BDA0002455169760000116
If at the same time satisfy
Figure BDA0002455169760000117
Then there is carbohydrate intake, otherwise there is no carbohydrate intake; wherein the content of the first and second substances,
Figure BDA0002455169760000118
is the forward difference of the external interference factor at time k,
Figure BDA0002455169760000119
is the forward difference of the external interference factor at the moment of k-1, Threshold is a set Threshold value, Flag k-i0 is an indication of no carbohydrate intake.
The blood glucose prediction module dynamically predicts the insulin concentration in the plasma of the target object in the future 30 minutes based on the real-time updated system state of the particle wave observer module
Figure BDA00024551697600001110
Glucose concentration in plasma
Figure BDA00024551697600001111
And the glucose concentration of the subcutaneous intercellular substance
Figure BDA00024551697600001112
Namely:
dynamically predicting the future 30-minute plasma insulin concentration of the target object based on the real-time updated system state of the particle wave observer module
Figure BDA00024551697600001113
Glucose concentration in plasma
Figure BDA00024551697600001114
And the glucose concentration of the subcutaneous intercellular substance
Figure BDA00024551697600001115
Namely:
Figure BDA0002455169760000121
Figure BDA0002455169760000122
Figure BDA0002455169760000123
Figure BDA0002455169760000124
Figure BDA0002455169760000125
Figure BDA0002455169760000126
Figure BDA0002455169760000127
Figure BDA0002455169760000128
Figure BDA0002455169760000129
Figure BDA00024551697600001210
in the embodiment of the invention, the method is verified by applying blood sugar simulation software certified by the Food and Drug Administration (FDA), the detailed physiological parameters of 30 patients are recorded in the system, and when external conditions (information such as insulin infusion amount and food intake) are input, the software simulates the blood sugar change condition of the patients on the basis of the physiological data. In this example, all patients had their insulin infusion as their own during the time period in which the patient parameters were collected, and the meals were taken and ingested three meals a day normally, resulting in 7-day changes in the blood glucose system for all patients. In the embodiment of the invention, the data are continuously input into the system designed by the scheme only by utilizing the time sequence of the blood glucose measurement value and the exogenous insulin infusion amount, the insulin concentration and the glucose concentration in the blood plasma are tracked and predicted, and the blood glucose value in the future 30 minutes is predicted at the same time. For one of the patients to be simulated, the results shown in fig. 2-5 were obtained by comparing the blood glucose measurements with the blood plasma insulin concentration, glucose concentration and blood glucose concentration generated by the blood glucose simulation software. The results of all patients were statistically analyzed, and the rms error of the prediction for plasma insulin concentrations was 0.1283, the rms error of the prediction for plasma glucose concentrations was 1.5602, and the rms error of the prediction for subcutaneous interstitial glucose concentrations was 0.0145. The accuracy rate of the food intake detection is 86.67%, the false report rate is 18.75%, and the false report rate is 13.33%.
Fig. 6 is a schematic flow chart of a method for predicting the concentration of adaptive glucose and insulin provided by the present invention, and as shown in fig. 6, the method for predicting the concentration of adaptive glucose and insulin provided by the present invention includes:
s601, collecting monitoring data of a person to be monitored; the monitoring data includes initial subcutaneous interstitial glucose concentration, exogenous insulin infusion rate, and body weight of the subject to be monitored.
S602, constructing a blood glucose-insulin model according to the monitoring data of the person to be monitored; the blood glucose-insulin model includes: an insulin delivery sub-model, a glucose insulin dynamics sub-model, and a glucose delivery sub-model.
The insulin transmission submodel specifically adopts the following formula:
Figure BDA0002455169760000131
wherein x is1(t) is the effective insulin amount in an insulin chamber, x2(t) the effective insulin amount in the diloinsulin chamber, u (t) the exogenous insulin infusion rate, tI(ii) is the time at which the concentration of available insulin reaches a maximum, x (t) is the concentration of insulin in plasma, W is the body weight of the subject to be monitored, M is the rate of insulin clearance from the human body, and M is 0.017 (l/kg/min);
the blood glucose insulin dynamics submodel specifically adopts the following formula:
Figure BDA0002455169760000132
wherein G (t) is the concentration of glucose in plasma, SIIs the sensitivity coefficient of insulin action, GbK is the rate of concentration self-regulation of glucose in plasma, and u (t) is the carbohydrate uptake factor that causes the concentration change of glucose in plasma.
The blood glucose transmission sub-model specifically adopts the following formula:
Figure BDA0002455169760000133
wherein IG (t) is the concentration of glucose in the subcutaneous intercellular matrix, and τ is a time-lag factor.
S603, converting the blood sugar-insulin model into a state space model.
Figure BDA0002455169760000141
Wherein, X is the position of the weight and the particle which can be adjusted according to the measurement result. Variable, X ═ X1,x2,G,IG,Si,tI,K,Gb,τ,U]，H＝[0,0,0,1,0,0,0,0,0,0]ω (t) is process noise, θ (t) is measurement noise, and z (t) is measurement state.
S604, converting the state space model into a discrete state space model.
S605, predicting and updating the state variables in the discrete state space model; the updated state variables include the effective insulin amount in one insulin chamber, the effective insulin amount in two insulin chambers, the concentration of glucose in plasma, the concentration of glucose in the subcutaneous intercellular substance, the insulin sensitivity coefficient, the basal effective insulin concentration, the basal blood glucose level, the rate of self-regulation of blood glucose, and the carbohydrate intake factor.
S606, determining the carbohydrate intake condition of the person to be monitored according to the updated carbohydrate intake factor; the intake profile includes intake with and without carbohydrate.
And S607, determining the insulin concentration in the blood plasma, the glucose concentration in the blood plasma and the glucose concentration in the subcutaneous intercellular substance in a future fixed time interval according to the updated state variables.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.

Claims (10)

1. An adaptive glucose and insulin concentration prediction system, the prediction system comprising:
the data acquisition module is used for acquiring monitoring data of a person to be monitored; the monitoring data includes an initial subcutaneous interstitial glucose concentration, an exogenous insulin infusion rate, and a body weight of the subject to be monitored;
the physiological model set construction module is used for constructing a blood sugar-insulin model according to the monitoring data of the person to be monitored; the blood glucose-insulin model comprises: an insulin transmission sub-model, a glycemic insulin dynamics sub-model, and a blood glucose transmission sub-model;
the state space conversion module is used for converting the blood glucose-insulin model into a state space model, and state variables in the state space model comprise state variables converted by system variables of the blood glucose-insulin model and state variables converted by system parameter extension in the blood glucose-insulin model;
the discrete state space model building module is used for converting the state space model into a discrete state space model;
the particle wave observer module is used for predicting and updating the state variables in the discrete state space model; the updated state variables include an effective insulin amount in one insulin chamber, an effective insulin amount in two insulin chambers, a concentration of glucose in plasma, a concentration of glucose in subcutaneous intercellular substance, an insulin sensitivity coefficient, a basal effective insulin concentration, a basal blood glucose level, a glycemic autoregulation rate, and a carbohydrate intake factor;
the automatic eating detection module is used for determining the carbohydrate intake condition of the person to be monitored according to the updated carbohydrate intake factor; intake profiles include intake with and without carbohydrate;
and the blood glucose prediction module is used for determining the insulin concentration in the blood plasma, the glucose concentration in the blood plasma and the glucose concentration in the subcutaneous intercellular substance within a future fixed time interval according to the updated state variable.
2. The adaptive glucose and insulin concentration prediction system of claim 1, wherein the insulin delivery submodel specifically uses the following equation:
Figure FDA0002455169750000011
wherein the content of the first and second substances,
Figure FDA0002455169750000012
wherein x is1(t) is the effective insulin amount in an insulin chamber, x2(t) the effective insulin amount in the diloinsulin chamber; u (t) exogenous insulin infusion rate, tI(ii) is the time at which the concentration of available insulin reaches a maximum, x (t) is the concentration of insulin in plasma, W is the body weight of the subject to be monitored, M is the rate of insulin clearance from the human body, and M is 0.017 (l/kg/min);
the blood glucose insulin dynamics submodel specifically adopts the following formula:
Figure FDA0002455169750000021
wherein G (t) is the concentration of glucose in plasma, SIIs the sensitivity coefficient of insulin action, Gb(ii) is the concentration level of glucose in the base plasma, K is the rate of concentration self-regulation of glucose in the plasma, u (t) is the carbohydrate uptake factor that causes the concentration change of glucose in the plasma;
the blood glucose transmission sub-model specifically adopts the following formula:
Figure FDA0002455169750000022
wherein IG (t) is the concentration of glucose in the subcutaneous intercellular matrix, and τ is a time-lag factor.
3. The adaptive glucose and insulin concentration prediction system of claim 2, wherein the state space transformation module is specifically configured to use the following equation:
Figure FDA0002455169750000023
wherein X is a state variable, X ═ X1,x2,G,IG,Si,tI,K,Gb,τ,U]，H＝[0,0,0,1,0,0,0,0,0,0]ω (t) is process noise, θ (t) is measurement noise, and z (t) is measurement state.
4. The adaptive glucose and insulin concentration prediction system of claim 1, wherein the discrete state space model construction module specifically employs the following equation:
Figure FDA0002455169750000024
wherein, Xk-1Is a state variable at time k-1, uk-1Exogenous insulin infusion Rate at time k-1, omegakFor process noise and obedience to be 0, the variance is σwkGaussian noise of (v)kTo measure noise and obey a desired 0, the variance σvkGaussian noise (g).
5. The adaptive glucose and insulin concentration prediction system of claim 4, wherein the particle observer module predicts and updates using the following equations:
Figure FDA0002455169750000031
wherein, XkIs a state variable at time k, Xk-1Is a state variable at time k-1, z1:k-1For all blood glucose data sequences acquired before the time point k-1, z1:kAll blood glucose data sequences acquired before the k-th time are acquired.
6. The adaptive glucose and insulin concentration prediction system of claim 1, wherein the automatic meal detection module specifically comprises:
a forward difference obtaining unit for obtaining the forward difference of the external interference factor at the time k
Figure FDA0002455169750000032
Wherein the content of the first and second substances,
Figure FDA0002455169750000033
Uk-1carbohydrate uptake factor, U, being the induced change in plasma glucose concentration at time k-1kA carbohydrate uptake factor that is the induced change in concentration of glucose in plasma at time k;
a judging unit for judging whether the forward difference is larger than the predetermined threshold value
Figure FDA0002455169750000034
And (4) judging whether the patient takes the carbohydrate or not.
7. The adaptive glucose and insulin concentration prediction system according to claim 6, wherein the determining unit specifically comprises:
judgment of
Figure FDA0002455169750000035
And
Figure FDA0002455169750000036
whether or not to simultaneously satisfy
Figure FDA0002455169750000037
If at the same time satisfy
Figure FDA0002455169750000038
Then there is carbohydrate intake, otherwise there is no carbohydrate intake; wherein the content of the first and second substances,
Figure FDA0002455169750000039
is the forward difference of the external interference factor at time k,
Figure FDA00024551697500000310
is the forward difference of the external interference factors at the moment of k-1, Threshold is a set Threshold value, Flagk-i0 is an indication of no carbohydrate intake.
8. The adaptive glucose and insulin concentration prediction system of claim 6, wherein the blood glucose prediction module specifically uses the following formula:
Figure FDA0002455169750000041
wherein the content of the first and second substances,
Figure FDA0002455169750000042
indicating the predicted value of the insulin concentration in the plasma after one period in the future,
Figure FDA0002455169750000043
represents the predicted value of the insulin concentration in plasma after two future sampling periods,
Figure FDA0002455169750000044
represents the predicted value of the glucose concentration in the plasma after one period in the future,
Figure FDA0002455169750000045
represents the predicted value of the glucose concentration in plasma after the next two cycles,
Figure FDA0002455169750000046
represents the predicted value of the subcutaneous interstitial glucose concentration after one period in the future,
Figure FDA0002455169750000047
the subcutaneous interstitial glucose concentration after two future cycles is shown.
9. An adaptive glucose and insulin concentration prediction method, the prediction method comprising:
collecting monitoring data of a person to be monitored; the monitoring data includes an initial subcutaneous interstitial glucose concentration, an exogenous insulin infusion rate, and a body weight of the subject to be monitored;
constructing a blood sugar-insulin model according to the monitoring data of the person to be monitored; the blood glucose-insulin model comprises: an insulin delivery sub-model, a glucose insulin dynamics sub-model, and a glucose delivery sub-model;
converting the blood glucose-insulin model to a state space model;
converting the state space model into a discrete state space model;
predicting and updating the state variables in the discrete state space model; the updated state variables include an effective insulin amount in one insulin chamber, an effective insulin amount in two insulin chambers, a concentration of glucose in plasma, a concentration of glucose in subcutaneous intercellular substance, an insulin sensitivity coefficient, a basal effective insulin concentration, a basal blood glucose level, a glycemic autoregulation rate, and a carbohydrate intake factor;
determining the carbohydrate intake condition of the person to be monitored according to the updated carbohydrate intake factor; intake profiles include intake with and without carbohydrate;
the insulin concentration in plasma, the glucose concentration in plasma and the glucose concentration in the subcutaneous intercellular substance are determined for a future fixed time interval based on the updated state variables.
10. The adaptive glucose and insulin concentration prediction method of claim 9, wherein the insulin delivery submodel specifically uses the following formula:
Figure FDA0002455169750000051
wherein the content of the first and second substances,
Figure FDA0002455169750000052
wherein x is1(t) is the effective insulin amount in an insulin chamber, x2(t) the effective insulin amount in the diloinsulin chamber; u (t) rate of exogenous insulin infusion,tI(ii) is the time at which the concentration of available insulin reaches a maximum, x (t) is the concentration of insulin in plasma, W is the body weight of the subject to be monitored, M is the rate of insulin clearance from the human body, and M is 0.017 (l/kg/min);
the blood glucose insulin dynamics submodel specifically adopts the following formula:
Figure FDA0002455169750000053
wherein G (t) is the concentration of glucose in plasma, SISensitivity coefficient of insulin action, Gb(ii) is the concentration level of glucose in the base plasma, K is the rate of concentration self-regulation of glucose in the plasma, u (t) is the carbohydrate uptake factor that causes the concentration change of glucose in the plasma;
the blood glucose transmission sub-model specifically adopts the following formula:
Figure FDA0002455169750000054
wherein IG (t) is the glucose concentration in the subcutaneous intercellular substance, and τ is the time lag factor.
CN202010304292.3A 2020-04-17 2020-04-17 Self-adaptive glucose and insulin concentration prediction system and method Active CN111540436B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010304292.3A CN111540436B (en) 2020-04-17 2020-04-17 Self-adaptive glucose and insulin concentration prediction system and method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010304292.3A CN111540436B (en) 2020-04-17 2020-04-17 Self-adaptive glucose and insulin concentration prediction system and method

Publications (2)

Publication Number Publication Date
CN111540436A CN111540436A (en) 2020-08-14
CN111540436B true CN111540436B (en) 2022-06-03

Family

ID=71978694

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010304292.3A Active CN111540436B (en) 2020-04-17 2020-04-17 Self-adaptive glucose and insulin concentration prediction system and method

Country Status (1)

Country Link
CN (1) CN111540436B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112402732B (en) * 2020-10-10 2023-05-05 广东食品药品职业学院 Insulin infusion quantity control method based on self-adaptive control weighting factor strategy
CN113506632B (en) * 2021-06-08 2023-12-12 山东第一医科大学附属省立医院(山东省立医院) Non-periodic sampling data-based estimation method for blood sugar of type I diabetes patient
CN116159208B (en) * 2021-11-24 2024-03-15 上海微创生命科技有限公司 Artificial pancreas control method, readable storage medium and blood glucose management system
CN116504355B (en) * 2023-04-27 2024-04-02 广东食品药品职业学院 Closed-loop insulin infusion control method, device and storage medium based on neural network

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1471373A (en) * 2001-02-05 2004-01-28 ��³��ɭ˹��˾ Methods of determining concentration of glucose in blood
CN1696931A (en) * 2004-05-11 2005-11-16 希森美康株式会社 Simulation system for simulating material concentration in a living body and storage medium
EP2236077A1 (en) * 2009-03-31 2010-10-06 Sensile Pat AG Medical device for measuring an analyte concentration
CN110869075A (en) * 2017-04-07 2020-03-06 生命扫描知识产权控股有限责任公司 Calculation of residual insulin Activity content in Artificial islet systems

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1471373A (en) * 2001-02-05 2004-01-28 ��³��ɭ˹��˾ Methods of determining concentration of glucose in blood
CN1696931A (en) * 2004-05-11 2005-11-16 希森美康株式会社 Simulation system for simulating material concentration in a living body and storage medium
EP2236077A1 (en) * 2009-03-31 2010-10-06 Sensile Pat AG Medical device for measuring an analyte concentration
CN110869075A (en) * 2017-04-07 2020-03-06 生命扫描知识产权控股有限责任公司 Calculation of residual insulin Activity content in Artificial islet systems

Also Published As

Publication number Publication date
CN111540436A (en) 2020-08-14

Similar Documents

Publication Publication Date Title
CN111540436B (en) Self-adaptive glucose and insulin concentration prediction system and method
Mirshekarian et al. Using LSTMs to learn physiological models of blood glucose behavior
CN105046042B (en) Guard technology for closed-loop insulin infusion system
EP2552313B1 (en) Calibration of glucose monitoring sensor and/or insulin delivery system
CN103959291B (en) Glucose predictions device based on the regularization network with adaptively selected core and regularization parameter
Allam et al. A recurrent neural network approach for predicting glucose concentration in type-1 diabetic patients
JP2014534483A (en) Method, system and computer-readable recording medium for adaptive advice control of diabetes
Facchinetti et al. An online self-tunable method to denoise CGM sensor data
JP2001204817A (en) System for extrapolation of glucose concentration
WO2011051922A2 (en) Method to recalibrate continuous glucose monitoring data on-line
JP2010521222A (en) Basic speed test using blood glucose input
KR102601111B1 (en) Closed-loop blood glucose control system and method
Zhao et al. Rapid model identification for online subcutaneous glucose concentration prediction for new subjects with type I diabetes
Vettoretti et al. Patient decision-making of CGM sensor driven insulin therapies in type 1 diabetes: In silico assessment
Kovatchev et al. Peculiarities of the continuous glucose monitoring data stream and their impact on developing closed-loop control technology
Chen et al. Committed moving horizon estimation for meal detection and estimation in type 1 diabetes
Vettoretti et al. Modeling the error of factory-calibrated continuous glucose monitoring sensors: application to Dexcom G6 sensor data
Prud'Homme et al. Preclinically assessed optimal control of postprandial glucose excursions for type 1 patients with diabetes
Othman et al. Optimization of identifying insulinaemic pharmacokinetic parameters using artificial neural network
Buckingham Clinical overview of continuous glucose monitoring
Chase et al. Active insulin infusion control of the blood glucose derivative
Ata et al. Model based control of artificial pancreas under meal disturbances
CN114145738A (en) Analyte concentration data generation method and device, and system for monitoring analyte level
Shuja et al. A switching based PID technique for blood glucose control
Wang et al. Adaptive model predictive control with particle filter for artificial pancreas

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant