Unknown exogenous input estimation-based postprandial long-time window blood glucose situation prediction method

Abstract

The invention discloses a postprandial long-time window blood glucose situation prediction method based on unknown exogenous input estimation, and belongs to the technical field of blood glucose prediction, and the method comprises the following steps: S1, establishing a blood glucose metabolism kinetic model for a glucose-insulin metabolism physiological process of a diabetic blood glucose patient; s2, building a fixed-order simultaneous input and state set value observer according to the blood glucose metabolism kinetic model; and S3, observing unknown exogenous input of the patient by adopting an observer, summarizing the living laws of the patient, predicting the unknown exogenous input of the current meal, and performing accurate blood glucose prediction in combination with known treatment information. According to the postprandial long-time window blood glucose situation prediction method based on unknown exogenous input estimation, the influence of factors difficult to quantify or measure, such as the living law and the body rhythm of the patient, on the blood glucose of the patient can be considered in the blood glucose prediction process, and the personalized blood glucose dynamic law of the patient is effectively captured; and postprandial long-time window blood glucose prediction in a small sample scene is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of blood sugar prediction, and in particular to a method for predicting blood sugar status in a long postprandial window based on unknown exogenous input estimation. Background Art

[0002] With the development of modern diabetes technology, intelligent blood sugar management methods have gradually become popular in clinical practice. In order to achieve effective blood sugar control in diabetic patients, the blood sugar of patients should be continuously monitored, and abnormal blood sugar phenomena of patients should be discovered and predicted in time, and corresponding intervention or preventive measures should be taken.

[0003] Accurate blood sugar prediction is the basis for achieving effective closed-loop blood sugar control. Based on known blood sugar monitoring information, the patient's blood sugar level and abnormal blood sugar events that may occur in the future can be predicted. According to the prediction results, corresponding blood sugar management measures can be implemented, such as adjusting the insulin injection dose or dietary intake to control the patient's blood sugar within a safe range and maintain the patient's health.

[0004] At present, the research on diabetes blood glucose prediction mainly focuses on the following three directions: model prediction, data-driven model prediction and hybrid model prediction. The model prediction method establishes a kinetic model based on the physiological laws of human blood glucose-insulin metabolism. Commonly used models include Bergman's minimum model, Hovorka model, etc. These models are composed of differential equations of varying degrees of complexity. The model-based prediction method can ensure that the prediction results are consistent with physiological laws, but it has weak flexibility during use. In order to achieve higher-precision predictions, it is usually necessary to use a larger-scale model, and there are many physiological parameters to be determined in the model, which are difficult to identify, calculate and use in real scenarios.

[0005] Data-driven models are built based on patients' blood glucose monitoring data, insulin treatment data, patient meal information, and other available patient information. They are often referred to as black box methods. Such models usually require the support of machine learning techniques. Common methods include time series models (ARX, ARMAX, ARIMAX, etc.), fuzzy logic models, Gaussian mixture models, regular learning, reinforcement learning models, Kalman filters, artificial neural network models, etc. Hybrid models combine two or more different methods in the data preprocessing, feature extraction, or learning stages. Many hybrid models combine physiological models with machine learning techniques to improve performance.

[0006] Although existing methods have explored blood glucose prediction from different aspects, the problem of long-term post-meal prediction under small sample conditions has not been fully solved. In particular, how to comprehensively consider the combined effects of multiple unknown interferences that are difficult to directly measure on blood glucose status remains a major challenge.

[0007] On the one hand, in actual clinical medical scenarios, it is often necessary to make blood glucose predictions with only a small amount of individualized blood glucose and other data from the patient. In current research, blood glucose model training often requires a large amount of training data. However, in actual clinical situations, the amount of blood glucose data for each patient is limited and it is difficult to meet the model training needs. Due to the insufficient amount of individual data, some studies choose to use group data sets for model training. However, in actual situations, the individual differences in blood glucose status between patients are large, and models trained with group data are difficult to achieve highly flexible personalized blood glucose predictions.

[0008] On the other hand, there are many factors that affect blood sugar in the long period after a meal, including many factors that are difficult to quantify, such as the patient's exercise activities, snacks, body rhythms, stress levels, etc. The impact of these factors on the patient's blood sugar is difficult to directly reflect in physiological models or data-driven models, which may lead to poor blood sugar prediction results. The above factors can be collectively referred to as unknown exogenous inputs that affect blood sugar, which are closely related to the patient's daily living habits. People's daily activities and body rhythms often follow relatively stable habits, and diabetic patients may have more regular living habits due to routine blood sugar management behaviors. In order to achieve self-blood sugar management, patients often have to eat, inject insulin, check blood sugar and other routine blood sugar management behaviors in a specific time interval. Their eating activities and daily exercise behaviors outside the fixed time of three meals are also limited by blood sugar control targets. The regularity of the above living habits enables us to summarize the patient's living habits more accurately in a small sample, and predict the unknown exogenous inputs in the future based on the above rules.

[0009] Based on the above content, it is proposed to integrate a variety of blood glucose influencing factors that are difficult to quantify and measure as unknown exogenous inputs in the prediction, and establish a long-term postprandial blood glucose prediction method based on a small sample suitable for clinical treatment scenarios. Summary of the invention

[0010] The purpose of the present invention is to provide a method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation to solve the problems in the background technology.

[0011] To achieve the above object, the present invention provides a method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation, comprising the following steps:

[0012] S1. A blood glucose metabolism kinetic model is established for the physiological process of glucose-insulin metabolism in diabetic patients to simulate the physiological response of blood glucose to carbohydrate intake and insulin treatment strategies;

[0013] S2. According to the structure of the blood glucose metabolism dynamics model in S1, a fixed-order simultaneous input and state set value observer is constructed, which can observe the unknown exogenous inputs that affect blood glucose within the meal time window based on the known patient blood glucose data, thereby reflecting the patient's life pattern and body rhythm, and observing and updating the unknown exogenous inputs that affect blood glucose within the meal time window;

[0014] S3. Use an observer to observe the unknown exogenous input of the patient's past data, summarize the patient's life pattern based on the observation results, use the weighted moving average method to predict the unknown exogenous input of the current meal, and combine the known treatment information to make accurate blood sugar predictions.

[0015] Preferably, the blood glucose metabolism kinetic model consists of a core sub-model, a dietary absorption sub-model and an insulin transport sub-model, wherein the core sub-model reflects the interaction between glucose and insulin, and the kinetic equation is:

[0016]

[0017] Where G(t) is the plasma glucose concentration, G b is the basal glucose concentration related to the patient's basal rate of insulin delivery, I p (t) is the mass of plasma insulin, I b is the basal plasma insulin concentration, R a (t) is the rate of plasma glucose appearance, X(t) is the insulin effect, S g Indicates glucose availability, S I is insulin sensitivity, which represents the ability of insulin to control glucose production and use, p2 is the delay between insulin concentration and insulin action, and V g and V I are the glucose and insulin distribution volumes, respectively, and BW is the subject’s body weight;

[0018] The dietary absorption submodel reflects the patient's dietary intake behavior, and the kinetic equation is:

[0019]

[0020] R a (t) = k q1 f Q1(t)+k q1 ·f·Q2(t);

[0021] Where Q1(t) and Q2(t) are the amounts of food in the first and second food absorption compartments, k q1 , k q12 and k q2 is the rate constant, f is the dietary bioavailability, u m(t) is dietary input;

[0022] The kinetic equation of the insulin transporter model is:

[0023]

[0024] In the formula, I sc1 (t) and I sc2 (t) represents the mass of insulin in the two subcutaneous transport compartments, I p (t) is the mass of plasma insulin, k c1 , k c2 and k c12 is the rate constant for subcutaneous insulin transport, k cl is the insulin clearance rate, u i (t) is subcutaneous insulin infusion.

[0025] Preferably, the specific steps of establishing the blood glucose metabolism kinetics model in S1 are:

[0026] 1) First, the core sub-model, dietary absorption sub-model and insulin transport sub-model are linearized. At the same time, there are many factors that affect blood sugar levels in real scenarios. In addition to insulin input and dietary input, a variety of blood sugar influencing factors that are difficult to quantify and measure are considered, such as patients' daily activities, body hormone fluctuations, etc. The way these factors affect blood sugar does not follow any specific deterministic physiological laws;

[0027] The above factors are combined into the unknown exogenous input d(t) and incorporated into the blood glucose model to consider its impact on blood glucose. The unknown exogenous input is not constrained to be any type of signal, and it is not assumed to follow any model. The continuous linear time-invariant blood glucose state space model after adding the unknown exogenous input is obtained as follows:

[0028]

[0029] y(t)=C c x(t);

[0030] Where d(t) is the unknown external input signal to be estimated, u i (t) and u m (t) are the known insulin input and meal carbohydrate intake, y(t) is the model output, i.e., blood glucose value, and x(t) is the state vector defined as:

[0031] x(t)=[G(t) X(t) I sc1 (t) I sc2 (t) I p (t) Q1(t) Q2(t)] T ;

[0032] Input matrix B ci , B cm , B cd , output matrix C c And the state transfer matrix A c The specific forms are:

[0033] B ci =[0 0 1 0 0 0 0] T ;

[0034] B cm =[0 0 0 0 0 1 0] T ;

[0035] B cd =[1 0 0 0 0 0 0] T ;

[0036] C c =[1 0 0 0 0 0 0];

[0037]

[0038] 2) Then, the continuous linear time-invariant blood glucose state space model after adding unknown exogenous input is discretized, and the discrete step length h=5min is preferably set to efficiently capture and monitor the patient's blood glucose behavior, and the discrete blood glucose model is obtained as follows:

[0039] x k+1 =Ax k +Bu k +Gd k +Wω k ;

[0040] y k =Cx k +Du k +Hd k +v k ;

[0041] Where k is the discrete time step, k = 1, 2, 3..., T, x k ,d k and k are the sampling sequences of x(t), d(t) and y(t), respectively. k+1 is the state vector at the k+1th sampling time; u k =[u ik ;u mk ] is a known input vector. The specific input time and value of the patient's insulin input and carbohydrate input are obtained by reading the patient's case. ik and u mk u i(t),u m (t) sampling sequence; ω k is the process noise, v k To measure the noise, we simulate the bounded noise in the actual blood glucose testing equipment and set ‖ω k ‖≤η ω =10 -4 , ‖v k ‖≤η v =10 -2 ;

[0042] The matrices A, B, G, W, C, D, and H in the discrete blood glucose model are:

[0043] A=h·A c +I7;B i =h·B ci ; B m =h·B cm ; B = [B i B m ]; G = h·B cd ;W=I7;C=C c ; D = 0 1×2 ; H = 0;

[0044] Where I7 represents the 7×7 unit matrix.

[0045] Preferably, in S2, the specific steps of building a fixed-order simultaneous input and state set-valued observer are:

[0046] 1) Perform singular value decomposition on the matrix H in S1:

[0047]

[0048] definition are two unitary matrices obtained after singular value decomposition of H, ∑ is a diagonal matrix obtained after singular value decomposition, and the rank of H matrix is ​​p H , U1 and V1 are the first p of the U matrix and V matrix respectively. H The matrix consists of columns, U2 and V2 are the pth columns of the U matrix and the V matrix respectively. H +1 to the last column of the matrix;

[0049] 2) The unknown external input is orthogonally decomposed by the matrix obtained by singular value decomposition, and decomposed into two components d 1,k and d 2,k , the expression is as follows:

[0050]

[0051] 3) Bring the two decomposed components into the discrete blood glucose model and define the matrix The output y of the discrete blood glucose model is transformed using non-singular transformation k Decoupling, after decoupling, two output components z are obtained 1,k and z 2,k , the decoupled blood glucose model system is as follows:

[0052]

[0053] In the formula,

[0054] The decoupled system structure separates the blood glucose model parts that are affected by unknown exogenous inputs and those that are not affected by unknown exogenous inputs, and the decoupled matrices will be used for subsequent observer calculations.

[0055] Preferably, the step of the observer performing observation update in S2 is:

[0056] S21, unknown exogenous input estimation, using the current actual blood glucose monitoring value and the estimated state set to estimate the unknown exogenous input set:

[0057]

[0058] Set at initialization x0 is the initial state of the system at time 0, where is the estimated value of a component of the unknown external input at step k, The initial value of is given by Calculated, M1 and M2 are the gain matrices of the observer, is the estimated value of the final updated state of the kth step based on the information of the kth step, and are the estimated values ​​of the two components of the unknown exogenous input at step k-1, is the estimated value of the final updated state of step k based on the information of step k-1, is the estimated value of the unknown exogenous input at step k-1;

[0059] S22, time update, transfer and update the state estimation process quantity according to the system dynamics:

[0060]

[0061] In the formula, is the estimated value of the final updated state of step k-1 based on the information of step k-1, u k-1 is the system input value at time k-1, is the estimated value of the transmission state of the kth step based on the information of the kth step;

[0062] S23, measurement update, estimate the final state set based on the current blood glucose measurement value:

[0063]

[0064] In the formula, L and is the gain matrix of the observer, and the minimum H is calculated according to the blood glucose model structure. ∞ Optimal observer gain in the norm sense;

[0065] S24, loop iterate S21, S22 and S23, and continuously observe the unknown external input.

[0066] Preferably, in S3, the calculation formula for predicting the unknown exogenous input of the current meal to be predicted using weighted moving average is:

[0067]

[0068] Where T is the time period, d i and d k-T+i are the unknown exogenous input sequences of the i-th meal and the k-T+i-th meal observed by the observer in the past data of the current meal type, respectively, and dWMA k+1 The weighted moving average method is used to predict the unknown exogenous input sequence of the k+1th meal. The weighted moving average method assigns higher weights to newer data, which can better reflect the latest changes in the unknown exogenous input that affects the patient's blood sugar.

[0069] Preferably, in said S3, the known treatment information includes: the patient's pre-meal blood sugar data, meal data and medication data;

[0070] The known treatment information and the unknown exogenous input results obtained by the prediction calculation formula are substituted into the blood glucose system for multi-step recursive calculation to obtain accurate blood glucose prediction results.

[0071] Therefore, the method for predicting blood sugar status in a long postprandial window based on unknown exogenous input estimation of the present invention has the following beneficial effects:

[0072] (1) The prediction method proposed in the present invention can take into account the impact of factors that are difficult to quantify or measure, such as the patient's lifestyle and body rhythm, on his or her blood sugar during the blood sugar prediction process, effectively capture the patient's blood sugar dynamics, and thus significantly improve the accuracy of postprandial long-term blood sugar window prediction.

[0073] (2) By summarizing the patient's life patterns through unknown exogenous input data and taking this factor into consideration during the prediction process, we can better capture the patient's individualized blood glucose dynamics and achieve more accurate blood glucose prediction.

[0074] (3) The present invention can predict blood glucose with only a small amount of individualized blood glucose and other data from the patient, overcoming the limitation of traditional methods that require a large amount of training data and is more suitable for clinical treatment scenarios.

[0075] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 It is a schematic diagram of a flow chart of an embodiment of the present invention;

[0077] Figure 2 This is a schematic diagram of observation results of unknown exogenous input of breakfast meal patterns of a patient for multiple consecutive days in an embodiment of the present invention;

[0078] Figure 3 Schematic diagram of the predicted results of blood glucose after breakfast for a patient for multiple consecutive days in an embodiment of the present invention, wherein (a)-(f) are schematic diagrams of the predicted results of blood glucose after breakfast for the patient from the fifth day to the tenth day, respectively. DETAILED DESCRIPTION

[0079] The technical solution of the present invention is further described below through the accompanying drawings and embodiments.

[0080] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0081] Example

[0082] The accuracy of blood glucose prediction is tested by using the method for predicting blood glucose status in a long postprandial window based on prediction of unknown exogenous input proposed by the present invention.

[0083] 1) Data collection and preprocessing: Data were generated using the FDA-approved UVA / Padova T1DM metabolic simulator, including 30-day blood glucose data from 10 virtual patients. Dietary intake was uniformly distributed within a reasonable range, and meal boluses were injected at the same time. In addition, the dietary dose was determined using a standard dose calculation formula.

[0084] Before model training, meal segments are created from the collected data. Blood glucose data measured from 2 hours before to 4 hours after a meal are extracted according to meal type to form a meal segment. Blood glucose data is sampled every 5 minutes, and the timestamp closest to the meal time is used as the anchor point to distinguish the blood glucose trajectories before and after a meal.

[0085] 2) Performance evaluation index: The root mean square error (RMSE) is used to verify the effectiveness of the algorithm. The calculation formula is as follows:

[0086]

[0087] Where n is the total number of blood sugar prediction points, and i Represent the predicted value and true value of blood glucose respectively. The root mean square error of prediction 2 hours and 4 hours after meal was calculated to evaluate the prediction effect of the algorithm in the long post-meal window.

[0088] 3) Model configuration: To verify the prediction performance of the algorithm in a small sample scenario, each patient was predicted from the 5th day. According to the algorithm operation mechanism of this prediction method, the data of each patient was divided into 3 sets according to breakfast, lunch and dinner, reflecting the blood glucose metabolism of breakfast, lunch and dinner respectively. The blood glucose metabolism dynamics model was established according to the meal type, and the fixed-order simultaneous input and state set value observer was built to perform observation updates and blood glucose prediction.

[0089] The subjects' body weight BW and insulin sensitivity S in the model I and plasma glucose concentration basal state G b is a personalized parameter, and the other parameters are determined parameters, using the group average obtained by experiment. BW is read according to the patient's physical examination results, G b According to the patient's HbA1c test data, the following is obtained:

[0090] G b =HbA1c·28.7-46.7;

[0091] If the HbA1c test data of some patients are missing, the average HbA1c value of the patient group will be used for calculation. I It is identified using the least squares method based on the patient's known blood sugar data and can more accurately reflect the patient's blood sugar change pattern.

[0092] After inspection, the discrete blood glucose model of each patient meets the strong detectability condition, that is, the unknown exogenous input and state of each patient meet the strong detectability condition.

[0093] This embodiment designs the observation time window of the patient's life pattern to be 7 days, and summarizes the life pattern within this time window, that is, the prediction of unknown exogenous input. The average root mean square error of blood glucose prediction for all patients and all meal types is used as the evaluation of the algorithm effect. In order to illustrate the effectiveness of introducing unknown exogenous input in the prediction, this embodiment compares the prediction effect of the blood glucose prediction method with the addition of unknown exogenous input and the prediction effect of blood glucose prediction using only the identified physiological model, and uses the proportion of samples with improved prediction effect and the average root mean square error of the algorithm as evaluation indicators.

[0094] 4) Result analysis: The comparison between the prediction effect of the proposed algorithm and the pure model is shown in the following table:

[0095]

[0096] A total of 780 meal data were used to verify the prediction algorithm. As shown in the table above, compared with the pure model prediction method without adding unknown exogenous input sequences, after adopting the proposed prediction method, the prediction effect of 98.85% of the sample data 2 hours after the meal was improved, and the prediction effect of 97.95% of the sample data 4 hours after the meal was improved, which fully demonstrated the scientificity and effectiveness of the proposed method in blood glucose prediction. Using the proposed blood glucose prediction method, the average RMSE of the prediction 2 hours after the meal was 12.745 mg / dL, and the average RMSE of the prediction 4 hours after the meal was 15.654 mg / dL, with high prediction accuracy. This method can achieve high-precision predictions in prediction tasks of different patients and different meal types, and the algorithm has strong stability and universality.

[0097] Figure 2 This is the observation result of unknown exogenous input of a patient's breakfast meal type for many consecutive days. As can be seen from the figure, the unknown exogenous input of a patient with the same meal type shows significant similarity characteristics in extreme value performance and change trend. Most of the unknown exogenous input sequences are tightly concentrated in a small data interval, which shows that the dynamic trend of the unknown exogenous input has high stability and concentration. The above similarities reflect the regularity of the occurrence of blood sugar events in the patient's life. Therefore, adding unknown exogenous input to the prediction can better reflect the impact of the patient's personalized living habits on blood sugar.

[0098] Figure 3 This is the effect diagram of using the proposed method to predict a patient's blood sugar after breakfast for 6 consecutive days starting from the 5th day data. The prediction result has a high degree of match with the actual blood sugar situation and can correctly reflect the blood sugar change trend.

[0099] Therefore, the method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation can take into account the influence of factors such as patient's life pattern and body rhythm that are difficult to quantify or measure on their blood glucose in the process of blood glucose prediction, effectively capture the dynamic law of blood glucose of patients, and thus significantly improve the accuracy of blood glucose prediction in a long postprandial window. The experimental results show that compared with the pure physiological model, the prediction effect of 98.85% of the sample data 2 hours after meal is improved, and the prediction effect of 97.95% of the sample data 4 hours after meal is improved. The average RMSE of the prediction 2 hours after meal is 12.745mg / dL, and the average RMSE of the prediction 4 hours after meal is 15.654mg / dL. The prediction accuracy is high, which fully demonstrates the scientificity and effectiveness of the proposed method in blood glucose prediction.

[0100] By summarizing the patient's life patterns through unknown exogenous input data and taking this factor into consideration during the prediction process, we can better capture the patient's individualized blood sugar dynamics and achieve more accurate blood sugar predictions.

[0101] The present invention can predict blood sugar levels with only a small amount of individualized blood sugar and other data from the patient, overcoming the limitation of traditional methods that require a large amount of training data and being more suitable for clinical treatment scenarios.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation, characterized in that: The following steps are involved: S1. Establish a blood glucose metabolism kinetic model based on the physiological process of glucose-insulin metabolism in diabetic patients; S2. According to the structure of the blood glucose metabolism dynamics model in S1, a fixed-order simultaneous input and state set value observer is constructed to observe and update the unknown exogenous inputs that affect blood glucose within the meal time window; S3. Use an observer to observe the unknown exogenous input of the patient's past data, summarize the patient's life pattern based on the observation results, use the weighted moving average method to predict the unknown exogenous input of the current meal, and combine the known treatment information to make accurate blood sugar predictions.

2. The method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation according to claim 1, characterized in that: In S1, the blood glucose metabolism kinetic model consists of a core sub-model, a dietary absorption sub-model and an insulin transport sub-model, wherein the kinetic equation of the core sub-model is: Where G(t) is the plasma glucose concentration, G b is the basal glucose concentration related to the patient's basal rate of insulin delivery, I p (t) is the mass of plasma insulin, I b is the basal plasma insulin concentration, R a (t) is the rate of plasma glucose appearance, X(t) is the insulin effect, S g Indicates glucose availability, S I is insulin sensitivity, which represents the ability of insulin to control glucose production and use, p2 is the delay between insulin concentration and insulin action, and V g and V I are the glucose and insulin distribution volumes, respectively, and BW is the subject’s body weight; The kinetic equation of the dietary absorption submodel is: R a (t)=k q1 ·f·Q1(t)+k q1 ·f·Q2(t); Where Q1(t) and Q2(t) are the amounts of food in the first and second food absorption compartments, k q1 , k q12 and k q2 is the rate constant, f is the dietary bioavailability, u m (t) is dietary input; The kinetic equation of the insulin transporter model is: In the formula, I sc1 (t) and I sc2 (t) represents the mass of insulin in the two subcutaneous transport compartments, I p (t) is the mass of plasma insulin, k c1 , k c2 and k c12 is the rate constant for subcutaneous insulin transport, k cl is the insulin clearance rate, u i (t) is subcutaneous insulin infusion.

3. The method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation according to claim 2, characterized in that: The specific steps of establishing the blood glucose metabolism kinetics model in S1 are: 1) First, the core sub-model, dietary absorption sub-model and insulin transport sub-model are linearized and unknown exogenous input is added. The continuous linear time-invariant blood glucose state space model after adding unknown exogenous input is obtained as follows: y(t)=C c x(t); In the formula, A c is the state transfer matrix, B ci , B cm and B cd is the input matrix, C c is the output matrix, x(t) is the state vector, u i (t) is subcutaneous insulin infusion, u m (t) is the dietary input, d(t) is the unknown exogenous input signal to be estimated, and y(t) is the blood glucose value output by the model; 2) Then, the continuous linear time-invariant blood glucose state space model after adding unknown exogenous input is discretized to obtain a discrete blood glucose model: x k+1 =Ax k +Bu k +GD k +Wω k ; y k =Cx k +You k +HD k +v k ; Where A, B, G, W, C, D, H are all discretized matrices, k is the discrete time step, k = 1, 2, 3..., T, x k ,d k and k are the sampling sequences of x(t), d(t) and y(t), respectively. k+1 is the state vector at the k+1th sampling time, ω k is the process noise, let ‖ω k ‖≤η ω =10 -4 , v k To measure the noise, let ‖v k ‖≤η v =10 -2 , where u k u is the specific input time and value of the patient's insulin input and carbohydrate input read from the patient's case. k =[u ik ;u mk ],u ik and u mk u i (t) and u m (t) is the sampling sequence.

4. The method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation according to claim 3, characterized in that: In S2, the specific steps of building a fixed-order simultaneous input and state set-valued observer are: 1) Perform singular value decomposition on the matrix H in S1: definition are two unitary matrices obtained after singular value decomposition of H, Σ is a diagonal matrix obtained after singular value decomposition, and the rank of H matrix is ​​p H , U1 and V1 are the first p of the U matrix and V matrix respectively. H The matrix consists of columns, U2 and V2 are the pth columns of the U matrix and the V matrix respectively. H +1 to the last column of the matrix; 2) Use the matrix obtained by singular value decomposition to perform orthogonal decomposition on the unknown external input and decompose it into two components d 1,k and d 2,k , the expression is as follows: 3) Bring the two decomposed components into the discrete blood glucose model and define the matrix The output y of the discrete blood glucose model is transformed using non-singular transformation k Decoupling, after decoupling, two output components z are obtained 1,k and z 2,k , the decoupled blood glucose model system is as follows: x k+1 =Ax k +Bu k +G1d 1,k +G2d 2,k +Wω k 4 y k =Cx k +From k +H1d 1,k +v k ; from 1,k =C1x k +D1u k +∑d 1,k +υ 1,k ; z 2,k =V2x k +D2u k +υ 2,k ; In the formula, 5. The method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation according to claim 4, characterized in that: The steps for the observer in S2 to perform observation update are: S21, unknown exogenous input estimation: using the current actual blood glucose monitoring value and the estimated state set to estimate the unknown exogenous input set; S22, time update: transfer and update the state estimation process quantity according to the system dynamics; S23, measurement update: estimating the final state set according to the current blood glucose measurement value; S24, loop iterate S21, S22 and S23, and continuously observe unknown external input.

6. The method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation according to claim 1, characterized in that: In S3, the calculation formula for predicting the unknown exogenous input of the current meal using weighted moving average is: Where T is the time period, d i and d k-T+i are the unknown exogenous input sequences of the i-th meal and the k-T+i-th meal observed by the observer in the past data of the current meal type, respectively, and dWMA k+1 is the unknown exogenous input sequence of the k+1th meal predicted by the weighted moving average method.

7. The method for predicting blood glucose status in a long postprandial window based on unknown exogenous input estimation according to claim 6, characterized in that: In S3, the known treatment information includes: the patient's pre-meal blood sugar data, meal data, and medication data; The known treatment information and the unknown exogenous input results obtained by the prediction calculation formula are substituted into the blood glucose model system for multi-step recursive calculation to obtain accurate blood glucose prediction results.

Images