Blood glucose trend prediction system and method based on digital twinning

Abstract

The invention provides a blood glucose trend prediction system and method based on digital twinning. The blood glucose trend prediction system based on digital twinning comprises a blood glucose prediction platform, detection equipment worn by a patient, a data transmission module and a patient mobile terminal, the detection equipment worn by the patient is used for collecting physiological signal parameters of the patient in real time; the data transmission module is used for sending the physiological signal parameters to a blood glucose prediction platform and feeding back a blood glucose trend result obtained by the blood glucose prediction platform to the patient mobile terminal; the blood glucose prediction platform is used for calling the trained and tested digital twin model after receiving the corresponding physiological signal parameters of the patient, and performing blood glucose trend prediction according to the corresponding physiological signal parameters of the patient to obtain a blood glucose trend result; the patient mobile terminal is used for receiving the blood glucose trend result in real time and visually displaying the blood glucose trend result. The method comprises steps corresponding to the system module.

Description

Technical Field

[0001] The present invention proposes a blood sugar trend prediction system and method based on digital twins, which belongs to the technical field of blood sugar prediction. Background Art

[0002] Diabetes is a global chronic disease that affects hundreds of millions of people. Its management and treatment are of great significance for improving the quality of life of patients and reducing the risk of complications. Traditional diabetes management mainly relies on patients to self-monitor their blood sugar levels and adjust their diet, exercise and medication according to the guidance of doctors. However, this approach has many limitations, such as the frequency and inconvenience of blood sugar monitoring, differences in patients' self-management capabilities, and the static and non-personalized nature of treatment plans. In recent years, with the rapid development of the Internet of Things, big data and artificial intelligence technologies, the field of diabetes management has ushered in new changes. In particular, the application of digital twin technology provides a more accurate and personalized solution for blood sugar trend prediction. Digital twin is a virtual model based on a physical entity or system that can reflect the state and behavior of the entity in real time and predict future changes through data analysis, simulation and optimization.

[0003] However, the existing digital twin-based blood sugar trend prediction systems and methods still have some shortcomings. On the one hand, most systems rely only on blood sugar data for prediction, ignoring the impact of other physiological signal parameters (such as heart rate, blood pressure, insulin delivery data, etc.) on blood sugar levels, resulting in incomplete and inaccurate prediction results. On the other hand, the imperfection of data transmission and processing mechanisms also limits the real-time and reliability of the system. In addition, patients lack an intuitive and convenient way to receive and understand the prediction results.

[0004] Therefore, there is an urgent need for a blood glucose trend prediction system and method based on digital twins that can comprehensively consider multiple physiological signal parameters, realize real-time data transmission and processing, and provide intuitive visual display. Such a system can not only improve the accuracy and personalization of blood glucose prediction, but also enhance the patient's self-management ability and participation, and provide more comprehensive and effective support for diabetes management. The present invention is proposed based on such background technology, and aims to achieve comprehensive, accurate and personalized management of patients' blood glucose conditions by integrating technical means such as multiple physiological signal parameter acquisition, real-time data transmission, blood glucose trend prediction based on digital twins, and visual display of results, bringing new breakthroughs to the field of diabetes treatment and management. Summary of the invention

[0005] The present invention provides a blood sugar trend prediction system and method based on digital twins to solve the technical problems in the prior art. The technical solutions adopted are as follows: A blood sugar trend prediction system based on digital twins, the blood sugar trend prediction system based on digital twins comprising a blood sugar prediction platform, a detection device worn by a patient, a data transmission module and a patient mobile terminal; The detection device worn by the patient is used to collect the patient's physiological signal parameters in real time; The data transmission module is used to send the physiological signal parameters to the blood glucose prediction platform, and feed back the blood glucose trend results obtained by the blood glucose prediction platform to the patient's mobile terminal; The blood sugar prediction platform is used to retrieve the trained and tested digital twin model after receiving the physiological signal parameters corresponding to the patient, perform blood sugar trend prediction according to the physiological signal parameters corresponding to the patient, and obtain the blood sugar trend results; The patient mobile terminal is used to receive blood sugar trend results in real time and visualize the blood sugar trend results.

[0006] Furthermore, the detection equipment worn by the patient includes a heart rate acquisition sensor, a blood pressure acquisition sensor, a blood sugar acquisition sensor and an insulin pump; The heart rate acquisition sensor is used to collect the patient's heart rate data and send the patient's heart rate data to the data transmission module; The blood pressure collection sensor is used to collect the patient's blood pressure data and send the patient's blood pressure data to the data transmission module; The blood sugar collection sensor is used to collect the patient's blood sugar data and send the patient's blood sugar data to the data transmission module; The insulin pump is used to collect the insulin delivery time and delivery measurement data of the patient's insulin injection, and send the patient's insulin delivery time and delivery measurement data to the data transmission module.

[0007] Further, the data transmission module includes a physiological signal parameter sending module and a blood sugar trend result sending module; The physiological signal parameter sending module is used to send the physiological signal parameters to the blood glucose prediction platform; The blood sugar trend result sending module is used to feed back the blood sugar trend result obtained by the blood sugar prediction platform to the patient's mobile terminal.

[0008] Furthermore, the data transmission module also includes a physiological signal parameter transmission adjustment module; The physiological signal parameter sending adjustment module is used to adjust the data sending frequency of the patient's blood sugar data to the blood sugar prediction platform according to the patient's heart rate data and blood pressure data.

[0009] Furthermore, the operation process of the physiological signal parameter sending and adjusting module includes: Receive user's heart rate data in real time; When the user's heart rate data exceeds a preset heart rate data threshold, the patient's blood sugar data and blood pressure data within a preset monitoring time period are retrieved; wherein the preset monitoring time period ranges from 1h to 6h; The first data transmission frequency adjustment coefficient is obtained according to the blood sugar data, blood pressure data and heart rate data of the patient within the preset monitoring time period; wherein the first data transmission frequency adjustment coefficient is obtained by the following formula: Among them, S 01 represents the first data transmission frequency adjustment coefficient; x, y and z represent the first adjustment parameter, the second adjustment parameter and the third adjustment parameter respectively, and the first adjustment parameter is obtained by the following formula: Wherein, x represents the first adjustment parameter; α represents the first adjustment constant, and the value range of the first adjustment constant is 0.12-0.54; n represents the number of times blood sugar data is collected; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP BP Indicates the overall systolic blood pressure value change rate corresponding to the m-time blood pressure data collection; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; BS i Indicates the blood sugar value corresponding to the i-th blood sugar collection; BS 0i BS represents the blood sugar reference value corresponding to the i-th blood sugar collection; z Indicates the median fasting blood glucose value corresponding to the collection of n blood glucose data; The second adjustment parameter is obtained by the following formula: Wherein, y represents the second adjustment parameter; β represents the second adjustment constant, and the value range of the second adjustment constant is 0.18-0.39; m represents the total number of blood pressure data collection; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP Indicates the overall systolic blood pressure value change rate corresponding to m blood pressure data collections; BP i represents the systolic blood pressure value corresponding to the i-th blood pressure collection; BP0 represents the systolic blood pressure reference value corresponding to the patient; BS z It represents the median value of systolic blood pressure corresponding to the collection of m blood pressure data; The third adjustment parameter is obtained by the following formula: Wherein, z represents the third adjustment parameter; γ represents the third adjustment constant, and the value range of the third adjustment constant is 0.15-0.32; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP Indicates the overall systolic blood pressure value change rate corresponding to m blood pressure data collections; HR i represents the heart rate value corresponding to the i-th heart rate collection; r represents the number of heart rate data collection; HR i-1 represents the heart rate value corresponding to the i-1th heart rate acquisition; HR0 represents the heart rate reference value corresponding to the patient; HR z Indicates the median heart rate value corresponding to the collection of r heart rate data; Using the first data transmission frequency adjustment coefficient, the frequency of sending blood glucose data to the blood glucose prediction platform after the end of the preset monitoring time period is adjusted; The frequency of sending the adjusted blood sugar data to the blood sugar prediction platform is obtained by the following formula: Among them, F t 01 represents the data transmission frequency of the adjusted blood sugar data obtained by the first data transmission frequency adjustment coefficient and sent to the blood sugar prediction platform; F0 represents the data transmission frequency of the blood sugar data before adjustment and sent to the blood sugar prediction platform; S 01 Indicates the first data sending frequency adjustment coefficient.

[0010] Furthermore, the operation process of the physiological signal parameter sending and adjusting module further includes: Receive users' blood pressure data in real time; When the blood pressure data of the user exceeds the preset blood pressure data threshold, the blood sugar data and heart rate data of the patient within the preset monitoring time period are retrieved; wherein the preset monitoring time period ranges from 1h to 6h; The second data transmission frequency adjustment coefficient is obtained according to the blood glucose data and heart rate data of the patient within the preset monitoring time period; wherein the second data transmission frequency adjustment coefficient is obtained by the following formula: Among them, S 02 represents the second data transmission frequency adjustment coefficient; A and B represent the fourth adjustment parameter and the fifth adjustment parameter respectively, and the fourth adjustment parameter is obtained by the following formula: Wherein, A represents the fourth adjustment parameter; α represents the first adjustment constant, and the value range of the first adjustment constant is 0.12-0.54; n represents the number of times blood sugar data is collected; BS i Indicates the blood sugar value corresponding to the i-th blood sugar collection; BS 0i BS represents the blood sugar reference value corresponding to the i-th blood sugar collection; z Indicates the median fasting blood glucose value corresponding to the collection of n blood glucose data; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP represents the overall systolic blood pressure value change rate corresponding to the m-time blood pressure data collection; BP represents the current systolic blood pressure value; BP0 represents the systolic blood pressure reference value corresponding to the patient; The fifth adjustment parameter is obtained by the following formula: Wherein, B represents the fifth adjustment parameter; BP represents the current systolic blood pressure value; BP0 represents the systolic blood pressure reference value corresponding to the patient; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP represents the overall systolic blood pressure value change rate corresponding to the m-time blood pressure data collection; BP represents the current systolic blood pressure value; HR represents the current heart rate value; HR0 represents the patient's corresponding heart rate reference value; HR z represents the middle heart rate value corresponding to the collection of r heart rate data; γ represents the third adjustment constant, and the value range of the third adjustment constant is 0.15-0.32; r represents the number of heart rate data collection; W represents the sign parameter, when HR-HR0 is not less than 1, W=1; when HR-HR0 is less than 1, W=-1; Using the second data transmission frequency adjustment coefficient, the frequency of sending blood glucose data to the blood glucose prediction platform after the end of the preset monitoring time period is adjusted; The frequency of sending the adjusted blood sugar data to the blood sugar prediction platform is obtained by the following formula: Among them, F t 02 represents the data transmission frequency of the blood glucose data after adjustment obtained by the second data transmission frequency adjustment coefficient and sent to the blood glucose prediction platform; F0 represents the data transmission frequency of the blood glucose data before adjustment and sent to the blood glucose prediction platform; S 02 Indicates the second data sending frequency adjustment coefficient.

[0011] Furthermore, the structure of the digital twin model that has completed training and testing is as follows: The input layer is used to process and integrate the input data from the heart rate sensor, blood pressure sensor, blood sugar sensor and insulin pump; The feature extraction layer is used to use the sliding time window to obtain the static feature data corresponding to the heart rate data, blood pressure data, blood sugar data, insulin delivery time and dosage data from the data provided by the heart rate acquisition sensor, blood pressure acquisition sensor, blood sugar acquisition sensor and insulin pump within the sliding time window; Multiple LSTM sub-networks are used to process heart rate data, blood pressure data, blood sugar data, insulin delivery time and dosage data, and obtain processed data information; The fusion layer is used to output and merge the processed data information obtained by multiple LSTM sub-networks; A regression layer, used to obtain blood sugar trend prediction results using one or more fully connected layers; The output layer is used to output the blood sugar trend prediction result.

[0012] Furthermore, the time length setting of the sliding time window includes: The data sending frequency of extracting the blood glucose data before adjustment corresponding to the blood glucose data and sending it to the blood glucose prediction platform; Extract the blood sugar data corresponding to the current adjusted data sending frequency sent to the blood sugar prediction platform; Acquire the time length of the adjusted sliding time window according to the initial data sending frequency and the currently corresponding adjusted data sending frequency; The time length of the adjusted sliding time window is obtained by the following formula: Wherein, T represents the time length of the sliding time window after adjustment; T0 represents the initial time length of the sliding time window; F0 represents the frequency of sending blood glucose data to the blood glucose prediction platform before adjustment; F t 01 represents the data transmission frequency of the adjusted blood sugar data obtained by the first data transmission frequency adjustment coefficient and sent to the blood sugar prediction platform; F t 02 represents the data transmission frequency of the adjusted blood sugar data obtained by the second data transmission frequency adjustment coefficient and sent to the blood sugar prediction platform; F max and F min They respectively represent the maximum and minimum values ​​of the data transmission frequency that occur during the data transmission frequency adjustment process.

[0013] Further, the multiple LSTM subnetworks include a heart rate data processing LSTM subnetwork, a blood pressure data processing LSTM subnetwork, a blood sugar data processing LSTM subnetwork, and an insulin delivery time and dosage data processing LSTM subnetwork; The heart rate data processing LSTM sub-network is used to process the heart rate data; The blood pressure data processing LSTM sub-network is used to process the blood pressure data; The blood sugar data processing LSTM sub-network is used to process the blood sugar data; The insulin administration time and dosage data processing LSTM sub-network is used to process the insulin administration time and dosage data; The data processing includes data cleaning, digital serialization and data dynamic feature extraction.

[0014] A blood sugar trend prediction method based on digital twins, the blood sugar trend prediction method based on digital twins comprising: The patient's physiological signal parameters are collected in real time through the detection equipment worn by the patient; The physiological signal parameters are sent to the blood glucose prediction platform using the data transmission module After receiving the physiological signal parameters corresponding to the patient, the blood sugar prediction platform calls the digital twin model that has been trained and tested, predicts the blood sugar trend based on the physiological signal parameters corresponding to the patient, and obtains the blood sugar trend results; Using the data transmission module to send the blood sugar trend result to the patient's mobile terminal; The patient's mobile terminal receives blood sugar trend results in real time and displays the blood sugar trend results visually.

[0015] Beneficial effects of the present invention: The blood sugar trend prediction system and method based on digital twin proposed by the present invention adopts a digital twin model. The system can comprehensively consider the patient's various physiological signal parameters and historical data, so as to more accurately predict the blood sugar trend. This helps patients take timely measures to avoid the health risks caused by high or low blood sugar. The detection equipment worn by the patient can collect data in real time and quickly send it to the blood sugar prediction platform through the data transmission module. At the same time, the blood sugar prediction platform can also quickly process the data and generate blood sugar trend results, and then display them to the patient in real time through the patient's mobile terminal. This real-time performance helps patients understand their blood sugar status at any time and make corresponding adjustments. The digital twin model can be customized and optimized according to the specific situation of each patient, so as to realize personalized blood sugar trend prediction. This helps to meet the needs of different patients and improve the prediction effect. The visual display function of the patient's mobile terminal enables patients to easily view their blood sugar trend results without having to frequently go to the hospital or clinic for testing. This not only improves the convenience of patients, but also reduces medical costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a system block diagram of the system of the present invention; Figure 2 The present invention is a flowchart of the method. DETAILED DESCRIPTION

[0017] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0018] The embodiment of the present invention proposes a blood sugar trend prediction system based on digital twins, such as Figure 1 As shown, the blood sugar trend prediction system based on digital twins includes a blood sugar prediction platform, a detection device worn by the patient, a data transmission module and a patient mobile terminal; The detection device worn by the patient is used to collect the patient's physiological signal parameters in real time; The data transmission module is used to send the physiological signal parameters to the blood glucose prediction platform, and feed back the blood glucose trend results obtained by the blood glucose prediction platform to the patient's mobile terminal; The blood sugar prediction platform is used to retrieve the trained and tested digital twin model after receiving the physiological signal parameters corresponding to the patient, perform blood sugar trend prediction according to the physiological signal parameters corresponding to the patient, and obtain the blood sugar trend results; The patient mobile terminal is used to receive blood sugar trend results in real time and visualize the blood sugar trend results.

[0019] The working principle of the above technical solution is: the detection equipment worn by the patient (such as the sensor of the continuous blood glucose monitoring system CGM) collects the patient's physiological signal parameters, such as blood glucose level, heart rate, etc. in real time. The collected physiological signal parameters are sent to the blood glucose prediction platform through the data transmission module. At the same time, this module is also responsible for feeding back the blood glucose trend results generated by the blood glucose prediction platform to the patient's mobile terminal. After receiving the patient's physiological signal parameters, the blood glucose prediction platform will call up the digital twin model that has been trained and tested. The model uses the patient's real-time data and historical data to predict blood glucose trends and generate blood glucose trend results. The patient's mobile terminal receives the blood glucose trend results sent by the blood glucose prediction platform in real time, and displays them to the patient through a visual interface so that the patient can understand his or her blood glucose status at any time.

[0020] The effect of the above technical solution is: by adopting the digital twin model, the system can comprehensively consider the patient's various physiological signal parameters and historical data, so as to more accurately predict the blood sugar trend. This helps patients take timely measures to avoid the health risks caused by high or low blood sugar. The detection equipment worn by the patient can collect data in real time and quickly send it to the blood sugar prediction platform through the data transmission module. At the same time, the blood sugar prediction platform can also quickly process the data and generate blood sugar trend results, and then display them to the patient in real time through the patient's mobile terminal. This real-time nature helps patients understand their blood sugar status at any time and make corresponding adjustments. The digital twin model can be customized and optimized according to the specific situation of each patient, so as to realize personalized blood sugar trend prediction. This helps to meet the needs of different patients and improve the prediction effect. The visual display function of the patient's mobile terminal enables patients to easily view their blood sugar trend results without having to frequently go to the hospital or clinic for testing. This not only improves the convenience of patients, but also reduces medical costs.

[0021] In summary, the blood sugar trend prediction system based on digital twins has technical effects such as improving prediction accuracy, real-time performance, personalization and convenience, and can provide better health management services for diabetic patients.

[0022] In one embodiment of the present invention, the detection device worn by the patient includes a heart rate acquisition sensor, a blood pressure acquisition sensor, a blood sugar acquisition sensor and an insulin pump; The heart rate acquisition sensor is used to collect the patient's heart rate data and send the patient's heart rate data to the data transmission module; The blood pressure collection sensor is used to collect the patient's blood pressure data and send the patient's blood pressure data to the data transmission module; The blood sugar collection sensor is used to collect the patient's blood sugar data and send the patient's blood sugar data to the data transmission module; The insulin pump is used to collect the insulin delivery time and delivery measurement data of the patient's insulin injection, and send the patient's insulin delivery time and delivery measurement data to the data transmission module.

[0023] Wherein, the data transmission module includes a physiological signal parameter sending module and a blood sugar trend result sending module; The physiological signal parameter sending module is used to send the physiological signal parameters to the blood glucose prediction platform; The blood sugar trend result sending module is used to feed back the blood sugar trend result obtained by the blood sugar prediction platform to the patient's mobile terminal.

[0024] The working principle of the above technical solution is: through skin contact or close-fitting wearing, the patient's heart rate data is collected in real time. These data reflect the frequency and intensity of the patient's heart activity and are one of the important indicators for evaluating health status. Similarly, through skin contact or a specific wearing method, the patient's blood pressure, including systolic and diastolic pressure, is measured continuously or intermittently to evaluate the pressure state of the patient's vascular system. Minimally invasive or non-invasive technology (such as a continuous glucose monitoring system) is used to monitor the patient's blood sugar level in real time, providing a direct basis for blood sugar trend prediction. It is connected to the patient's insulin injection system to record and transmit the insulin delivery time and delivery metering data, which are crucial for predicting blood sugar trends and adjusting treatment strategies.

[0025] Summarize the physiological signal parameters received from the heart rate sensor, blood pressure sensor, blood sugar sensor and insulin pump, and send these data to the blood sugar prediction platform safely and efficiently through wireless communication technology (such as Bluetooth, Wi-Fi). Receive the blood sugar trend results calculated by the blood sugar prediction platform based on the digital twin model, and then send these results to the patient's mobile terminal through the same communication channel. After receiving the physiological signal parameters, use the pre-trained digital twin model (trained based on the patient's historical data, physiological characteristics, living habits and other information) to analyze and predict, and generate blood sugar trend results, including blood sugar change trends in the future, possible blood sugar peaks and valleys, etc. Receive and display blood sugar trend results, which may include charts, graphical interfaces, etc., so that patients can intuitively understand their blood sugar status and adjust their lifestyles such as diet, exercise or insulin use according to the predicted results.

[0026] The effect of the above technical solution is: a comprehensive analysis combining multiple physiological signal parameters (heart rate, blood pressure, blood sugar, insulin delivery data) is more accurate than a single parameter prediction and can better reflect the patient's overall health status. The real-time collected data can be fed back to the blood sugar prediction platform immediately, and the platform responds quickly to generate personalized blood sugar trend forecasts, providing patients with immediate and targeted health management advice. By predicting blood sugar trends, patients can take measures in advance to prevent the occurrence of hypoglycemia or hyperglycemia, thereby reducing the risk of diabetes-related complications. Visualizing blood sugar trend results allows patients to understand their health status more intuitively, increase their enthusiasm for self-management, and improve their quality of life. Through accurate predictions, unnecessary hospital visits and examinations can be reduced, medical resources can be reasonably allocated, and the burden on the medical system can be reduced.

[0027] In summary, this technical solution achieves comprehensive, accurate and personalized management of patients' blood sugar status by integrating multiple physiological signal parameter acquisition, real-time data transmission, blood sugar trend prediction based on digital twins, and visual display of results, effectively improving the efficiency and effectiveness of diabetes management.

[0028] In one embodiment of the present invention, the data transmission module further comprises a physiological signal parameter transmission adjustment module; The physiological signal parameter sending adjustment module is used to adjust the data sending frequency of the patient's blood sugar data to the blood sugar prediction platform according to the patient's heart rate data and blood pressure data.

[0029] The operation process of the physiological signal parameter sending and adjusting module includes: Receive user's heart rate data in real time; When the user's heart rate data exceeds a preset heart rate data threshold, the patient's blood sugar data and blood pressure data within a preset monitoring time period are retrieved; wherein the preset monitoring time period ranges from 1h to 6h; The first data transmission frequency adjustment coefficient is obtained according to the blood sugar data, blood pressure data and heart rate data of the patient within the preset monitoring time period; wherein the first data transmission frequency adjustment coefficient is obtained by the following formula: Among them, S 01 represents the first data transmission frequency adjustment coefficient; x, y and z represent the first adjustment parameter, the second adjustment parameter and the third adjustment parameter respectively, and the first adjustment parameter is obtained by the following formula: Wherein, x represents the first adjustment parameter; α represents the first adjustment constant, and the value range of the first adjustment constant is 0.12-0.54; n represents the number of times blood sugar data is collected; ΔPBS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP BP Indicates the overall systolic blood pressure value change rate corresponding to the m-time blood pressure data collection; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; BS i Indicates the blood sugar value corresponding to the i-th blood sugar collection; BS 0i BS represents the blood sugar reference value corresponding to the i-th blood sugar collection; z Indicates the median fasting blood glucose value corresponding to the collection of n blood glucose data; The second adjustment parameter is obtained by the following formula: Wherein, y represents the second adjustment parameter; β represents the second adjustment constant, and the value range of the second adjustment constant is 0.18-0.39; m represents the total number of blood pressure data collection; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP Indicates the overall systolic blood pressure value change rate corresponding to m blood pressure data collections; BP i represents the systolic blood pressure value corresponding to the i-th blood pressure collection; BP0 represents the systolic blood pressure reference value corresponding to the patient; BS z It represents the median value of systolic blood pressure corresponding to the collection of m blood pressure data; The third adjustment parameter is obtained by the following formula: Wherein, z represents the third adjustment parameter; γ represents the third adjustment constant, and the value range of the third adjustment constant is 0.15-0.32; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP Indicates the overall systolic blood pressure value change rate corresponding to m blood pressure data collections; HR i represents the heart rate value corresponding to the i-th heart rate collection; r represents the number of heart rate data collection; HR i-1 represents the heart rate value corresponding to the i-1th heart rate acquisition; HR0 represents the heart rate reference value corresponding to the patient; HR z Indicates the median heart rate value corresponding to the collection of r heart rate data; Using the first data transmission frequency adjustment coefficient, the frequency of sending blood glucose data to the blood glucose prediction platform after the end of the preset monitoring time period is adjusted; The frequency of sending the adjusted blood sugar data to the blood sugar prediction platform is obtained by the following formula: Among them, F t 01 represents the data transmission frequency of the adjusted blood sugar data obtained by the first data transmission frequency adjustment coefficient and sent to the blood sugar prediction platform; F0 represents the data transmission frequency of the blood sugar data before adjustment and sent to the blood sugar prediction platform; S 01 Indicates the first data sending frequency adjustment coefficient.

[0030] The working principle of the above technical solution is as follows: the physiological signal parameter transmission adjustment module first receives the patient's heart rate data in real time. When the received heart rate data exceeds the preset heart rate data threshold, the module believes that the patient may be in a certain physiological abnormal state and needs to monitor blood glucose data more intensively. The module will retrieve the blood glucose data, blood pressure data and heart rate data within the preset monitoring time period (1h-6h). By analyzing these data, the first data transmission frequency adjustment coefficient (S01) is calculated.

[0031] The first adjustment parameter (x): is calculated based on the number of blood glucose data collection times (n), the overall blood glucose value change rate (ΔPBS), the overall systolic blood pressure value change rate (ΔPBP), the overall heart rate value change rate (ΔPHR), the blood glucose value of each blood glucose collection (BSi), the blood glucose reference value (BS0i) and the fasting blood glucose median value (BSz).

[0032] The second adjustment parameter (y): calculated based on the overall blood glucose value change rate (ΔPBS), the overall heart rate value change rate (ΔPHR), the overall systolic blood pressure value change rate (ΔPBP), the systolic blood pressure value (BPi) of each blood pressure collection, the systolic blood pressure reference value (BP0) and the systolic blood pressure median value (BSz, which should be BPz here to avoid confusion, but to maintain consistency in the original text, it will not be changed for the time being).

[0033] The third adjustment parameter (z): calculated based on the overall blood glucose value change rate (ΔPBS), the overall heart rate value change rate (ΔPHR), the overall systolic blood pressure value change rate (ΔPBP), the heart rate value of each heart rate collection (HRi), the number of heart rate data collections (r), the heart rate reference value (HR0) and the heart rate median value (HRz).

[0034] Data transmission frequency adjustment: using the calculated first data transmission frequency adjustment coefficient (S01), the data transmission frequency (Ft01) of sending blood glucose data to the blood glucose prediction platform after the end of the preset monitoring time period is adjusted according to a preset formula.

[0035] The effect of the above technical solution is: by monitoring the patient's heart rate data in real time and dynamically adjusting the frequency of sending blood glucose data when the threshold is exceeded, the system can respond to the patient's physiological changes more flexibly, ensuring more intensive data support at critical moments. Combining multiple physiological signal parameters (blood glucose, blood pressure, heart rate) for calculation makes the adjustment of data sending frequency more personalized and accurate, which helps to provide patients with more personalized health management services. By dynamically adjusting the frequency of data transmission, while ensuring data quality, unnecessary data transmission can be reduced, thereby optimizing the utilization of data transmission resources. More intensive blood glucose data monitoring helps the blood glucose prediction platform obtain richer data samples, thereby improving the accuracy and reliability of blood glucose trend predictions. For patients who may be in an abnormal physiological state, more frequent blood glucose monitoring can help to promptly detect and deal with potential blood glucose problems, thereby enhancing patient safety.

[0036] On the other hand, by receiving the user's heart rate data in real time, the technical solution can immediately respond to changes in the user's health status. When the heart rate data exceeds the preset heart rate data threshold, the system can quickly retrieve the blood sugar and blood pressure data within the preset monitoring time period to provide comprehensive health monitoring for the user. According to the user's blood sugar, blood pressure and heart rate data, the solution dynamically adjusts the data transmission frequency of the blood sugar data sent to the blood sugar prediction platform by calculating the first data transmission frequency adjustment coefficient (S01). This dynamic adjustment mechanism helps to increase the frequency of data transmission when necessary, thereby more accurately reflecting the changes in the user's health status and improving the timeliness of monitoring. The technical solution takes into account multiple physiological parameters such as blood sugar, blood pressure and heart rate, and comprehensively evaluates the user's health status through complex calculation formulas (involving the first adjustment parameter x, the second adjustment parameter y and the third adjustment parameter z). This comprehensive evaluation method is more comprehensive than single indicator monitoring and can more accurately reflect the user's health status. By adjusting the data transmission frequency, the solution can provide personalized health management services based on the user's actual health status. For example, when the user is in poor health or has an increased risk of disease, the system can automatically increase the data transmission frequency so that doctors or health managers can obtain the user's health data in a timely manner and take corresponding intervention measures. Through automated and intelligent data monitoring and transmission, this technical solution can reduce the burden on medical staff and improve the efficiency and quality of medical services. At the same time, it also helps to reduce the waste of medical resources and improve the efficiency of medical resource utilization. Through continuous monitoring and dynamic adjustment of data transmission frequency, this solution helps to timely identify potential health problems and take corresponding preventive measures. This helps to reduce the incidence of diseases and improve the overall health level of users.

[0037] In summary, this technical solution provides users with a comprehensive, accurate and efficient health management service through the advantages of real-time health monitoring, dynamic adjustment of data transmission frequency, comprehensive evaluation of health indicators, personalized health management and improved medical efficiency. At the same time, this technical solution realizes the dynamic adjustment of the patient's blood sugar data transmission frequency by introducing the physiological signal parameter transmission adjustment module, which not only improves the accuracy and personalization of blood sugar trend prediction, but also optimizes the utilization of data transmission resources, and enhances the safety of patients and the health management effect.

[0038] In one embodiment of the present invention, the operation process of the physiological signal parameter sending and adjusting module further includes: Receive users' blood pressure data in real time; When the blood pressure data of the user exceeds the preset blood pressure data threshold, the blood sugar data and heart rate data of the patient within the preset monitoring time period are retrieved; wherein the preset monitoring time period ranges from 1h to 6h; The second data transmission frequency adjustment coefficient is obtained according to the blood glucose data and heart rate data of the patient within the preset monitoring time period; wherein the second data transmission frequency adjustment coefficient is obtained by the following formula: Among them, S 02 represents the second data transmission frequency adjustment coefficient; A and B represent the fourth adjustment parameter and the fifth adjustment parameter respectively, and the fourth adjustment parameter is obtained by the following formula: Wherein, A represents the fourth adjustment parameter; α represents the first adjustment constant, and the value range of the first adjustment constant is 0.12-0.54; n represents the number of times blood sugar data is collected; BS i Indicates the blood sugar value corresponding to the i-th blood sugar collection; BS 0i BS represents the blood sugar reference value corresponding to the i-th blood sugar collection; z Indicates the median fasting blood glucose value corresponding to the collection of n blood glucose data; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP represents the overall systolic blood pressure value change rate corresponding to the m-time blood pressure data collection; BP represents the current systolic blood pressure value; BP0 represents the systolic blood pressure reference value corresponding to the patient; The fifth adjustment parameter is obtained by the following formula: Wherein, B represents the fifth adjustment parameter; BP represents the current systolic blood pressure value; BP0 represents the systolic blood pressure reference value corresponding to the patient; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP represents the overall systolic blood pressure value change rate corresponding to the m-time blood pressure data collection; BP represents the current systolic blood pressure value; HR represents the current heart rate value; HR0 represents the patient's corresponding heart rate reference value; HR z represents the middle heart rate value corresponding to the collection of r heart rate data; γ represents the third adjustment constant, and the value range of the third adjustment constant is 0.15-0.32; r represents the number of heart rate data collection; W represents the sign parameter, when HR-HR0 is not less than 1, W=1; when HR-HR0 is less than 1, W=-1; Using the second data transmission frequency adjustment coefficient, the frequency of sending blood glucose data to the blood glucose prediction platform after the end of the preset monitoring time period is adjusted; The frequency of sending the adjusted blood sugar data to the blood sugar prediction platform is obtained by the following formula: Among them, F t 02 represents the data transmission frequency of the blood glucose data after adjustment obtained by the second data transmission frequency adjustment coefficient and sent to the blood glucose prediction platform; F0 represents the data transmission frequency of the blood glucose data before adjustment and sent to the blood glucose prediction platform; S 02 Indicates the second data sending frequency adjustment coefficient.

[0039] The working principle of the above technical solution is: the module receives the user's heart rate and blood pressure data in real time.

[0040] When the received heart rate data exceeds the preset heart rate data threshold, the module will retrieve the blood glucose data and blood pressure data within the preset monitoring time period (1h-6h), calculate the first data transmission frequency adjustment coefficient (S01), and then adjust the data transmission frequency according to the coefficient (this part is the same as the previous description). When the received blood pressure data exceeds the preset blood pressure data threshold, the module will also retrieve the blood glucose data and heart rate data within the preset monitoring time period, but this time it is to calculate the second data transmission frequency adjustment coefficient (S02).

[0041] The fourth adjustment parameter (A): is calculated based on the number of blood glucose data collection times (n), the blood glucose value of each blood glucose collection (BSi), the blood glucose reference value (BS0i), the fasting blood glucose median value (BSz), the overall blood glucose value change rate (ΔPBS), the overall heart rate value change rate (ΔPHR), the overall systolic blood pressure value change rate (ΔPBP), the current systolic blood pressure value (BP) and the systolic blood pressure reference value (BP0).

[0042] The fifth adjustment parameter (B): calculated based on the current systolic blood pressure value (BP), the systolic blood pressure reference value (BP0), the overall blood sugar value change rate (ΔPBS), the overall heart rate value change rate (ΔPHR), the overall systolic blood pressure value change rate (ΔPBP), the current heart rate value (HR), the heart rate reference value (HR0), the heart rate median value (HRz), the third adjustment constant (γ) and the sign parameter (W). The sign parameter (W) is determined based on the difference between the current heart rate and the heart rate reference value, and is used to reflect the direction of change of the heart rate.

[0043] The calculated second data transmission frequency adjustment coefficient (S02) is used to adjust the data transmission frequency (Ft02) of the blood glucose data after the end of the preset monitoring time period to the blood glucose prediction platform according to a preset formula.

[0044] The effect of the above technical solution is: by monitoring heart rate and blood pressure data at the same time, and adjusting the frequency of sending blood sugar data when any data exceeds the threshold, the system can provide patients with more comprehensive physiological status monitoring and ensure more intensive data support at critical moments. Combining multiple physiological signal parameters (blood sugar, blood pressure, heart rate) for calculation, the adjustment of data sending frequency is more personalized and accurate, which helps to provide patients with more personalized health management services and improve the accuracy of blood sugar trend prediction. Dynamic adjustment of data sending frequency can not only reduce unnecessary data transmission while ensuring data quality, but also improve the efficiency of data transmission and resource utilization. More intensive blood sugar data monitoring helps to timely discover and deal with potential blood sugar problems. Combined with blood pressure data monitoring, it can further enhance patient safety and the reliability of health management. Through intelligent adjustment of data sending frequency, the system can reduce unnecessary interruptions while ensuring sufficient data support when needed, thereby improving the user experience.

[0045] At the same time, the technical solution not only monitors heart rate and blood sugar, but also adds real-time monitoring of blood pressure data. This provides users with a more comprehensive physiological parameter monitoring system that can more accurately reflect the overall health status of users. When the user's blood pressure data exceeds the preset blood pressure data threshold, the system can respond quickly and retrieve the blood sugar and heart rate data within the preset monitoring time period. This dynamic response mechanism helps to timely discover changes in the user's health status and take corresponding early warning measures. According to the user's blood sugar, heart rate and blood pressure data, the solution dynamically adjusts the data transmission frequency of blood sugar data sent to the blood sugar prediction platform by calculating the second data transmission frequency adjustment coefficient (S02). This intelligent adjustment mechanism can ensure that the frequency of data transmission is increased when necessary, thereby providing more timely and accurate health data. The technical solution comprehensively evaluates the user's health status through a complex calculation formula (involving the fourth adjustment parameter A and the fifth adjustment parameter B). This comprehensive evaluation method takes into account the changes in multiple physiological parameters and can provide users with more personalized health management services. Through automated and intelligent data monitoring and transmission, the technical solution can reduce the burden on medical staff and improve the efficiency and quality of medical services. At the same time, it also helps to reduce the waste of medical resources and improve the utilization efficiency of medical resources. This technical solution helps to timely detect potential health problems and take corresponding preventive measures by continuously monitoring the user's physiological parameters and dynamically adjusting the data transmission frequency according to data changes. This helps to reduce the incidence of diseases and improve the overall health level of users. By providing comprehensive physiological parameter monitoring and intelligent data transmission frequency adjustment, this technical solution can enhance the user experience. Users can more easily understand their health status and take corresponding health management measures according to data changes.

[0046] In summary, this technical solution provides users with a more intelligent, efficient and personalized health management service through comprehensive physiological parameter monitoring, dynamic response and early warning, intelligent adjustment of data transmission frequency, comprehensive evaluation and personalized management, improving the efficiency and quality of medical services, and promoting health management and disease prevention. At the same time, by introducing the function of adjusting the frequency of blood sugar data transmission according to blood pressure data, this technical solution further enhances the comprehensiveness and personalization of the system, improves the accuracy of blood sugar trend prediction, optimizes the utilization of data transmission resources, and enhances the safety and health management effect of patients.

[0047] In one embodiment of the present invention, the structure of the digital twin model that has completed training and testing is as follows: The input layer is used to process and integrate the input data from the heart rate sensor, blood pressure sensor, blood sugar sensor and insulin pump; The feature extraction layer is used to use the sliding time window to obtain the static feature data corresponding to the heart rate data, blood pressure data, blood sugar data, insulin delivery time and dosage data from the data provided by the heart rate acquisition sensor, blood pressure acquisition sensor, blood sugar acquisition sensor and insulin pump within the sliding time window; Multiple LSTM sub-networks are used to process heart rate data, blood pressure data, blood sugar data, insulin delivery time and dosage data, and obtain processed data information; The fusion layer is used to output and merge the processed data information obtained by multiple LSTM sub-networks; A regression layer, used to obtain blood sugar trend prediction results using one or more fully connected layers; The output layer is used to output the blood sugar trend prediction result.

[0048] The multiple LSTM subnetworks include a heart rate data processing LSTM subnetwork, a blood pressure data processing LSTM subnetwork, a blood sugar data processing LSTM subnetwork, and an insulin delivery time and dosage data processing LSTM subnetwork; The heart rate data processing LSTM sub-network is used to process the heart rate data; The blood pressure data processing LSTM sub-network is used to process the blood pressure data; The blood sugar data processing LSTM sub-network is used to process the blood sugar data; The insulin administration time and dosage data processing LSTM sub-network is used to process the insulin administration time and dosage data; The data processing includes data cleaning, digital serialization and data dynamic feature extraction.

[0049] The working principle of the above technical solution is as follows: Input layer: collects data from heart rate sensors, blood pressure sensors, blood sugar sensors and insulin pumps, integrates these data and passes them to the next layer.

[0050] Feature extraction layer: Using sliding time window technology, static feature data of heart rate, blood pressure, blood sugar, and insulin delivery time and dosage are extracted within the set time window. These static feature data reflect the stable state of each physiological parameter within a specific time period.

[0051] Multiple LSTM sub-networks: Each LSTM sub-network processes one type of data (heart rate, blood pressure, blood sugar, insulin delivery time and dosage). Data processing includes data cleaning (removing outliers, filling missing values, etc.), digital serialization (converting non-numeric data to numerical data), and data dynamic feature extraction (identifying trends, cycles, etc. in data). LSTM (Long Short-Term Memory) networks can capture the time dependency in data, thereby effectively processing time series data.

[0052] Fusion layer: Merges the data information processed by multiple LSTM sub-networks. This information is integrated together in some way (such as weighted summation, concatenation, etc.) so that it can be processed by the subsequent fully connected layer.

[0053] Regression layer: The fused data is processed using one or more fully connected layers. By training these fully connected layers, the model can learn the complex relationship between the input data and blood sugar trends.

[0054] Finally, the regression layer outputs the predicted blood sugar trend results.

[0055] Output layer: Output the predicted blood sugar trend results for users or subsequent systems to use.

[0056] The effect of the above technical solution is: by integrating multiple physiological parameters (heart rate, blood pressure, blood sugar, insulin delivery time and dosage), the model can provide comprehensive physiological state monitoring and analysis. By utilizing the ability of LSTM network to process time series data, the model can capture the time dependency in the data, thereby improving the accuracy of blood sugar trend prediction. Since the model is trained and tested based on the user's actual data, it can provide each user with a personalized blood sugar management plan. By predicting blood sugar trends, the model can issue an early warning before abnormal blood sugar fluctuations occur, thereby helping users to adjust their diet, exercise or insulin delivery strategies in a timely manner. The blood sugar trend prediction results provided by the model can be used as a decision support tool for doctors or patients to formulate diabetes management plans. Through accurate blood sugar predictions and personalized management plans, the model helps reduce the risk of complications in diabetic patients, thereby improving their quality of life.

[0057] In summary, this technical solution achieves accurate prediction and personalized management of blood sugar trends by integrating multiple physiological parameters and using the LSTM network for time series analysis, providing an effective health management tool for diabetic patients.

[0058] In one embodiment of the invention, the time length setting of the sliding time window includes: The data sending frequency of extracting the blood glucose data before adjustment corresponding to the blood glucose data and sending it to the blood glucose prediction platform; Extract the blood sugar data corresponding to the current adjusted data sending frequency sent to the blood sugar prediction platform; Acquire the time length of the adjusted sliding time window according to the initial data sending frequency and the currently corresponding adjusted data sending frequency; The time length of the adjusted sliding time window is obtained by the following formula: Wherein, T represents the time length of the sliding time window after adjustment; T0 represents the initial time length of the sliding time window; F0 represents the frequency of sending blood glucose data to the blood glucose prediction platform before adjustment; F t 01 represents the data transmission frequency of the adjusted blood sugar data obtained by the first data transmission frequency adjustment coefficient and sent to the blood sugar prediction platform; F t 02 represents the data transmission frequency of the adjusted blood sugar data obtained by the second data transmission frequency adjustment coefficient and sent to the blood sugar prediction platform; F max and F min They respectively represent the maximum and minimum values ​​of the data transmission frequency that occur during the data transmission frequency adjustment process.

[0059] The working principle of the above technical solution is: first, set the initial time length T0 of the sliding time window. This length is a value preset according to system requirements or experience. At the same time, determine the data sending frequency F0 of the blood glucose data before adjustment to the blood glucose prediction platform, which is also an initial, fixed frequency. As the system runs, the data sending frequency may need to be adjusted for various reasons (such as network conditions, device performance, user needs, etc.). The data sending frequency of the adjusted blood glucose data sent to the blood glucose prediction platform can be obtained through the first data sending frequency adjustment coefficient Ft01 and the second data sending frequency adjustment coefficient Ft02. These two adjustment coefficients may represent different adjustment strategies or conditions.

[0060] According to the adjusted data transmission frequency, the time length T of the adjusted sliding time window is calculated using the given formula. The formula takes into account the maximum value Fmax and the minimum value Fmin of the data transmission frequency, which helps to ensure that the adjusted sliding time window length is within a reasonable range. The adjusted sliding time window is applied to the feature extraction layer to extract static feature data of heart rate, blood pressure, blood sugar, and insulin delivery time and dosage within the set time window. These feature data will be used as input for subsequent LSTM sub-network processing.

[0061] The effect of the above technical solution is: by dynamically adjusting the time length of the sliding time window, the technical solution can adapt to different data sending frequencies, thereby improving the flexibility and adaptability of the system. Adjusting the length of the sliding time window can ensure that a sufficient number of data points are obtained in each time window, which helps to improve the accuracy of feature extraction, thereby improving the accuracy of blood sugar trend prediction. Adjusting the length of the sliding time window according to changes in the data sending frequency can avoid wasting computing resources in unnecessary time periods, thereby optimizing the overall performance of the system. For users, the dynamically adjusted sliding time window can ensure that accurate blood sugar prediction results are obtained in a timely manner when needed, thereby providing a better user experience. By reasonably setting and adjusting the length of the sliding time window, it can be ensured that the system maintains a stable operating state when facing different data sending frequencies, avoiding system crashes or performance degradation due to changes in data sending frequencies.

[0062] In summary, this technical solution adapts to changes in data sending frequency by dynamically adjusting the time length of the sliding time window, improves the flexibility, accuracy and resource optimization capabilities of the system, and provides users with a better experience and system stability.

[0063] The embodiment of the present invention proposes a blood sugar trend prediction method based on digital twins, such as Figure 2 As shown, the blood sugar trend prediction method based on digital twins includes: S1. Real-time collection of physiological signal parameters of the patient through the detection equipment worn by the patient; S2. Using a data transmission module to send the physiological signal parameters to a blood glucose prediction platform; S3. After receiving the physiological signal parameters corresponding to the patient, the blood sugar prediction platform retrieves the digital twin model that has been trained and tested, predicts the blood sugar trend according to the physiological signal parameters corresponding to the patient, and obtains the blood sugar trend results; S4, using the data transmission module to send the blood sugar trend result to the patient's mobile terminal; S5. The patient's mobile terminal receives blood sugar trend results in real time and displays the blood sugar trend results visually.

[0064] The working principle of the above technical solution is: the detection equipment worn by the patient (such as the sensor of the continuous blood glucose monitoring system CGM) collects the patient's physiological signal parameters, such as blood glucose level, heart rate, etc. in real time. The collected physiological signal parameters are sent to the blood glucose prediction platform through the data transmission module. At the same time, this module is also responsible for feeding back the blood glucose trend results generated by the blood glucose prediction platform to the patient's mobile terminal. After receiving the patient's physiological signal parameters, the blood glucose prediction platform will call up the digital twin model that has been trained and tested. The model uses the patient's real-time data and historical data to predict blood glucose trends and generate blood glucose trend results. The patient's mobile terminal receives the blood glucose trend results sent by the blood glucose prediction platform in real time, and displays them to the patient through a visual interface so that the patient can understand his or her blood glucose status at any time.

[0065] The effect of the above technical solution is: by adopting the digital twin model, the system can comprehensively consider the patient's various physiological signal parameters and historical data, so as to more accurately predict the blood sugar trend. This helps patients take timely measures to avoid the health risks caused by high or low blood sugar. The detection equipment worn by the patient can collect data in real time and quickly send it to the blood sugar prediction platform through the data transmission module. At the same time, the blood sugar prediction platform can also quickly process the data and generate blood sugar trend results, and then display them to the patient in real time through the patient's mobile terminal. This real-time nature helps patients understand their blood sugar status at any time and make corresponding adjustments. The digital twin model can be customized and optimized according to the specific situation of each patient, so as to realize personalized blood sugar trend prediction. This helps to meet the needs of different patients and improve the prediction effect. The visual display function of the patient's mobile terminal enables patients to easily view their blood sugar trend results without having to frequently go to the hospital or clinic for testing. This not only improves the convenience of patients, but also reduces medical costs.

[0066] In summary, the blood sugar trend prediction system based on digital twins has technical effects such as improving prediction accuracy, real-time performance, personalization and convenience, and can provide better health management services for diabetic patients.

[0067] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A blood sugar trend prediction system based on digital twins, characterized in that: The blood sugar trend prediction system based on digital twins includes a blood sugar prediction platform, a detection device worn by the patient, a data transmission module and a patient mobile terminal; The detection device worn by the patient is used to collect the patient's physiological signal parameters in real time; The data transmission module is used to send the physiological signal parameters to the blood glucose prediction platform, and feed back the blood glucose trend results obtained by the blood glucose prediction platform to the patient's mobile terminal; The blood sugar prediction platform is used to retrieve the trained and tested digital twin model after receiving the physiological signal parameters corresponding to the patient, perform blood sugar trend prediction according to the physiological signal parameters corresponding to the patient, and obtain the blood sugar trend results; The patient mobile terminal is used to receive blood sugar trend results in real time and visualize the blood sugar trend results.

2. The blood sugar trend prediction system based on digital twin according to claim 1 is characterized in that: The detection equipment worn by the patient includes a heart rate acquisition sensor, a blood pressure acquisition sensor, a blood sugar acquisition sensor and an insulin pump; The heart rate acquisition sensor is used to collect the patient's heart rate data and send the patient's heart rate data to the data transmission module; The blood pressure collection sensor is used to collect the patient's blood pressure data and send the patient's blood pressure data to the data transmission module; The blood sugar collection sensor is used to collect the patient's blood sugar data and send the patient's blood sugar data to the data transmission module; The insulin pump is used to collect the insulin delivery time and delivery measurement data of the patient's insulin injection, and send the patient's insulin delivery time and delivery measurement data to the data transmission module.

3. The blood sugar trend prediction system based on digital twin according to claim 1 is characterized in that: The data transmission module includes a physiological signal parameter sending module and a blood sugar trend result sending module; The physiological signal parameter sending module is used to send the physiological signal parameters to the blood glucose prediction platform; The blood sugar trend result sending module is used to feed back the blood sugar trend result obtained by the blood sugar prediction platform to the patient's mobile terminal.

4. The blood sugar trend prediction system based on digital twin according to claim 1 or 3, characterized in that: The data transmission module also includes a physiological signal parameter transmission adjustment module; The physiological signal parameter sending adjustment module is used to adjust the data sending frequency of the patient's blood sugar data to the blood sugar prediction platform according to the patient's heart rate data and blood pressure data.

5. The blood sugar trend prediction system based on digital twin according to claim 4 is characterized in that: The operation process of the physiological signal parameter sending and adjusting module includes: Receive user's heart rate data in real time; When the user's heart rate data exceeds a preset heart rate data threshold, the patient's blood sugar data and blood pressure data within a preset monitoring time period are retrieved; wherein the preset monitoring time period ranges from 1h to 6h; The first data transmission frequency adjustment coefficient is obtained according to the patient's blood sugar data, blood pressure data and heart rate data within the preset monitoring time period; wherein the first data transmission frequency adjustment coefficient is obtained by the following formula: Among them, S 01 represents the first data transmission frequency adjustment coefficient; x, y and z represent the first adjustment parameter, the second adjustment parameter and the third adjustment parameter respectively, and the first adjustment parameter is obtained by the following formula: Wherein, x represents the first adjustment parameter; α represents the first adjustment constant, and the value range of the first adjustment constant is 0.12-0.54; n represents the number of times blood sugar data is collected; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP BP Indicates the overall systolic blood pressure value change rate corresponding to the m-time blood pressure data collection; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; BS i Indicates the blood sugar value corresponding to the i-th blood sugar collection; BS 0i BS represents the blood sugar reference value corresponding to the i-th blood sugar collection; z Indicates the median fasting blood glucose value corresponding to the collection of n blood glucose data; The second adjustment parameter is obtained by the following formula: Wherein, y represents the second adjustment parameter; β represents the second adjustment constant, and the value range of the second adjustment constant is 0.18-0.39; m represents the total number of blood pressure data collection; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP Indicates the overall systolic blood pressure value change rate corresponding to m blood pressure data collections; BP i represents the systolic blood pressure value corresponding to the i-th blood pressure collection; BP0 represents the systolic blood pressure reference value corresponding to the patient; BS z It represents the median value of systolic blood pressure corresponding to the collection of m blood pressure data; The third adjustment parameter is obtained by the following formula: Wherein, z represents the third adjustment parameter; γ represents the third adjustment constant, and the value range of the third adjustment constant is 0.15-0.32; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP Indicates the overall systolic blood pressure value change rate corresponding to m blood pressure data collections; HR i represents the heart rate value corresponding to the i-th heart rate collection; r represents the number of heart rate data collection; HR i-1 represents the heart rate value corresponding to the i-1th heart rate acquisition; HR0 represents the heart rate reference value corresponding to the patient; HR z Indicates the median heart rate value corresponding to the collection of r heart rate data; Using the first data transmission frequency adjustment coefficient, the frequency of sending blood glucose data to the blood glucose prediction platform after the end of the preset monitoring time period is adjusted; The frequency of sending the adjusted blood sugar data to the blood sugar prediction platform is obtained by the following formula: Among them, F t 01 represents the data transmission frequency of the adjusted blood sugar data obtained by the first data transmission frequency adjustment coefficient and sent to the blood sugar prediction platform; F0 represents the data transmission frequency of the blood sugar data before adjustment and sent to the blood sugar prediction platform; S 01 Indicates the first data sending frequency adjustment coefficient.

6. The blood sugar trend prediction system based on digital twin according to claim 4 is characterized in that: The operation process of the physiological signal parameter sending and adjusting module also includes: Receive users' blood pressure data in real time; When the blood pressure data of the user exceeds the preset blood pressure data threshold, the blood sugar data and heart rate data of the patient within the preset monitoring time period are retrieved; wherein the preset monitoring time period ranges from 1h to 6h; The second data transmission frequency adjustment coefficient is obtained according to the blood glucose data and heart rate data of the patient within the preset monitoring time period; wherein the second data transmission frequency adjustment coefficient is obtained by the following formula: Among them, S 02 represents the second data transmission frequency adjustment coefficient; A and B represent the fourth adjustment parameter and the fifth adjustment parameter respectively, and the fourth adjustment parameter is obtained by the following formula: Wherein, A represents the fourth adjustment parameter; α represents the first adjustment constant, and the value range of the first adjustment constant is 0.12-0.54; n represents the number of times blood sugar data is collected; BS i Indicates the blood sugar value corresponding to the i-th blood sugar collection; BS 0i BS represents the blood sugar reference value corresponding to the i-th blood sugar collection; z Indicates the median fasting blood glucose value corresponding to the collection of n blood glucose data; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP represents the overall systolic blood pressure value change rate corresponding to the m-time blood pressure data collection; BP represents the current systolic blood pressure value; BP0 represents the systolic blood pressure reference value corresponding to the patient; The fifth adjustment parameter is obtained by the following formula: Wherein, B represents the fifth adjustment parameter; BP represents the current systolic blood pressure value; BP0 represents the systolic blood pressure reference value corresponding to the patient; ΔP BS Indicates the overall blood sugar value change rate corresponding to n blood sugar data collections; ΔP HR Indicates the overall heart rate value change rate corresponding to r heart rate data collections; ΔP BP represents the overall systolic blood pressure value change rate corresponding to the m-time blood pressure data collection; BP represents the current systolic blood pressure value; HR represents the current heart rate value; HR0 represents the patient's corresponding heart rate reference value; HR z represents the middle heart rate value corresponding to the collection of r heart rate data; γ represents the third adjustment constant, and the value range of the third adjustment constant is 0.15-0.32; r represents the number of heart rate data collection; W represents the sign parameter, when HR-HR0 is not less than 1, W=1; when HR-HR0 is less than 1, W=-1; Using the second data transmission frequency adjustment coefficient, the frequency of sending blood glucose data to the blood glucose prediction platform after the end of the preset monitoring time period is adjusted; The frequency of sending the adjusted blood sugar data to the blood sugar prediction platform is obtained by the following formula: Among them, F t 02 represents the data transmission frequency of the blood glucose data after adjustment obtained by the second data transmission frequency adjustment coefficient and sent to the blood glucose prediction platform; F0 represents the data transmission frequency of the blood glucose data before adjustment and sent to the blood glucose prediction platform; S 02 Indicates the second data sending frequency adjustment coefficient.

7. The blood sugar trend prediction system based on digital twin according to claim 1 is characterized in that: The structure of the digital twin model that has been trained and tested is as follows: The input layer is used to process and integrate the input data from the heart rate sensor, blood pressure sensor, blood sugar sensor and insulin pump; The feature extraction layer is used to use the sliding time window to obtain the static feature data corresponding to the heart rate data, blood pressure data, blood sugar data, insulin delivery time and dosage data from the data provided by the heart rate acquisition sensor, blood pressure acquisition sensor, blood sugar acquisition sensor and insulin pump within the sliding time window; Multiple LSTM sub-networks are used to process heart rate data, blood pressure data, blood sugar data, insulin delivery time and dosage data, and obtain processed data information; The fusion layer is used to output and merge the processed data information obtained by multiple LSTM sub-networks; A regression layer, used to obtain blood sugar trend prediction results using one or more fully connected layers; The output layer is used to output the blood sugar trend prediction result.

8. The blood sugar trend prediction system based on digital twin according to claim 7 is characterized in that: The time length setting of the sliding time window includes: The data sending frequency of extracting the blood glucose data before adjustment corresponding to the blood glucose data and sending it to the blood glucose prediction platform; Extract the blood sugar data corresponding to the current adjusted data sending frequency sent to the blood sugar prediction platform; Acquire the time length of the adjusted sliding time window according to the initial data sending frequency and the currently corresponding adjusted data sending frequency; The time length of the adjusted sliding time window is obtained by the following formula: Wherein, T represents the time length of the sliding time window after adjustment; T0 represents the initial time length of the sliding time window; F0 represents the frequency of sending blood glucose data to the blood glucose prediction platform before adjustment; F t 01 represents the data transmission frequency of the adjusted blood sugar data obtained by the first data transmission frequency adjustment coefficient and sent to the blood sugar prediction platform; F t 02 represents the data transmission frequency of the adjusted blood sugar data obtained by the second data transmission frequency adjustment coefficient and sent to the blood sugar prediction platform; F max and F min They respectively represent the maximum and minimum values ​​of the data transmission frequency that occur during the data transmission frequency adjustment process.

9. The blood sugar trend prediction system based on digital twin according to claim 7 is characterized in that: The multiple LSTM subnetworks include a heart rate data processing LSTM subnetwork, a blood pressure data processing LSTM subnetwork, a blood sugar data processing LSTM subnetwork, and an insulin delivery time and dosage data processing LSTM subnetwork; The heart rate data processing LSTM sub-network is used to process the heart rate data; The blood pressure data processing LSTM sub-network is used to process the blood pressure data; The blood sugar data processing LSTM sub-network is used to process the blood sugar data; The insulin administration time and dosage data processing LSTM sub-network is used to process the insulin administration time and dosage data; The data processing includes data cleaning, digital serialization and data dynamic feature extraction.

10. A blood sugar trend prediction method based on digital twins, characterized in that: The blood sugar trend prediction method based on digital twins includes: The patient's physiological signal parameters are collected in real time through the detection equipment worn by the patient; The physiological signal parameters are sent to the blood glucose prediction platform using the data transmission module After receiving the physiological signal parameters corresponding to the patient, the blood sugar prediction platform calls the digital twin model that has been trained and tested, predicts the blood sugar trend based on the physiological signal parameters corresponding to the patient, and obtains the blood sugar trend results; Using the data transmission module to send the blood sugar trend result to the patient's mobile terminal; The patient's mobile terminal receives blood sugar trend results in real time and displays the blood sugar trend results visually.

Images