A method for shortening the time of non-invasive blood glucose sensor head and skin heat exchange

Abstract

The present invention relates to the technical field of blood glucose detection, and specifically discloses a method for shortening the heat exchange time between a non-invasive blood glucose sensing probe and the skin; the method includes: S100, emitting near-infrared light through the non-invasive blood glucose sensing probe and receiving and identifying the wavelength data of the diffuse reflection light; S200, performing data processing on the wavelength data of the diffuse reflection light, and inputting the processed data into a steady baseline prediction model to obtain a predicted blood glucose value; the steady baseline prediction model is obtained by training with a neural network; through the prediction of blood glucose changes by the steady baseline prediction model, the present invention can determine the position where the blood glucose change curve tends to be stable, and thus does not require the user to wait for the time point when the temperature of the human body and the blood glucose detection device is stable, that is, it can predict the value detected under the stable state, which can significantly improve the accuracy of the blood glucose detection result and shorten the time of non-invasive blood glucose detection.

Description

Technical Field

[0001] The present invention relates to the technical field of blood glucose detection, and specifically to a method for shortening the heat exchange time between a non-invasive blood glucose sensing probe and the skin. Background Art

[0002] For diabetic patients, blood glucose detection needs to be carried out for a long time. Traditional blood glucose detectors mainly use the method of pricking to draw blood. Although this method has high detection accuracy, it will cause wounds to patients. Long-term detection processes will lead to problems such as wound infections in patients. Therefore, non-invasive blood glucose detection is the development direction of blood glucose detection.

[0003] In the existing non-invasive blood glucose detection, near-infrared wavelengths are used as the sensing source for detecting blood glucose. According to the human skin, a near-infrared wavelength transmitter and receiver are designed. When starting the detection, the user wears the detector on the back of the hand. At this time, the transmitter emits light with specified wavelengths in the range of 1300 - 1800 nm, and the receiver receives the diffusely reflected light and performs digital processing; during this process, since the temperature of the skin epidermis is dynamically stable, while the detection head of the infrared wave transmitter has no temperature, the propagation path of the infrared wavelength will be affected by temperature. Therefore, during the test, there will be a heat exchange process between the epidermis temperature and the infrared detection head, which will affect the propagation of the infrared wave, thus affecting the measurement result, resulting in the measured data showing a slow downward trend. At the same time, it takes about 2 hours for this trend to stabilize. Only when the data after stabilization is measured can the accuracy of the data be ensured. Therefore, it is difficult to fully implement this process in wearable devices, so there is a deviation between the detected data and the actual value. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for shortening the heat exchange time between a non-invasive blood glucose sensing probe and the skin, and to solve the following technical problems:

[0005] How to shorten the time of non-invasive blood glucose detection and ensure the accuracy of the detection.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] A method for shortening the heat exchange time between a non-invasive blood glucose sensing probe and the skin, the method comprising:

[0008] S100. Emitting near-infrared light through a non-invasive blood glucose sensing probe and receiving and identifying the wavelength data of the diffusely reflected light;

[0009] S200. Performing data processing on the wavelength data of the diffusely reflected light, and inputting the processed data into a steady baseline prediction model to obtain a predicted blood glucose value;

[0010] The stable baseline prediction model is obtained by training a neural network.

[0011] Furthermore, the process of establishing the stable baseline prediction model is as follows:

[0012] Collect data of the human body from the start stage of the test to the stage of stable values under the same test conditions to generate a training data set;

[0013] Build a model using a one-dimensional convolutional neural network and train the built model with the training data set.

[0014] Furthermore, step S102 also includes:

[0015] Add random noise and negative samples to train the built model.

[0016] Furthermore, before step S200, the method also includes:

[0017] Detect the current environmental temperature value, and select the corresponding stable baseline prediction model according to the environmental temperature range in which the current environmental temperature value falls.

[0018] Furthermore, the process of establishing the stable baseline prediction model also includes:

[0019] Set different environmental temperature ranges for the environmental temperature, and collect data of the human body from the start stage of the test to the stage of stable values under the environmental state at the central value of the environmental temperature range to generate a training data set;

[0020] Build and train a convolutional neural network according to the training data sets of different environmental temperature ranges to obtain stable baseline prediction models corresponding to different environmental temperature ranges.

[0021] Advantages of the present invention:

[0022] (1) Through the prediction of blood glucose changes by the stable baseline prediction model of the present invention, the position where the blood glucose change curve tends to be stable can be judged. Furthermore, it is not necessary for the user to wait for the time point when the temperature of the human body and the blood glucose detection device is stable, that is, the value detected under the stable state can be predicted. Compared with the directly detected result, the present invention can significantly improve the accuracy of the blood glucose detection result. At the same time, the present invention can obtain the blood glucose value under the stable state in a relatively short time, thereby achieving the effect of shortening the time of non-invasive blood glucose detection and ensuring the detection accuracy. Description of the Drawings

[0023] The present invention will be further described below with reference to the drawings.

[0024] Figure 1 is a flowchart of the steps of the method for shortening the non-invasive blood glucose sensor and skin heat exchange time of the present invention;

[0025] Figure 2 It is a graph showing the change of the detected current value of the test data over time during the establishment of the stable baseline prediction model of the present invention;

[0026] Figure 3 It is a fitting curve showing the change of the detected current over time of the test data during the establishment of the stable baseline prediction model of the present invention. Specific Embodiments

[0027] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0028] Please refer to Figure 1 As shown, in one embodiment, a method for shortening the skin heat exchange time of a non-invasive blood glucose sensor head is provided. The method includes:

[0029] S100. Emit near-infrared light through a non-invasive blood glucose sensor head and receive and identify the wavelength data of the diffuse reflected light;

[0030] S200. Process the wavelength data of the diffuse reflected light, input the processed data into the stable baseline prediction model, and obtain the predicted blood glucose value;

[0031] The stable baseline prediction model is obtained by training with a neural network.

[0032] Through the above technical solutions, the present invention provides a way to shorten the time of non-invasive blood glucose detection. Specifically, first, the detection data of the user is collected through near-infrared light. Secondly, the detection data of the user is input into the stable baseline prediction model. Through the prediction of the blood glucose change by the stable baseline prediction model, the position where the blood glucose change curve tends to be stable can be judged. Furthermore, it is not necessary for the user to wait for the time point when the temperature of the human body and the blood glucose detection device is stable, that is, the value detected in the stable state can be predicted. Compared with the directly detected result, the present invention can significantly improve the accuracy of the blood glucose detection result. At the same time, the present invention can obtain the blood glucose value in the stable state in a shorter time, thereby achieving the effect of shortening the time of non-invasive blood glucose detection and ensuring the detection accuracy.

[0033] As an implementation manner of the present invention, the process of establishing the stable baseline prediction model is as follows:

[0034] Collect data of the human body from the start stage to the stable stage of the test under the same test conditions to generate a training data set;

[0035] A one-dimensional convolutional neural network is used to build a model, and the built model is trained with a training dataset.

[0036] Through the above technical solution, this embodiment provides a method for establishing a stable baseline prediction model. Specifically, please refer to Figure 2 , Figure 3 As shown, in the same test environment, data of the human body from the start stage of the test to the stage of stable value is collected to generate a training dataset. A one-dimensional convolutional neural network is used to build a model, and the built model is trained with the training dataset. Furthermore, a model for obtaining changes in blood glucose detection values can be obtained based on various test data. Therefore, through the establishment process of the stable baseline prediction model, after the blood glucose detection device obtains real-time blood glucose data, the stable baseline prediction model is used to adjust the real-time blood glucose data, and then more accurate measurement data can be obtained.

[0037] As an implementation manner of the present invention, step S102 further includes:

[0038] Adding random noise and negative samples to train the built model.

[0039] Furthermore, since step S102 requires a large number of samples to be collected and the data scale is large, this embodiment adds random noise and negative samples during the training process, which can improve the accuracy of model analysis.

[0040] As an implementation manner of the present invention, the method further includes before step S200:

[0041] Detecting the current environmental temperature value, and selecting a corresponding stable baseline prediction model according to the environmental temperature range in which the current environmental temperature value falls.

[0042] In order to further improve the accuracy of the stable baseline prediction model analysis, this embodiment has established corresponding stable baseline prediction models for different environmental temperature values; at the same time, the temperature value of the current environment is detected by the blood glucose detection device, and the corresponding stable baseline prediction model is selected according to the environmental temperature range where the environmental temperature value is located. Since the temperature change of the human body is small, while the range of the environmental temperature is large and the temperature of the blood glucose detection device is the same as the environmental temperature, therefore, by detecting the environmental temperature and selecting the corresponding stable baseline prediction model, the curve of blood glucose value change can be predicted more accurately, and then the position where the detected blood glucose curve region is stable can be judged more accurately, and the detected blood glucose value can be determined more accurately.

[0043] As an implementation manner of the present invention, the process of establishing the stable baseline prediction model further includes:

[0044] Set different environmental temperature ranges according to a specific range of ambient temperatures, and collect data of the human body from the start stage of the test to the stage of stable values under the environmental conditions at the central values of the environmental temperature ranges to generate a training dataset;

[0045] Build and train a convolutional neural network based on the training datasets of different environmental temperature ranges to obtain a stable baseline prediction model corresponding to each environmental temperature range.

[0046] Through the above technical solution, in this embodiment, the stable baseline prediction model is established for different environmental ranges as follows: First, set different environmental temperature ranges according to a specific range of ambient temperatures, and collect data of the human body from the start stage of the test to the stage of stable values under the environmental conditions at the central values of the environmental temperature ranges to generate a training dataset; then build and train a convolutional neural network based on the training datasets of different environmental temperature ranges to obtain a stable baseline prediction model corresponding to each environmental temperature range. Since the training datasets in this embodiment are all obtained by detection within the corresponding environmental temperature ranges, the model established based on the training datasets can accurately predict the trend of blood glucose changes in this environmental range, thus achieving the effect of prediction according to different temperature adaptabilities.

[0047] The working principle of the present invention: Through the prediction of blood glucose changes by the stable baseline prediction model of the present invention, it is possible to determine the position where the blood glucose change curve tends to be stable, so that it is not necessary for the user to wait for the time point when the temperature of the human body and the blood glucose detection device is stable, that is, the value detected in the stable state can be predicted. Compared with the directly detected result, the present invention can significantly improve the accuracy of the blood glucose detection result. At the same time, the present invention can obtain the blood glucose value in the stable state in a shorter time, thus achieving the effect of shortening the time of non-invasive blood glucose detection and ensuring the detection accuracy; the present invention selects the corresponding stable baseline prediction model according to the environmental temperature range where the environmental temperature value is located, can more accurately predict the curve of blood glucose value changes, and thus more accurately determine the position where the detected blood glucose curve is stable, and further more accurately determine the detected blood glucose value.

[0048] The above has described a specific embodiment of the present invention in detail, but the content described is only the preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the application of the present invention should still fall within the scope covered by the patent of the present invention.

Claims

1. A method for shortening the time of non-invasive blood glucose sensor head and skin heat exchange, characterized in that The method includes: S100. Emitting near-infrared light through a non-invasive blood glucose sensor and receiving and identifying the wavelength data of diffuse reflection light; S200. Processing the wavelength data of diffuse reflection light, inputting the processed data into a steady baseline prediction model, and obtaining a predicted blood glucose value; The steady baseline prediction model is obtained by training with a neural network; Before step S200, the method further includes: Detecting the current environmental temperature value, and selecting a corresponding steady baseline prediction model according to the environmental temperature range in which the current environmental temperature value falls; The process of establishing the steady baseline prediction model further includes: Setting different environmental temperature ranges for the environmental temperature, collecting data of the human body from the start stage of the test to the stage of stable value under the environmental state at the central value of the environmental temperature range, and generating a training data set; Building and training a convolutional neural network according to the training data sets of different environmental temperature ranges, and obtaining steady baseline prediction models corresponding to different environmental temperature ranges.

2. A method for shortening the skin heat exchange time of a non-invasive blood glucose sensor probe according to claim 1, characterized in that, The process of establishing the steady baseline prediction model is: Collecting data of the human body from the start stage of the test to the stage of stable value under the same test conditions, and generating a training data set; Building a model using a one-dimensional convolutional neural network, and training the built model with the training data set.

3. A method for shortening the skin thermal exchange time of a non-invasive blood glucose sensor probe according to claim 2, characterized in that Training the built model by adding random noise and negative samples.

Images