WO2018214298A1

WO2018214298A1 - Non-invasive blood glucose detection system and method

Info

Publication number: WO2018214298A1
Application number: PCT/CN2017/098190
Authority: WO
Inventors: 张贯京; 葛新科; 高伟明; 张红治; 梁昊原; 陈琦; 周亮
Original assignee: 深圳市前海安测信息技术有限公司
Priority date: 2017-05-20
Filing date: 2017-08-19
Publication date: 2018-11-29
Also published as: CN107361776A

Abstract

A non-invasive blood glucose detection system and method. The method comprises the steps of: controlling an infrared light source (13) so that same emits near-infrared light to irradiate a human body part to be detected, and acquiring, by means of a photoelectric sensor (14), a pulse wave signal from the human body part to be detected and under the irradiation of the near-infrared light; pre-processing the pulse wave signal by means of a signal processing circuit (21); extracting a characteristic value of the pulse wave signal; using a predictive neural network to detect the characteristic value of the pulse wave signal so as to obtain an initial detection value of a blood glucose concentration; using a classification neural network to detect the characteristic value of the pulse wave signal so as to obtain a detection interval of the blood glucose concentration; when the initial detection value is within the detection interval, displaying the initial detection value on a display screen (24) as a human body blood glucose concentration value of a person to be detected; and sending, by means of a communication unit (25), the human body blood glucose concentration value and a blood glucose measurement period of the person to be detected to a health management platform (02). The system and method can improve the accuracy of blood glucose concentration detection and enable a person to be detected to accurately understand his/her blood glucose condition.

Description

Non-invasive blood glucose detecting system and method

Technical field

[0001] The present invention relates to the field of non-invasive blood glucose detecting technology, and in particular, to a non-invasive blood glucose detecting system and method.

Background technique

[0002] With the development of social economy, diabetes has become one of the major diseases that endanger human health in modern society. Hyperglycemia is too high or too low to affect not only the patient's metabolism, but also some complications, such as cardiovascular disease and neuropathy, which pose a great threat to the health of patients. According to the World Health Organization report, there will be 300 million people with diabetes worldwide by 2035, and a large proportion of people with diabetes in China will also be present. In recent years, patients with diabetes have not only existed in some elderly people, but also in some young people. Diabetes is a chronic disease that is difficult to achieve with a single treatment, so people with diabetes need to know exactly what their blood sugar levels are.

[0003] However, at present, for the blood glucose detecting method, in the hospital or the patient's own home, an invasive blood glucose detecting method is adopted, that is, the patient's blood is directly taken, and the blood sugar level of the patient is detected according to an electrochemical method. This method of detection causes certain physical pain to the patient, and repeated blood draws are likely to cause infection. Further, electrochemical reaction test papers are expensive, and are also a large economic burden for diabetic patients. Non-invasive blood glucose testing can eliminate the pain of patient detection, can be frequently detected, improve the quality of life of patients

[0004] At present, there are many non-invasive blood glucose detecting methods, and infrared light-based methods for detecting glucose concentration in human blood are widely used in the research of non-invasive blood glucose detecting. However, there are many other components in the blood other than glucose, which limits the accuracy of blood glucose detection. How to improve the accuracy of blood glucose detection has become an urgent problem to be solved.

technical problem

[0005] A primary object of the present invention is to provide a non-invasive blood glucose detecting system and method, which aims to solve the technical problem that the existing non-invasive blood glucose detecting method has low accuracy for blood glucose detecting.

Problem solution Technical solution

[0006] In order to achieve the above object, the present invention provides a non-invasive blood glucose detecting system, which is operated in a non-invasive blood glucose detecting device, the non-invasive blood glucose detecting device includes a signal collector and a non-invasive blood glucose detecting device, and the signal collector includes an infrared light source. And the photoelectric sensor, the non-invasive blood glucose detecting device comprises a signal processing circuit, a display screen and a communication unit, and the non-invasive blood glucose detecting system comprises: a signal acquiring module, configured to control the infrared light source to emit near-infrared light to be irradiated on the body to be tested, and pass The photoelectric sensor obtains a pulse wave signal from a part to be tested of the human body under the illumination of the near-infrared light; the signal processing module is configured to pre-process the acquired pulse wave signal by the signal processing circuit to obtain a digital signal of the pulse wave signal, and from the pulse The characteristic value of the pulse wave signal is extracted from the digital signal of the wave signal; the initial detection module is configured to detect the characteristic value of the pulse wave signal by using the predictive neural network to obtain the initial detection value of the blood glucose concentration; the interval detecting module is configured to detect by using the classification neural network Pulse wave signal The eigenvalue obtains a detection interval of the blood glucose concentration; the blood glucose determination module is configured to determine whether the initial detection value belongs to the detection interval, and when the initial detection value is within the detection interval, the initial detection value is displayed on the display screen as the test subject The blood glucose concentration value of the human body; the blood glucose management module is configured to display a segment selection box on the display screen for the person to be tested to select the blood glucose measurement segment, receive the blood glucose measurement segment selected by the test subject, and measure the blood glucose concentration of the human body to be tested. The value and blood glucose measurement segments are sent to the health management platform via the communication unit for the physician to determine the blood glucose condition of the subject.

[0007] Preferably, the blood glucose determining module is further configured to: when the initial detection value is not in the detection interval, the non-invasive blood glucose detecting device discards the initial detection value and detects the next pulse wave signal until the initial detection value is in the detection interval. So far.

[0008] Preferably, the signal processing circuit is configured to filter the pulse wave signal acquired by the photoelectric sensor to remove noise and DC components in the pulse wave signal, leaving a required AC component, and amplifying the pulse wave signal. Analog to digital conversion to obtain a digital signal of the pulse wave signal.

[0009] Preferably, the signal processing module is configured to perform wavelet transform on the acquired pulse wave signal to obtain a wavelet transform sequence, and search for a modulus maximum value that meets a preset threshold value in the wavelet transform sequence according to a preset threshold value.

And extracting a characteristic value of the pulse wave signal according to a modulus maximum value.

[0010] Preferably, the infrared light source comprises at least a plurality of near-infrared light-emitting tubes in a wavelength range of 800 _n m to 1500 nm, and is configured to emit near-infrared light including a band of 800 _n m to 1500 nm to a portion to be tested of a human body.

[0011] The present invention also provides a non-invasive blood glucose detecting method, which is applied to a non-invasive blood glucose detecting device, which is non-invasive The blood glucose detecting device comprises a signal collector and a non-invasive blood glucose detecting device, the signal collector comprises an infrared light source and a photoelectric sensor, the non-invasive blood glucose detecting device comprises a signal processing circuit, a display screen and a communication unit, and the non-invasive blood glucose detecting method comprises the steps : controlling the infrared light source to emit near-infrared light to be irradiated on the part to be tested, and obtaining a pulse wave signal from the body to be tested under the near-infrared light by the photoelectric sensor; pre-processing the acquired pulse wave signal by the signal processing circuit Obtaining the digital signal of the pulse wave signal, and extracting the characteristic value of the pulse wave signal from the digital signal of the pulse wave signal; using the predicted neural network to detect the characteristic value of the pulse wave signal to obtain the initial detection value of the blood glucose concentration; using the classification neural network to detect the pulse value The characteristic value of the wave signal obtains the detection interval of the blood glucose concentration; determines whether the initial detection value belongs to the detection interval, and when the initial detection value is within the detection interval, the initial detection value is displayed on the display screen as the human body blood glucose concentration value of the test subject On the display Displaying a segment selection box for the candidate to select the blood glucose measurement segment, and receiving the blood glucose measurement segment selected by the test subject; and transmitting the human body blood glucose concentration value and the blood glucose measurement segment of the test subject to the health management platform through the communication unit For the doctor to determine the blood sugar status of the person to be tested.

[0012] Preferably, the non-invasive blood glucose detecting method further comprises the step of: discarding the initial detection value and detecting the next pulse wave signal until the initial detection value is within the detection interval if the initial detection value is not within the detection interval.

[0013] Preferably, the step of pre-processing the acquired pulse wave signal by the signal processing circuit to obtain the digital signal of the pulse wave signal comprises the following steps: filtering the pulse wave signal to remove noise in the pulse wave signal and a DC component, leaving a desired AC component; amplifying and analog-to-digital converting the pulse wave signal to obtain a digital signal of the pulse wave signal, and transmitting the digital signal to the non-invasive blood glucose detecting device

[0014] Preferably, the step of extracting the characteristic value of the pulse wave signal from the digital signal of the pulse wave signal comprises the following steps: performing wavelet transformation on the acquired pulse wave signal to obtain a wavelet transform sequence; and wavelet according to a preset threshold value Finding a modulus maxima corresponding to the preset threshold in the transform sequence; extracting the feature value of the pulse wave signal according to the modulus maxima.

[0015] Preferably, the infrared light source comprises at least a plurality of the near-infrared light emitting tube 800 _n m-1500nm band, a human body part to be measured for transmitting to at least near-infrared light 800 _n m-1500nm band.

Advantageous effects of the invention Beneficial effect

[0016] Compared with the prior art, the non-invasive blood glucose detecting system and method of the present invention uses a pulse wave signal obtained by predicting a neural network to obtain a detection interval to which a blood glucose concentration belongs, and determines a pulse wave signal obtained by the classification neural network detection. Whether the initial detection value of the blood glucose concentration belongs to the detection interval, and when the initial detection value is within the detection interval, the initial detection value is a blood glucose concentration value, and by determining the interval to which the initial detection value belongs, the human blood can be effectively reduced. The interference of other components in the blood glucose concentration improves the accuracy and accuracy of blood glucose concentration detection. In addition, the human blood glucose concentration value and the blood glucose measurement segment of the test subject are sent to the health management platform for the doctor to determine the blood glucose condition of the test subject. Brief description of the drawing

DRAWINGS

1 is a schematic diagram of an application environment of a preferred embodiment of the non-invasive blood glucose detecting system of the present invention;

2 is a schematic structural view of a preferred embodiment of the non-invasive blood glucose detecting device of FIG. 1;

3 is a functional block diagram of a non-invasive non-invasive blood glucose detecting system of the present invention;

4 is a flow chart of a preferred embodiment of the non-invasive blood glucose detecting method of the present invention;

[0021] FIG. 5 is a waveform diagram of a pulse wave signal;

6 is a detailed flowchart of the feature value of the extracted pulse wave signal in step S33 of FIG. 4.

[0023] The implementation, functional features, and advantages of the present invention will be further described with reference to the accompanying drawings.

BEST MODE FOR CARRYING OUT THE INVENTION

The specific embodiments, structures, features, and effects of the present invention are described in detail below with reference to the accompanying drawings and preferred embodiments. It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0025] Referring to FIG. 1, FIG. 1 is a schematic diagram of an application environment of a preferred embodiment of the non-invasive blood glucose detecting system of the present invention.

In the present embodiment, the non-invasive blood glucose detecting system 20 is applied to the non-invasive blood glucose detecting device 01, and the non-invasive blood glucose detecting device 01 establishes a communication connection with the health management platform 02 via the communication network 03. The non-invasive blood glucose detecting device 01 includes a signal collector 1 and a non-invasive blood glucose detecting device 2, wherein: the signal collector 1 is configured to collect and output a pulse wave signal. The non-invasive blood glucose detecting device 2 is connected to the signal collector 1 The non-invasive blood glucose detecting system 20 detects the pulse wave signal outputted by the signal collector 1 to obtain a human blood glucose concentration value of the test subject, and sends the detected blood glucose concentration value to the health management platform 02.

In the embodiment, the health management platform 02 is a server, a computer or a doctor workstation installed in a medical management institution of a third party. For the doctor to pass the health management platform 02, according to the human body blood glucose concentration value of the test subject, the blood glucose level of the test subject such as high blood sugar concentration, low blood sugar level and normal blood sugar concentration result can be given. The doctor can also give the testee's blood glucose management recommendations based on the blood glucose status of the test subject through the health management platform 02, such as dietary advice, lifestyle recommendations and exercise recommendations. Thus, the health management platform 02 transmits the blood glucose condition of the subject and the blood glucose management advice to the communication device (e.g., mobile phone) of the test subject via the communication network 03, so that the test subject manages his or her blood sugar condition. The communication network 03 may be a wired communication network or a wireless communication network. The communication network 03 in this embodiment is preferably a wireless communication network, including but not limited to, a GSM network, a GPRS network, a CDMA network, a TD-SC DMA network, and a WiMAX. Wireless transmission networks such as networks, TD-LTE networks, and FDD-LTE networks.

Referring to FIG. 2, FIG. 2 is a schematic structural view of a preferred embodiment of the non-invasive blood glucose detecting device 01 of FIG. In the present embodiment, the signal collector 1 includes a support frame 10, an upper detection plate 11, a lower detection plate 12, an infrared light source 13, and a photosensor 14. The upper end of the support frame 10 is provided with a sliding slot 100, and one end of the upper detecting plate 11 is movably connected with the sliding slot 100, that is, the upper detecting plate 11 can be moved up and down in the sliding slot 100, so that the upper detecting board 11 can be adjusted. The distance between the plates 12 is suitable for placing blood pressure detection on different parts of the human body to be tested, which improves the flexibility and applicability of the device. The lower detecting plate 12 is fixed to the lower end of the support frame 10 and disposed in parallel with the axis of the upper detecting plate 11. The upper surface of the lower detecting plate 12 is provided with a hinge mechanism 16, and the hinge mechanism 16 is provided with return springs 17 on both sides thereof, and is connected to the upper detecting plate 11 via a return spring 17. The lower end of the hinge mechanism 16 is provided with a support rod 18 which is fixed to the upper surface of the lower detecting plate 12 by a support rod 18.

[0028] In the embodiment, the upper detecting plate 11 is located at a position directly above the hinge mechanism 16 and is provided with a screw hole 110, and the screw hole 110 is provided with an internal thread. The upper end of the hinge mechanism 16 is fixed with a screw 111. The outer surface of the screw 111 is provided with an external thread. The screw 111 passes through the screw hole 110 and the external thread of the screw 111 matches the internal thread of the screw hole 110. Since the hinge mechanism 16 is connected to the upper detecting plate 11 through the return spring 17, the user can manually rotate the screw 111 to move the upper detecting plate 11 up and down to bring the return spring 17 up and down and compress, so that the upper detecting plate 11 can be moved along the sliding groove 100. Move up and down, so you can adjust the upper detection board 11 and the next inspection The distance between the plates 12. In this embodiment, the infrared light source 13 is embedded in the lower surface of the front end of the upper detecting plate 11, and the photoelectric sensor 14 is embedded on the front surface of the front end of the lower detecting plate 12, because the upper detecting plate 11 and the lower detecting plate 12 are arranged in parallel. Therefore, the infrared light source 13 and the photosensor 14 are disposed coaxially.

[0029] In the embodiment, the photosensor 14 is disposed in the detecting portion 15, and the detecting portion 15 is provided with a front surface of the front end of the lower detecting plate 12 for providing a blood sugar concentration detecting place, which can be used for placing the human body. a portion to be tested, such as a human finger capillary tip, an ear lobe, or a wrist, such as a human capillary-dense body tissue; the infrared light source 13 is a near-infrared light emitting tube for transmitting an optical signal including at least near-infrared light to the detecting portion 15 as In a preferred embodiment, the infrared light source 13 may include a plurality of near-infrared light-emitting tubes in a band of 800 _n m to 1500 nm, and the peak wavelength deviation of the near-infrared light-emitting tube is ±10 nm, and the radiation power is greater than 3 mW; the photosensor 14 is configured to receive the portion after passing through the detecting portion. The optical signal is converted into an electrical signal output. In a specific embodiment, the wavelength band received by the photosensor 14 can be set such that the optical signal received by the photosensor 14 is in the near-infrared band, specifically, the photoelectric sensor. 14 The received peak wavelength deviation is ±10nm, the photocurrent is greater than 10uA, and the peak wavelength deviation received by photosensor 14 is less than ± 10nm. It should be noted that, in a preferred embodiment, after the wavelength band received by the photosensor 14 is set, the infrared light source 13 may include optical signals of other wavelength bands, but the optical signal sent by the infrared light source 13 needs to meet at least the near Infrared light, used to emit near-infrared light to be irradiated on the part to be tested. The photosensor 14 is configured to acquire a pulse wave signal from a portion to be tested of a human body irradiated by near-infrared light. The pulse wave signal may be a photoelectric volume pulse wave signal, or may be a bio-impedance signal or a pressure sensing signal. In this embodiment, the pulse wave signal collected by the photosensor 14 is a photoplethysmographic pulse signal (PPG signal).

In the present embodiment, the PPG signal is acquired to select the earlobe or fingertip of the human body as a portion for extracting the PPG signal, and the earlobe or fingertip is placed on the detecting portion 15 of the signal collector 1. The blood of the fingertips and the earlobe is relatively abundant. With the periodic circulation of the heart, the photoelectric signal can be detected periodically by the photoelectric sensor. In order to obtain a stable photoplethysmographic pulse wave, the external influence factors need to be minimized or changed. For controllability, such as ambient temperature and humidity, in summary, the earlobe or fingertip is the most suitable site for extracting PPG signals. The photoelectric volume pulse wave is obtained by transmitting the human skin tissue through the near-infrared spectrum or by reflecting the human skin tissue.

[0031] Preferably, the support frame 10 is embedded with a micro power source 101 for providing working power for the signal collector 1 (for example, the infrared light source 13 and the photoelectric sensor 14). The upper surface of the upper detecting plate 11 is further provided with a power source At the gate 19, the power switch 19 is connected to the infrared light source 13 and the photosensor 14 through a power line for activating the infrared light source 13 to emit near-infrared light, or turning off the infrared light source 13 to stop emitting near-infrared light.

In a preferred embodiment, the non-invasive blood glucose detecting device 2 includes a signal processing circuit 21, a microprocessor 2, a memory 23, a display screen 24, and a communication unit 25. The signal processing circuit 21 is connected to the microprocessor 22 via a signal line, which is connected to the infrared light source 13 and the photosensor 14 of the signal collector 1 via a signal line, the memory 23, the display screen 24 and the communication Units 25 are each connected to microprocessor 22 via signal lines. The signal processing circuit 21 performs signal conversion, preamplification, and filtering on the pulse wave signal to obtain a digital signal of the pulse wave signal. Preferably, the signal processing circuit 21 can also provide operating power to the non-invasive blood glucose detecting device 2.

[0033] The microprocessor 22 may be a central processing unit (CPU), a microcontroller (MCU), a data processing chip, or an information processing unit having a data processing function. The memory 23 can be a read only memory unit ROM, an electrically erasable memory unit EEPROM, a flash memory unit FLASH or a solid state hard disk. The display screen 24 is a small-sized LCD or LED display unit that is embedded in the outer surface of the casing of the non-invasive blood glucose detecting device 2, since the measured blood sugar concentration value of the human body is displayed. The communication unit 25 is a wireless communication interface with remote wireless communication function, for example, supports communication such as GSM, GPRS, WCDMA, CDMA, TD-SCDMA, WiMAX, TD-LTE, FDD-LTE, 3G\4G\5G, etc. Technical communication interface.

Referring to FIG. 3, FIG. 3 is a functional block diagram of the non-invasive non-invasive blood glucose detecting system 20 of the present invention. In the present embodiment, the non-invasive non-invasive blood glucose detecting system 20 includes, but is not limited to, a signal acquiring module 201, a signal processing module 202, an initial detecting module 203, a section detecting module 204, a blood glucose determining module 205, and a blood glucose managing module 205. A module referred to in the present invention refers to a series of computer program instructions executable by the microprocessor 22 and capable of performing a fixed function, which are stored in the memory 23. The present embodiment will specifically describe the functions of the various modules in the non-invasive blood glucose detecting system 20 in conjunction with Figs. 4, 5 and 6.

[0035] Referring to FIG. 4, a flow chart of a preferred embodiment of the non-invasive blood glucose detecting method of the present invention. In the present embodiment, the non-invasive blood glucose detecting method of the present invention is applied to the above non-invasive blood glucose detecting device 01, and as shown in Fig. 1, Fig. 2 and Fig. 3, the non-invasive blood glucose detecting method comprises the following steps:

[0036] Step S31, when the user turns on the power of the signal collector 1 , the signal acquisition module 201 controls the infrared light source 13 to emit near-infrared light to be irradiated on the body to be tested, and passes the photoelectric sensor 14 from the near infrared. The pulse wave signal is acquired by the body to be tested under light irradiation. In this embodiment, the infrared light source 13 can emit a plurality of near-infrared light-emitting tubes in the 800 _n -1500 nm band, and the signal acquisition module 201 controls the infrared light source 13 to emit near-infrared light of different slopes and illuminate the part to be tested, and the photoelectric The sensor 14 acquires a pulse wave signal from a portion of the human body to be measured illuminated by human near-infrared light. The pulse wave signal referred to in this embodiment is preferably a photoplethysmographic pulse wave signal carrying blood glucose concentration information of the sample to be tested. For example, the glucose formula contains a plurality of 0-H, CH chemical bonds, and there are absorption peaks and absorption peaks and valleys in the 800 nm-1500 nm band, and the absorption peak wavelength is a key wavelength, which is the peak wavelength of blood glucose absorption to near-infrared light, which can reflect blood sugar. For the absorption of near-infrared light, the absorption peak-to-valley wavelength is used as the reference wavelength. The photoplethysmographic pulse wave generated by the critical wavelength contains not only the absorption information of the near-infrared light by the blood glucose, but also the absorption information of the near-infrared light by other substances in the blood. Modeling the reference wavelength and the critical wavelength together can effectively reduce the effects of other substances on near-infrared light absorption. In this embodiment, another important reason for selecting near-infrared light with a wavelength of less than 1500 nm is that these wavelengths are easily acquired, and some common near-infrared wavelengths, such as typical gallium arsenide diodes, can meet the demand and reduce non-invasive blood glucose. The cost of testing.

[0037] Step S32, the signal processing module 202 performs pre-processing on the acquired pulse wave signal by the signal processing circuit 21 to obtain a digital signal of the pulse wave signal. In the present embodiment, the signal processing module 202 performs signal processing such as signal conversion, preamplification, and filtering on the pulse wave signal through the signal processing circuit 21. Specifically, after filtering the pulse wave signal, the pulse wave signal can be filtered by the filter in the signal processing circuit 21, and the noise and DC components in the pulse wave signal can be removed, leaving the required AC component; The pulse wave signal is amplified and analog-digital converted, and the pulse wave signal can be amplified by the signal amplifier in the signal processing circuit 21, and the sampling is performed by using a 12-bit analog-to-digital converter (A DC) in the signal processing circuit 21, and the sampling frequency is used. For example, it may be 1 ΚΗζ to obtain a digital signal of the pulse wave signal, and the digital signal is sent to the non-invasive blood glucose detecting device 2 for subsequent processing.

[0038] Step S33, the signal processing module 202 extracts the feature value of the pulse wave signal from the digital signal of the pulse wave signal. In a specific embodiment, the pulse wave signal referred to in this embodiment carries the blood glucose concentration information of the sample to be tested. The characteristic value of the pulse wave signal may be a magnitude in a unit period of the pulse wave signal, or may be a ratio of a main wave peak to a main wave rising day and a secondary wave main wave relative height. In a preferred embodiment, please refer to FIG. 5. FIG. 5 is a waveform diagram of a pulse wave signal. In one waveform period, the characteristic value is the main peak P of the pulse wave signal, the main trough A, the secondary peak T, and the main peak amplitude hP of the secondary trough V, the main Valley amplitude hA, secondary peak amplitude hT, secondary valley amplitude hV, main peak-sub-valley interval tl, main peak-second peak inter-turn interval t2, main peak-main valley inter-tank interval t3, and adjacent main peak inter-turn interval T4. In a specific embodiment, the signal processing module 202 may use the wavelet transform method to extract the characteristic values of the pulse wave signal: the main peak amplitude hP, the main valley amplitude hA, the secondary peak amplitude hT, the secondary valley amplitude hV, the main peak - The inter-valley interval tl, the main peak-sub-peak interval t2, the main peak-main valley interval t3, and the adjacent main peak interval t4.

[0039] Step S34, the initial detection module 203 uses the predicted neural network to detect the characteristic value of the pulse wave signal to obtain an initial detection value of the blood sugar concentration. In a specific embodiment, the initial detection module 203 may perform a first detection on the acquired characteristic value of the pulse wave signal by predicting the neural network, thereby obtaining an initial detection value of the blood glucose concentration. In the specific implementation, the predictive neural network should be trained first. The training of the predicted neural network may be online or offline. In this embodiment, offline training is preferred, and the standard PPG eigenvalue training signal may be used. Training, For predictive neural networks, the input is the characteristic value of the photoplethysmographic pulse wave, and the corresponding invasively detected blood glucose concentration value (pre-acquired blood glucose sample value) is used as an output, and then the predicted neural network is trained using, for example, a MATLAB neural network.

[0040] Step S35: The interval detecting module 204 detects the characteristic value of the pulse wave signal by using the classification neural network to obtain a detection interval to which the blood sugar concentration belongs. In a specific embodiment, the interval detecting module 204 may perform a second detection on the acquired pulse wave signal through the classification neural network to obtain an interval in which the acquired pulse wave signal is located. Specifically, taking the human body as an example, the blood glucose concentration can be divided into sections [3, 4], [4, 5], [5, 6] ... [24, 25] according to a step size of 1, which will cover the human body. The blood glucose concentration range of 3 to 25 is divided into a plurality of sections. In the training of the classification neural network, the input is the characteristic value of the photoplethysmographic pulse wave, and the corresponding blood glucose concentration value is classified. For example, the blood glucose concentration value belongs to the interval [3, 4] and is recorded as the first category, belonging to the interval [4, 5 ] is recorded as the second category, which belongs to the interval [5, 6] as the third category, and so on as the output until the interval [3, 25] of all blood glucose values appearing during the training is included, and then used, for example. The MATLAB neural network trains the classification neural network.

[0041] In step S36, the blood glucose determining module 205 compares the initial detection value with the detection interval to determine whether the initial detection value belongs to the detection interval. If the initial detection value is in the detection interval, step S37 is performed, then the initial detection value is the blood glucose concentration value carried in the pulse wave signal, and the blood glucose determination module 205 displays the initial detection value on the display screen 24 as the human blood glucose concentration. value. If the initial detection value is not within the detection interval, step S38 is performed and then proceeds to step S32, that is, the blood glucose determination module 205 discards the initial detection value and the next The pulse wave signals are detected until the initial detection value is within the detection interval. For example, if the initial detection value obtained by using the predictive neural network is 4.6, and the detection interval obtained by the classification neural network is [4, 5], the initial detection value belongs to the detection interval; otherwise, if the classification neural network is used The obtained detection interval is [5, 6], [3, 4 domain [9, 10], etc., indicating that the initial detection value does not belong to the detection interval. If the initial detection value is not within the detection interval, if the initial detection value differs greatly from the actual blood glucose concentration value, the initial detection value is discarded and the next pulse wave signal is detected.

[0042] In this embodiment, the feature values of the PPG acquisition signals are extracted and sent to the prediction neural network and the classification neural network for the first detection and the second detection, respectively. The initial detection value (blood glucose value R1) is obtained by the first detection using the predictive neural network, and the detection interval (blood glucose interval R2) is obtained by the second detection using the classification neural network. Use R1 to judge the interval to which the blood glucose level belongs. If the interval where R1 belongs to R2, then R1 is considered correct, and the detection result R1 is retained as the blood glucose concentration value of the human body; otherwise, the detection result is considered incorrect, and the detection result R1 is discarded and the next pulse is discarded. The wave signal is detected until the detection result R1 belongs to the interval belonging to R2, and will be used as the blood glucose concentration value of the human body. Such repeated detection can effectively reduce the interference of other components (such as moisture) on the blood glucose concentration, and improve the accuracy of blood glucose concentration detection. And accuracy.

[0043] Step S39, the blood glucose management module 206 displays a segment selection box on the display screen 24 for the subject to select the blood glucose measurement segment, and receives the blood glucose measurement segment selected by the test subject. In this embodiment, the blood glucose measuring section includes two measuring segments, such as a fasting segment and two small meals after a meal. Since the blood glucose concentration measured by the fasting section of the human body is lower than the blood glucose concentration measured after the meal, and the blood glucose concentration measured after two hours of the meal is significantly higher, the selection box of the sacral segment provided by the embodiment is selected by the subject to be tested. The blood glucose measurement section can actually reflect the actual measurement of the blood glucose concentration of the person to be tested, thereby further ensuring the accuracy of the blood glucose concentration of the human body.

[0044] Step S40, the blood glucose management module 206 sends the human blood glucose concentration value and the blood glucose measurement segment of the test subject to the health management platform 02 through the communication unit 25 for the doctor to determine the blood sugar condition of the test subject. In this embodiment, the doctor can view the blood glucose concentration value of the test subject and the blood glucose measurement section through the health management platform 02, and can give the blood sugar condition of the test subject, for example, the blood sugar concentration is high, the blood sugar is low, and the blood sugar is low. The concentration is normal. In addition, the doctor can also give blood sugar management advice, such as dietary advice, lifestyle recommendations and exercise recommendations, to the person in charge of the test according to the blood glucose status of the test subject. Thereby, the health management platform 02 sends the blood sugar condition and the blood sugar management recommendation of the test subject to the test through the communication network 03. The communication device (such as a mobile phone) allows the person to be tested to understand their blood sugar status and manage their blood sugar status.

As shown in FIG. 6, FIG. 6 is a detailed flowchart of extracting the characteristic values of the pulse wave signals in step S33 in FIG. Specifically, the signal processing module 202 extracts the characteristic values of the pulse wave signal by means of wavelet transform, including the following steps:

[0046] Step S331, the signal processing module 202 performs wavelet transform on the acquired pulse wave signal to obtain a wavelet transform sequence. Before the wavelet transform, the obtained pulse wave signal (for example, PPG signal) may be denoised first, then the smoothed wavelet transform is performed on the denoised pure signal, and the wavelet transform sequence is obtained according to the obtained value after stationary wavelet transform.

[0047] Step S332, the signal processing module 202 searches for a modulus maxima corresponding to the preset threshold in the wavelet transform sequence according to the preset threshold. After obtaining the wavelet transform sequence, a suitable preset threshold may be determined to find a modulus maxima that meets a preset threshold. In this implementation, the modulus maxima in the wavelet transform sequence includes a positive modulus maxima, a negative The modulus maxima and associated submodule maxima.

[0048] Step S333, the signal processing module 202 extracts the feature value of the pulse wave signal according to the modulus maximum value. In this example, as shown in FIG. 4, the characteristic values of the pulse wave signal include a main peak amplitude hP, a main valley amplitude hA, a secondary peak amplitude hT, a secondary valley amplitude hV, and according to the main peak P, the main trough A The position of the secondary peak T and the secondary trough V is obtained as an eigenvalue main peak-sub-valley interval 11, a main peak-sub-peak inter-turn interval t2, a main peak-main valley inter-turn interval t3, and an adjacent main peak inter-turn interval t4.

[0049] The non-invasive blood glucose detecting system and method disclosed in the present embodiment obtains a blood glucose concentration by detecting a pulse wave signal obtained by using a predictive neural network to detect a pulse wave signal, and determining a pulse wave signal obtained by the classification neural network. Whether the initial detection value belongs to the detection interval, and when the initial detection value is within the detection interval, the initial detection value is a blood glucose concentration value, and by determining the interval to which the initial detection value belongs, the other components can effectively reduce the blood glucose concentration. The resulting interference improves the accuracy and accuracy of blood glucose concentration detection.

[0050] Those skilled in the art may understand that all or part of the steps of the various methods in the above embodiments may be completed by a program to instruct related hardware, and the program may be stored in a computer readable storage medium, and the storage medium may include: Memory, random access memory, disk or CD, etc.

The above is only a preferred embodiment of the present invention, and thus does not limit the scope of the patent of the present invention. The equivalent structure or equivalent flow of the invention and the equivalents of the drawings are directly or indirectly applied to other related technical fields, and are included in the scope of patent protection of the present invention.

Industrial applicability

Compared with the prior art, the non-invasive blood glucose detecting system and method of the present invention uses a pulse wave signal obtained by predicting a neural network to obtain a detection interval to which a blood glucose concentration belongs, and determines a pulse wave signal obtained by the classification neural network to obtain a blood glucose concentration. Whether the initial detection value belongs to the detection interval, and when the initial detection value is within the detection interval, the initial detection value is a blood glucose concentration value, and by determining the interval to which the initial detection value belongs, the other components in the human blood can be effectively reduced. The interference caused by blood glucose concentration improves the accuracy and accuracy of blood glucose concentration detection. In addition, the human blood glucose concentration value and the blood glucose measurement segment of the test subject are sent to the health management platform for the doctor to determine the blood glucose condition of the test subject.

Claims

Claim

[Claim 1] A non-invasive blood glucose detecting system, which is operated in a non-invasive blood glucose detecting device, the non-invasive blood glucose detecting device includes a signal collector and a non-invasive blood glucose detecting device, the signal collector includes an infrared light source and a photoelectric sensor, and the non-invasive The blood glucose detecting device comprises a signal processing circuit, a display screen and a communication unit, wherein the non-invasive blood glucose detecting system comprises: a signal acquiring module, configured to control the infrared light source to emit near-infrared light to be irradiated on the body to be tested, and pass the photoelectric sensor Obtaining a pulse wave signal from a part of the human body to be tested under near-infrared light; a signal processing module for pre-processing the acquired pulse wave signal by a signal processing circuit to obtain a digital signal of the pulse wave signal, and from the pulse wave signal The digital signal extracts the characteristic value of the pulse wave signal; the initial detection module is configured to detect the pulse wave signal characteristic value by using the prediction neural network to obtain the initial detection value of the blood glucose concentration; the interval detection module is configured to detect the pulse wave by using the classification neural network Signal characteristics The value obtains the detection interval of the blood glucose concentration; the blood glucose determination module is configured to determine whether the initial detection value belongs to the detection interval, and when the initial detection value is within the detection interval, the initial detection value is displayed on the display screen as the human blood glucose of the test subject Concentration value; blood glucose management module, used to display a segment selection box on the display screen for the person to be tested to select the blood glucose measurement segment, receive the blood glucose measurement segment selected by the test subject, and the human body blood glucose concentration value of the test subject and The blood glucose measurement section is sent to the health management platform through the communication unit for the doctor to determine the blood glucose condition of the test subject.

[Claim 2] The non-invasive blood glucose detecting system according to claim 1, wherein the blood glucose determining module is further configured to: when the initial detected value is not within the detecting interval, the non-invasive blood glucose detecting device discards the initial detected value and The next pulse wave signal is detected until the initial detection value is within the detection interval.

[Claim 3] The non-invasive blood glucose detecting system according to claim 1, wherein the signal processing circuit is configured to filter a pulse wave signal acquired by the photoelectric sensor to remove noise and a direct current component in the pulse wave signal. The desired AC component is left, and the pulse wave signal is amplified and analog-digital converted to obtain a digital signal of the pulse wave signal.

[Claim 4] The non-invasive blood glucose detecting system according to claim 1, wherein the signal processing module is specifically configured to perform wavelet transform on the acquired pulse wave signal to obtain a wavelet transform sequence. And searching for a modulus maxima corresponding to the preset threshold in the wavelet transform sequence according to the preset threshold, and extracting the feature value of the pulse wave signal according to the modulus maxima.

The non-invasive blood glucose detecting system according to any one of claims 1 to 4, wherein the infrared light source comprises at least a plurality of near-infrared light-emitting tubes in a band of 800 _n m to 1500 nm for emitting to a part to be tested of a human body. At least the near-infrared light in the 800 nm to 1500 nm band is included.

A non-invasive blood glucose detecting method is applied to a non-invasive blood glucose detecting device, the non-invasive blood glucose detecting device includes a signal collector and a non-invasive blood glucose detecting device, the signal collector includes an infrared light source and a photoelectric sensor, and the non-invasive blood glucose detecting device includes a signal The processing circuit, the display screen and the communication unit are characterized in that: the non-invasive blood glucose detecting method comprises the steps of: controlling the infrared light source to emit near-infrared light to be irradiated on the part to be tested of the human body, and detecting the human body under the illumination of the near-infrared light through the photoelectric sensor The pulse wave signal is obtained by the position; the digital signal of the pulse wave signal is obtained by pre-processing the acquired pulse wave signal by the signal processing circuit; the characteristic value of the pulse wave signal is extracted from the digital signal of the pulse wave signal; and the predictive neural network is used for detecting The characteristic value of the pulse wave signal is used to obtain the initial detection value of the blood glucose concentration; the classification value of the pulse wave signal is detected by the classification neural network to obtain the detection interval of the blood glucose concentration; whether the initial detection value belongs to the detection interval, and the initial detection value is in the detection interval.吋, the initial detection value is displayed on the display screen as the blood glucose concentration value of the human body to be tested; a segment selection box is displayed on the display screen for the person to be tested to select the blood glucose measurement section, and receives the blood glucose measurement selected by the test subject The human blood glucose concentration value and the blood glucose measurement section of the test subject are sent to the health management platform through the communication unit for the doctor to determine the blood sugar condition of the test subject.

The non-invasive blood glucose detecting method according to claim 6, wherein the method further comprises the step of: discarding the initial detected value and detecting the next pulse wave signal until the initial detected value if the initial detected value is not within the detection interval; Until the detection interval.

The non-invasive blood glucose detecting method according to claim 6, wherein the step of pre-processing the acquired pulse wave signal by the signal processing circuit to obtain a digital signal of the pulse wave signal comprises the following steps: Filtering removes noise and DC components in the pulse wave signal, leaving the desired AC component; amplifying and analog-to-digital converting the pulse wave signal to obtain a digital signal of the pulse wave signal.

[Claim 9] The non-invasive blood glucose detecting method according to claim 6, wherein the step of extracting the characteristic value of the pulse wave signal from the digital signal of the pulse wave signal comprises the steps of: acquiring the pulse wave The signal is wavelet transformed to obtain a wavelet transform sequence; the modulus maximum value corresponding to the preset threshold is searched in the wavelet transform sequence according to the preset threshold; and the characteristic value of the pulse wave signal is extracted according to the modulus maximum value.

The non-invasive blood glucose detecting method according to any one of claims 6 to 9, wherein the method further comprises the step of: the infrared light source comprising at least a plurality of near infrared rays in a band of 800 _n m to 1500 nm The light-emitting tube is configured to emit near-infrared light including at least 800 _n m-1500 nm into the body to be tested.