CN115530816A - Non-invasive blood glucose monitoring device and method based on ballistocardiography - Google Patents

Abstract

The invention relates to a noninvasive blood glucose monitoring device and method based on a ballistocardiogram, which comprises a photoelectric transceiving module, a sensing module, an MCU and a terminal, wherein the photoelectric transceiving module is connected with the sensing module; the photoelectric transceiving module comprises a light source and a photoelectric detector; the light emitted by the light source is incident to one port of the sensing module through the optical fiber or the optical fiber connector; the light output by the sensing module reaches the photoelectric detector from the other port through an optical fiber or an optical fiber connector; the photoelectric detector transmits data to the MCU; and the MCU is connected with the terminal through the wireless module. According to the invention, the Ballistocardiogram (BCG) signals are acquired through the optical fiber sensor, and the deep learning algorithm analysis is combined, so that the continuous blood sugar value of the human body is obtained, and the noninvasive, non-sensitive and continuous blood sugar level monitoring is realized.

Description

Translated fromChinese

基于心冲击图的无创血糖监测装置及方法Non-invasive blood glucose monitoring device and method based on ballistocardiography

技术领域technical field

本发明涉及血糖监测仪领域，具体涉及一种基于心冲击图的无创血糖监测装置及方法。The invention relates to the field of blood glucose monitors, in particular to a non-invasive blood glucose monitoring device and method based on ballistocardiography.

背景技术Background technique

心电图（ECG）是通过电流信号来记录人体心脏每次心动的电位变化，心脏产生的电流大小变化因人而异，因此从心电图的特征中可反映出个人心脏的健康状况。根据发表的论文（Tobore I, Li J, Kandwal A, et al. Statistical and spectral analysis ofECG signal towards achieving non-invasive blood glucose monitoring[J]. BMCMed Inform Decis Mak,19, 266, 2019.），ECG信号也可以反应出人体血糖的信息。如当低血糖时，反映在ECG信号上的是QT间隔延长，同时可观察到心率增加。然而，ECG信号的采集需要对身体多个部位进行贴片，与人体皮肤直接接触并且需要专业的医护人员来操作，所以不利于患者血糖的无感、实时监测。An electrocardiogram (ECG) uses current signals to record the potential changes of each heartbeat of the human heart. The changes in the magnitude of the current produced by the heart vary from person to person, so the characteristics of the ECG can reflect the health status of the individual heart. According to the published paper (Tobore I, Li J, Kandwal A, et al. Statistical and spectral analysis of ECG signal towards achieving non-invasive blood glucose monitoring[J]. BMCMed Inform Decis Mak,19, 266, 2019.), ECG signal It can also reflect the information of human blood sugar. For example, when hypoglycemia occurs, the QT interval is prolonged as reflected in the ECG signal, and an increase in heart rate can be observed at the same time. However, the collection of ECG signals requires patches on multiple parts of the body, direct contact with human skin and requires professional medical staff to operate, so it is not conducive to the non-sensing and real-time monitoring of blood sugar of patients.

心冲击图（BCG）是一种描述心脏在跳动过程中对人体表面产生力作用的方法。由于心脏在正常工作时血液会随着其收缩与舒张对血管造成压力，随着心脏的跳动过程，血液在流动时引起人体质心的改变使人体表面发生了微小的运动。血糖的变化会影响这些运动。如图1所示，在这段BCG信号中存在H、I、J、K、L、M和N波。一段标准的BCG信号形状与ECG信号形状大体一致，如J波与ECG信号中的R波对应，但在具体波形细节会有一些不同。一段稳定的BCG信号中包含了大量的心脏健康信号以及其他生命体征信号包括血糖信息。Ballistocardiography (BCG) is a method of describing the force that the heart exerts on the surface of the body during its beating. When the heart is working normally, the blood will cause pressure on the blood vessels with its contraction and relaxation. With the beating process of the heart, the blood will cause changes in the center of mass of the human body when it flows, causing tiny movements on the surface of the human body. Changes in blood sugar can affect these exercises. As shown in Figure 1, there are H, I, J, K, L, M, and N waves in this segment of the BCG signal. The shape of a standard BCG signal is generally consistent with the shape of the ECG signal, such as the J wave corresponds to the R wave in the ECG signal, but there are some differences in the details of the specific waveform. A stable BCG signal contains a large number of heart health signals and other vital signs including blood sugar information.

目前心电信号检测设备体积较大，操作比较复杂；可穿戴设备尽管操作简单但是不能做到无感；并且基于BCG信号的装置的应用只限于心跳与呼吸信号的采集和监测，尚未探究完全由光纤传感器记录的BCG波的特征与血糖的相关应用。At present, the ECG signal detection equipment is large in size and the operation is relatively complicated; although the wearable device is simple to operate, it cannot be non-inductive; and the application of devices based on BCG signals is limited to the collection and monitoring of heartbeat and respiratory signals, and has not yet been explored. Characterization of BCG waves recorded by fiber optic sensors and blood glucose related applications.

发明内容Contents of the invention

有鉴于此，本发明的目的在于提供一种基于心冲击图的无创、无感血糖监测装置及方法，通过光纤传感器采集BCG信号，结合深度学习算法分析，获得了人体连续的血糖值，实现无创、无感和连续的血糖水平监测。In view of this, the purpose of the present invention is to provide a non-invasive, non-inductive blood sugar monitoring device and method based on ballistocardiography, which collects BCG signals through optical fiber sensors, and combines deep learning algorithm analysis to obtain continuous blood sugar levels of the human body, realizing non-invasive , sensorless and continuous blood glucose level monitoring.

为实现上述目的，本发明采用如下技术方案：To achieve the above object, the present invention adopts the following technical solutions:

一种基于心冲击图的无创血糖监测装置，包括光电收发模块、传感模块、MCU以及终端；所述光电收发模块包括光源和光电探测器；所述光源发射的光通过光纤或光纤连接器入射到传感模块一个端口；所述传感模块输出的光从另一个端口通过光纤或光纤连接器到达光电探测器；所述光电探测器将数据传送至MCU；所述MCU通过无线模块与终端连接。A non-invasive blood glucose monitoring device based on ballistocardiography, including a photoelectric transceiver module, a sensing module, an MCU, and a terminal; the photoelectric transceiver module includes a light source and a photodetector; the light emitted by the light source is incident through an optical fiber or an optical fiber connector to one port of the sensing module; the light output by the sensing module reaches the photodetector through the optical fiber or fiber optic connector from the other port; the photodetector transmits the data to the MCU; the MCU is connected to the terminal through the wireless module .

进一步的，所述传感模块包括两个端口和一个模式干涉光纤组件。Further, the sensing module includes two ports and a mode interference fiber optic assembly.

进一步的，所述光源采用激光光源或发光二极管。Further, the light source is a laser light source or a light emitting diode.

进一步的，所述传感模块采用具有近周期滤波光谱特性的光纤干涉仪或弯曲构件产生的光谱特性器件。Further, the sensing module adopts an optical fiber interferometer with near-period filtering spectral characteristics or a spectral characteristic device produced by a bending member.

进一步的，所述光电探测器采用PIN光电二极管探测器、APD光电探测器或单光子光电探测器。Further, the photodetector adopts PIN photodiode detector, APD photodetector or single photon photodetector.

一种基于心冲击图的无创血糖监测装置的血糖监测方法，包括以下步骤：A blood glucose monitoring method based on a ballistocardiogram-based noninvasive blood glucose monitoring device, comprising the following steps:

步骤S1:待测人员躺或背靠在传感模块上，心脏搏动会对传感模块产生压力，使得光在传输过程中的能量发生损耗变化；Step S1: The person to be tested lies or leans against the sensing module, and the heartbeat will generate pressure on the sensing module, causing the energy loss of light during transmission to change;

步骤S2:光电探测器中接收到带有患者特征信息的光信号,并传送至MCU；Step S2: The light signal with patient characteristic information is received in the photodetector and sent to the MCU;

步骤S3：在MCU中，光电流经过放大、滤波、模数转换，计算处理后，通过无线模块传输至终端，得到心冲击图信号；Step S3: In the MCU, the photocurrent is amplified, filtered, converted from analog to digital, calculated and processed, and transmitted to the terminal through the wireless module to obtain a shock cardiogram signal;

步骤S4:结合深度学习算法进行建模分析，获得相应的血糖值。Step S4: Perform modeling analysis in combination with deep learning algorithms to obtain corresponding blood glucose values.

进一步的，所述步骤S4具体为：Further, the step S4 is specifically:

步骤S41:对基于心冲击图信号进行去基线漂移、去工频干扰的预处理工作；Step S41: Carry out the preprocessing work of removing baseline drift and removing power frequency interference based on the ballistocardiogram signal;

步骤S42:对预处理后的信号进行分段处理，获得每一分段预处理信号的时域图像特征；Step S42: Carry out segment processing to the preprocessed signal, obtain the time-domain image feature of each segment preprocessing signal;

步骤S43:根据时域图像特征，结合深度学习算法进行建模的方式获取不同波形图下对应的血糖值，从心冲击图中提取出患者的血糖值信号；Step S43: According to the time-domain image features, in combination with the deep learning algorithm for modeling, the corresponding blood glucose values under different waveform diagrams are obtained, and the patient's blood glucose level signal is extracted from the shockcardiogram;

步骤S44:在终端软件上显示出来患者的具体血糖值，并依据血糖值的大小可分为三类，即低血糖、正常或高血糖。Step S44: The patient's specific blood sugar level is displayed on the terminal software, and can be divided into three categories according to the blood sugar level, namely hypoglycemia, normal or hyperglycemia.

进一步的，所述时域图像特征由9个直接连接不同基准点的基准距离及其斜率组成的19个特征,即HI波段长度、HI波段斜率、HJ波段长度、HJ波段斜率、HK波段长度、HK波段斜率、HL波段长度、HL波段斜率、IJ波段长度、IJ波段斜率、IK波段长度、IK波段斜率、IL波段长度、IL波段斜率、JK波段长度、JK波段斜率、JL波段长度、JL波段斜率;依据两相邻JJ波段的距离作为另一个判断特征。Further, the time-domain image features are 19 features consisting of 9 reference distances directly connected to different reference points and their slopes, namely HI band length, HI band slope, HJ band length, HJ band slope, HK band length, HK Band Slope, HL Band Length, HL Band Slope, IJ Band Length, IJ Band Slope, IK Band Length, IK Band Slope, IL Band Length, IL Band Slope, JK Band Length, JK Band Slope, JL Band Length, JL Band Slope; according to the distance between two adjacent JJ bands as another judgment feature.

并利用深度学习算法对目标数据进行建模分析，进而从预处理的心冲击信号中得到患者的血糖信息，如下公式所示：And use the deep learning algorithm to model and analyze the target data, and then obtain the patient's blood sugar information from the preprocessed cardiac shock signal, as shown in the following formula:

BS=X1*W1+X2*W2+...+X19*W19BS=X1*W1+X2*W2+...+X19*W19

X1到X19是心冲击图19个特征值，W1到W19是需要在学习过程中调整的参数，BS是输出的血糖值。X1 to X19 are the 19 eigenvalues of the shockcardiogram, W1 to W19 are the parameters that need to be adjusted during the learning process, and BS is the output blood glucose value.

本发明与现有技术相比具有以下有益效果：Compared with the prior art, the present invention has the following beneficial effects:

本发明通过光纤传感器采集BCG信号，结合深度学习算法分析，获得了人体连续的血糖值，实现无创、无感和连续的血糖水平监测。The present invention collects BCG signals through an optical fiber sensor, combines with deep learning algorithm analysis, obtains continuous blood sugar levels of the human body, and realizes non-invasive, non-inductive and continuous blood sugar level monitoring.

附图说明Description of drawings

图1为本发明一实施例在线性工作区解调的心冲击图示意图；Fig. 1 is a schematic diagram of the ballistocardiogram demodulated in the linear working area according to an embodiment of the present invention;

图2为本发明装置结构示意图；Fig. 2 is the structural representation of device of the present invention;

图3为本发明光纤组件示意图；3 is a schematic diagram of an optical fiber assembly of the present invention;

图4为本发明一实施例中的心冲击图；Fig. 4 is the shock diagram in an embodiment of the present invention;

图5为本发明一实施例中的心冲击图特征提取示意图；Fig. 5 is a schematic diagram of feature extraction of ballistocardiogram in an embodiment of the present invention;

图6为本发明一实施例中的数据处理流程示意图；6 is a schematic diagram of a data processing flow in an embodiment of the present invention;

图7为本发明方法流程示意图；Fig. 7 is a schematic flow chart of the method of the present invention;

图8为本发明一实施例中与医院糖耐量测试结果比较结果示意图。Fig. 8 is a schematic diagram of the results compared with the results of the glucose tolerance test in a hospital in an embodiment of the present invention.

具体实施方式detailed description

下面结合附图及实施例对本发明做进一步说明。The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

请参照图1-8，本发明提供一种基于心冲击图的无创血糖监测装置，包括光电收发模块、传感模块、MCU以及终端；所述光电收发模块包括光源和光电探测器；光源LD发射的光通过光纤或光纤连接器11入射到传感模块5端口1；入射到传感模块5端口1 的光经过传感回路13，34，24再从端口2输出；该输出的光经过端口2或者光纤连接器22，到达光电探测器PD。传感模块5有2个端口1和2，一个模式干涉光纤组件，通过模式干涉或者其他方法形成干涉滤波光谱。当光束通过光电探测器后，变成光电流；在MCU中，光电流经过放大、滤波、模数转换，计算等一系列处理后通过无线比如WiFi、蓝牙等传输到终端。Please refer to Figures 1-8, the present invention provides a non-invasive blood glucose monitoring device based on ballistocardiography, including a photoelectric transceiver module, a sensing module, an MCU, and a terminal; the photoelectric transceiver module includes a light source and a photodetector; the light source LD emits The light incident on thesensor module 5 port 1 through the optical fiber or fiberoptic connector 11; the light incident on thesensor module 5 port 1 passes through thesensing circuit 13, 34, 24 and then output from theport 2; the output light passes through theport 2 Or fiberoptic connector 22, to the photodetector PD. Thesensing module 5 has twoports 1 and 2, a mode interference optical fiber assembly, and forms an interference filter spectrum through mode interference or other methods. When the light beam passes through the photodetector, it becomes a photocurrent; in the MCU, the photocurrent is amplified, filtered, analog-to-digital converted, and calculated, and then transmitted to the terminal through wireless such as WiFi and Bluetooth.

优选的，传感模块包括两个端口和一个模式干涉光纤组件。光纤组件中的光纤绕在一圆柱体上。圆柱体可以是金属或塑料。Preferably, the sensing module includes two ports and a mode interference fiber optic assembly. The optical fiber in the fiber optic assembly is wound on a cylinder. The cylinder can be metal or plastic.

优选的，光源采用激光光源或发光二极管。Preferably, the light source is a laser light source or a light emitting diode.

优选的，传感模块采用具有近周期滤波光谱特性的光纤干涉仪或弯曲构件产生的光谱特性器件。Preferably, the sensing module adopts an optical fiber interferometer with near-period filtering spectral characteristics or a spectral characteristic device produced by a bending member.

优选的，光电探测器采用PIN光电二极管探测器、APD光电探测器或单光子光电探测器。Preferably, the photodetector is a PIN photodiode detector, an APD photodetector or a single photon photodetector.

优选的，MCU微控制器处理用来对PIN光电探测器PD做电放大、滤波、模数转换和计算等一系列处理，蓝牙传输到终端上位机。上位机软件将接收到的结果进行应用方面的各种处理、分析、显示和报警。Preferably, the MCU microcontroller is used to perform a series of processing such as electrical amplification, filtering, analog-to-digital conversion and calculation on the PIN photodetector PD, and the bluetooth is transmitted to the terminal host computer. The upper computer software performs various processing, analysis, display and alarm on the received results.

在本实施例中，该装置的实施步骤如图6所示，包括：In this embodiment, the implementation steps of the device are shown in Figure 6, including:

（1）采集待监测的BCG信号；当患者躺或背靠在传感模块5上时，心脏搏动会对传感模块产生压力，使得光在传输过程中的能量发生损耗变化，从而在光电探测器中可以接收到带有患者特征信息的光信号；(1) Collect the BCG signal to be monitored; when the patient is lying or leaning against thesensing module 5, the heartbeat will exert pressure on the sensing module, causing the energy loss of the light during transmission to change, so that in the photoelectric detection The optical signal with patient characteristic information can be received in the device;

（2）对待监测的BCG信号进行滤波处理，去除待监测的BCG信号的干扰部分；从心冲击图信号中获取目标数据，首先进行预处理环节，即进行去基线漂移、去工频干扰的预处理工作。然后对所述预处理信号进行分段处理，之后获得每一分段预处理信号的时域图像特征，如图3所示。从图中可以看到BCG信号存在H、I、J、K、L、M和N波。(2) Perform filtering processing on the BCG signal to be monitored to remove the interference part of the BCG signal to be monitored; to obtain the target data from the ballistocardiogram signal, first perform the preprocessing link, that is, perform preprocessing to remove baseline drift and power frequency interference. Process work. Then segment the preprocessed signal, and then obtain the time-domain image features of each segmented preprocessed signal, as shown in FIG. 3 . It can be seen from the figure that there are H, I, J, K, L, M and N waves in the BCG signal.

（3）数据处理过程如图4-5所示，从心冲击图中经筛选后可获得目标数据，对所述目标数据进行分段处理，之后获得每一分段预处理信号的时域图像特征，如HI波段长度、HI波段斜率、HJ波段长度、HJ波段斜率、HK波段长度、HK波段斜率、HL波段长度、HL波段斜率、IJ波段长度、IJ波段斜率、IK波段长度、IK波段斜率、IL波段长度、IL波段斜率、JK波段长度、JK波段斜率、JL波段长度、JL波段斜率，以及相邻心冲击图的JJ间隔。然后利用深度学习算法对目标数据进行建模分析，进而从预处理的心冲击信号中得到患者的血糖信息，如下公式所示：(3) The data processing process is shown in Figure 4-5. The target data can be obtained after screening from the shock-cardiogram, and the target data is segmented, and then the time domain image of each segmental preprocessing signal is obtained. Features such as HI Band Length, HI Band Slope, HJ Band Length, HJ Band Slope, HK Band Length, HK Band Slope, HL Band Length, HL Band Slope, IJ Band Length, IJ Band Slope, IK Band Length, IK Band Slope , IL-band length, IL-band slope, JK-band length, JK-band slope, JL-band length, JL-band slope, and the JJ interval of adjacent ballistocardiograms. Then use the deep learning algorithm to model and analyze the target data, and then obtain the patient's blood sugar information from the preprocessed cardiac shock signal, as shown in the following formula:

BS=X1*W1+X2*W2+...+X19*W19BS=X1*W1+X2*W2+...+X19*W19

X1到X19是心冲击图19个特征值，W1到W19是需要在学习过程中调整的参数，BS是输出的血糖值。X1 to X19 are the 19 eigenvalues of the shockcardiogram, W1 to W19 are the parameters that need to be adjusted during the learning process, and BS is the output blood glucose value.

（4）将所提取出的所有信号特征输入到预先训练好的分类模型中；终端软件显示血糖值，并依据血糖值对血糖水平进行分类。(4) Input all the extracted signal features into the pre-trained classification model; the terminal software displays the blood sugar level, and classifies the blood sugar level according to the blood sugar level.

采用上述方法，与医院糖耐量测试结果比较，采用本方法的预测值与医院检查结果比较，误差在允许标准内，如图7所示。Using the above method, compared with the results of the glucose tolerance test in the hospital, and comparing the predicted value using this method with the results of the hospital examination, the error is within the allowable standard, as shown in Figure 7.

以上所述仅为本发明的较佳实施例，凡依本发明申请专利范围所做的均等变化与修饰，皆应属本发明的涵盖范围。The above descriptions are only preferred embodiments of the present invention, and all equivalent changes and modifications made according to the scope of the patent application of the present invention shall fall within the scope of the present invention.

Claims (8)

1. A noninvasive blood glucose monitoring device based on a ballistocardiogram is characterized by comprising a photoelectric transceiving module, a sensing module, an MCU and a terminal; the photoelectric transceiving module comprises a light source and a photoelectric detector; the light emitted by the light source is incident to one port of the sensing module through the optical fiber or the optical fiber connector; the light output by the sensing module reaches the photoelectric detector from the other port through an optical fiber or an optical fiber connector; the photoelectric detector transmits data to the MCU; and the MCU is connected with the terminal through the wireless module.

2. The ballistocardiogram-based noninvasive blood glucose monitoring device of claim 1, wherein the sensing module comprises two ports and a modal interference fiber optic assembly.

3. The ballistocardiogram-based noninvasive blood glucose monitoring device of claim 1, wherein the light source is a laser light source or a light emitting diode.

4. The ballistocardiogram-based noninvasive blood glucose monitoring device of claim 1, wherein the sensing module employs a spectral feature device produced by a fiber optic interferometer or a flexure member with a near-periodic filtered spectral feature.

5. The ballistocardiogram-based noninvasive blood glucose monitoring device of claim 1, wherein the photodetector is a PIN photodiode detector, an APD photodetector, or a single photon photodetector.

6. The method for blood glucose monitoring of a ballistocardiogram-based noninvasive blood glucose monitoring device of any one of claims 1-5, comprising the steps of:

s1, when a person to be detected lies or leans against a sensing module, the heart beat can generate pressure on the sensing module, so that the energy of light in the transmission process is lost and changed;

s2, receiving an optical signal with patient characteristic information in the photoelectric detector and transmitting the optical signal to the MCU;

and step S3: in the MCU, the photocurrent is amplified, filtered, subjected to analog-to-digital conversion, calculated and processed, and then transmitted to a terminal through a wireless module to obtain a ballistocardiogram signal;

and S4, modeling analysis is carried out by combining a deep learning algorithm to obtain the corresponding blood sugar value.

7. The ballistocardiogram-based noninvasive blood glucose monitoring method of claim 6, wherein the step S4 is specifically:

s41, preprocessing work of removing baseline drift and power frequency interference is carried out on the ballistocardiogram-based signal;

s42, carrying out segmentation processing on the preprocessed signals to obtain the time domain image characteristics of each segmented preprocessed signal;

s43, obtaining corresponding blood sugar values under different oscillograms by combining a deep learning algorithm for modeling according to the time domain image characteristics, and extracting blood sugar value signals of the patient from the ballistocardiogram;

and S44, displaying the specific blood sugar value of the patient on the terminal software, and dividing the blood sugar value into three categories, namely hypoglycemia, normal blood sugar or hyperglycemia according to the size of the blood sugar value.

8. The ballistocardiogram-based noninvasive blood glucose monitoring method of claim 7, wherein the time-domain image features are 19 features consisting of 9 reference distances and slopes thereof directly connecting different reference points, namely HI band length, HI band slope, HJ band length, HJ band slope, HK band length, HK band slope, HL band length, HL band slope, IJ band length, IJ band slope, IK band length, IK band slope, IL band length, IL band slope, JK band length, JL band slope, and JL band slope;

and modeling and analyzing the target data by utilizing a deep learning algorithm, and further obtaining the blood sugar information of the patient from the preprocessed heart attack signal, wherein the formula is as follows:

BS=X1*W1+X2*W2+...+X19*W19

x1 to X19 are the values of 19 features of the ballistocardiogram, W1 to W19 are the parameters to be adjusted in the learning process, and BS is the output blood glucose value.

Images