Health condition monitoring system and preparation method thereof

Abstract

The invention belongs to the technical field of intelligent monitoring, and particularly relates to a health condition monitoring system and a preparation method thereof. The system comprises a health condition monitoring mask and an upper computer, the health condition monitoring mask comprises a mask and a respiration monitoring unit, a heartbeat monitoring unit and a health condition monitoring unit which are arranged in the mask, and the upper computer comprises a display interface and a data processing module which is arranged in the display interface. The respiration monitoring unit comprises thermoelectric devices prepared from P-type and N-type semiconductor thermoelectric particles, the heartbeat monitoring unit is composed of MAX30102, a voltage stabilizing circuit, a decoupling capacitor and a pull-up resistor which are integrated on a flexible PCB, and the health condition monitoring unit is sewn in a mask and comprises an ADC conversion unit I, a microcontroller chip with a Bluetooth module and an integrated circuit bus. The upper computer comprises a display interface and a data processing module which are used for receiving monitoring data of the respiration monitoring unit and the heartbeat monitoring unit, matching the data of the respiration monitoring unit and the heartbeat monitoring unit into groups for multi-source fusion, and identifying and judging the health condition of a human body according to the data after multi-source fusion.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent monitoring, and in particular relates to a health status monitoring system and a preparation method thereof. Background Art

[0002] Heart rate and respiratory rate are two important physiological parameters for human health monitoring. At present, many people are already in sub-health due to stress, staying up late, irregular diet, etc. In addition to regular physical examinations, daily real-time monitoring of heart rate and respiration is also required for judgment and early warning of human health.

[0003] At present, there are many shortcomings in the monitoring equipment for heart rate and respiration. Usually, some small portable devices do not have the function of monitoring heart rate and respiration at the same time. And some devices that can monitor heart rate and respiration at the same time are often expensive, not easy to carry, and not suitable for daily real-time monitoring. At present, with the improvement of people's living standards and the emphasis on health, a low-cost, portable and easy-to-operate device is needed for daily monitoring of heart rate and respiration. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a health status monitoring system and a preparation method thereof.

[0005] A health status monitoring system comprises a health status monitoring mask and a host computer. The health status monitoring mask comprises the mask and a breathing monitoring unit, a heartbeat monitoring unit and a health status monitoring unit arranged therein. The host computer comprises a display interface and a data processing module built into the host computer.

[0006] The health monitoring unit is sewn into the mask, including ADC conversion unit I, a microcontroller chip with a built-in Bluetooth module and an integrated circuit bus (I2C), and the ADC conversion unit I includes channel 0 and channel 1;

[0007] The respiratory monitoring unit includes a thermoelectric device made of P-type semiconductor thermoelectric particles and N-type semiconductor thermoelectric particles. The external positive and negative wires of the thermoelectric device are respectively connected to channel 0 and channel 1 of the health monitoring unit. Channel 0 and channel 1 form a differential signal, which is transmitted to ADC conversion unit I. ADC conversion unit I converts the differential signal into a digital signal, which is transmitted to a microcontroller chip (model: ESP32) in the health monitoring unit, and then uploaded to the host computer in real time through the Bluetooth module of the microcontroller chip.

[0008] The heartbeat monitoring unit is composed of MAX30102, voltage stabilizing circuit, decoupling capacitor and pull-up resistor integrated on the flexible PCB as the substrate. MAX30102 converts the collected optical signal into an electrical signal through its own photodiode (PD), and the electrical signal is converted into a digital signal through its own ADC. The digital signal is transmitted to the microcontroller chip in the health monitoring unit through the integrated circuit bus (I2C) of the microcontroller chip (model: ESP32), and then uploaded to the host computer in real time through the Bluetooth module of the microcontroller chip.

[0009] The data processing module included in the host computer receives the data transmitted by the respiratory monitoring unit and the heartbeat monitoring unit, matches the data of the two into groups for multi-source fusion, and identifies and judges the real-time health status of the human body based on the multi-source fused data.

[0010] Preferably, the data processing module includes a data recording module, a data preprocessing module, a data training module, and a data identification module which are sequentially signal-connected; the data recording module collects and stores historical data through a respiratory monitoring unit and a heartbeat monitoring unit; the data preprocessing module preprocesses the data collected by the data recording module; in the data training module, an XGBoost framework is used to establish an XGBoost model, the preprocessed data is used to train the XGBoost model, the hyperparameters of the model are adjusted and optimized, and an optimized XGBoost model is obtained; the optimized XGBoost model is loaded into the data identification module to identify the health status of the human body.

[0011] Preferably, in the data preprocessing module, the specific method of preprocessing is: matching the four types of data into groups for multi-source fusion, the first type of historical data is multi-source fused and defined as positive, the second, third, and fourth types of historical data are multi-source fused and defined as negative; the data processed as above are divided into a training set and a test set according to a ratio of 9:1, with the training set being 90% and the test set being 10%; wherein the four types of data are matched into groups for multi-source fusion respectively, which means: matching the voltage waveform data reflecting the phrase transmitted by the respiratory monitoring unit with the waveform data reflecting the heartbeat transmitted by the heartbeat monitoring unit, and dividing them into Four types: the first type: when a person says an abnormal phrase such as "help", the voltage waveform reflecting the abnormal phrase is sent by the respiratory monitoring unit to the data recording module, and the corresponding waveform reflecting the abnormal heartbeat is sent by the heartbeat monitoring unit to the data recording module when the person is in a dangerous state such as falling and losing weight at that moment. The second type: when a person says an abnormal phrase such as "help", the voltage waveform reflecting the abnormal phrase is sent by the respiratory monitoring unit to the data recording module, and the corresponding waveform reflecting the normal heartbeat is sent by the heartbeat monitoring unit to the data recording module when the person is in a safe state such as sitting, walking, jogging, etc. The third type: when a person says "how The fourth type: when a person says normal phrases such as "how are you", "yes", "no", etc., the voltage waveform reflecting the normal phrases sent by the respiratory monitoring unit to the data recording module and the corresponding waveform reflecting the abnormal heartbeat sent by the heartbeat monitoring unit to the data recording module when the person is in a dangerous state such as falling and losing weight at that moment. The fourth type: when a person says normal phrases such as "how are you", "yes", "no", etc., the voltage waveform reflecting the normal phrases sent by the respiratory monitoring unit to the data recording module and the corresponding waveform reflecting the normal heartbeat sent by the heartbeat monitoring unit to the data recording module when the person is in a safe state such as sitting, walking, jogging, etc. at that moment.

[0012] The present invention also provides a method for preparing the above health status monitoring system, comprising the following steps:

[0013] (1.1) Mask processing

[0014] A commercial cotton mask was selected as the flexible substrate, and a hole array was prepared on one side of the mask corresponding to the area between the mouth and nostrils by laser cutting. The holes in the hole array were 1.4 mm long, 1.4 mm wide, and the spacing between them was 2.3 mm.

[0015] (1.2) Construction of health status monitoring unit

[0016] The health status monitoring unit is sewn inside the mask. The health status monitoring unit includes ADC conversion unit I, a microcontroller chip with built-in Bluetooth module and integrated circuit bus, and ADC conversion unit I includes channel 0 and channel 1.

[0017] (2) Preparation of respiratory monitoring unit

[0018] The respiratory monitoring unit includes a thermoelectric device, which includes P-type semiconductor thermoelectric particles and N-type semiconductor thermoelectric particles. The specific preparation process is as follows:

[0019] (2.1) Preparation of P-type and N-type semiconductor thermoelectric particles: Sb2Te3 is used as the raw material of P-type semiconductor thermoelectric particles, and Bi2Te3 is used as the raw material of N-type semiconductor thermoelectric particles. Both P-type and N-type semiconductor thermoelectric particles are formed into blocks, and their dimensions are 1.4 mm in length, 1.4 mm in width, and 2.5 mm in height. The spacing between N-type semiconductor thermoelectric particles and P-type semiconductor thermoelectric particles is 2.3 mm.

[0020] (2.2) Preparation of electrodes: The conductive fabric was cut into rectangular pieces as electrodes by laser cutting. The size of the electrodes was length × width = 4.9 mm × 1.4 mm.

[0021] (2.3) placing the electrode prepared in step 2.2 on a metal steel mesh, then transferring it to a thermal release tape, and coating the electrode with tin paste;

[0022] (2.4) Lay the mask obtained in step 1 flat on a heating plate, then alternately embed the P-type and N-type semiconductor thermoelectric particles obtained in step 2.1 into the holes of the hole array prepared in step 1, and place the electrodes prepared in step 2.3 on the outer ends of the semiconductor thermoelectric particles. After placement, press a glass plate on it to prevent the P-type and N-type semiconductor thermoelectric particles from moving during heating; adjust the heating temperature of the heating plate to 200° C., and weld at 200° C. for 10 minutes to complete welding; after welding is completed, naturally cool to room temperature and then remove the glass plate;

[0023] (2.5) Turning over the mask after completing step 2.4, and laying it flat on the heating plate again; then, placing the electrode prepared in step 2.3 on the unwelded end of the semiconductor thermoelectric particle, and after placement, pressing a glass plate on it to prevent the P-type and N-type semiconductor thermoelectric particles from moving during the heating process; adjusting the heating temperature of the heating plate to 200° C., welding at 200° C. for 10 minutes to complete welding; after welding is completed, naturally cool to room temperature and then remove the glass plate to obtain the thermoelectric device; wherein the P-type and N-type semiconductor thermoelectric particles form an electrically series and thermally parallel structure;

[0024] (2.6) The external positive and negative wires of the thermoelectric device are connected to channel 0 and channel 1 of the health monitoring unit respectively. Channel 0 and channel 1 form a differential signal, which is input into ADC conversion unit I. ADC conversion unit I converts the differential signal into a digital signal, which is transmitted to the microcontroller chip in the health monitoring unit and then uploaded to the host computer in real time through the Bluetooth module of the microcontroller chip.

[0025] (3) Preparation of heart rate monitoring unit

[0026] MAX30102 is selected as the heartbeat sensor module. A voltage regulator circuit is designed to maintain it at 1.8V as the power supply of MAX30102, and a decoupling capacitor is added between the power supply and ground to filter out high-frequency noise. Pull-up resistors are added to the data line and clock line of the integrated circuit bus communication between MAX30102 and the health monitoring unit to prevent overshoot.

[0027] The MAX30102, voltage stabilization circuit, decoupling capacitor and pull-up resistor are all integrated on a flexible PCB as the base of the heartbeat monitoring unit; the size of the heartbeat monitoring unit is 2cm 2 , is sewn into the mask.

[0028] MAX30102 converts the collected optical signals into electrical signals through its built-in photodiode, and the electrical signals are converted into digital signals through its built-in ADC. The digital signals are transmitted to the microcontroller chip in the health monitoring unit through the integrated circuit bus, and then uploaded to the host computer in real time through the Bluetooth module built into the microcontroller chip.

[0029] (4) Construction of data processing module

[0030] The data processing module is built into the host computer, and the data processing module includes a data recording module, a data preprocessing module, a data training module, and a data recognition module which are sequentially connected by signals;

[0031] (4.1) The data recording module collects historical data through the respiratory monitoring unit and the heartbeat monitoring unit, and the historical data includes the first type: when a person says an abnormal phrase such as "help", the voltage waveform reflecting the abnormal phrase is sent to the data recording module by the respiratory monitoring unit, and the corresponding waveform reflecting the abnormal heartbeat is sent to the data recording module by the heartbeat monitoring unit when the person is in a dangerous state such as falling and losing weight at that moment; the second type: when a person says an abnormal phrase such as "help", the voltage waveform reflecting the abnormal phrase is sent to the data recording module by the respiratory monitoring unit, and the corresponding waveform reflecting the normal heartbeat is sent to the data recording module by the heartbeat monitoring unit when the person is in a safe state such as sitting, walking, jogging, etc. at that moment; the third type: when a person says normal phrases such as "how are you", "yes", "no", the voltage waveform reflecting the normal phrase is sent to the data recording module by the respiratory monitoring unit, and the corresponding waveform reflecting the abnormal heartbeat is sent to the data recording module by the heartbeat monitoring unit when the person is in a dangerous state such as falling and losing weight at that moment; the fourth type: when a person says "how are the voltage waveform reflecting the normal phrases sent by the breathing monitoring unit to the data recording module when the person is in a safe state such as sitting, walking, jogging, etc. at that moment, and the waveform reflecting the normal heartbeat sent by the heartbeat monitoring unit to the data recording module;

[0032] (4.2) The data preprocessing module preprocesses the data collected by the data recording module in step 4.1; the specific method of preprocessing is: the four types of data in step 4.1 are matched into groups for multi-source fusion, the first type of historical data is multi-source fused and defined as positive, the second, third, and fourth types of historical data are multi-source fused and defined as negative; the data processed as above is divided into a training set and a test set in a ratio of 9:1, with the training set being 90% and the test set being 10%;

[0033] (4.3) In the data training module, the XGBoost framework is used to establish an XGBoost model. The training set of step 4.2 is used to train the XGBoost model, and the hyperparameters of the model are adjusted and optimized according to the training results. The trained and tuned XGBoost model is evaluated through the test set of step 4.2, and the recognition results of the XGBoost model are judged according to the evaluation results. The accuracy and generalization ability of the model are evaluated by comparing with the true values ​​of the test set to judge whether the model achieves the expected effect. Assuming that the test set is used for evaluation, the test set contains a set of sample data with known true values. These sample data can be input into the trained and tuned XGBoost model and then compared with the true values. We compare these recognition values ​​with the true values ​​and use some evaluation indicators to measure the performance of the model, such as "accuracy" and "precision". If the evaluation results show that the XGBoost model has a high accuracy, that is, the values ​​of "accuracy" and "precision" are high, then it can be considered that the model has good generalization ability and can accurately identify; conversely, if the evaluation results show that the recognition error of the XGBoost model is large, then it may be necessary to further optimize the model, adjust the model parameters or retrain the model to improve its accuracy and performance. In the XGBoost model, the number of classifiers is 100, the maximum depth of the tree is 6, the learning rate is 0.1, and the regularization coefficient is 1e -5 , the downsampling rate is 0.8.

[0034] (4.4) The optimized XGBoost model obtained in step 4.3 is loaded into the data recognition module to identify the health status of the human body, and the recognition results are displayed on the display interface of the host computer.

[0035] The host computer is a desktop computer or a mobile phone.

[0036] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows:

[0037] 1. The health monitoring mask of the present invention simultaneously identifies two physiological indicators, breathing and heartbeat, through the breathing monitoring unit and the heartbeat monitoring unit. The two are combined through the data processing module in the host computer to realize the multi-source fusion recognition of abnormal phrases and abnormal heartbeats, and can accurately identify the critical situation of the wearer. The recognition rate is as high as 98.67%, which has great potential in medical rescue; wherein, the working principle of the breathing monitoring unit is: according to the Seebeck principle, based on the temperature difference between the inner and outer surfaces of the breathing monitoring unit of the thermoelectric device, a temperature difference is formed at both ends of the semiconductor thermoelectric particles that are thermally parallel and electrically connected in series, and the thermoelectric device outputs a voltage. The human body exhales gas and inhales air. The exhaled gas takes out the heat of the human body, causing the air temperature to rise, and the inhaled air takes away the air heat, causing the air temperature to drop. Therefore, the breathing monitoring unit can monitor the breathing changes of the human body through the thermoelectric device. Similarly, people's speech is also accompanied by the exhalation and inhalation of gas, resulting in an increase and decrease in air temperature. In the respiratory monitoring unit of the present invention, when a person says different phrases, the exhaled gas and the inhaled gas are different, which makes the air temperature rise and fall by different amplitudes, so that the voltage output by the thermoelectric device is different, that is, the generated voltage waveform is different. Therefore, the respiratory monitoring unit can recognize phrases; the working principle of the heartbeat monitoring unit is: based on the PPG principle, our heartbeat monitoring unit will transmit light through the skin into the blood vessels. Each heartbeat of the human body will cause the expansion and contraction of the blood vessels, thereby causing the light intensity reflected by the blood vessels to change. The heartbeat monitoring unit outputs a changing digital signal according to the change in light intensity to monitor the heartbeat.

[0038] 2. The health monitoring mask device of the present invention has high integration, wireless transmission, and a flexible PCB of only 2cm 2 , easy to wear.

[0039] 3. The thermoelectric performance of the thermoelectric device in the health monitoring mask of the present invention is excellent and the response speed is fast.

[0040] 4. People often release distress signals and emergency reactions in critical situations. For example, in critical scenarios where weightlessness is accompanied by pain, people will call for help (say phrases such as "Help") and have abnormal heartbeats (heartbeat intensity suddenly increases). In order to help people deal with such critical situations and facilitate rescue, the health monitoring mask of the present invention is combined with deep learning to realize abnormal phrases (people say different phrases, corresponding to different exhaled gases and inhaled gases, different breathing brings different heat changes, different heat changes cause different temperature differences, and different voltage waveforms are output through thermoelectric devices) and abnormal heartbeat multi-source fusion abnormal situation recognition and alarm. When people say abnormal phrases accompanied by abnormal heartbeat waveforms, our system will identify the abnormal situation, the host computer will display that it is in an abnormal situation, and use a loudspeaker to warn. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a physical picture of the health status monitoring system of the present invention.

[0042] Figure 2 This is the relationship diagram between the output voltage and power of the respiratory monitoring unit.

[0043] Figure 3 is a graph of the response speed of the respiratory monitoring unit.

[0044] Figure 4 It is a breathing and heart rate monitoring diagram of the health status monitoring system of the present invention under different motion states of the human body.

[0045] Figure 5 It is a construction diagram of the data processing module in the present invention. DETAILED DESCRIPTION

[0046] The present invention is further described below in conjunction with the accompanying drawings and embodiments. The present invention is not limited by the following embodiments, and the specific implementation method can be determined based on the technical solution of the present invention and actual conditions.

[0047] Example 1

[0048] like Figure 1 As shown, a health status monitoring system includes a health status monitoring mask and a host computer. The health status monitoring mask includes the mask and a respiratory monitoring unit, a heartbeat monitoring unit, and a health status monitoring unit arranged therein. The host computer includes a display interface and a data processing module built into the host computer.

[0049] The health monitoring unit is sewn into the mask, including ADC conversion unit I, a microcontroller chip with a built-in Bluetooth module and an integrated circuit bus. The ADC conversion unit I includes channel 0 and channel 1;

[0050] The respiratory monitoring unit includes a thermoelectric device made of P-type semiconductor thermoelectric particles and N-type semiconductor thermoelectric particles. The external positive and negative wires of the thermoelectric device are respectively connected to channel 0 and channel 1 of the health monitoring unit. Channel 0 and channel 1 form a differential signal, which is input into the ADC conversion unit I. The ADC conversion unit I converts the differential signal into a digital signal, which is transmitted to the microcontroller chip (model: ESP32) in the health monitoring unit, and then uploaded to the host computer in real time through the Bluetooth module of the microcontroller chip.

[0051] The heartbeat monitoring unit is composed of MAX30102, voltage stabilizing circuit, decoupling capacitor and pull-up resistor integrated on a flexible PCB (printed circuit board) as a substrate. MAX30102 converts the collected optical signal into an electrical signal through its own photodiode (PD), and the electrical signal is converted into a digital signal through its own ADC. The digital signal is transmitted to the microcontroller chip in the health monitoring unit through the integrated circuit bus (I2C) of the microcontroller chip (model: ESP32) in the health monitoring unit, and then uploaded to the host computer in real time through the Bluetooth module of the microcontroller chip.

[0052] The host computer has a built-in data processing module, which receives data transmitted by the respiratory monitoring unit and the heartbeat monitoring unit to identify and judge the real-time health status of the human body; wherein the data processing module includes a data recording module, a data preprocessing module, a data training module, and a data recognition module which are sequentially connected by signals; the data recording module collects and stores historical data through the respiratory monitoring unit and the heartbeat monitoring unit; the data preprocessing module preprocesses the data collected by the data recording module; in the data training module, the XGBoost framework is adopted to establish an XGBoost model, the preprocessed data is used to train the XGBoost model, the hyperparameters of the model are adjusted and optimized, and an optimized XGBoost model is obtained; the optimized XGBoost model is loaded into the data recognition module to identify the health status of the human body.

[0053] In the data preprocessing module, the specific method of preprocessing is: matching the four types of data into groups for multi-source fusion, the first type of historical data is multi-source fused and defined as positive, the second, third, and fourth types of historical data are multi-source fused and defined as negative; the data processed as above are divided into a training set and a test set according to a ratio of 9:1, with the training set being 90% and the test set being 10%; wherein, the four types of data are matched into groups for multi-source fusion, which means: matching the voltage waveform data reflecting the phrase transmitted by the respiratory monitoring unit with the voltage waveform data reflecting the heartbeat transmitted by the heartbeat monitoring unit, and dividing them into four types :The first type: when a person says an abnormal phrase such as "help", the voltage waveform reflecting the abnormal phrase is sent by the breathing monitoring unit to the data recording module, and the corresponding waveform reflecting the abnormal heartbeat is sent by the heartbeat monitoring unit to the data recording module when the person is in a dangerous state such as falling and losing weight at that moment; the second type: when a person says an abnormal phrase such as "help", the voltage waveform reflecting the abnormal phrase is sent by the breathing monitoring unit to the data recording module, and the corresponding waveform reflecting the normal heartbeat is sent by the heartbeat monitoring unit to the data recording module when the person is in a safe state such as sitting, walking, jogging, etc.; the third type: when a person says "how The fourth type: when a person says normal phrases such as "how are you", "yes", "no", etc., the voltage waveform reflecting the normal phrases sent by the respiratory monitoring unit to the data recording module and the corresponding waveform reflecting the abnormal heartbeat sent by the heartbeat monitoring unit to the data recording module when the person is in a dangerous state such as falling and losing weight at that moment. The fourth type: when a person says normal phrases such as "how are you", "yes", "no", etc., the voltage waveform reflecting the normal phrases sent by the respiratory monitoring unit to the data recording module and the corresponding waveform reflecting the normal heartbeat sent by the heartbeat monitoring unit to the data recording module when the person is in a safe state such as sitting, walking, jogging, etc. at that moment.

[0054] Example 2

[0055] like Figure 1-5 , a method for preparing a health status monitoring system, comprising the following steps:

[0056] (1.1) Mask processing

[0057] A commercial cotton mask was selected as the flexible substrate, and a hole array was prepared on one side of the mask corresponding to the area between the mouth and nostrils by laser cutting. The holes in the hole array were 1.4 mm long, 1.4 mm wide, and the spacing between them was 2.3 mm.

[0058] (1.2) Construction of health status monitoring unit

[0059] The health status monitoring unit is sewn inside the mask. The health status monitoring unit includes ADC conversion unit I, a microcontroller chip with built-in Bluetooth module and integrated circuit bus, and ADC conversion unit I includes channel 0 and channel 1.

[0060] (2) Preparation of respiratory monitoring unit

[0061] The respiratory monitoring unit includes a thermoelectric device, which includes P-type semiconductor thermoelectric particles and N-type semiconductor thermoelectric particles. The specific preparation process is as follows:

[0062] (2.1) Preparation of P-type and N-type semiconductor thermoelectric particles: Sb2Te3 is used as the raw material of P-type semiconductor thermoelectric particles, and Bi2Te3 is used as the raw material of N-type semiconductor thermoelectric particles. Both P-type and N-type semiconductor thermoelectric particles are formed into blocks, and their dimensions are 1.4 mm in length, 1.4 mm in width, and 2.5 mm in height. The spacing between N-type semiconductor thermoelectric particles and P-type semiconductor thermoelectric particles is 2.3 mm.

[0063] (2.2) Preparation of electrodes: The conductive fabric was cut into rectangular pieces as electrodes by laser cutting. The size of the electrodes was length × width = 4.9 mm × 1.4 mm.

[0064] (2.3) placing the electrode prepared in step 2.2 on a metal steel mesh, then transferring it to a thermal release tape, and coating the electrode with tin paste;

[0065] (2.4) Lay the mask obtained in step 1 flat on a heating plate, then alternately embed the P-type and N-type semiconductor thermoelectric particles obtained in step 2.1 into the holes of the hole array prepared in step 1, and place the electrodes prepared in step 2.3 on the outer ends of the semiconductor thermoelectric particles. After placement, press a glass plate on it to prevent the P-type and N-type semiconductor thermoelectric particles from moving during heating; adjust the heating temperature of the heating plate to 200° C., and weld at 200° C. for 10 minutes to complete welding; after welding is completed, naturally cool to room temperature and then remove the glass plate;

[0066] (2.5) Turning over the mask after completing step 2.4, and laying it flat on the heating plate again; then, placing the electrode prepared in step 2.3 on the unwelded end of the semiconductor thermoelectric particle, and after placement, pressing a glass plate on it to prevent the P-type and N-type semiconductor thermoelectric particles from moving during the heating process; adjusting the heating temperature of the heating plate to 200° C., welding at 200° C. for 10 minutes to complete welding; after welding is completed, naturally cool to room temperature and then remove the glass plate to obtain the thermoelectric device; wherein the P-type and N-type semiconductor thermoelectric particles form an electrically series and thermally parallel structure;

[0067] (2.6) The external positive and negative wires of the thermoelectric device are connected to channel 0 and channel 1 of the health monitoring unit respectively. Channel 0 and channel 1 form a differential signal, which is input into ADC conversion unit I. ADC conversion unit I converts the differential signal into a digital signal, which is transmitted to the microcontroller chip (model: ESP32) in the health monitoring unit, and then uploaded to the host computer in real time through the Bluetooth module of the microcontroller chip.

[0068] (3) Preparation of heart rate monitoring unit

[0069] MAX30102 is selected as the heartbeat sensor module. A voltage regulator circuit is designed to maintain it at 1.8V as the power supply of MAX30102, and a decoupling capacitor is added between the power supply and ground to filter out high-frequency noise. Pull-up resistors are added to the data line (SDA) and clock line (SCL) that communicate between MAX30102 and the integrated circuit bus (I2C) of the health monitoring unit to prevent overshoot.

[0070] The MAX30102, voltage stabilization circuit, decoupling capacitor and pull-up resistor are all integrated on a flexible PCB as the base of the heartbeat monitoring unit; the size of the heartbeat monitoring unit is 2cm 2 , is sewn into the mask.

[0071] MAX30102 converts the collected optical signals into electrical signals through its built-in photodiode (PD), and the electrical signals are converted into digital signals through its built-in ADC. The digital signals are transmitted to the microcontroller chip (model: ESP32) in the health monitoring unit through the integrated circuit bus (IIC), and then uploaded to the host computer in real time through the Bluetooth module built into the microcontroller chip.

[0072] (4) Construction of data processing module

[0073] The data processing module is loaded in the host computer, and the data processing module includes a data recording module, a data preprocessing module, a data training module, and a data recognition module which are sequentially connected by signals;

[0074] (4.1) The data recording module collects historical data through the respiratory monitoring unit and the heartbeat monitoring unit, and the historical data includes the first type: when a person says an abnormal phrase such as "help", the voltage waveform reflecting the abnormal phrase is sent to the data recording module by the respiratory monitoring unit, and the corresponding waveform reflecting the abnormal heartbeat is sent to the data recording module by the heartbeat monitoring unit when the person is in a dangerous state such as falling and losing weight at that moment; the second type: when a person says an abnormal phrase such as "help", the voltage waveform reflecting the abnormal phrase is sent to the data recording module by the respiratory monitoring unit, and the corresponding waveform reflecting the normal heartbeat is sent to the data recording module by the heartbeat monitoring unit when the person is in a safe state such as sitting, walking, jogging, etc. at that moment; the third type: when a person says normal phrases such as "how are you", "yes", "no", the voltage waveform reflecting the normal phrase is sent to the data recording module by the respiratory monitoring unit, and the corresponding waveform reflecting the abnormal heartbeat is sent to the data recording module by the heartbeat monitoring unit when the person is in a dangerous state such as falling and losing weight at that moment; the fourth type: when a person says "how are the voltage waveform reflecting the normal phrases sent by the breathing monitoring unit to the data recording module when the person is in a safe state such as sitting, walking, jogging, etc. at that moment, and the waveform reflecting the normal heartbeat sent by the heartbeat monitoring unit to the data recording module;

[0075] (4.2) The data preprocessing module preprocesses the data collected by the data recording module in step 4.1; the specific method of preprocessing is: the four types of data in step 4.1 are matched into groups for multi-source fusion, the first type of historical data is multi-source fused and defined as positive, the second, third, and fourth types of historical data are multi-source fused and defined as negative; the data processed as above is divided into a training set and a test set in a ratio of 9:1, with the training set being 90% and the test set being 10%;

[0076] (4.3) In the data training module, the XGBoost framework is used to establish an XGBoost model. The training set of step 4.2 is used to train the XGBoost model, and the hyperparameters of the model are adjusted according to the training results. The trained and tuned XGBoost model is evaluated through the test set of step 4.2, and the recognition results of the XGBoost model are judged according to the evaluation results. The accuracy and generalization ability of the model are evaluated by comparing with the true values ​​of the test set to judge whether the model achieves the expected effect. Assuming that the test set is used for evaluation, the test set contains a set of sample data with known true values. These sample data can be input into the trained and tuned XGBoost model and then compared with the true values. We compare these recognition values ​​with the true values ​​and use some evaluation indicators to measure the performance of the model, such as "accuracy" and "precision". If the evaluation results show that the XGBoost model has a high accuracy, that is, the values ​​of "accuracy" and "precision" are high, then it can be considered that the model has good generalization ability and can accurately identify; conversely, if the evaluation results show that the XGBoost model has a low accuracy, then it may be necessary to further optimize the model, optimize the model parameters or retrain the model to improve its accuracy and performance. In the XGBoost model, the number of classifiers is 100, the maximum depth of the tree is 6, the learning rate is 0.1, and the regularization coefficient is 1e -5 , the downsampling rate is 0.8.

[0077] (4.4) The optimized XGBoost model obtained in step 4.3 is loaded into the data recognition module to identify the health status of the human body, and the recognition results are displayed on the display interface of the host computer.

[0078] After actual operation, the recognition accuracy of the above data processing module is 98.67%.

[0079] Performance test: Test the prepared health status monitoring system performance. Place the prepared thermoelectric device on the test bench, apply different temperatures to the upper and lower surfaces of the thermoelectric device to form a temperature difference, and use Keithley 2400 to measure the voltage generated under different temperature differences. The measured data is as follows Figure 2 As shown. When the temperature difference is 30K, the open circuit voltage of the thermoelectric device is 140mV. The airflow velocity of 3.2m / s is used to simulate human breathing and applied to one side of the thermoelectric device to test its reaction rate. Figure 3 As shown, the response time of the thermoelectric device is 1.4s. The respiratory and heart rate monitoring of the present invention tests the respiration and heart rate under different motion states, such as Figure 4 shown.

[0080] The above are only preferred embodiments of the present invention. It should be pointed out that a person skilled in the art can make several modifications and improvements without departing from the inventive concept of the present invention, and these all fall within the protection scope of the present invention.

Claims

1. A health monitoring system, the system comprising a health monitoring mask and a host computer, the health monitoring mask comprising a mask and a breathing monitoring unit, a heartbeat monitoring unit, and a health monitoring unit arranged therein, the host computer comprising a display interface and a data processing module built into the host computer; The health monitoring unit is sewn into the mask, including ADC conversion unit I, a microcontroller chip with a built-in Bluetooth module and an integrated circuit bus. The ADC conversion unit I includes channel 0 and channel 1; The respiratory monitoring unit includes a thermoelectric device made of P-type semiconductor thermoelectric particles and N-type semiconductor thermoelectric particles. The external positive and negative wires of the thermoelectric device are respectively connected to channel 0 and channel 1 of the health monitoring unit. Channel 0 and channel 1 form a differential signal, which is transmitted to ADC conversion unit I. ADC conversion unit I converts the differential signal into a digital signal, which is transmitted to the microcontroller chip in the health monitoring unit and then uploaded to the host computer in real time through the Bluetooth module of the microcontroller chip. The heartbeat monitoring unit is composed of MAX30102, voltage regulator circuit, decoupling capacitor and pull-up resistor integrated on the flexible PCB as the substrate. MAX30102 converts the collected optical signal into an electrical signal through the built-in photodiode, and the electrical signal is converted into a digital signal through the built-in ADC. The digital signal is transmitted to the microcontroller chip in the health monitoring unit through the integrated circuit bus, and then uploaded to the host computer in real time through the Bluetooth module built into the microcontroller chip. The data processing module included in the host computer receives the data transmitted by the respiratory monitoring unit and the heartbeat monitoring unit, matches the data of the two into groups for multi-source fusion, and identifies and judges the real-time health status of the human body based on the multi-source fused data.

2. A mask-type health status monitoring device according to claim 1, characterized in that: The data processing module includes a data recording module, a data preprocessing module, a data training module, and a data identification module which are sequentially connected by signals; the data recording module collects and stores historical data through a respiratory monitoring unit and a heartbeat monitoring unit; The data preprocessing module preprocesses the data collected by the data recording module; in the data training module, the XGBoost framework is used to establish the XGBoost model, the preprocessed data is used to train and test the XGBoost model, the hyperparameters of the model are adjusted and optimized, and the optimized XGBoost model is obtained; the optimized XGBoost model is loaded into the data recognition module to identify the health status of the human body.

3. A mask-type health status monitoring device according to claim 2, characterized in that: In the data preprocessing module, the specific method of preprocessing is: matching the four types of data into groups for multi-source fusion, the first type of historical data is multi-source fused and defined as positive, the second, third, and fourth types of historical data are multi-source fused and defined as negative; the data processed as above are divided into a training set and a test set according to a ratio of 9:1, with the training set being 90% and the test set being 10%; wherein, the multi-source fusion of the four types of data refers to: matching the voltage waveform data reflecting the phrase transmitted by the respiratory monitoring unit with the waveform data reflecting the heartbeat transmitted by the heartbeat monitoring unit, which are divided into four types: the first type: when a person says an abnormal phrase, the voltage waveform reflecting the abnormal phrase sent to the data recording module by the respiratory monitoring unit and the corresponding voltage waveform reflecting the abnormal phrase at that moment sent by the heartbeat monitoring unit to the data recording module The first type is a waveform reflecting an abnormal heartbeat sent to the data recording module, the second type is a voltage waveform reflecting an abnormal phrase sent to the data recording module by the respiratory monitoring unit when a person speaks an abnormal phrase, and the corresponding waveform reflecting a normal heartbeat sent to the data recording module by the heartbeat monitoring unit when the person is in a safe state at that moment, the third type is a voltage waveform reflecting a normal phrase sent to the data recording module by the respiratory monitoring unit when a person speaks a normal phrase, and the corresponding waveform reflecting an abnormal heartbeat sent to the data recording module by the heartbeat monitoring unit when the person is in a dangerous state at that moment, the fourth type is a voltage waveform reflecting a normal phrase sent to the data recording module by the respiratory monitoring unit when a person speaks a normal phrase, and the corresponding waveform reflecting a normal heartbeat sent to the data recording module by the heartbeat monitoring unit when the person is in a safe state at that moment.

4. A method for preparing a health status monitoring system according to any one of claims 1 to 3, comprising the following steps: (1.1) Mask processing A commercial cotton mask was selected as the flexible substrate, and a hole array was prepared on one side of the mask corresponding to the area between the mouth and nostrils by laser cutting according to the design requirements; (1.2) Construction of health status monitoring unit The health monitoring unit is sewn into the mask, and includes an ADC conversion unit I, a microcontroller chip with a built-in Bluetooth module and an integrated circuit bus, and the ADC conversion unit I includes channel 0 and channel 1; (2) Preparation of respiratory monitoring unit The respiratory monitoring unit includes a thermoelectric device, which includes P-type semiconductor thermoelectric particles and N-type semiconductor thermoelectric particles. The specific preparation process is as follows: (2.1) Preparation of P-type and N-type semiconductor thermoelectric particles: Sb2Te3 is used as the raw material of P-type semiconductor thermoelectric particles, and Bi2Te3 is used as the raw material of N-type semiconductor thermoelectric particles. Both P-type and N-type semiconductor thermoelectric particles are formed into blocks, and their length, width, and spacing between semiconductor thermoelectric particles match the length and width of the holes in the hole array and the hole spacing; (2.2) Preparation of electrodes: Cut the conductive fabric into rectangular pieces as electrodes by laser cutting; (2.3) placing the electrode prepared in step 2.2 on a metal steel mesh for alignment, then transferring it to a thermal release tape and coating the electrode with tin paste; (2.4) Lay the mask obtained in step 1 flat on the heating plate, then alternately embed the P-type and N-type semiconductor thermoelectric particles obtained in step 2.1 into the holes of the hole array prepared in step 1, and place the electrodes prepared in step 2.3 on the outer ends of the semiconductor thermoelectric particles. After placement, press a glass plate on it to prevent the P-type and N-type semiconductor thermoelectric particles from moving during heating; heat the heating plate to a certain temperature and weld it; after welding, naturally cool it to room temperature and then remove the glass plate; (2.5) Turn over the mask after completing step 2.4 and lay it flat on the heating plate again; then, place the electrode prepared in step 2.3 on the unwelded end of the semiconductor thermoelectric particle, and after placement, press a glass plate on it to prevent the P-type and N-type semiconductor thermoelectric particles from moving during the heating process; heat the heating plate to a certain temperature and weld; after welding, naturally cool to room temperature and then remove the glass plate to obtain the thermoelectric device; wherein the P-type and N-type semiconductor thermoelectric particles form an electrically series and thermally parallel structure; (2.6) Connecting the external positive and negative wires of the thermoelectric device to channel 0 and channel 1 of the health monitoring unit respectively, channel 0 and channel 1 form a differential signal, the differential signal is input into ADC conversion unit I, ADC conversion unit I converts the differential signal into a digital signal, the digital signal is transmitted to the microcontroller chip in the health monitoring unit, and then uploaded to the host computer in real time through the Bluetooth module of the microcontroller chip; (3) Preparation of heart rate monitoring unit Select MAX30102 as the heartbeat sensor module; design a voltage regulator circuit to maintain 1.8V as the power supply of MAX30102, add a decoupling capacitor between the power supply and the ground to filter out high-frequency noise, and add pull-up resistors on the SDA data line and SCL clock line of the integrated circuit bus communication between MAX30102 and the health monitoring unit to prevent overshoot; The MAX30102, voltage stabilization circuit, decoupling capacitor and pull-up resistor are all integrated on a flexible PCB serving as the base of the heartbeat monitoring unit; the heartbeat monitoring unit is sewn into the mask; MAX30102 converts the collected optical signals into electrical signals through its built-in photodiode, and the electrical signals are converted into digital signals through its built-in ADC. The digital signals are transmitted to the microcontroller chip in the health monitoring unit through the integrated circuit bus I2C, and then uploaded to the host computer in real time through the Bluetooth module built into the microcontroller chip. (4) Construction of data processing module The data processing module is loaded in the host computer, and the data processing module includes a data recording module, a data preprocessing module, a data training module, and a data recognition module which are sequentially connected by signals; (4.1) The data recording module collects historical data through the respiratory monitoring unit and the heartbeat monitoring unit, and the historical data includes the first type: the voltage waveform reflecting the abnormal phrase sent to the data recording module by the respiratory monitoring unit when the person utters the abnormal phrase and the corresponding waveform reflecting the abnormal heartbeat sent to the data recording module by the heartbeat monitoring unit when the person is in a dangerous state at that moment; the second type: the voltage waveform reflecting the abnormal phrase sent to the data recording module by the respiratory monitoring unit when the person utters the abnormal phrase and the corresponding waveform reflecting the normal heartbeat sent to the data recording module by the heartbeat monitoring unit when the person is in a safe state at that moment; the third type: the voltage waveform reflecting the normal phrase sent to the data recording module by the respiratory monitoring unit when the person utters the normal phrase and the corresponding waveform reflecting the abnormal heartbeat sent to the data recording module by the heartbeat monitoring unit when the person is in a dangerous state at that moment; the fourth type: the voltage waveform reflecting the normal phrase sent to the data recording module by the respiratory monitoring unit when the person utters the normal phrase and the corresponding waveform reflecting the normal heartbeat sent to the data recording module by the heartbeat monitoring unit when the person is in a safe state at that moment; (4.2) The data preprocessing module preprocesses the data collected by the data recording module in step 4.1; the specific method of preprocessing is: the four types of data in step 4.1 are matched into groups for multi-source fusion, the first type of historical data is multi-source fused and defined as positive, the second, third, and fourth types of historical data are multi-source fused and defined as negative; the data processed as above are divided into a training set and a test set in a ratio of 9:1, with the training set accounting for 90% and the test set accounting for 10%; (4.3) In the data training module, the XGBoost framework is used to establish an XGBoost model, and the training set of step 4.2 is used to train the XGBoost model, and the hyperparameters of the model are adjusted and optimized according to the training results; the trained and tuned XGBoost model is evaluated through the test set of step 4.2, and the recognition results of the XGBoost model are judged according to the evaluation results to determine whether the model achieves the expected effect; the recognition value is compared with the true value, and "accuracy" and "precision" are used as evaluation indicators to measure the performance of the model; (4.4) The optimized XGBoost model obtained in step 4.3 is loaded into the data recognition module to identify the health status of the human body, and the recognition results are displayed on the display interface of the host computer.

5. The preparation method according to claim 4, characterized in that: In step (2), the P-type and N-type semiconductor thermoelectric particles are both formed into blocks, and their dimensions are 1.4 mm in length, 1.4 mm in width, and 2.5 mm in height. The spacing between the N-type semiconductor thermoelectric particles and the P-type semiconductor thermoelectric particles is 2.3 mm. The dimensions of the electrode are length × width = 4.9 mm × 1.4 mm.

6. The preparation method according to claim 4, characterized in that: In steps (2.4) and (2.5), the step of heating the heating plate to a certain temperature for welding is specifically as follows: adjusting the heating temperature of the heating plate to 200° C., and welding at 200° C. for 10 minutes to complete welding.

7. The preparation method according to claim 4, characterized in that: In step (3), the size of the heart rate monitoring unit is 2 cm 2 .

8. The preparation method according to claim 4, characterized in that: The host computer is a desktop computer or a mobile phone including a display interface.

9. The preparation method according to claim 4, characterized in that: The XGBoost model described in step (3) has 100 classifiers, a maximum tree depth of 6, a learning rate of 0.1, and a regularization coefficient of 1e -5 , the downsampling rate is 0.8.

Images