Expiratory gas detection device for dynamically adjusting gas flow rate based on machine learning

Abstract

The invention discloses a machine learning-based expired gas detection device capable of dynamically adjusting gas flow rate, and relates to the technical field of expired gas detection.The machine learning-based expired gas detection device comprises a sucking pump and a multiple-reflection pool, the output end of the sucking pump is connected with the multiple-reflection pool, and the multiple-reflection pool is provided with a light source inlet and a light source outlet; a light source module is arranged at the light source inlet; the light source outlet is connected with a spectral analysis module through an optical fiber; the flow change of exhaled air is predicted in real time through a sensor module in the intelligent air pump flow rate control module, control parameters of an air pump are dynamically adjusted, the stability and accuracy of exhaled air detection are improved, data processing, machine learning model operation and PID control are achieved through an STM32 single chip microcomputer in the machine learning module, and the intelligent air pump flow rate control system is suitable for large-scale popularization and application. The expired gas detection device can quickly respond and stably control the flow of the expired gas, can adapt to a dynamic environment, ensures the real-time performance and the high efficiency, and improves the stability and the accuracy of expired gas detection.

Description

Technical Field

[0001] The present invention relates to the technical field of exhaled breath detection, and in particular to an exhaled breath detection device with dynamic adjustment of gas flow rate based on machine learning. Background Art

[0002] Exhaled breath testing is a non-invasive method for diagnosing and monitoring diseases by analyzing the components in human exhaled breath using various analytical techniques, such as gas chromatograph (GC), mass spectrometer (MS), etc., to quantitatively and qualitatively analyze volatile organic compounds (VOCs) in exhaled breath. This test can reveal the physiological and pathological state of the human body. Because exhaled breath contains a variety of volatile organic compounds, changes in the concentration of these compounds may be related to specific health conditions, so exhaled breath testing is of great significance in disease diagnosis and health monitoring.

[0003] Traditional optical detection devices generally use an air pump control method when measuring the gas concentration of human exhaled breath, mainly relying on PID control. However, in the actual process of detecting the type and concentration of human exhaled breath, since the flow rate of the exhaled gas of the measured population is uneven, it is often necessary to frequently manually adjust the PID parameters. This air pump flow rate adjustment method cannot adapt to the dynamic environment, resulting in unstable exhaled breath detection results and reducing the accuracy of exhaled breath detection. Therefore, the present invention proposes an exhaled breath detection device with dynamic gas flow rate adjustment based on machine learning to solve the problems existing in the prior art. Summary of the invention

[0004] In view of the above problems, the purpose of the present invention is to propose an exhaled breath detection device with dynamic adjustment of gas flow rate based on machine learning, so as to solve the problem that traditional optical detection devices using air pump flow rate adjustment cannot adapt to dynamic environments, resulting in unstable exhaled breath detection results and low accuracy of exhaled breath detection.

[0005] In order to achieve the purpose of the present invention, the present invention is implemented through the following technical solutions: an exhaled breath detection device with dynamic adjustment of gas flow rate based on machine learning, comprising an air pump and a multiple reflection pool, the output end of the air pump is connected to the multiple reflection pool, the multiple reflection pool is provided with a light source entrance port for light source injection and a light source exit port for light source emission, a light source module is provided at the position of the light source entrance port, the light source exit port is connected to a spectrum analysis module through an optical fiber, the input end of the air pump is connected to an intelligent air pump control module, and the input end of the intelligent air pump control module is connected to a filtering and dehumidification module;

[0006] The intelligent air pump control module includes a sensor module for collecting exhaled gas data in real time, a machine learning control module for processing and analyzing the collected exhaled gas data, and an air pump control module for adjusting the working state of the air pump according to the data processing and analysis results.

[0007] A further improvement is that a dust filter and a physical dehumidification element are integrated on the filtering and dehumidifying module to filter out dust and moisture in the exhaled gas.

[0008] A further improvement is that the sensor module detects the flow rate data, pressure data, temperature data and humidity data of the exhaled gas passing through the filtering and dehumidification module in real time, and sends the detected data to the machine learning control module for processing and analysis.

[0009] A further improvement is that the machine learning control module is integrated with an STM32 single-chip microcomputer and a PID control module, and the machine learning control module uses the STM32 single-chip microcomputer to pre-process the data collected by the sensor module, including data cleaning and normalization.

[0010] A further improvement is that the specific steps of the STM32 microcontroller preprocessing the data collected by the sensor module are: first, a data cleaning algorithm is used to remove erroneous, incomplete and incorrectly formatted data from the collected data, and then a normalization algorithm is used to scale the refined data.

[0011] Further improvements are as follows: the machine learning control module predicts the current flow rate of exhaled gas and dynamically adjusts the PID parameters in real time based on the data preprocessed by the STM32 single-chip microcomputer, combined with historical data and model training results; the STM32 single-chip microcomputer sends instructions to the air pump control module according to the control signal output by the PID control module; after receiving the instructions from the STM32 single-chip microcomputer, the air pump control module adjusts the working state of the air pump according to the instructions.

[0012] A further improvement is that the spectral analysis module uses spectroscopy to measure gas concentration, and the absorbance expression is:

[0013]

[0014] Where I0 is the incident light intensity, I t is the intensity of transmitted light, T is the transmittance, K is the absorption coefficient, L is the thickness of the absorbing medium, and C is the concentration of the gas to be measured.

[0015] A further improvement is that the vacuum pump is a DC brushless diaphragm air pump, and the flow rate is controlled by adjusting the rotation speed of the brushless motor.

[0016] The beneficial effects of the present invention are as follows: the present invention predicts the flow rate change of exhaled gas in real time through the sensor module in the intelligent air pump flow rate control module, dynamically adjusts the control parameters of the suction pump, improves the stability and accuracy of exhaled gas detection, and realizes data processing, machine learning model operation and PID control through the STM32 single-chip microcomputer in the machine learning module, which can quickly respond to and stably control the exhaled gas flow rate, can adapt to dynamic environments, ensure the real-time and high efficiency of the exhaled gas detection device, improve the stability and accuracy of exhaled gas detection, have broad application prospects in the field of exhaled gas detection, and provide new ideas and solutions for intelligent exhaled gas detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a structural schematic diagram of an exhaled breath detection device with dynamic gas flow rate regulation based on machine learning of the present invention;

[0019] Figure 2 It is a structural schematic diagram of the intelligent air pump control module of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Embodiment 1

[0022] See also Figure 1 , Figure 2The present embodiment provides an exhaled gas detection device with dynamic gas flow rate adjustment based on machine learning, comprising an air pump for collecting exhaled gas and a multiple reflection pool for providing a reflection function for light, wherein the air outlet at the output end of the air pump is connected to the multiple reflection pool so as to inject the collected exhaled gas into the multiple reflection pool, and the multiple reflection pool is also provided with a light source entrance port and a light source exit port, wherein a light source module is provided at the position of the light source entrance port, and light emitted by the light source module is emitted into the multiple reflection pool through the light source entrance port, and the light source exit port is connected to a spectrum analysis module through an optical fiber, and after the light emitted by the light source module passes through the multiple reflection pool and is fully absorbed by the exhaled gas in the multiple reflection pool, it is transmitted to the spectrum analysis module through the optical fiber, and the composition and concentration of the exhaled gas are analyzed by the spectrum analysis module, and the input end of the air pump is connected to an intelligent air pump control module for adjusting the flow rate of the air pump, and the intelligent air pump control module is the core part of the exhaled gas detection device, and the input end of the intelligent air pump control module is connected to a filtering and dehumidification module for filtering out moisture in the exhaled gas to reduce the influence of moisture on the differential optical density of the exhaled gas;

[0023] The intelligent air pump control module includes a sensor module, a machine learning control module and an air pump control module. The sensor module is used to collect exhaled gas data in real time, including gas flow rate, gas pressure, gas temperature and gas humidity. The machine learning control module is used to process and analyze the collected exhaled gas data. The air pump control module is used to adjust the working state of the air pump according to the data processing and analysis results.

[0024] The filter dehumidification module integrates a dust filter and a physical dehumidification element, wherein the dust filter is used to filter out dust in the exhaled gas, and the physical dehumidification element is used to filter out moisture in the exhaled gas.

[0025] The sensor module detects the flow rate data, pressure data, temperature data and humidity data of the exhaled gas passing through the filtering and dehumidification module in real time, and sends the detected data to the machine learning control module for processing and analysis.

[0026] The machine learning control module of this embodiment is integrated with an STM32 single-chip microcomputer and a PID control module. The STM32 single-chip microcomputer is used to realize the operation of the machine learning model, obtain the data of the sensor module in real time, predict the change of the exhaled air flow rate, and dynamically adjust the parameters of the PID control module according to the prediction results, calculate the control signal, and adjust the flow rate of the air pump to ensure that the gas flows stably in the spectral analysis module, thereby overcoming the limitations of the traditional fixed flow rate adjustment method, thereby ensuring the accuracy and real-time performance of the detection results;

[0027] The machine learning control module uses the STM32 microcontroller to pre-process the data collected by the sensor module, including data cleaning and normalization. The specific steps are: first use the data cleaning algorithm to remove erroneous, incomplete and incorrectly formatted data from the collected data, and then use the normalization algorithm to scale the refined data.

[0028] The PID control module of this embodiment is composed of three parts: proportional, integral and derivative. Its control signal u(t) is given by the following formula:

[0029]

[0030] Where, e(t) is the current error, which is equal to the difference between the set value and the measured value, K P is the proportional gain, which determines the contribution of the current error to the control signal, K i is the integral gain, which determines the influence of the accumulated overdue error on the control signal, K d is the differential gain, which determines the impact of the error change rate on the control signal.

[0031] The machine learning control module predicts the current exhaled gas flow rate and dynamically adjusts the PID parameters in real time based on the data preprocessed by the STM32 single-chip microcomputer, combined with historical data and model training results. According to the control signal output by the PID control module, the STM32 single-chip microcomputer sends instructions to the air pump control module. After receiving the instructions from the STM32 single-chip microcomputer, the air pump control module adjusts the working state of the air pump according to the instructions;

[0032] The specific strategy for dynamically adjusting PID parameters is as follows:

[0033] According to the predicted change in exhaled air flow, the proportional gain is appropriately increased or decreased to enhance or weaken the impact of the current error on the control signal. According to the predicted steady-state error trend, the integral gain is adjusted to eliminate the steady-state error more quickly. According to the predicted error change rate, the differential gain is dynamically adjusted to reduce the overshoot and oscillation of the control signal.

[0034] The implementation and real-time control and optimization process of PID control algorithm are as follows:

[0035] First, the sensor module collects the data of exhaled gas in real time, and the machine learning control module processes the data to calculate the current error e(t). Then, the cumulative integral of the error and the rate of change of the error are calculated. Based on these calculation results, the machine learning model is used to predict the change of exhaled gas flow and dynamically adjust the parameter K of the PID control module. P , K i , K d, and then calculate the control signal u(t) through the adjusted PID parameters, the formula is:

[0036]

[0037] The STM32 microcontroller generates a PWM signal based on the calculated control signal, sets the appropriate frequency and duty cycle, and transmits the generated PWM signal to the vacuum pump drive circuit to control the operation of the brushless motor. This process is cyclically executed in each control cycle. Through the combination of machine learning and PID control modules, it ensures that the detection system can quickly respond and stably control the exhaled air flow, thereby achieving high precision and stability in exhaled air detection.

[0038] The spectrum analysis module of this embodiment uses a spectrum method to measure gas concentration. The principle is the Beer-Lambert law. When a light beam is transmitted in a medium, it is weakened due to interaction with the substance. The absorbance expression is:

[0039]

[0040] Where I0 is the incident light intensity, I t is the intensity of transmitted light, T is the transmittance, K is the absorption coefficient, L is the thickness of the absorbing medium, and C is the concentration of the gas to be measured;

[0041] The absorption coefficient of gas molecules is a function of the absorption wavelength and is independent of the incident light intensity, the concentration of the absorbing substance, and the thickness. According to the above absorbance expression, if the absorption coefficient of the measured gas is known, then the concentration of the measured gas can be calculated by measuring the incident light intensity, the transmitted light intensity and the thickness of the measured gas. This method can realize the rapid and accurate detection of various gas components in the exhaled breath.

[0042] The vacuum pump in this embodiment is a DC brushless diaphragm air pump (the diaphragm pump has good corrosion resistance and can handle gases of various properties, including corrosive and volatile gases), which ensures the diversity and reliability of the gases that can be detected by the exhaled breath detection device, and controls the flow rate by adjusting the speed of the brushless motor.

[0043] Embodiment 2

[0044] Based on the first embodiment, this embodiment provides a detection method of an exhaled breath detection device with dynamic adjustment of gas flow rate based on machine learning, comprising the following steps: starting an air pump to absorb the gas exhaled by the human body, the exhaled gas first passes through a filtering and dehumidification module to remove moisture therein, and then uses a sensor module in an intelligent air pump control module to collect the flow rate, pressure, temperature and humidity data of the exhaled gas in real time, and sends the data to a machine learning control module in the intelligent air pump control module, and the STM32 single-chip microcomputer in the machine learning control module performs preprocessing including data cleaning and normalization to improve high-quality input data for the machine learning model, and the machine learning control module then collects the data according to the preprocessing. The processed data is combined with historical data and model training results to predict the current exhaled gas flow rate and dynamically adjust the PID parameters in real time. According to the control signal output by the PID control module, the STM32 microcontroller sends an instruction. After receiving the instruction, the air pump control module in the machine learning control module adjusts the working state of the air pump to ensure the stability of the exhaled gas flow rate. When the air pump stably injects the exhaled gas into the multiple reflection pool, the light source module is used to irradiate the multiple reflection pool. The light emitted by the light source module passes through the multiple reflection pool and is fully absorbed by the gas to be tested therein, and then is transmitted to the spectral analysis module through the optical fiber to analyze the composition and concentration of the exhaled gas, thereby realizing exhaled gas detection.

[0045] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An exhaled breath detection device with dynamic gas flow rate adjustment based on machine learning, comprising an air pump and a multiple reflection pool, characterized in that: The output end of the air extraction pump is connected to a multiple reflection pool, and the multiple reflection pool is provided with a light source entrance for light source injection and a light source exit for light source emission, a light source module is provided at the position of the light source entrance, and the light source exit is connected to a spectrum analysis module via an optical fiber, and the input end of the air extraction pump is connected to an intelligent air pump control module, and the input end of the intelligent air pump control module is connected to a filtering and dehumidification module; The intelligent air pump control module includes a sensor module for collecting exhaled gas data in real time, a machine learning control module for processing and analyzing the collected exhaled gas data, and an air pump control module for adjusting the working state of the air pump according to the data processing and analysis results.

2. The exhaled breath detection device with dynamic gas flow rate adjustment based on machine learning according to claim 1, characterized in that: The filtering and dehumidifying module is integrated with a dust filter and a physical dehumidifying element to filter out dust and moisture in the exhaled gas.

3. The exhaled breath detection device with dynamic gas flow rate adjustment based on machine learning according to claim 1, characterized in that: The sensor module detects the flow rate data, pressure data, temperature data and humidity data of the exhaled gas passing through the filtering and dehumidification module in real time, and sends the detected data to the machine learning control module for processing and analysis.

4. The exhaled breath detection device with dynamic gas flow rate adjustment based on machine learning according to claim 1, characterized in that: The machine learning control module is integrated with an STM32 single-chip microcomputer and a PID control module. The machine learning control module uses the STM32 single-chip microcomputer to pre-process the data collected by the sensor module, including data cleaning and normalization.

5. The exhaled breath detection device with dynamic gas flow rate adjustment based on machine learning according to claim 4, characterized in that: The specific steps of preprocessing the data collected by the sensor module by the STM32 single chip microcomputer are: firstly, using a data cleaning algorithm to remove erroneous, incomplete and incorrectly formatted data from the collected data, and then using a normalization algorithm to scale the refined data.

6. The exhaled breath detection device with dynamic gas flow rate adjustment based on machine learning according to claim 4, characterized in that: The machine learning control module predicts the current flow rate of exhaled gas and dynamically adjusts the PID parameters in real time based on the data preprocessed by the STM32 single-chip microcomputer, in combination with historical data and model training results. According to the control signal output by the PID control module, the STM32 single-chip microcomputer sends an instruction to the air pump control module. After receiving the instruction from the STM32 single-chip microcomputer, the air pump control module adjusts the working state of the air pump according to the instruction.

7. The exhaled breath detection device with dynamic gas flow rate adjustment based on machine learning according to claim 1, characterized in that: The spectrum analysis module uses spectroscopy to measure gas concentration, and the absorbance expression is: Where I0 is the incident light intensity, I t is the intensity of transmitted light, T is the transmittance, K is the absorption coefficient, L is the thickness of the absorbing medium, and C is the concentration of the gas to be measured.

8. The exhaled breath detection device with dynamic gas flow rate adjustment based on machine learning according to claim 1, characterized in that: The vacuum pump is a DC brushless diaphragm air pump, and the flow rate is controlled by adjusting the rotation speed of the brushless motor.

Images