Dust-respiratory dual-modal detection system and detection method integrated with flexible electronic skin

Abstract

A flexible electronic skin-integrated dust-respiration dual-modal detection system and method are disclosed. The dual-modal sensing module, signal processing module, wireless transmission module, data application module, anomaly warning module, and power supply are integrated on a composite flexible substrate. The dual-modal sensing module monitors the dust concentration characteristics and breathing frequency characteristics of workers in the workplace in real time. The signal processing module receives the signals and performs noise reduction and data fusion processing to obtain dataset D. The wireless transmission module transmits dataset D to the data application module in real time for real-time data display, historical data storage, report generation, and data export functions, supporting synchronization to an occupational health management platform. Simultaneously, the data application module processes the data in dataset D and feeds the processing results back to the anomaly warning module, which triggers a warning signal based on the feedback data. This invention can continuously monitor targets and provides high-precision detection results.

Description

Technical Field

[0001] This invention relates to a dust concentration detection system, specifically a flexible electronic skin integrated dust-breathing dual-modal detection system and method, belonging to the field of dust concentration monitoring and early warning technology. It is particularly suitable for dust exposure environments such as construction sites, mines, and chemical workshops, enabling real-time synchronous detection and data linkage analysis of environmental dust concentration and human respiratory parameters. Background Technology

[0002] In the field of occupational health protection, dust is the core pathogenic factor that induces legally defined occupational diseases such as pneumoconiosis in occupational exposure scenarios. Abnormal fluctuations in respiratory parameters can serve as an early physiological manifestation of the lungs being affected by dust. The synergistic monitoring of the two plays a key supporting role in early warning of occupational health risks.

[0003] Currently, dust concentration detection and respiratory parameter monitoring are separate detection systems, requiring differentiated solutions. Data cannot be linked in real time, making it impossible to achieve dynamic correlation analysis between dust exposure intensity and human respiratory physiological state. Furthermore, traditional detection solutions are not well-suited for long-term continuous operation scenarios, making it difficult to achieve low-interference, high-stability continuous monitoring. Data processing often relies on offline export and analysis, resulting in significant delays in early warning information feedback for excessive dust concentrations or abnormal respiratory parameters, leading to poor timeliness of protective responses. In addition, dust detection has limited accuracy in distinguishing between particles of characteristic sizes such as PM2.5 and PM10, with measurement errors generally exceeding 10%, and is easily affected by fluctuations in environmental temperature and humidity, making it difficult to guarantee the reliability of detection results. The existing technical system can no longer meet the needs of real-time, accurate, and collaborative monitoring of dust and respiratory parameters in occupational health protection scenarios. Summary of the Invention

[0004] The purpose of this invention is to provide a flexible electronic skin integrated dust-breathing dual-modal detection system and method, which can realize a synchronous real-time detection system for dust concentration and breathing parameters, continuously monitor targets, and realize the linkage acquisition and dynamic correlation analysis of the two types of data, with high-precision detection results.

[0005] To achieve the above objectives, this invention provides a flexible electronic skin integrated dust-respiration dual-modal detection system, including a dual-modal sensing module, a signal processing module, a wireless transmission module, a data application module, an anomaly warning module, and a power supply. The dual-modal sensor module includes an independent dust sensor array and a respiration sensor array. The dust sensor array is used to detect the dust concentration signal in the workplace in real time and is supported on a second flexible substrate. A detachable medical Velcro adhesive layer is provided at the edge of the second flexible substrate to fit the contour of the human wrist or neck, and the adhesive layer avoids the detection area of ​​the dust sensor array to prevent obstruction. The respiration sensor array is used to detect the breathing rate signal of workers in the workplace in real time and is supported on a first flexible substrate. A medical-grade silicone adhesive layer is provided at the edge of the first flexible substrate to fit the human chest and ensure a tight fit with the chest skin. The arrays are connected and integrated into the same system via flexible wires. The signal processing module receives dust concentration signals from the dual-modal sensing module and breathing frequency signals from workers, and performs noise reduction, AD conversion, and data fusion processing to obtain dataset D. The wireless transmission module transmits the processed dataset D to the terminal platform in real time. The terminal platform then performs real-time data display, historical data storage, report generation, and data export functions, and supports synchronization to the occupational health management platform. It can achieve data-free transmission within a 10m transmission distance, ensuring data real-time performance. The data application module processes the data in dataset D transmitted by the wireless transmission module in real time and feeds back the processing results to the anomaly warning module. The anomaly warning module presets dust concentration and breathing frequency warning thresholds, which trigger warning signals accordingly. The power supply provides power to the dual-modal sensing module, signal processing module, wireless transmission module, data application module, and anomaly warning module.

[0006] The first and second flexible substrates of this invention are both PDMS / graphene composite flexible materials with a thickness of 0.1-0.3 mm and an elongation of ≥50%. They are provided with a medical-grade silicone adhesive layer at the edge and are adapted to the human chest. They have strong resistance to environmental temperature and humidity interference and can achieve long-term continuous detection without interference.

[0007] The dust sensor array of this invention consists of three surface acoustic wave (SAW) sensors with a ZnO nanofilm as the sensitive layer, operating at a frequency of 2.4-2.5 GHz, detecting particle sizes ranging from 0.3-10 μm, and having a detection error of ≤5%. The respiration sensor array of this invention consists of two piezoresistive sensors with a graphene / PDMS composite film as the sensitive layer, detecting respiration rates of 8-60 breaths / minute and having a response time of ≤100 ms. The respiration sensor array and the dust sensor array are connected to the same signal processing module via flexible wires (made of polyimide-coated copper core wire with a diameter of ≤0.5 mm) with a length of 30-50 cm. The wires can bend with human movement without affecting the detection stability.

[0008] The sensor array of the present invention is fabricated as follows: Dust sensor array: A 500nm thick ZnO nanofilm was deposited in the middle of a flexible substrate using radio frequency magnetron sputtering. SAW sensor electrodes were fabricated using photolithography. The electrodes were made of Al material with a linewidth of 50μm, forming three SAW sensors with a spacing of 5mm. Breathing sensor array: Graphene / PDMS composite slurry (graphene content 1wt%) is coated on both sides of a flexible substrate. After drying, two piezoresistive sensors are formed by laser etching, with a size of 2mm×5mm and a spacing of 8mm.

[0009] The signal processing module of the present invention includes a micro MCU, a signal amplification circuit, and a filtering circuit. The micro MCU, the signal amplification circuit, and the filtering circuit are connected and soldered onto a flexible PCB board with a thickness of 0.1 mm. The flexible PCB board is integrated on a first flexible substrate (chest) and receives signals from the respiratory sensor array and the dust sensor array through wires to achieve synchronous processing.

[0010] The above-mentioned module integration of the present invention is as follows: the respiratory sensor array on the first flexible substrate is electrically connected to the signal processing module integrated on the substrate through conductive silver paste; the dust sensor array on the second flexible substrate is electrically connected to one end of a flexible wire through conductive silver paste, and the other end of the flexible wire is electrically connected to the signal processing module integrated on the first flexible substrate; a Bluetooth 5.0 module (wireless transmission module) with a size of 3mm×5mm, a flexible lithium polymer battery (power supply) with a thickness of 0.3mm and a micro vibration motor (abnormal warning module execution component) with a diameter of 3mm are respectively connected to the signal processing module integrated on the first flexible substrate through wires; a medical-grade silicone adhesive layer with a thickness of 0.1mm (suitable for chest fitting) is pasted on the edge of the first flexible substrate, and a detachable medical Velcro adhesive layer with a thickness of 0.1mm (suitable for wrist / neck fitting, and avoiding the detection area of ​​the dust sensor array) is pasted on the edge of the second flexible substrate.

[0011] A detection method for a flexible electronic skin-integrated dust-breathing dual-modal detection system includes the following steps: S1: Wearing: The user attaches the integrated first flexible electronic skin to the middle of the chest and fixes it with a silicone adhesive layer; the second flexible base is attached to the wrist / neck with a detachable medical Velcro. S2: Dual-modal sensor module acquires data: The dust sensor array detects the characteristic signal of dust concentration in the environment through the SAW sensor, and the breathing sensor array detects the characteristic signal of breathing frequency through the pressure change generated by the deformation of the thoracic cavity, thus obtaining the original dust concentration characteristic signal and breathing frequency characteristic signal; S3: Signal processing module processes the original signal: The original dust concentration characteristic signal and breathing frequency characteristic signal from step S2 are transmitted to the signal processing module. After noise reduction, AD conversion and data fusion processing, the dataset D is obtained. S4: Wireless transmission and feedback: Data set D is transmitted to the data application module through the wireless transmission module, and the data application module processes data set D in real time; S5: Data storage and analysis: The data application module processes the data transmitted by the wireless transmission module in real time and feeds the processing results back to the abnormal warning module. When the parameters exceed the preset threshold range, the vibration motor is triggered until the parameters return to normal. S6: Finally, all historical testing data will be exported to the occupational health management platform for health risk assessment and work environment optimization.

[0012] The processing flow of the signal processing module of this invention is as follows: S31: Raw signal preprocessing: Preprocessing of raw dust concentration characteristic signal: The raw dust concentration characteristic signal (analog quantity) is amplified by a gain of 100-500 times by the signal processing module, then high-frequency noise is removed by a 10Hz low-pass filter (cutoff frequency 10Hz), and finally converted into a digital signal by a 12-bit AD converter. Raw respiratory frequency characteristic signal preprocessing: The strain / pressure analog signal of the respiratory sensing unit (respiratory sensor array) of the dual-modal sensing module is also amplified, filtered by a 10Hz low-pass filter and converted into a digital signal by a 12-bit AD conversion as described above; S32: 12-bit AD conversion calibration: The 12-bit AD conversion accuracy is "full scale / 4096". Combined with a detection error ≤5%, a calibration formula of "AD value - physical quantity" is established through standard source calibration: Dust concentration: Respiratory rate: S33: Outlier Clearance: Thresholds are set based on the detection range of the sensing unit: dust concentrations exceeding the reasonable concentration range corresponding to 0.3-10μm are excluded; breathing rates exceeding 8-60 breaths / minute are excluded. S34: Timing synchronization (ensures time alignment of dual-modal data) Based on the characteristic of the dual-modal sensing module having a response time ≤100ms, time axis unification is achieved: Set the sampling frequency to 10Hz and add a precise timestamp to each set of data; For discrete points with a time deviation > 10ms, linear interpolation is used to map them to a standard time axis, forming a synchronized dataset. , where t i For timestamps, C i For dust concentration, F i This refers to the respiratory rate.

[0013] In step S4 of this invention, the data application module processes the data in the synchronous dataset D transmitted by the wireless transmission module in real time according to the following steps, and feeds back the processing results to the anomaly warning module. Finally, all historical detection data are exported to the occupational health management platform for health risk assessment and work environment optimization. S41: Extracting multi-dimensional features from the synchronized dataset D: Dust feature set : Instantaneous characteristics: real-time dust concentration fine particulate matter concentration ; Short-term statistical characteristics: 1-minute mean ; Dynamic trend characteristics: 5-second concentration change rate Concentration change within 1 minute 150 Duration ; respiratory feature set : Instantaneous characteristics: Real-time respiratory rate ; Short-term statistical characteristics: mean frequency over 5 respiratory cycles ; Dynamic trend characteristics: 5-second frequency change rate ; S42: Constructing a multiple linear regression model: using dust characteristics as independent variables and respiration parameters as dependent variables, quantifying the influence weights: Respiratory rate regression model: in: This is a constant term (baseline respiratory rate, respiratory rate without dust exposure). ω1-ω4 are regression coefficients (reflecting the weight of the influence of the corresponding dust characteristics on the breathing rate). The error is random (following a normal distribution where all values ​​are 0). This is a predicted value for respiratory rate; S43: Judging the Influence Strength Based on Regression Models The impact of dust on respiration is assessed from two dimensions: the "weight of individual features" and the "goodness of fit" of the model, based on the "weight of individual features" and the "intensity of overall impact." Influence weight of single dust characteristics: Sort by the absolute value of the regression coefficient; the larger the absolute value, the stronger the influence of the dust characteristic on respiration parameters. Assuming the respiratory rate model The influence of weight ranking is as follows: ; Determine the direction of influence by combining the signs of the coefficients: For example, a coefficient ω1>0 indicates that an increase in dust characteristics will lead to an increase in respiratory parameters; S44: Explanatory power of dust characteristics for respiration parameters based on model "goodness of fit". Classification into levels, The closer the value is to 1, the stronger the overall impact of dust on respiration. in: The total deviation squared (TDO) refers to the degree of deviation between the actual value of the dependent variable and its mean, reflecting the magnitude of fluctuation in the dependent variable itself. Y1=F avg-5cyc , It is the average of the mean respiratory rates of all samples. ; This is the sum of squared residuals. The residuals are the differences between the actual values ​​of the dependent variable and the model's predicted values, reflecting the variation that the model did not capture. , ; For a strong influence, For the sake of influence, It has a weak impact.

[0014] In step S5 of this invention, the abnormal early warning module converts the model output into an alarm judgment basis, and the early warning signal trigger state continues until the detection parameters recover to the preset threshold range; S51 Result Processing Flow (1) Abnormal threshold comparison (basic judgment) The actual monitored value is compared with the preset threshold of the anomaly warning module, and marked as a basic anomaly state: Dust anomaly marker: , recorded as (Triggered dust threshold), otherwise ; Abnormal breathing markers: , recorded as (Triggering the respiratory threshold), otherwise ; (2) Correlation strength verification (excluding non-dust-induced factors) Combined with the model correlation strength R 2 To determine whether abnormal breathing is caused by dust, and to rule out false alarms caused by non-dust factors such as exercise or emotions: like (Abnormal breathing) requires verification: Model R 2 0.3 (medium impact or above) to determine whether abnormal breathing is caused by dust; This is recorded as (Abnormal breathing is strongly correlated with dust); otherwise (Non-dust-induced respiratory abnormalities); (3) Forecasting and predicting trends (early warning) Based on respiratory rate prediction It can determine whether an anomaly will be triggered in the near future (e.g., within 10 seconds) and provide early warnings. If the current <25 times / minute, but ≥25 times / minute, recorded as (Abnormal breathing is about to occur); If the current But T over-150 The continued increase, combined with model predictions, may lead to An increase, denoted as (Dust exposure will be exacerbated); S52 alarm triggering conditions are linked with the early warning module. Based on the processed judgment indicators , , , This triggers the corresponding signal (1Hz or 2Hz) of the anomaly warning module, according to the following rules: (1) Level 1 alarm (double anomaly + strong correlation, highest priority) Triggering condition: Flag C =1 (dust concentration ≥150μg / m³) 3 )+ Flag F=1 (Respiratory rate ≥25 breaths / minute) + Flag FR =1 (abnormal breathing is mainly caused by dust) + R 2 >0.6 (strong impact); Warning signal: Triggers the "1Hz+2Hz" combined signal (continuous alarm) of the abnormal warning module and simultaneously pushes the text prompt "Dust concentration exceeds the standard and has caused abnormal breathing, intervene immediately"; Basis: Both thresholds were triggered, and the model proved that the abnormal breathing was mainly caused by dust, requiring an emergency response; (2) Level 2 alarm (single anomaly + high risk, medium priority) Triggering condition 1 (abnormal dust + respiratory risk): Flag C =1+Flag F =0 (No abnormal breathing) + Flag F−future =1 (predicting impending respiratory abnormalities) + C 0.3-2.5 ≥100μg / m 3 (High-impact characteristics exceed the standard); Warning signals: 1Hz signal (dust warning) + 2Hz intermittent signal (respiratory risk warning), indicating "dust concentration exceeds the standard, respiratory parameters are about to become abnormal, it is recommended to stay away"; Triggering condition 2 (abnormal breathing + dust association): Flag F =1 + Flag C =0 (Dust levels are not currently exceeding the standard) + Flag FR =1 (Respiratory abnormalities are associated with dust) + (Concentration rises rapidly); Warning signals: 2Hz signal (respiratory warning) + 1Hz intermittent signal (dust risk warning), indicating that "abnormal breathing may be caused by a rapid increase in dust concentration, and attention should be paid to changes in dust." (3) Level 3 warning (potential risk, low priority) Triggering condition: Flag C =0 and Flag F =0, but R 2 ≥0.3 (Medium impact) + Dust concentration change rate (Concentration rises rapidly); Warning signal: 1Hz intermittent signal (dust risk warning), indicating that "dust concentration is rising rapidly, which may affect breathing. It is recommended to strengthen monitoring."

[0015] The power supply described in this invention adopts a low-power design and supports wireless power supply.

[0016] Compared with existing technologies, this invention integrates a dual-modal sensing module, a signal processing module, a wireless transmission module, a data application module, an anomaly warning module, and a power supply onto a composite flexible substrate. The dual-modal sensing module monitors the dust concentration characteristic signals and the breathing frequency characteristic signals of workers in the workplace in real time. The signal processing module receives the raw dust concentration characteristic signals and breathing frequency characteristic signals from the dual-modal sensing module and performs noise reduction and data fusion processing to obtain a fused dataset D. The wireless transmission module transmits the processed dataset D to the data application module in real time. The data application module's terminal platform then performs real-time data display, historical data storage, report generation, and data export functions, supporting synchronization to an occupational health management platform. Simultaneously, the data application module processes the data in dataset D transmitted by the wireless transmission module in real time and feeds the processing results back to the anomaly warning module. The anomaly warning module presets dust concentration and breathing frequency warning thresholds, triggering corresponding warning signals. This invention achieves this through dual-modal sensing... The modal sensing module enables synchronous real-time detection of dust concentration and breathing parameters. The data processing module facilitates the linked acquisition and analysis of these two types of data. The data application module processes the data in dataset D transmitted by the wireless transmission module in real time and feeds the processing results back to the anomaly warning module. The anomaly warning module presets warning thresholds for dust concentration and breathing frequency, triggering corresponding warning signals. The alarm signals of this invention are divided into three levels: Level 1 alarm, the highest priority, alerts staff to intervene immediately when dust concentration exceeds the standard and has caused abnormal breathing; Level 2 alarm, a medium priority, alerts staff to move away when dust concentration exceeds the standard and breathing parameters are about to become abnormal; and Level 3 alarm, a low priority, recommends enhanced monitoring when dust concentration rises rapidly and may affect breathing. This invention enables a synchronous real-time detection system for dust concentration and breathing parameters, continuously monitors targets, and achieves linked acquisition and dynamic correlation analysis of the two types of data, providing highly accurate detection results. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the module connections of the present invention. Detailed Implementation

[0018] The invention will now be further described with reference to the accompanying drawings.

[0019] like Figure 1 As shown, a flexible electronic skin-integrated dust-breathing dual-modal detection system includes an integrated dual-modal sensing module, a signal processing module, a wireless transmission module, a data application module, an anomaly warning module, and a power supply; all of the above modules are integrated into one system. The dual-modal sensor module includes an independent dust sensor array and a respiration sensor array. The dust sensor array is used to detect the dust concentration signal in the workplace in real time and is supported on a second flexible substrate. A detachable medical Velcro adhesive layer is provided at the edge of the second flexible substrate to fit the contour of the human wrist or neck, and the adhesive layer avoids the detection area of ​​the dust sensor array to avoid obstruction. The respiration sensor array is used to detect the breathing rate signal of the workers in the workplace in real time and is supported on a first flexible substrate. A medical-grade silicone adhesive layer is provided at the edge of the first flexible substrate to fit the human chest and ensure a tight fit with the chest skin. The dust sensor array and the respiration sensor array are connected and integrated through flexible wires. Within the same system, the dust sensor array of this invention consists of three surface acoustic wave (SAW) sensors with a ZnO nanofilm as the sensitive layer, operating at a frequency of 2.4-2.5 GHz, detecting particle sizes ranging from 0.3-10 μm, and having a detection error ≤5%. The respiration sensor array of this invention consists of two piezoresistive sensors with a graphene / PDMS composite film as the sensitive layer, detecting respiratory rates of 8-60 breaths / minute and having a response time ≤100 ms. The respiration sensor array and the dust sensor array of this invention are connected to the same signal processing module via flexible wires (made of polyimide-coated copper core wire with a diameter ≤0.5 mm) of 30-50 cm in length. The wires can bend with human movement without affecting detection stability.

[0020] The first and second flexible substrates of this invention are both PDMS / graphene composite flexible materials with a thickness of 0.1-0.3 mm, an elongation of ≥50%, and strong resistance to environmental temperature and humidity interference, enabling long-term continuous detection without interference. The preparation process of the flexible substrate of this invention is as follows: PDMS prepolymer and curing agent are mixed at a mass ratio of 10:1, 0.5 wt% of graphene powder (particle size 5-10 nm) is added, ultrasonically dispersed for 30 min, poured into a mold with a thickness of 0.2 mm, and cured in an oven at 80 °C for 2 h to obtain the PDMS / graphene composite flexible substrate.

[0021] The sensor array of the present invention is fabricated as follows: Dust sensor array: A 500nm thick ZnO nanofilm was deposited in the middle of a flexible substrate using radio frequency magnetron sputtering. SAW sensor electrodes were fabricated using photolithography. The electrodes were made of Al material with a linewidth of 50μm, forming three SAW sensors with a spacing of 5mm. Breathing sensor array: Graphene / PDMS composite slurry (graphene content 1wt%) is coated on both sides of a flexible substrate. After drying, two piezoresistive sensors are formed by laser etching, with a size of 2mm×5mm and a spacing of 8mm.

[0022] The dual-modal sensing module acquires two types of core raw signals: Dust concentration characteristic signal: Real-time dust concentration in the particle size range of 0.3-10μm (unit: μg / m³) 3 This includes the concentration contribution from different particle size ranges (such as the 0.3-2.5μm and 2.5-10μm sub-range concentrations, which are used for subsequent refinement analysis). The respiratory rate characteristic signal of the staff: the dynamic signal generated by periodic respiratory movements (strain / pressure changes of the flexible electronic skin), can be analyzed as: Respiratory rate (F, breaths / minute): calculated from the signal cycle.

[0023] The signal processing module receives the dust concentration characteristic signal from the dual-modal sensing module and the breathing frequency characteristic signal from the worker, and performs noise reduction, AD conversion, and data fusion processing to obtain dataset D. The signal processing module of this invention includes a micro MCU, a signal amplification circuit, and a filtering circuit. The micro MCU, signal amplification circuit, and filtering circuit are connected and soldered onto a flexible PCB board with a thickness of 0.1mm. The flexible PCB board is integrated on a first flexible substrate (chest) and receives signals from the breathing sensor array and the dust sensor array through wires to achieve synchronous processing.

[0024] The wireless transmission module transmits the processed dataset D to the terminal platform of the data application module in real time. The terminal platform then performs real-time data display, historical data storage, report generation, and data export functions, and supports synchronization with the occupational health management platform. The wireless transmission module of this invention can achieve data loss-free transmission within a 10m transmission distance, ensuring data real-time performance. The data application module of this invention can also process the data in dataset D transmitted by the wireless transmission module in real time and transmit the processing results back to the anomaly warning module. The anomaly warning module presets dust concentration and breathing frequency warning thresholds, which trigger warning signals accordingly. The power supply is used to power the dual-modal sensing module, signal processing module, wireless transmission module, data application module, and anomaly warning module.

[0025] The above-mentioned modules of the present invention are integrated as follows: Within the breathing sensing module, a breathing sensor array on a first flexible substrate is electrically connected to a signal processing module integrated on the substrate via conductive silver paste; within the dust sensing module, a dust sensor array on a second flexible substrate is electrically connected to one end of a flexible wire via conductive silver paste, and the other end of the flexible wire is electrically connected to the signal processing module (integrated on the first flexible substrate); a 3mm×5mm Bluetooth 5.0 module, a 0.3mm thick flexible lithium polymer battery, and a 3mm diameter micro-vibration motor (an execution component of the abnormal warning module) are respectively connected to the signal processing module integrated on the first flexible substrate via wires; a 0.1mm thick medical-grade silicone adhesive layer (suitable for chest fitting) is pasted onto the edge of the first flexible substrate, and a 0.1mm thick detachable medical Velcro adhesive layer (suitable for wrist / neck fitting, avoiding the dust sensor array detection area) is pasted onto the edge of the second flexible substrate, forming an electronic flexible material.

[0026] Integrated packaging: A 0.05mm thick polyethylene terephthalate (PET) film is used to encapsulate the surface of the electronic flexible material to protect the internal modules from dust and sweat corrosion.

[0027] The power supply of this invention adopts a low-power design and supports wireless power supply.

[0028] A detection method for a flexible electronic skin-integrated dust-breathing dual-modal detection system includes the following steps: S1: Wearing: The user attaches the integrated first flexible electronic material (flexible electronic skin) to the middle of the chest and fixes it with a silicone adhesive layer; the second flexible base is attached to the wrist / neck with a detachable medical Velcro. S2: Dual-modal sensor module acquires data: The dust sensor array detects dust characteristic signals in the environment through the SAW sensor, and the breathing sensor array detects breathing characteristic signals through pressure changes caused by chest deformation, obtaining the original dust characteristic signals and breathing characteristic signals; S3: Signal processing module processes the original signal: The original dust characteristic signal and respiration characteristic signal from step S2 are transmitted to the signal processing module. After noise reduction, AD conversion and data fusion processing, the dataset D is obtained. S4: Wireless transmission and feedback: Data set D is transmitted to the data application module through the wireless transmission module, and the data application module processes data set D in real time; S5: Data storage and analysis: The data application module processes the data transmitted by the wireless transmission module in real time and feeds the processing results back to the abnormal warning module. When the parameters exceed the preset threshold range, the vibration motor is triggered until the parameters return to normal. S6: Finally, all historical testing data will be exported to the occupational health management platform for health risk assessment and work environment optimization.

[0029] The processing flow of the signal processing module of this invention is as follows: S31: Raw signal preprocessing: Preprocessing of raw dust concentration characteristic signal: The raw dust concentration characteristic signal (analog quantity) is amplified by a gain of 100-500 times by the signal processing module (dynamically adjusted according to the ambient dust concentration: 500 times gain for low concentration and 100 times gain for high concentration to avoid saturation), then filtered by a 10Hz low-pass filter (cutoff frequency 10Hz) to remove high-frequency noise (such as environmental vibration interference), and finally converted into a digital signal by a 12-bit AD converter (accuracy 0.024% FS, full scale corresponds to the maximum concentration under a particle size of 0.3-10μm). Raw respiratory frequency characteristic signal preprocessing: The strain / pressure analog signal of the respiratory sensing unit is also amplified as described above (adjusting the gain according to the respiratory intensity: 500 times for shallow breathing and 100 times for deep breathing), 10Hz low-pass filtered (to filter out high-frequency interference other than respiratory motion), and 12-bit AD conversion to be converted into a digital signal (the period of the periodic signal corresponding to the respiratory frequency is calculated). S32: 12-bit AD conversion calibration: The 12-bit AD conversion accuracy is "full scale / 4096". Combined with a detection error ≤5%, calibration is performed using a standard source (such as a dust generator with known concentration or a standard breathing simulator) to establish a calibration formula of "AD value - physical quantity": Dust concentration: Respiratory rate: S33: Outlier Clearance: Thresholds are set based on the detection range of the sensing unit: dust concentrations exceeding the reasonable concentration range of 0.3-10μm are excluded (in conjunction with an error ≤5%, abnormal points exceeding ±5% deviation are excluded); breathing rates exceeding 8-60 breaths / minute are excluded (considered as sensor false triggering or extreme interference). S34: Timing synchronization (ensures time alignment of dual-modal data) Based on the characteristic of the dual-modal sensing module having a response time ≤100ms, time axis unification is achieved: Set the sampling frequency to 10Hz (generate 1 set of data every 100ms), and attach a precise timestamp to each set of data (accuracy ≤ 10ms, synchronized by the terminal platform clock). For discrete points with a time deviation > 10ms, linear interpolation is used to map them to a standard time axis, forming a synchronized dataset. , where t i For timestamps, Ci For dust concentration, F i This refers to the respiratory rate.

[0030] In step S4 of this invention, the data application module processes the data in the synchronous dataset D transmitted by the wireless transmission module in real time according to the following steps, and feeds back the processing results to the anomaly warning module. Finally, all historical detection data are exported to the occupational health management platform for health risk assessment and work environment optimization. S41: Extracting multi-dimensional features from the synchronized dataset D: Dust feature set : Instantaneous characteristics (dust concentration distribution at a specific moment): real-time dust concentration fine particulate matter concentration ; Short-term statistical characteristics (concentration levels and fluctuations over a short period): 1-minute average ; Dynamic trend characteristics (concentration rise / fall rate and cumulative duration of exceedance (related to anomaly warning threshold)): 5-second concentration change rate Concentration change within 1 minute 150 Duration ; respiratory feature set : Instantaneous characteristics (respiratory state at a specific moment): real-time respiratory rate ; Short-term statistical characteristics (respiratory stability (the smaller the coefficient of variation, the more stable)): mean frequency over 5 respiratory cycles ; Dynamic trend characteristics (rate of increase and decrease in respiratory rate, and ventilation per unit frequency (efficiency indicator)): 5-second rate of change of respiratory rate ; S42: Constructing a multiple linear regression model: using dust characteristics as independent variables and respiration parameters as dependent variables, quantifying the influence weights: Respiratory rate regression model: in: ω1-ω4 are constant terms (baseline respiratory rate, respiratory rate without dust exposure); ω1-ω4 are regression coefficients (reflecting the weight of the influence of the corresponding dust characteristics on the respiratory rate). The error is random (following a normal distribution where all values ​​are 0). This is a predicted value for respiratory rate; S43: Judging the Influence Strength Based on Regression Models The impact of dust on respiration is assessed from two dimensions: the "weight of individual features" and the "goodness of fit" of the model, based on the "weight of individual features" and the "intensity of overall impact." Influence weight of single dust characteristics: Sort by the absolute value of the regression coefficient; the larger the absolute value, the stronger the influence of the dust characteristic on respiration parameters. Example: In the respiratory rate model The influence of weight ranking is as follows: ; Determine the direction of influence by combining the coefficient sign: For example, a coefficient ω1>0 indicates that an increase in dust characteristics will lead to an increase in respiratory parameters (e.g., the higher the concentration, the faster the breathing rate). S44: Explanatory power of dust characteristics for respiration parameters based on model "goodness of fit". Classification into levels, The closer the value is to 1, the stronger the overall impact of dust on respiration. in: The total deviation squared (TDO) refers to the degree of deviation between the actual value of the dependent variable and its mean, reflecting the magnitude of fluctuation in the dependent variable itself. , It is the average of the mean respiratory rates of all samples. ; This is the sum of squared residuals. The residuals are the differences between the actual values ​​of the dependent variable and the model's predicted values, reflecting the variation that the model did not capture (caused by other unconsidered factors). , ; For a strong influence, For the sake of influence, It has a weak impact.

[0031] In step S5 of this invention, the abnormal early warning module converts the model output into an alarm judgment basis, and the early warning signal trigger state continues until the detection parameters recover to the preset threshold range; S51 Result Processing Flow (1) Abnormal threshold comparison (basic judgment) The actual monitored value is compared with the preset threshold of the early warning module and marked as a basic abnormal state: Dust anomaly marker: , recorded as (Triggered dust threshold), otherwise ; Abnormal breathing markers: , recorded as (Triggering the respiratory threshold), otherwise ; (2) Correlation strength verification (excluding non-dust-induced factors) Combined with the model correlation strength R 2 To determine whether abnormal breathing is caused by dust, and to rule out false alarms caused by non-dust factors such as exercise or emotions: like (Abnormal breathing) requires verification: Model R 2 0.3 (medium impact or above) to determine whether abnormal breathing is caused by dust; This is recorded as (Abnormal breathing is strongly correlated with dust); otherwise (Non-dust-induced respiratory abnormalities); (3) Forecasting and predicting trends (early warning) Based on respiratory rate prediction It can determine whether an anomaly will be triggered in the near future (e.g., within 10 seconds) and provide early warnings. If the current <25 times / minute, but ≥25 times / minute, recorded as (Abnormal breathing is about to occur); If the current But T over-150 The continued increase, combined with model predictions, may lead to An increase, denoted as (Dust exposure will be exacerbated); S52 alarm triggering conditions are linked with the early warning module. Based on the processed judgment indicators , , , This triggers the corresponding signal (1Hz or 2Hz) of the anomaly warning module, according to the following rules: (1) Level 1 alarm (double anomaly + strong correlation, highest priority) Triggering condition: Flag C =1 (dust concentration ≥150μg / m³) 3 )+ Flag F =1 (Respiratory rate ≥25 breaths / minute) + Flag FR =1 (abnormal breathing is mainly caused by dust) + R 2> 0.6 (strong impact); Warning signal: Triggers the "1Hz+2Hz" combined signal (continuous alarm) of the abnormal warning module and simultaneously pushes the text prompt "Dust concentration exceeds the standard and has caused abnormal breathing, intervene immediately"; Basis: Both thresholds were triggered, and the model proved that the abnormal breathing was mainly caused by dust, requiring an emergency response; (2) Level 2 alarm (single anomaly + high risk, medium priority) Triggering condition 1 (abnormal dust + respiratory risk): Flag C =1+Flag F =0 (Breathing is currently normal)+ (Predicting impending respiratory abnormality) + C 0.3-2.5 ≥100μg / m 3 (High-impact characteristics exceed the standard); Warning signals: 1Hz signal (dust warning) + 2Hz intermittent signal (respiratory risk warning), indicating "dust concentration exceeds the standard, respiratory parameters are about to become abnormal, it is recommended to stay away"; Triggering condition 2 (abnormal breathing + dust association): Flag F =1+ Flag C =0 (Dust levels are not currently exceeding the standard) + Flag FR =1 (Respiratory abnormalities are associated with dust) + (Concentration rises rapidly); Warning signals: 2Hz signal (respiratory warning) + 1Hz intermittent signal (dust risk warning), indicating that "abnormal breathing may be caused by a rapid increase in dust concentration, and attention should be paid to changes in dust." (3) Level 3 warning (potential risk, low priority) Triggering condition: Flag C =0 and Flag F =0, but R 2 ≥0.3 (Medium impact) + Dust concentration change rate (Concentration rises rapidly); Warning signal: 1Hz intermittent signal (dust risk warning), indicating that "dust concentration is rising rapidly, which may affect breathing. It is recommended to strengthen monitoring."

[0032] The following are embodiments of the present invention. In a certain underground coal mine excavation face, the system of this invention is used for real-time monitoring: dust characteristics are collected by a dust sensor (sampling frequency 1Hz), breathing parameters are collected by a chest strap breathing sensor (sampling frequency 5Hz), and the data is synchronously transmitted to the data application module for multivariate linear regression model calculation. Finally, the system is linked to the early warning module (supporting 1Hz / 2Hz signal output) to realize risk warning.

[0033] The system's data collection and calculation results at a certain time (t=10:23:15) are as follows: Dust characteristics: C inst =160 μg / m 3 , C avg-1min =160 μg / m 3 , C 0.3-2.5 =120 μg / m 3 , , T over-150 =40s; Respiratory characteristics: F inst =26 times / minute, F avg-5cyc =26 times / minute, F pred =25.9 times / minute; Regression coefficients: Goodness of fit: R 2 =0.72 (Strong Influence) Prediction error: The prediction is valid.

[0034] According to the result processing flow of this invention: Dust concentration: C inst =160≥150μg / m 3 Flag C =1 (Dust abnormality); Respiratory rate: F inst =26 ≥ 25 times / minute, mark with Flag F =1 (abnormal breathing).

[0035] Because of Flag F =1, verify association: R 2 =0.72≥0.3, indicating that the abnormal breathing is caused by dust, and that the dust has the strongest effect. Regression coefficient (Increased dust levels lead to a faster breathing rate), hence the flag. FR =1 (abnormal breathing is dominated by dust).

[0036] The current respiratory rate is already above normal; there is no need to predict future abnormalities. =0, dust concentration exceeds the standard. =0.

[0037] Meets Level 1 alarm criteria: Flag C =1+FlagF =1+Flag FR =1+R 2 =0.72≥0.6, therefore the abnormal warning module's "1Hz+2Hz" combined signal is triggered.

[0038] Terminal notification: The mine monitoring terminal displays "10:23:15, dust concentration at the excavation surface is 160 μg / m³". 3 "(Exceeding the standard), respiratory rate 26 breaths / minute (abnormal)", correlation analysis showed that the abnormality was caused by the accumulation of fine particles and high concentrations, and a push notification message was sent: "Dust concentration exceeds the standard and has caused abnormal breathing. Intervene immediately."

Claims

1. A flexible electronic skin-integrated dust-respiration dual-modal detection system, characterized in that, The system includes a dual-modal sensing module, a signal processing module, a wireless transmission module, a data application module, an anomaly warning module, and a power supply. The dual-modal sensing module comprises two independent arrays: a dust sensor array and a respirator sensor array. The dust sensor array detects dust concentration signals in the workplace in real time and is mounted on a second flexible substrate. A detachable medical Velcro adhesive layer is located at the edge of the second flexible substrate, adapting to the contours of the wrist or neck, and the adhesive layer avoids the detection area of ​​the dust sensor array. The respirator sensor array detects the breathing rate signals of workers in the workplace in real time and is mounted on a first flexible substrate. A medical-grade silicone adhesive layer is located at the edge of the first flexible substrate, adapting to the human chest. The dust sensor array and the respirator sensor array are connected and integrated into the same system via flexible wires. The signal processing module receives the dust concentration characteristic signals and the workers' breathing rate characteristic signals from the dual-modal sensing module and performs noise reduction, AD conversion, and data fusion processing to obtain a dataset. D The wireless transmission module will process the dataset. D The data is transmitted in real time to the terminal platform, where it is displayed in real time, historical data is stored, reports are generated, and data is exported. It also supports synchronization with the occupational health management platform. The data application module can also process datasets transmitted by the wireless transmission module in real time. D The system processes the data and transmits the results back to the anomaly warning module. The anomaly warning module has preset warning thresholds for dust concentration and breathing frequency, and triggers warning signals accordingly based on the feedback data. The power supply is used to power the dual-modal sensing module, signal processing module, wireless transmission module, data application module, and anomaly warning module. The dust sensor array consists of three surface acoustic wave sensors with a ZnO nanofilm as the sensitive layer. The operating frequency is 2.4-2.5 GHz, the detection particle size range is 0.3-10 μm, and the detection error is ≤5%. The respiration sensor array consists of two piezoresistive sensors with a graphene / PDMS composite film as the sensitive layer. The detection range includes a breathing rate of 8-60 breaths / minute, and the response time is ≤100 ms. The respiration sensor array and the dust sensor array are connected to the same signal processing module via flexible wires of 30-50 cm in length.

2. The dust-respiration dual-modal detection system integrated with flexible electronic skin according to claim 1, characterized in that, Both the first and second flexible substrates are PDMS / graphene composite flexible materials with a thickness of 0.1-0.3 mm and an elongation of ≥50%.

3. The dust-respiration dual-modal detection system integrated with flexible electronic skin according to claim 2, characterized in that, The signal processing module includes a micro MCU, a signal amplification circuit, and a filtering circuit. The micro MCU, signal amplification circuit, and filtering circuit are connected and soldered onto a flexible PCB board with a thickness of 0.1mm. The flexible PCB board is integrated into the first flexible substrate and receives signals from the breathing sensor array and the dust sensor array through wires to achieve synchronous processing.

4. The dust-respiration dual-modal detection system integrated with flexible electronic skin according to claim 3, characterized in that, The module is integrated as follows: a respiratory sensor array on a first flexible substrate is electrically connected to a signal processing module integrated on the substrate via conductive silver paste; a dust sensor array on a second flexible substrate is electrically connected to one end of a flexible wire via conductive silver paste, and the other end of the flexible wire is electrically connected to the signal processing module integrated on the first flexible substrate; a Bluetooth 5.0 module with dimensions of 3mm×5mm, a flexible lithium polymer battery with a thickness of 0.3mm, and a micro vibration motor with a diameter of 3mm are respectively connected to the signal processing module integrated on the first flexible substrate via wires; a medical-grade silicone adhesive layer with a thickness of 0.1mm is pasted on the edge of the first flexible substrate, and a detachable medical Velcro adhesive layer with a thickness of 0.1mm is pasted on the edge of the second flexible substrate.

5. A detection method for a dust-breathing dual-modal detection system integrated with flexible electronic skin according to claim 4, characterized in that, Includes the following steps: S1: Wearing: The user attaches the integrated first flexible electronic skin to the middle of the chest and fixes it with a silicone adhesive layer; the second flexible base is attached to the wrist / neck with a detachable medical Velcro. S2: Dual-modal sensing module acquires data: The dust sensor array detects the characteristic signal of dust concentration in the environment through the SAW sensor, and the breathing sensor array detects the characteristic signal of the worker's breathing rate through the pressure change generated by the deformation of the thoracic cavity, thus obtaining the original dust concentration characteristic signal and breathing rate characteristic signal; S3: Signal processing module's processing of the original signal: The original dust concentration characteristic signal and breathing frequency characteristic signal from step S2 are transmitted to the signal processing module. After noise reduction, AD conversion, and data fusion processing, a dataset is obtained. D ; S4: Wireless Transmission and Feedback: Dataset D The data is transmitted wirelessly to the data application module, which then processes the dataset in real time. D ; S5: Data storage and analysis: The data application module processes the data transmitted by the wireless transmission module in real time and feeds the processing results back to the abnormal warning module. When the parameters exceed the preset threshold range, the vibration motor is triggered until the parameters return to normal. S6: Finally, all historical testing data will be exported to the occupational health management platform for health risk assessment and work environment optimization.

6. The detection method according to claim 5, characterized in that, The signal processing module's processing flow is as follows: S31: Raw signal preprocessing: Preprocessing of raw dust concentration characteristic signal: The raw dust concentration characteristic signal is amplified by a gain of 100-500 times by the signal processing module, then high-frequency noise is removed by a 10Hz low-pass filter, and finally converted into a digital signal by a 12-bit AD converter; Original respiratory rate characteristic signal preprocessing: The strain / pressure analog signal of the respiratory sensing unit is also amplified, filtered by a 10Hz low-pass filter and converted into a digital signal by a 12-bit AD conversion as described above; S32: 12-bit AD conversion calibration: The 12-bit AD conversion accuracy is "full scale / 4096". Combined with a detection error ≤5%, a calibration formula of "AD value - physical quantity" is established through standard source calibration: Dust concentration: Respiratory rate: S33: Outlier Clearance: Thresholds are set based on the detection range of the sensing module: dust concentrations exceeding the reasonable concentration range corresponding to 0.3-10μm are excluded; breathing rates exceeding 8-60 breaths / minute are excluded. S34: Timing Synchronization Based on the characteristic of the dual-modal sensing module having a response time ≤100ms, time axis unification is achieved: Set the sampling frequency to 10Hz and add a precise timestamp to each set of data; For discrete points with a time deviation > 10ms, linear interpolation is used to map them to a standard time axis, forming a synchronized dataset. ,in t i For timestamps, C i Dust concentration, F i This refers to the respiratory rate.

7. The detection method according to claim 6, characterized in that, In step S4, the data application module processes the synchronized dataset transmitted by the wireless transmission module in real time according to the following steps. D The data is processed and the results are fed back to the anomaly warning module. Finally, all historical detection data is exported to the occupational health management platform for health risk assessment and work environment optimization. S41: From synchronized dataset D Extracting multi-dimensional features: Dust feature set : Instantaneous characteristics: real-time dust concentration fine particulate matter concentration ; Short-term statistical characteristics: 1-minute mean ; dynamic trend Feature: Concentration change rate in 5 seconds Concentration change within 1 minute 150 Duration ; respiratory feature set : Instantaneous characteristics: Real-time respiratory rate ; short term Statistical characteristics: mean frequency over 5 respiratory cycles ; Dynamic trend characteristics: 5-second frequency change rate ; S42: Constructing a multiple linear regression model: using dust characteristics as independent variables and respiration parameters as dependent variables, quantifying the influence weights: Respiratory rate regression model: in: ω1-ω4 is the constant term; ω1-ω4 is the regression coefficient. This is random error; This is a predicted value for respiratory rate; S43: Judging the Influence Strength Based on Regression Models The impact of dust on respiration is assessed from two dimensions: the "weight of individual features" and the "goodness of fit" of the regression coefficients. Influence weight of single dust characteristics: Sort by the absolute value of the regression coefficient; the larger the absolute value, the stronger the influence of the dust characteristic on respiration parameters. Assuming the respiratory rate model | ω 2|=0.5>| ω 1 | = 0.1 > | ω 4 | = 0.08 > | ω If 3| = 0.06, then the weighting order is influenced by: C 0.3-2.5 > C avg-1min > T over-150 >∆ C / ∆ t ; Judging the direction of influence by combining the coefficient sign: coefficient ω1>0 indicates that an increase in dust characteristics will lead to an increase in respiratory parameters; S44: Explanatory power of dust characteristics for respiration parameters based on model "goodness of fit". Classification into levels, The closer the value is to 1, the stronger the overall impact of dust on respiration. in: The total deviation squared (TDO) refers to the degree of deviation between the actual value of the dependent variable and its mean, reflecting the magnitude of fluctuation in the dependent variable itself. , Y 1= F avg-5cyc , It is the average of the mean respiratory rates of all samples. ; This is the sum of squared residuals. The residuals are the differences between the actual values ​​of the dependent variable and the model's predicted values, reflecting the variation that the model did not capture. , ; For a strong influence, For the sake of influence, It has a weak impact.

8. The detection method according to claim 6, characterized in that, The abnormal warning module in step S5 converts the model output into alarm judgment criteria, and the warning signal trigger state continues until the detection parameters recover to the preset threshold range; S51 Result Processing Flow (1) Abnormal threshold comparison The actual monitored value is compared with the preset threshold of the early warning module and marked as a basic abnormal state: Dust anomaly marker: , recorded as ,otherwise ; Abnormal breathing markers: , recorded as ,otherwise ; (2) Correlation strength verification Combined with model correlation strength R 2 To determine whether abnormal breathing is caused by dust, and to rule out false alarms caused by non-dust factors such as exercise or emotional state: like Verification required: Model R 2 0.3, to determine whether abnormal breathing is caused by dust; This is recorded as ;otherwise ; (3) Trend prediction Based on respiratory rate prediction To determine whether anomalies will be triggered in the near future and to provide early warnings: If the current <25 times / minute, but ≥25 times / minute, recorded as ; If the current ,but T over-150 The continued increase, combined with model predictions, may lead to An increase, denoted as ; S52 alarm triggering conditions are linked with the early warning module. Based on the processed judgment indicators , , , This triggers the corresponding signal from the anomaly warning module, with the specific rules as follows: (1) Level 1 alarm Triggering conditions: Flag C =1+ Flag F =1+ Flag FR =1+ R 2 >0.6; Warning signal: Triggers the "1Hz+2Hz" combined signal of the abnormal warning module and simultaneously pushes the text prompt "Dust concentration exceeds the standard and has caused abnormal breathing, intervene immediately"; Basis: Both thresholds were triggered, and the model proved that the abnormal breathing was mainly caused by dust, requiring an emergency response; (2) Level 2 alarm Triggering condition 1: Flag C =1+ Flag F =0+ Flag F-future =1+ C 0.3-2.5 ≥100 μg / m 3 ; Warning signal: 1Hz signal + 2Hz intermittent signal, indicating "dust concentration exceeds the standard, respiratory parameters are about to become abnormal, it is recommended to stay away"; Triggering condition 2: Flag F =1+ Flag C =0+ Flag FR =1+ ∆C / ∆t ≥5 μg / m 3 · s ; Warning signal: 2Hz signal + 1Hz intermittent signal, indicating "abnormal breathing may be caused by a rapid increase in dust concentration, and attention should be paid to changes in dust." (3) Level III early warning Triggering conditions: Flag C =0 and Flag F =0, but R 2 ≥0.3+ dust concentration change rate ∆C / ∆t ≥3 μg / m 3 · s ; Warning signal: 1Hz intermittent signal, indicating "dust concentration is rising rapidly, which may affect breathing. It is recommended to strengthen monitoring".

9. The detection method according to claim 6, characterized in that, The power supply adopts a low-power design and supports wireless power supply.

Images