Mask internal pressure control method and system and respirator

Abstract

A mask internal pressure control method and system and a respirator, relating to the technical field of respirators. The control method comprises: providing a fan on a mask, and when the fan is started, continuously supplying airflow into the mask; determining a correlation between the rotating speed of the fan, a pressure value in the mask, and a respiratory behavior of a user; and in the working process, controlling the rotating speed of the fan on the basis of the pressure value collected in real time and the correlation. When the fan is started, the airflow is continuously supplied into the mask, and the rotating speed of the fan is dynamically adjusted on the basis of the collected pressure value and the determined correlation, so that the system can adjust airflow supply according to a real-time requirement of the user so as to reduce the respiratory resistance to the maximum extent. The pressure control method has the advantages of providing a personalized and real-time ventilation solution that can be adjusted according to a specific respiratory requirement of the user, thereby effectively handling the respiratory resistance.

Description

A method and system for controlling internal pressure of a mask and a respirator Technical Field

[0001] The present invention belongs to the technical field of respirators, and in particular relates to a method and system for controlling the internal pressure of a mask, and a respirator. Background Art

[0002] As protective equipment, respirators filter the air users breathe by setting up filter elements to prevent harmful substances from entering the respiratory tract. The core of the respirator is the filter element, which uses different materials and technologies to filter particulate matter, gas or vapor in the air. Different filter elements can selectively filter different types of pollutants according to the design of the equipment and the use environment. Respirators are widely used in various places such as various types of mining, stone and wood processing, agricultural production, loading and unloading of cargo yards at ports, and many processing companies, chemical production companies, laboratories, etc.

[0003] Different filter materials are often used for different types of filter elements, and these materials have different air permeability. Some high-efficiency filter materials may create greater resistance to the flow of gas, so this resistance needs to be overcome when breathing; in addition, the design of the filter element will also affect the breathing resistance. Some filter elements with more compact and dense designs may cause greater breathing resistance.

[0004] During the use of the above-mentioned respirator, as time goes by, the filter element will gradually be adsorbed by particulate matter or gas, resulting in a gradual increase in filtration resistance. When the filter element reaches a certain degree of adsorption saturation, a new filter element needs to be replaced, otherwise the user will face greater breathing resistance.

[0005] With respect to existing respirator masks, how to solve the problem caused by the above-mentioned breathing resistance has become a technical problem that needs to be solved. Summary of the Invention

[0006] The present invention provides a mask internal pressure control method, system and respirator, which can effectively solve the problems in the background technology.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is:

[0008] A method for controlling internal pressure of a mask, comprising:

[0009] A fan is provided on the mask, and after the fan is started, the air flow is continuously supplied to the interior of the mask;

[0010] Determining the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior;

[0011] During operation, the fan speed is controlled according to the pressure value and the correlation collected in real time.

[0012] Furthermore, determining the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior includes:

[0013] Collecting the air pressure value inside the mask in real time;

[0014] The process of increasing the air pressure value corresponds to an intake process, which includes the user exhaling gas and the blower delivering gas into the mask, and the process of decreasing the air pressure value corresponds to an exhaust process, which includes the user inhaling gas and the blower delivering gas into the mask, wherein the increase or decrease of the air pressure value is determined by comparing with a set reference air pressure value;

[0015] The correlation is set such that during the intake process, the fan speed is reduced, and during the exhaust process, the fan speed is increased.

[0016] Furthermore, the reference air pressure value is introduced into a PID controller, and the PID controller uses the reference air pressure value as a set point, and controls the fan speed according to the degree to which the currently acquired air pressure value deviates from the set point, so as to reduce the degree of deviation.

[0017] Furthermore, the control model of the PID controller is:

[0018]

[0019] in,

[0020] u(t) is the fan speed output by the PID controller;

[0021] e(t) is the error between the set point and the currently acquired air pressure value;

[0022] Kp is the gain parameter of the proportional term;

[0023] Ki is the gain parameter of the integral term;

[0024] Kd is the gain parameter of the differential term;

[0025] is the integral of e(t) over time;

[0026] Expressed as the derivative of e(t) with time;

[0027] Kf is the user's dynamic gain parameter;

[0028] f(t) is a fan speed revision function, the input of which is the user's motion data and the air pressure data inside the mask, and the output is a fan speed revision value.

[0029] Furthermore, the method for determining the fan speed revision function f(t) includes:

[0030] A motion sensor is provided on the mask to collect and gather the user's motion data;

[0031] Collecting the speed data of the fan through the controller of the fan;

[0032] Collecting air pressure data inside the mask;

[0033] Feature extraction is performed based on the motion data, speed data and air pressure data, and the fan speed revision function f(t) is established using the extracted features through a machine learning algorithm.

[0034] Furthermore, feature extraction is performed based on the motion data, speed data, and air pressure data, and the fan speed revision function f(t) is established by a machine learning algorithm using the extracted features, including:

[0035] Extracting features associated with fan speed from the motion data and air pressure data, and extracting features associated with motion and air pressure changes from the speed data;

[0036] Merge the features extracted from different data sources and define the fan speed value to be predicted as the target variable;

[0037] The dataset is divided into training set and test set in proportion and used for training the decision tree model;

[0038] The trained decision tree model is used as the fan speed revision function f(t).

[0039] Furthermore, determining the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior includes:

[0040] Collecting experimental data related to the fan speed, the pressure value inside the mask, and the user's breathing behavior;

[0041] Extracting features from the collected data to obtain a data set, wherein the features are used to reflect the relationship between the fan speed, the pressure value inside the mask, and the user's breathing behavior;

[0042] Establishing a neural network model that describes the relationship between the fan speed, the pressure value inside the mask, and the user's breathing behavior;

[0043] Dividing the data set into a training set and a test set in proportion, and using the training set for the neural network model;

[0044] The trained neural network model is used as a correlation model of the fan speed, the pressure value inside the mask, and the user's breathing behavior.

[0045] Furthermore, the neural network model includes:

[0046] The input layer has three nodes corresponding to the fan speed, the pressure value inside the mask, and the characteristics of the user's breathing behavior;

[0047] Hidden layers, including at least one LSTM layer, are used to capture temporal information in sequence data;

[0048] Output layer, outputs the predicted fan speed;

[0049] Activation function, added after each hidden layer and output layer, introduces nonlinearity;

[0050] When the neural network model is used after training, the currently collected pressure value inside the mask is input and the predicted value of the fan speed is output. The predicted value is used to control the fan speed.

[0051] A mask internal pressure control system adopts the mask internal pressure control method as described above, comprising:

[0052] A fan is provided on the mask and continuously supplies airflow to the interior of the mask after being started;

[0053] a correlation analysis module, determining the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior;

[0054] The control module controls the fan speed according to the pressure value and the correlation collected in real time during operation.

[0055] A respirator comprises a mask and the mask internal pressure control system as described above, wherein the mask internal pressure control system controls the pressure inside the mask.

[0056] The technical solution of the present invention can achieve the following technical effects:

[0057] In this invention, the fan continuously provides airflow into the mask after activation. Based on the collected pressure values ​​and the determined correlation, the fan speed is dynamically adjusted. The system can adjust the airflow supply according to the user's real-time needs to minimize breathing resistance. The advantage of this pressure control method is that it provides a personalized, real-time ventilation solution that can be adjusted according to the specific breathing needs of the user, thereby ensuring effective response to breathing resistance. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0059] FIG1 is a flow chart of a method for controlling internal pressure of a mask;

[0060] FIG2 is a flow chart for determining the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior;

[0061] Figure 3 is a flow chart of PID control;

[0062] Figure 4 is an optimization flow chart of PID control;

[0063] FIG5 is a flow chart of a method for determining a fan speed revision function f(t);

[0064] FIG6 is a flowchart of extracting features based on motion data, speed data, and air pressure data, and establishing a fan speed revision function f(t) using the extracted features through a machine learning algorithm;

[0065] FIG. 7 is a flow chart of another embodiment for determining the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior. DETAILED DESCRIPTION

[0066] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0067] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification are for the purpose of describing specific embodiments only and are not intended to limit the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items. Example 1

[0068] A method for controlling internal pressure of a mask, as shown in FIG1 , comprises:

[0069] A1: Install a fan on the mask. After the fan is started, it continuously supplies air to the inside of the mask.

[0070] A2: Determine the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior;

[0071] A3: During operation, the fan speed is controlled based on the real-time collected pressure values ​​and their correlation.

[0072] In the present invention, by introducing a fan on the mask, the breathing resistance caused by the filter element can be reduced by introducing external airflow. After starting, the fan continuously provides airflow to the inside of the mask to ensure that there is a certain amount of airflow during the user's breathing process, which helps to maintain ventilation. During the actual breathing process, the speed of the fan is dynamically adjusted according to the collected pressure value and the determined correlation. The system can adjust the airflow supply according to the real-time needs of the user to minimize the breathing resistance. The advantage of the above-mentioned pressure control method is that it provides a personalized, real-time ventilation solution that can be adjusted according to the specific breathing needs of the user, thereby ensuring effective response to breathing resistance.

[0073] During the implementation process, when the correlation needs to be determined by the actual working process after the fan is installed, for example, it needs to rely on the collection of actual working data, the order between steps A1 and A2 cannot be changed; when the correlation does not depend on the actual working process, for example, through a set algorithm, the determined correlation can be directly output after the relevant parameters such as the fan and mask are input, then the order between steps A1 and A2 can be reversed, or performed separately.

[0074] As a preferred embodiment of the above, as shown in FIG2 , determining the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior includes:

[0075] B1: Real-time collection of air pressure inside the mask. In this step, a sensor or other device may be used to collect the air pressure inside the mask in real time. The sensor may be placed inside the mask or on a channel connecting the inside and outside of the mask to sense changes in air flow.

[0076] B2: The process of increasing the air pressure value corresponds to the intake process, which includes the user's exhaled gas and the blower's delivery of gas into the mask, and the process of decreasing the air pressure value corresponds to the exhaust process, which includes the user's inhaled gas and the blower's delivery of gas into the mask, wherein the increase or decrease of the air pressure value is determined by comparing it with a set reference air pressure value;

[0077] B3: Set the correlation to reduce the fan speed during intake and increase it during exhaust. This process sets the correlation rules so that the fan's operating state can be adjusted in real time based on changes in air pressure to meet the user's breathing needs.

[0078] Through the above optimization scheme, the system can more accurately follow the user's breathing pattern and adjust the fan speed to provide a more natural and comfortable breathing experience. Such a system can more effectively reduce breathing resistance and improve the user's comfort under the respirator.

[0079] In this preferred embodiment, by setting the reference air pressure value, the system can more flexibly adapt to the needs of different users. The normal breathing patterns of different people may be different, so setting the reference air pressure value can make the system more personalized and adapt to different physiological differences; the setting of the reference air pressure value makes the system adaptable and can cope with changes in breathing needs in different working environments and activity levels. For example, when exercising, the user's breathing pattern may be different from that when at rest. The setting of the reference air pressure value can enable the system to better adapt to these changes.

[0080] As a preferred embodiment of the above, a reference air pressure value is introduced into a PID controller. The PID controller uses the reference air pressure value as a set point and controls the fan speed according to the degree to which the currently collected air pressure value deviates from the set point to reduce the degree of deviation.

[0081] During the implementation process, as shown in Figure 3, the reference air pressure value is set to P1 and the collected air pressure value is P2. In this preferred embodiment, P1 and P2 are used as inputs of the PID controller, and a control signal is output, usually a duty cycle signal, which is used to control the motor drive of the fan to adjust the motor speed, thereby achieving regulation of the internal pressure of the mask.

[0082] As a specific approach, the control model of a traditional PID controller is:

[0083]

[0084] in,

[0085] u(t) is the fan speed output by the PID controller;

[0086] e(t) is the current deviation value, that is, the error between the set point and the currently acquired air pressure value;

[0087] K p is the gain parameter of the proportional term;

[0088] K i is the gain parameter of the integral term;

[0089] K d is the gain parameter of the differential term;

[0090] The above three gain parameters are used to adjust the system's response to the deviation value, integral term and rate of change respectively;

[0091] is the integral of e(t) over time, which is used to consider the long-term cumulative error of the system;

[0092] Expressed as the time derivative of e(t).

[0093] During implementation, the initial resistance value of the filter used in the respirator is often determined. The initial resistance value of the filter is set to P3, as shown in Figure 4. It can also be used as the input of the PID controller so that the system can take this factor into account when adjusting the fan speed.

[0094] In the specific implementation process, the user of the respirator often has dynamic movements, which obviously has a more significant impact on the breathing behavior. In order to fully consider the above influence, as a preferred control model, the control model of the PID controller is:

[0095]

[0096] in,

[0097] u(t) is the fan speed output by the PID controller;

[0098] e(t) is the error between the set point and the currently acquired air pressure value;

[0099] Kp is the gain parameter of the proportional term;

[0100] Ki is the gain parameter of the integral term;

[0101] Kd is the gain parameter of the differential term;

[0102] is the integral of e(t) over time;

[0103] Expressed as the derivative of e(t) with time;

[0104] Kf is the user's dynamic gain parameter;

[0105] f(t) is the fan speed correction function. Its input is the user's motion data and the air pressure data inside the mask, and its output is the fan speed correction value.

[0106] In this preferred embodiment, when Kf and f(t) are introduced into the PID controller, the dynamic gain mechanism is taken into consideration, where f(t) represents a dynamic factor related to the user, which may affect the change of air pressure in the respirator.

[0107] Kf is an adjustment factor that represents the strength of the added dynamic gain. By adjusting the value of Kf, the sensitivity of the system to dynamic factors can be controlled. When selecting Kf, it is necessary to balance the sensitivity and stability of the system. Specifically, the value of Kf needs to be adjusted in actual use based on user feedback and system performance. Specifically:

[0108] A small Kf value makes the system less sensitive to changes in dynamic factors, potentially resulting in a slower system response. However, it provides better stability and is less likely to experience oscillation or unstable behavior. A medium Kf value is moderately sensitive to changes in dynamic factors, balancing system performance in most situations and achieving a good balance between sensitivity and stability, making it suitable for general use cases. A large Kf value makes the system very sensitive to changes in dynamic factors, potentially causing the system to be overly aggressive, resulting in oscillation or unstable behavior. However, it may perform better in scenarios that require rapid adaptation to user changes, but caution is required to prevent system instability.

[0109] In practical applications, the optimal value of Kf can be determined through system simulation, experiments, and user feedback. This adjustment process can help ensure that the system can balance sensitivity and stability in various situations and provide users with comfortable and reliable breathing support.

[0110] As a preferred embodiment of the above embodiment, as shown in FIG5 , a method for determining the fan speed revision function f(t) includes:

[0111] C1: Install a motion sensor on the mask to collect and gather the user's motion data. In this step, selecting a motion sensor suitable for the mask is a key step. This may include an accelerometer, gyroscope, or even a visual sensor to collect the user's motion data in real time.

[0112] C2: Collect fan speed data through the fan controller. Ensure that fan speed data is acquired frequently enough to better capture the dynamic changes of the system. Depending on the response speed of the system, high-frequency data collection can be performed.

[0113] C3: Collecting air pressure data inside the mask; i.e., the data collected in step B1 in the above embodiment;

[0114] C4: Extract features based on motion data, speed data, and air pressure data, and use the extracted features to establish a revised fan speed function f(t) through a machine learning algorithm. During the feature extraction stage, it is necessary to ensure that the selected features can effectively reflect the relationship between motion data, fan speed, and air pressure. Specifically, time series characteristics, frequency domain characteristics, and possible interaction characteristics can be considered. During the data collection process, it is also necessary to ensure that the motion data, speed data, and air pressure data are synchronized in time. This is very important for building an accurate model. If data collection is not performed simultaneously, interpolation or other synchronization methods can be used.

[0115] The traditional PID control model is linear and may have difficulty handling certain nonlinear system responses. In this preferred solution, by introducing the fan speed revision function, the system can better adapt to nonlinear relationships and improve the modeling ability of system complexity; in addition, the traditional PID control model usually relies on pre-set parameters, and the introduction of dynamic gain parameters can be adjusted according to actual conditions during runtime, which can better balance the real-time and stability requirements, especially in real-time systems such as respirators.

[0116] During implementation, the dynamic gain parameter Kf allows the system to adjust its sensitivity based on real-time motion data, air pressure, and other factors. This improves the system's adaptability to varying operating conditions, ensuring stable and effective respiratory support in various usage scenarios. Machine learning models are better able to handle uncertainty and unstructured data, which can be encountered in respirator systems, such as sudden changes in the user's motion pattern or environmental uncertainty.

[0117] As a preferred embodiment of the above, feature extraction is performed based on motion data, speed data, and air pressure data, and the extracted features are used to establish a fan speed revision function f(t) through a machine learning algorithm, as shown in FIG6 , including:

[0118] C41: Extract features associated with fan speed from motion data and air pressure data, and extract features related to motion and air pressure changes from speed data;

[0119] Among them, motion data feature extraction includes motion amplitude and frequency, motion direction and change rate, motion average speed or acceleration, motion dynamic characteristics in time series analysis, etc.; air pressure value data feature extraction includes air pressure value trend and change rate, air pressure value peak and valley value, air pressure value statistical characteristics at different time scales, etc.; speed data feature extraction includes average speed and change rate, speed data spectrum analysis (possibly using Fourier transform, etc.), speed trend and periodic characteristics, etc.

[0120] C42: Merge the features extracted from different data sources and define the fan speed value to be predicted as the target variable. Specifically, merge the features from motion, air pressure, and speed into a single feature vector.

[0121] C43: Divide the dataset into training and test sets in proportion and use them to train the decision tree model;

[0122] C44: Use the trained decision tree model as the fan speed revision function f(t).

[0123] During the training process, it is necessary to ensure that the training set is used for model training and the test set is used to evaluate model performance. You can try to adjust the hyperparameters of the decision tree and perform cross-validation and other methods to improve model performance. Use the test set to evaluate the model to ensure the generalization performance of the model. Integrate the trained and evaluated decision tree model into the system and perform real-time testing to ensure that the model meets real-time requirements in actual use, and monitor model performance regularly.

[0124] In this preferred embodiment, the decision tree model is adopted because it has a clear structure and is easy to understand and explain. This is very important for professionals and end users who need to have a clear understanding of the working principle of the model in the respirator system. The decision tree can effectively capture and process nonlinear relationships. The decision tree has strong adaptability to the interaction between features. In the respirator system, there may be complex relationships between motion data, speed data and air pressure data. The decision tree can better handle this variable relationship, and the decision tree model can output the importance of features to help determine which features are most critical for predicting fan speed, which helps to optimize feature selection and model interpretation. It can handle mixed data types, including numerical and categorical data, which is beneficial for the different types of data that may be contained in the fan speed revision function.

[0125] As another optimization method, as shown in FIG7 , the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior is determined, including:

[0126] D1: Collect experimental data related to the fan speed, the pressure inside the mask, and the user's breathing behavior. This may include the internal pressure of the mask measured at different fan speeds, and data related to the user's breathing behavior, such as respiratory rate and tidal volume.

[0127] D2: Extract features from the collected data to obtain a dataset. The features are used to reflect the relationship between the fan speed, the pressure value inside the mask, and the user's breathing behavior.

[0128] D3: Build a neural network model that describes the relationship between the fan speed, the pressure inside the mask, and the user's breathing behavior;

[0129] D4: Divide the dataset into training and test sets in proportion and use them for training the neural network model;

[0130] D5: The trained neural network model is used as a correlation model between the fan speed, the pressure value inside the mask, and the user's breathing behavior.

[0131] Preferably, the neural network model includes:

[0132] The input layer has three nodes corresponding to the fan speed, the pressure value inside the mask, and the characteristics of the user's breathing behavior;

[0133] Hidden layers, including at least one LSTM layer, are used to capture temporal information in sequence data;

[0134] Output layer, outputs the predicted fan speed;

[0135] Activation function, added after each hidden layer and output layer, introduces nonlinearity;

[0136] When the neural network model is trained and used, the currently collected pressure value inside the mask is input and the predicted value of the fan speed is output. The predicted value is used to control the fan speed.

[0137] In the actual use of respirators, seasonal changes can cause variations in respiratory behavior. For example, changes in temperature and humidity can affect airway comfort and resistance. The LSTM layer can learn and capture these seasonal variations, helping the model better adapt to the dynamic changes in respiratory behavior over time. Internal respiratory resistance within the mask can be affected by multiple factors, and the impact of these factors can have long-term correlations. The LSTM layer can capture long-term dependencies, enabling the model to better understand and predict the long-term dynamic changes in internal respiratory resistance within the mask.

[0138] After the training of the above-mentioned neural network model is completed, the fan speed can be directly predicted by using the currently collected pressure value inside the mask as input. When the model training process uses the three features of fan speed, pressure value inside the mask and user's breathing behavior to establish an association model, the purpose of training is to allow the model to learn the relationship between these three features; when using the model, in order to reduce the cost of the respirator, only the pressure value inside the mask is collected during implementation, because the model has learned the correlation between the pressure value inside the mask and the fan speed during the training process, and the LSTM layer helps the model remember previous information and better adapt to the input at the current moment. The model can predict the fan speed based on the relationship it learned during training. Example 2

[0139] A mask internal pressure control system, using the mask internal pressure control method as described in Example 1, comprising:

[0140] The fan is installed on the mask and continuously supplies air to the inside of the mask after being started;

[0141] Correlation analysis module, which determines the correlation between the fan speed, the pressure value inside the mask, and the user's breathing behavior;

[0142] The control module controls the fan speed according to the real-time collected pressure values ​​and correlations during operation. Example 3

[0143] A respirator includes a mask and the mask internal pressure control system as described in the second embodiment, wherein the mask internal pressure control system controls the pressure inside the mask.

[0144] The technical effects achieved by Examples 2 and 3 are the same as those of Example 1 and will not be described in detail here.

[0145] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for controlling the internal pressure of a face mask, characterized in that, Comprising: A blower is provided on the face mask, and after the blower is started, it continuously supplies air flow into the face mask; determining the correlation among the blower speed, the pressure value inside the face mask, and the user's breathing behavior; During the working process, controlling the blower speed according to the pressure value collected in real time and the correlation; Among them, determining the correlation among the blower speed, the pressure value inside the face mask, and the user's breathing behavior includes: collecting the air pressure value inside the face mask in real time; corresponding the rising process of the air pressure value to the intake process, where the intake process includes the user inhaling gas and the blower sending gas into the face mask, and corresponding the falling process of the air pressure value to the exhaust process, where the exhaust process includes the user exhaling gas and the blower discharging gas from the face mask, and where the rise or fall of the air pressure value is judged by comparing with a set reference air pressure value; setting the correlation as reducing the blower speed during the intake process and increasing the blower speed during the exhaust process; Introducing the reference air pressure value into a PID controller, the PID controller uses the reference air pressure value as a set point and controls the blower speed according to the degree of deviation of the currently collected air pressure value from the set point to reduce the degree of deviation; the control model of the PID controller is: ； Wherein, u(t) is the blower speed output by the PID controller; e(t) is the error between the set point and the currently collected air pressure value; Kp is the gain parameter of the proportional term; Ki is the gain parameter of the integral term; Kd is the gain parameter of the differential term; is the integral of e(t) over time; is expressed as the derivative of e(t) over time; Kf is the dynamic gain parameter of the user; f(t) is a blower speed revision function, with the input being the user's motion data and the air pressure value data inside the face mask, and the output being the blower speed revision value; The determination method of the blower speed revision function f(t) includes: setting a motion sensor on the face mask to collect and gather the user's motion data; collecting the speed data of the blower through the controller of the blower; collecting the air pressure value data inside the face mask; performing feature extraction based on the motion data, speed data, and air pressure value data, and establishing the blower speed revision function f(t) using the extracted features through a machine learning algorithm; Feature extraction is performed based on the motion data, rotational speed data, and air pressure value data, and the fan rotational speed revision function f(t) is established using the extracted features through a machine learning algorithm, including: extracting features associated with the fan rotational speed from the motion data and air pressure value data, and extracting features related to motion and air pressure changes from the rotational speed data; the motion data feature extraction includes motion amplitude and frequency, motion direction and change rate, average velocity or acceleration of motion, and motion dynamic features in time series analysis; the air pressure value data feature extraction includes the trend and change rate of the air pressure value, peak and valley values of the air pressure value, and statistical features of the air pressure value at different time scales; the rotational speed data feature extraction includes average rotational speed and change rate, spectral analysis of the rotational speed data, and trend and periodic features of the rotational speed; the features extracted from different data sources are combined, and the fan rotational speed value to be predicted is defined as the target variable; the data set is divided into a training set and a test set according to a ratio and used for the training of the decision tree model; the trained decision tree model is used as the fan rotational speed revision function f(t). During the training process, it is necessary to ensure that the training set is used for model training, the test set is used for evaluating model performance, the hyperparameters of the decision tree are adjusted, and cross-validation is performed; determine the correlation between the fan rotational speed, the pressure value inside the mask, and the user's breathing behavior, including: collecting experimental data related to the fan rotational speed, the pressure value inside the mask, and the user's breathing behavior; extracting features from the collected data to obtain a data set, and the features are used to reflect the relationship between the fan rotational speed, the pressure value inside the mask, and the user's breathing behavior; establishing a neural network model describing the relationship between the fan rotational speed, the pressure value inside the mask, and the user's breathing behavior; dividing the data set into a training set and a test set according to a ratio and used for the training of the neural network model; the trained neural network model is used as the correlation model between the fan rotational speed, the pressure value inside the mask, and the user's breathing behavior. The neural network model includes: a hidden layer, at least including one LSTM layer, which is used to capture the temporal information in the sequence data.

2. The method for controlling the internal pressure of the face mask according to claim 1, characterized in that The neural network model further includes: An input layer, with three nodes corresponding to the features of the fan rotational speed, the pressure value inside the mask, and the user's breathing behavior respectively. An output layer, which outputs the predicted fan rotational speed; an activation function, added after each hidden layer and output layer, introducing nonlinearity. During the process of using the neural network model after training, the currently collected pressure value inside the mask is input, and the predicted value of the fan rotational speed is output, and the predicted value is used to control the fan rotational speed.

3. A face mask internal pressure control system, characterized in that, Adopt the method for controlling the pressure inside the mask as described in claim 1, including: A fan, arranged on the mask, continuously supplies air flow into the mask after being started. A correlation analysis module, which determines the correlation between the fan rotational speed, the pressure value inside the mask, and the user's breathing behavior; a control module, during the working process, controls the fan rotational speed according to the real-time collected pressure value and the correlation.

4. A respirator, characterized in that, Comprising a face mask, and an internal pressure control system for the face mask as described in claim 3, said internal pressure control system for the face mask controlling the pressure inside the face mask.

Images