A method, system, terminal and medium for controlling a bi-level breathing machine

Abstract

This invention discloses a bilevel ventilator control method, system, terminal, and medium, comprising: after the bilevel ventilator is started, detecting the first differential pressure value and the first pressure value inside the bilevel ventilator in real time at a preset number of detection moments; within each prediction cycle, inputting all the first differential pressure values ​​and all the first pressure values ​​within the current prediction cycle into a respiratory flow prediction model, so that the respiratory flow prediction model outputs the first respiratory flow prediction result corresponding to the current prediction cycle, and controlling and adjusting the ventilation pressure of the bilevel ventilator according to the user's current respiratory status and the first respiratory flow prediction result corresponding to the current prediction cycle. By comprehensively considering pressure data and differential pressure data, this invention provides a more comprehensive understanding of respiratory flow changes, enabling more accurate prediction of respiratory flow and thus optimizing the control of the bilevel ventilator.

Description

Technical Field

[0001] This invention relates to the field of ventilator control, and more particularly to a bilevel ventilator control method, system, terminal, and medium. Background Technology

[0002] In bilevel ventilator operation, the ventilator continuously delivers air via a fan to assist the user's breathing. However, the air pressure required by the user differs between the inhalation and exhalation phases. Therefore, as the user's breathing alternates, the ventilator needs to promptly switch the ventilation pressure level to ensure the user can breathe more smoothly and easily while maintaining sufficient ventilation. Existing trigger control methods typically involve weighted analysis of the ventilator's input parameters and controlling the ventilator based on the analysis results. These results can be trigger outcomes or disease types, allowing the ventilator to provide appropriate treatment based on the trigger outcome or disease type. Alternatively, flow trigger thresholds, pressure trigger thresholds, or time trigger thresholds can be set, switching the ventilation pressure level when the monitored data (flow, pressure, or time) exceeds the threshold. This approach struggles to guarantee the ventilation needs and user comfort of different users. Summary of the Invention

[0003] This invention provides a bilevel ventilator control method, system, terminal, and medium. By comprehensively considering both pressure data and differential pressure data, it is possible to gain a more comprehensive understanding of changes in respiratory flow, thereby more accurately predicting the value of respiratory flow and optimizing the control and adjustment of ventilation pressure on the bilevel ventilator.

[0004] To address the aforementioned technical problems, embodiments of the present invention provide a bilevel ventilator control method, comprising:

[0005] After the bilevel ventilator is started, the first differential pressure value and the first pressure value inside the bilevel ventilator are detected in real time at a number of preset detection times.

[0006] In each prediction cycle, all the first differential pressure values ​​and all the first pressure values ​​in the current prediction cycle are input into the respiratory flow prediction model so that the respiratory flow prediction model outputs the first respiratory flow prediction result corresponding to the current prediction cycle, and controls and adjusts the ventilation pressure of the bilevel ventilator according to the user's current breathing state and the first respiratory flow prediction result corresponding to the current prediction cycle.

[0007] The respiratory flow prediction model is obtained by training a pre-built artificial intelligence model, which includes multiple linear perceptron layers.

[0008] In implementing this embodiment of the invention, after the bilevel ventilator is started, the first differential pressure value and the first pressure value inside the bilevel ventilator are detected in real time at several preset detection moments. The pressure data reflects the airflow intensity within the respiratory system, while the differential pressure data represents the difference between two points during breathing. Then, within each prediction cycle, both the pressure data and the differential pressure data are considered comprehensively. All first differential pressure values ​​and all first pressure values ​​are simultaneously input into the respiratory flow prediction model, providing a more comprehensive understanding of respiratory flow changes and thus more accurately predicting the respiratory flow value, optimizing the control of the bilevel ventilator's ventilation pressure. Furthermore, since each person's breathing situation may differ, the ventilation pressure of the bilevel ventilator is controlled and adjusted based on the user's current breathing state and the first respiratory flow prediction result corresponding to the current prediction cycle. Adjusting the ventilator's ventilation pressure can change the dynamic characteristics of the airflow, such as the magnitude of flow rate and pressure changes, making ventilation smoother, more effective, and more suited to the user's breathing situation. This effectively reduces user discomfort and the risk of potential ventilator-related complications, providing the user with safer and more comfortable respiratory support. Furthermore, multilayer linear perceptrons have stronger expressive power in feature extraction and can better solve nonlinear problems. Therefore, by using an artificial intelligence model that includes multilayer linear perceptrons, relevant features can be extracted from the input first differential pressure value and first pressure value, and respiratory flow can be predicted based on the extracted features.

[0009] As a preferred embodiment, the step of controlling and adjusting the ventilation pressure of the bilevel ventilator based on the user's current respiratory status and the first respiratory flow prediction result corresponding to the current prediction cycle specifically involves:

[0010] When the user is in an inhalation state, it is determined whether the first respiratory flow prediction result corresponding to the current prediction cycle meets the preset expiratory ventilation pressure triggering condition; wherein, the expiratory ventilation pressure triggering condition is that the first respiratory flow prediction result corresponding to the current prediction cycle is less than a first threshold, or the first respiratory flow prediction result corresponding to the current prediction cycle is negative and the first respiratory flow prediction result corresponding to the previous prediction cycle is positive.

[0011] If so, the bilevel ventilator is controlled to trigger expiratory ventilation pressure, and the ventilation pressure of the bilevel ventilator is adjusted to the second threshold.

[0012] If not, the bilevel ventilator is controlled to continue maintaining the current ventilation pressure.

[0013] In a preferred embodiment of the present invention, when the user is in an inspiratory state, if the predicted first respiratory flow rate for the current prediction period is less than the first threshold, or if the predicted first respiratory flow rate for the current prediction period is negative and the predicted first respiratory flow rate for the previous prediction period is positive, it indicates that the user's respiratory state has changed. Therefore, in order to meet the user's breathing needs, the bilevel ventilator also triggers expiratory ventilation pressure and adjusts the ventilation pressure of the bilevel ventilator to the second threshold.

[0014] As a preferred embodiment, when the user is in an inspiratory state, determining whether the predicted first respiratory flow rate for the current prediction cycle meets the preset expiratory ventilation pressure triggering condition specifically involves:

[0015] When the user is in an inhalation state, determine whether the time interval between the current moment and the last time the expiratory ventilation pressure was triggered is less than the fourth threshold.

[0016] If so, the bilevel ventilator is controlled to maintain the current ventilation pressure;

[0017] If not, determine whether the predicted first respiratory flow rate for the current prediction cycle meets the preset expiratory ventilation pressure triggering condition.

[0018] In a preferred embodiment of the present invention, when the user is in an inhalation state, if it is determined that the time interval between the current moment and the last time the expiratory ventilation pressure was triggered is less than a fourth threshold, the bilevel ventilator is controlled to continue to maintain the current ventilation pressure to avoid frequent false triggering.

[0019] As a preferred embodiment, the step of controlling and adjusting the ventilation pressure of the bilevel ventilator based on the user's current respiratory status and the first respiratory flow prediction result corresponding to the current prediction cycle specifically involves:

[0020] When the user is in the exhalation state, it is determined whether the predicted first respiratory flow rate of the current prediction cycle meets the preset inspiratory and ventilatory pressure triggering condition; wherein, the inspiratory and ventilatory pressure triggering condition is that the predicted first respiratory flow rate of the current prediction cycle is greater than a third threshold, or the gradient change of the predicted first respiratory flow rate of the current prediction cycle within a first interval and within a preset time period meets the preset requirements, wherein the preset time period refers to a time period starting from the current moment and including several prediction cycles;

[0021] If so, the bilevel ventilator is controlled to trigger inspiratory and ventilatory pressure, and the ventilatory pressure of the bilevel ventilator is adjusted to the fourth threshold; wherein the second threshold is less than the fourth threshold;

[0022] If not, the bilevel ventilator is controlled to continue maintaining the current ventilation pressure.

[0023] In a preferred embodiment of the present invention, when the user is in an exhalation state, if the predicted first respiratory flow rate corresponding to the current prediction period is greater than the third threshold, or if the gradient change of the predicted first respiratory flow rate corresponding to the current prediction period within the first interval and the preset time period meets the preset requirements, it indicates that the user's current respiratory state has changed. Therefore, in order to meet the user's breathing needs, the bilevel ventilator also triggers the inspiratory and ventilatory pressure, adjusting the ventilatory pressure of the bilevel ventilator to a fourth threshold that is greater than the second threshold.

[0024] As a preferred embodiment, when the user is in an exhalation state, determining whether the predicted first respiratory flow rate corresponding to the current prediction cycle meets the preset inspiratory ventilation pressure triggering condition specifically involves:

[0025] When the user is in an exhalation state, it is determined whether the time interval between the current moment and the last time the inspiratory ventilation pressure was triggered is less than the fourth threshold and whether all the first respiratory flow prediction results between the current detection moment and the last time the inspiratory ventilation pressure was triggered are less than the third threshold.

[0026] If so, the bilevel ventilator is controlled to maintain the current ventilation pressure;

[0027] If not, determine whether the predicted first respiratory flow rate for the current prediction cycle meets the preset inspiratory and ventilatory pressure triggering conditions.

[0028] In a preferred embodiment of the present invention, when the user is in an expiratory state, for inspiratory ventilation pressure triggering, considering that the user has a certain probability of having a relatively long end-expiratory period, during which the user's respiratory flow data fluctuates around 0, if it is determined that the time interval between the current moment and the last time the inspiratory ventilation pressure was triggered is less than a fourth threshold and all first respiratory flow prediction results between the current detection moment and the last time the inspiratory ventilation pressure was triggered are less than a third threshold, then the bilevel ventilator is controlled to continue to maintain the current ventilation pressure, thereby preventing premature triggering and ensuring timely triggering, thus optimizing the ventilation performance of the ventilator.

[0029] As a preferred embodiment, after the bilevel ventilator is started, the first differential pressure value and the first pressure value inside the bilevel ventilator are detected in real time at a preset number of detection moments, specifically as follows:

[0030] After the bilevel ventilator is started, at a number of preset detection times, the internal pressure of the bilevel ventilator is monitored in real time by a differential pressure sensor installed inside the bilevel ventilator to obtain the first differential pressure value inside the bilevel ventilator.

[0031] After the bilevel ventilator is started, at a number of preset detection times, the pressure sensor installed inside the bilevel ventilator is used to detect the internal pressure of the bilevel ventilator in real time and obtain the first internal pressure value of the bilevel ventilator.

[0032] In this case, the time interval between adjacent detection times is the same.

[0033] In a preferred embodiment of the present invention, after the bilevel ventilator is started, at several preset, equally spaced detection times, the internal components of the bilevel ventilator are simultaneously monitored in real time by a differential pressure sensor and a pressure sensor installed inside the bilevel ventilator. The first differential pressure value and the first pressure value inside the bilevel ventilator are obtained. This allows for real-time assessment of whether the ventilation tubing of the bilevel ventilator is unobstructed, whether the valves are working properly, and whether there are any leaks. Furthermore, by observing the changes in the first differential pressure value and the first pressure value, the user's response to the ventilator can be monitored and evaluated, thereby assisting medical staff in judging the current treatment effect and adjusting the treatment plan.

[0034] As a preferred embodiment, the acquisition of the respiratory flow prediction model is specifically as follows:

[0035] The simulated lung is controlled to simulate several breathing modes, and during the simulation, the respiratory flow data of the simulated lung in each breathing mode and the second pressure difference value and second pressure value of the bilevel ventilator in each breathing mode are collected as training datasets.

[0036] Using the training dataset, the pre-built artificial intelligence model is iteratively trained. During each iteration, all the second differential pressure values ​​and all the second pressure values ​​of the bilevel ventilator in each of the breathing modes are input into the current artificial intelligence model, so that the current artificial intelligence model outputs the second respiratory flow prediction result corresponding to each of the breathing modes. Based on the output result of the current artificial intelligence model and the respiratory flow data of the simulated lung in each of the breathing modes, the prediction error value of the current artificial intelligence model is calculated. The iteration ends when the prediction error value of the current artificial intelligence model is less than a fifth threshold, and the current artificial intelligence model is used as the respiratory flow prediction model.

[0037] In a preferred embodiment of the present invention, a simulated lung is used to simulate several breathing modes. During the simulation, respiratory flow data of the simulated lung under various breathing modes, as well as the second differential pressure value and second pressure value of the bilevel ventilator under various breathing modes, are collected to enrich the data diversity of the training dataset. Furthermore, the pre-built artificial intelligence model is iteratively trained using the training dataset until the prediction error value of the artificial intelligence model during the training process is less than a fifth threshold. The iteration ends then, and the current artificial intelligence model is used as the respiratory flow prediction model. This improves the prediction accuracy and prediction efficiency of the respiratory flow prediction model from multiple aspects.

[0038] To address the same technical problem, embodiments of the present invention also provide a dual-level ventilator control system, comprising:

[0039] The detection module is used to detect the first differential pressure value and the first pressure value inside the bilevel ventilator in real time at a preset number of detection times after the bilevel ventilator is started.

[0040] The control module is used to input all the first differential pressure values ​​and all the first pressure values ​​in the current prediction period into the respiratory flow prediction model in each prediction period, so that the respiratory flow prediction model outputs the first respiratory flow prediction result corresponding to the current prediction period, and controls and adjusts the ventilation pressure of the bilevel ventilator according to the user's current respiratory state and the first respiratory flow prediction result corresponding to the current prediction period; wherein, the respiratory flow prediction model is obtained by training a pre-built artificial intelligence model, and the artificial intelligence model includes multiple linear perceptron layers.

[0041] To address the same technical problem, the present invention also provides a terminal, including a processor, a memory, and a computer program stored in the memory; wherein the computer program can be executed by the processor to implement the bilevel ventilator control method.

[0042] To address the same technical problem, the present invention also provides a computer-readable storage medium comprising a stored computer program; wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform the bilevel ventilator control method. Attached Figure Description

[0043] Figure 1 This is a flowchart illustrating a bilevel ventilator control method provided in Embodiment 1 of the present invention.

[0044] Figure 2 This is a schematic diagram of a dual-level ventilator control system provided in Embodiment 1 of the present invention. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Example 1:

[0047] Please refer to Figure 1 This invention provides a bilevel ventilator control method, which includes steps S1 to S2, as detailed below:

[0048] Step S1: After the bilevel ventilator is started, the first differential pressure value and the first pressure value inside the bilevel ventilator are detected in real time at several preset detection times.

[0049] In this embodiment, when the user puts on the mask and turns on the bilevel ventilator, the bilevel ventilator will first maintain a low pressure level for ventilation, and then generate a first pressure difference value and a first pressure value inside the bilevel ventilator every 20ms.

[0050] It should be noted that a BiPAP ventilator mainly consists of three parts: a power adapter, a main control system, and a motor control system. The power adapter converts 220V AC to 24V DC for the entire system. The main control system includes an interactive system, buttons and a display screen, flow detection, and pressure detection. The motor control system primarily controls the brushless motor. Flow and pressure monitoring in the main control system monitor the pressure difference and pressure data of the circulating air. The main control system uses the data obtained from flow and pressure detection to determine the inspiratory and expiratory trigger points and interacts with the motor control system to increase or decrease the pressure. Specifically, the BiPAP ventilator's power supply provides 24V, 12V, 5V, 3.3V, and 2.5V power. The main control system signal flow analysis is as follows: Input signals from buttons and a rotary encoder are detected and displayed on an SDRAM and RGB screen, showing different interfaces. The temperature sensor determines the current humidifier status. Data collected by the pressure and flow sensors determines whether the system is in an inspiratory or expiratory state. The signal flow analysis of the motor control system is as follows: The microcontroller determines the current motor speed through Hall sensors and three-phase current. Based on the target speed, the microcontroller achieves the target speed by controlling the MOSFET driver.

[0051] As a preferred embodiment, step S1 includes steps S11 to S12, each of which is detailed below:

[0052] Step S11: After the bilevel ventilator is started, at a number of preset detection times, the internal pressure of the bilevel ventilator is detected in real time by a differential pressure sensor installed inside the bilevel ventilator to obtain the first differential pressure value inside the bilevel ventilator.

[0053] Step S12: After the bilevel ventilator is started, at a number of preset detection times, the pressure sensor installed inside the bilevel ventilator is used to detect the internal pressure of the bilevel ventilator in real time to obtain the first internal pressure value of the bilevel ventilator.

[0054] In this case, the time interval between adjacent detection times is the same.

[0055] In this embodiment, after the bilevel ventilator is started, a time step is set every 20ms, and each time step is used as a detection moment.

[0056] Step S2: In each prediction cycle, input all first differential pressure values ​​and all first pressure values ​​in the current prediction cycle into the respiratory flow prediction model so that the respiratory flow prediction model outputs the first respiratory flow prediction result corresponding to the current prediction cycle, and controls and adjusts the ventilation pressure of the bilevel ventilator according to the user's current respiratory status and the first respiratory flow prediction result corresponding to the current prediction cycle.

[0057] The respiratory flow prediction model is obtained by training a pre-built artificial intelligence model, which includes multiple linear perceptron layers.

[0058] In this embodiment, the artificial intelligence model is first compiled using the Python language and then built using the PyTorch library. Please refer to... Figure 2 This artificial intelligence model consists of eight neural network layers. Its input consists of eight floating-point numbers, representing the differential pressure and pressure values ​​inside the ventilator at four consecutive time steps. Layers fc1, fc2, fc3, fc4, fc5, and fc6 represent linear perceptron layers (or fully connected layers) in the AI ​​model. ReLU represents an activation function that transforms the output of the previous neural network layer by setting values ​​less than zero to zero and leaving values ​​greater than or equal to zero unchanged. Specifically, fc1 and fc2 output 16 floating-point numbers, fc3 and fc4 output 32 floating-point numbers, fc5 outputs 8 floating-point numbers, and f6 outputs 1 floating-point number. This final floating-point number is then used as the user's predicted respiratory flow.

[0059] As an example, the input is X = [△P1, P1, △P2, P2, △P3, P3, △P4, P4], where △P is the pressure difference, P is the pressure, and X is a one-dimensional array composed of pressure difference and pressure data from four time steps. After passing through a linear sensing layer, the calculation Y = W*XT + B is performed, where W and B are the corresponding weight matrices, and Y is still a one-dimensional array. After passing through multiple linear sensing layers and activation functions, the model outputs a number Y, which is the predicted respiratory flow rate of the user at that moment.

[0060] As a preferred approach, the process for obtaining the respiratory flow prediction model includes steps S01 to S02, each of which is detailed below:

[0061] Step S01: Control the simulated lung to simulate several breathing modes respectively, and collect the respiratory flow data of the simulated lung in various breathing modes, as well as the second pressure difference value and second pressure value of the bilevel ventilator in various breathing modes, as training datasets during the simulation.

[0062] In this embodiment, the simulated lung simulates 18 different breathing patterns.

[0063] It should be noted that the BiPAP machine uses an internal fan to transfer outside air to the bellows. During this transfer, differential pressure and pressure sensors monitor the pressure difference and pressure data of the airflow. After passing through the bellows, the air goes through a perforated tube connected to the external atmospheric pressure. As the air passes through this tube, some air flows out to the outside through the perforation, while the rest is inhaled by the simulated lung (user). At this time, the flow sensor in the simulated lung records the flow rate data of the inhaled air (inhalation is positive, exhalation is negative). The bellows can be a 1.5m medical bellows, the perforated tube can be a plastic tube with a 1cm orifice diameter, a length of 4cm, and a perforation diameter of 2mm, and the simulated lung can be an ASL5000 simulated lung, which is used to simulate the breathing state of different users.

[0064] Step S02: Using the training dataset, iteratively train the pre-built artificial intelligence model. During each iteration, input all second differential pressure values ​​and all second pressure values ​​of the bilevel ventilator in each breathing mode into the current artificial intelligence model so that the current artificial intelligence model outputs the second respiratory flow prediction results corresponding to each breathing mode. Based on the output results of the current artificial intelligence model and the respiratory flow data of the simulated lung in each breathing mode, calculate the prediction error value of the current artificial intelligence model. The iteration ends when the prediction error value of the current artificial intelligence model is less than the fifth threshold, and the current artificial intelligence model is used as the respiratory flow prediction model.

[0065] In this embodiment, the fifth threshold is 2.

[0066] As a preferred embodiment, step S2 includes steps S21 to S27, each of which is detailed below:

[0067] Step S21: In each prediction cycle, input all first differential pressure values ​​and all first pressure values ​​in the current prediction cycle into the respiratory flow prediction model so that the respiratory flow prediction model outputs the first respiratory flow prediction result corresponding to the current prediction cycle.

[0068] In this embodiment, a prediction cycle includes four detection times, which means that during the respiratory flow prediction process, every four consecutive sets of first differential pressure values ​​and first pressure values ​​should be used as inputs to the respiratory flow prediction model.

[0069] Step S22: When the user is in an inhalation state, determine whether the prediction result of the first respiratory flow corresponding to the current prediction cycle meets the preset expiratory ventilation pressure trigger condition; if yes, proceed to step S23; if no, proceed to step S24.

[0070] As a preferred embodiment, step S22 includes steps S221 to S223, each of which is detailed below:

[0071] Step S221: When the user is in an inhalation state, determine whether the time interval between the current moment and the last time the expiratory ventilation pressure was triggered is less than the fourth threshold; if yes, proceed to step S222; if no, proceed to step S223.

[0072] Step S222: Control the bilevel ventilator to maintain the current ventilation pressure.

[0073] In this embodiment, the fourth threshold is 200ms. That is, if the number of time steps between the current time step and the time step corresponding to the last time the expiratory ventilation pressure was triggered is less than 10 (one time step occupies 20ms), then no trigger determination is made, and the bilevel ventilator continues to maintain the current ventilation pressure.

[0074] Step S223: Determine whether the predicted first respiratory flow rate for the current prediction cycle meets the preset expiratory ventilation pressure triggering condition.

[0075] The expiratory ventilation pressure triggering condition is that the predicted first respiratory flow rate for the current prediction period is less than the first threshold, or the predicted first respiratory flow rate for the current prediction period is negative and the predicted first respiratory flow rate for the previous prediction period is positive.

[0076] In this embodiment, the first threshold is -3 (L / min).

[0077] Step S23: Control the bilevel ventilator to trigger expiratory ventilation pressure and adjust the ventilation pressure of the bilevel ventilator to the second threshold (i.e., when the user changes from inhalation to exhalation, the bilevel ventilator switches to low level ventilation pressure).

[0078] Step S24: Control the bilevel ventilator to maintain the current ventilation pressure.

[0079] Step S25: When the user is in the exhalation state, determine whether the prediction result of the first respiratory flow corresponding to the current prediction cycle meets the preset inspiratory and ventilatory pressure triggering condition; if yes, proceed to step S26; if no, proceed to step S27.

[0080] As a preferred embodiment, step S25 includes steps S251 to S253, each of which is detailed below:

[0081] Step S251: When the user is in the exhalation state, determine whether the time interval between the current time and the last time the inspiratory ventilation pressure was triggered is less than the fourth threshold and whether all the first respiratory flow prediction results between the current detection time and the last time the inspiratory ventilation pressure was triggered are less than the third threshold; if yes, then proceed to step S252; if no, then proceed to step S253.

[0082] Step S252: Control the bilevel ventilator to maintain the current ventilation pressure.

[0083] In this embodiment, the third threshold is 3 (L / min), and the fourth threshold is 200ms. That is, if the number of time steps between the current time step and the time step corresponding to the last inspiratory pressure trigger is less than 10 (one time step occupies 20ms), and all first respiratory flow prediction results between the current detection time and the last inspiratory pressure trigger are less than 3 (L / min), then no trigger determination is made, and the bilevel ventilator continues to maintain the current ventilation pressure.

[0084] Step S253: Determine whether the predicted first respiratory flow rate corresponding to the current prediction cycle meets the preset inspiratory ventilation pressure triggering condition.

[0085] The inspiratory ventilation pressure triggering condition is that the predicted first respiratory flow rate for the current prediction cycle is greater than the third threshold, or the gradient change of the predicted first respiratory flow rate for the current prediction cycle within the first interval and within the preset time period meets the preset requirements. The preset time period refers to the time period starting from the current moment and including several prediction cycles.

[0086] Step S26: Control the bilevel ventilator to trigger the inspiratory and ventilatory pressure, and adjust the ventilatory pressure of the bilevel ventilator to the fourth threshold.

[0087] The second threshold is less than the fourth threshold. Specifically, when the user is inspiratory, if the bilevel ventilator is triggered by expiratory pressure, the bilevel ventilator will lower the ventilation pressure to the second threshold and maintain it. When the user is expiratory, if the bilevel ventilator is triggered by inspiratory pressure, the bilevel ventilator will raise the ventilation pressure to the fourth threshold and maintain it. Following this process, inspiratory and expiratory triggers alternate, continuously supplying air to the user and helping them breathe more easily and smoothly.

[0088] Step S27: Control the bilevel ventilator to maintain the current ventilation pressure.

[0089] As an example, the inspiratory ventilation pressure triggering process is as follows: When the user is in the expiratory state, if the predicted first respiratory flow rate for the current prediction cycle is greater than 3 (L / min), the bilevel ventilator triggers inspiratory ventilation pressure. If the predicted first respiratory flow rate for the current prediction cycle is greater than 0.5 (L / min) and less than or equal to 3 (L / min), the gradient change of the 13 consecutive time steps from the current time step is observed. If the gradient change of the last two time steps is greater than that of the first 11 time steps and the gradient change of the 11th time step is greater than that of the first 10 time steps, it indicates that the user's respiratory flow rate is increasing, and the ventilator triggers inspiratory ventilation pressure. If the number of time steps between the current time step and the time step corresponding to the last inspiratory ventilation pressure trigger is less than 10 (one time step occupies 20ms), and all predicted first respiratory flow rates between the current detection time and the last inspiratory ventilation pressure trigger are less than 3 (L / min), then no triggering determination is made.

[0090] Please refer to Figure 2 This is a schematic diagram of a bilevel ventilator control system provided in an embodiment of the present invention. The system includes a detection module M1 and a control module M2, and the specific details of each module are as follows:

[0091] The detection module M1 is used to detect the first differential pressure value and the first pressure value inside the bilevel ventilator in real time at a number of preset detection times after the bilevel ventilator is started.

[0092] The control module M2 is used to input all first differential pressure values ​​and all first pressure values ​​within the current prediction period into the respiratory flow prediction model in each prediction period, so that the respiratory flow prediction model outputs the first respiratory flow prediction result corresponding to the current prediction period, and controls and adjusts the ventilation pressure of the bilevel ventilator according to the user's current respiratory status and the first respiratory flow prediction result corresponding to the current prediction period; wherein, the respiratory flow prediction model is obtained by training a pre-built artificial intelligence model, and the artificial intelligence model includes multiple linear perceptron layers.

[0093] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0094] Additionally, embodiments of the present invention also provide a computer-readable storage medium, which includes a stored computer program; wherein, when the computer program is running, it controls the device where the computer-readable storage medium is located to execute a bilevel ventilator control method according to Embodiment 1.

[0095] Additionally, this embodiment of the invention also provides a terminal, including a processor, a memory, and a computer program stored in the memory; wherein the computer program can be executed by the processor to implement a bilevel ventilator control method of Embodiment 1.

[0096] Preferably, the computer program can be divided into one or more modules / units (such as a computer program, a computer program), and one or more modules / units are stored in memory and executed by a processor to complete the present invention. One or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in a terminal.

[0097] The processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor, or any conventional processor. The processor is the control center of the terminal, connecting various parts of the terminal through various interfaces and lines.

[0098] The memory mainly consists of a program storage area and a data storage area. The program storage area can store the operating system, applications required for at least one function, etc., while the data storage area can store related data. Furthermore, the memory can be high-speed random access memory, or non-volatile memory, such as plug-in hard drives, smart media cards (SMC), secure digital cards (SD), and flash cards, or other volatile solid-state storage devices.

[0099] It should be noted that the above-mentioned terminal may include, but is not limited to, processors and memory. Those skilled in the art will understand that the above-mentioned terminal is merely an example and does not constitute a limitation on the terminal. It may include more or fewer components, or combine certain components, or different components.

[0100] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0101] This invention provides a bilevel ventilator control method, system, terminal, and medium. After the bilevel ventilator is activated, the first differential pressure value and the first pressure value inside the bilevel ventilator are detected in real time at several preset detection moments. The pressure data reflects the airflow intensity within the respiratory system, while the differential pressure data represents the difference between two points during breathing. Then, within each prediction cycle, both the pressure data and the differential pressure data are comprehensively considered, and all the first differential pressure values ​​and all the first pressure values ​​are simultaneously input into the respiratory flow prediction model. This allows for a more comprehensive understanding of respiratory flow changes, thereby more accurately predicting the respiratory flow value and optimizing the control of the bilevel ventilator's ventilation pressure. Furthermore, since each person's breathing situation may differ, the ventilation pressure of the bilevel ventilator is controlled and adjusted based on the user's current breathing state and the first respiratory flow prediction result corresponding to the current prediction cycle. Adjusting the ventilator's ventilation pressure can change the dynamic characteristics of the airflow, such as the magnitude of flow rate and pressure changes, making ventilation smoother, more effective, and more suited to the user's breathing situation. This effectively reduces user discomfort and the risk of potential ventilator-related complications, providing users with safer and more comfortable respiratory support. Furthermore, multilayer linear perceptrons have stronger expressive power in feature extraction and can better solve nonlinear problems. Therefore, by using an artificial intelligence model that includes multilayer linear perceptrons, relevant features can be extracted from the input first differential pressure value and first pressure value, and respiratory flow can be predicted based on the extracted features.

[0102] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A bilevel ventilator control system, characterized in that, Includes a detection module and a control module: The detection module is used to detect the first differential pressure value and the first pressure value inside the bilevel ventilator in real time at a preset number of detection times after the bilevel ventilator is started. The control module is used to input all the first differential pressure values ​​and all the first pressure values ​​within the current prediction period into the respiratory flow prediction model in each prediction period, so that the respiratory flow prediction model outputs the first respiratory flow prediction result corresponding to the current prediction period, and controls and adjusts the ventilation pressure of the bilevel ventilator according to the user's current respiratory state and the first respiratory flow prediction result corresponding to the current prediction period; wherein, the respiratory flow prediction model is obtained by training a pre-built artificial intelligence model, and the artificial intelligence model includes multiple linear perceptron layers; The system also includes a respiratory flow prediction model acquisition module, which comprises a ninth unit and a tenth unit. The ninth unit is used to control a simulated lung to simulate several breathing modes, and during the simulation, it collects respiratory flow data of the simulated lung in each breathing mode, as well as the second differential pressure value and the second pressure value of the bilevel ventilator in each breathing mode, as a training dataset. The tenth unit is used to iteratively train a pre-built artificial intelligence model using the training dataset. In each iteration, it inputs all the second differential pressure values ​​and all the second pressure values ​​of the bilevel ventilator in each breathing mode into the current artificial intelligence model, so that the current artificial intelligence model outputs the second respiratory flow prediction result corresponding to each breathing mode. Based on the output result of the current artificial intelligence model and the respiratory flow data of the simulated lung in each breathing mode, it calculates the prediction error value of the current artificial intelligence model. The iteration ends when the prediction error value of the current artificial intelligence model is less than a fifth threshold, and the current artificial intelligence model is used as the respiratory flow prediction model.

2. The bilevel ventilator control system according to claim 1, characterized in that, The control module includes a first unit, a second unit, and a third unit; The first unit is used to determine whether the predicted first respiratory flow rate of the current prediction cycle meets the preset expiratory ventilation pressure triggering condition when the user is in an inhalation state; wherein, the expiratory ventilation pressure triggering condition is that the predicted first respiratory flow rate of the current prediction cycle is less than a first threshold, or the predicted first respiratory flow rate of the current prediction cycle is negative and the predicted first respiratory flow rate of the previous prediction cycle is positive. The second unit is used to control the bilevel ventilator to trigger expiratory ventilation pressure if the condition is met, and adjust the ventilation pressure of the bilevel ventilator to a second threshold. The third unit is used to control the bilevel ventilator to continue maintaining the current ventilation pressure if not.

3. The bilevel ventilator control system according to claim 2, characterized in that, The first unit includes a first subunit, a second subunit, and a third subunit; The first subunit is used to determine whether the time interval between the current moment and the last time the expiratory ventilation pressure was triggered is less than the fourth threshold when the user is in an inhalation state. The second subunit is configured to, if so, control the bilevel ventilator to continue maintaining the current ventilation pressure; The third subunit is used to determine, if not, whether the prediction result of the first respiratory flow corresponding to the current prediction cycle meets the preset expiratory ventilation pressure triggering condition.

4. A bilevel ventilator control system according to claim 2, characterized in that, The control module further includes a fourth unit, a fifth unit, and a sixth unit; The fourth unit is used to determine whether the predicted first respiratory flow rate of the current prediction cycle meets the preset inspiratory and ventilatory pressure triggering conditions when the user is in an exhalation state. The inspiratory and ventilatory pressure triggering conditions are that the predicted first respiratory flow rate of the current prediction cycle is greater than a third threshold, or the gradient change of the predicted first respiratory flow rate of the current prediction cycle within a first interval and a preset time period meets the preset requirements. The preset time period refers to a time period starting from the current moment and including several prediction cycles. The fifth unit is used to, if so, control the bilevel ventilator to trigger the inspiratory and ventilatory pressure, and adjust the ventilatory pressure of the bilevel ventilator to a fourth threshold; wherein the second threshold is less than the fourth threshold; The sixth unit is used to control the bilevel ventilator to continue maintaining the current ventilation pressure if not.

5. A bilevel ventilator control system according to claim 4, characterized in that, The fourth unit includes a fourth subunit, a fifth subunit, and a sixth subunit; The fourth subunit is used to determine, when the user is in an exhalation state, whether the time interval between the current moment and the last time the inspiratory ventilation pressure was triggered is less than a fourth threshold and whether all the first respiratory flow prediction results between the current detection moment and the last time the inspiratory ventilation pressure was triggered are less than the third threshold. The fifth subunit is used to control the bilevel ventilator to continue maintaining the current ventilation pressure if the condition is met. The sixth subunit is used to determine, if not, whether the prediction result of the first respiratory flow corresponding to the current prediction cycle meets the preset inspiratory ventilation pressure triggering condition.

6. A bilevel ventilator control system according to claim 1, characterized in that, The detection module includes a seventh unit and an eighth unit; The seventh unit is used to detect the interior of the bilevel ventilator in real time at several preset detection times after the bilevel ventilator is started, by using a differential pressure sensor installed inside the bilevel ventilator, to obtain the first differential pressure value inside the bilevel ventilator. The eighth unit is used to detect the internal pressure of the bilevel ventilator in real time at several preset detection times after the bilevel ventilator is started, by means of a pressure sensor installed inside the bilevel ventilator, and to obtain the first pressure value inside the bilevel ventilator; wherein the time interval between adjacent detection times is the same.

7. A terminal, characterized in that, It includes a processor, a memory, and a computer program stored in the memory; wherein the computer program can be executed by the processor to perform the functions of each module in the bilevel ventilator control system as described in any one of claims 1 to 6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program; wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform the functions of each module in the bilevel ventilator control system as described in any one of claims 1 to 6.

Images