Continuous positive pressure ventilation system and method based on self-adaptive airflow compensation

Abstract

The invention discloses a continuous positive pressure ventilation system and method based on self-adaptive airflow compensation, and relates to the technical field of medical breathing auxiliary equipment, and the system comprises a pressure generation module which is used for generating continuous positive pressure gas flow; the multi-mode sensor group is used for monitoring respiratory airflow parameters of a patient in real time; the data processing module is connected with the multi-mode sensor group and is used for carrying out breathing waveform analysis according to the breathing airflow parameters, calculating a required airflow pressure compensation value according to an analysis result and generating a corresponding pressure compensation instruction; and the self-adaptive pressure regulation and control module is respectively connected with the data processing module and the pressure generation module and is used for dynamically regulating the pressure of the positive pressure gas flow output by the pressure generation module according to the pressure compensation instruction. The breathing airflow change of the patient can be sensed in real time, the airflow demand is predicted, the output airflow pressure is dynamically adjusted, the treatment effect is optimized, and the comfort of the patient is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical respiratory assistance devices, and particularly to a continuous positive airway pressure system and method based on adaptive airflow compensation. Background Art

[0002] In the related art, continuous positive airway pressure (CPAP) devices usually adopt a fixed pressure or a static pressure regulation mode based on simple breathing events (such as snoring, airflow limitation), and have the following defects: pressure delay, there is a lag in the response to breathing events, resulting in limited treatment effects; airflow fluctuation interference, when the patient's breathing pattern changes dynamically (such as body position change, sleep stage transition), the fixed pressure is likely to cause overshoot or insufficiency of airflow, resulting in patient discomfort; lack of personalized compensation, the dynamic changes of individual physiological parameters such as the patient's airway resistance and lung compliance are not considered.

[0003] Therefore, there is an urgent need for an intelligent continuous positive airway pressure system that can sense the changes in the patient's breathing airflow in real time, predict the airflow demand and dynamically adjust the output pressure, so as to optimize the treatment effect and improve the patient's comfort. Summary of the Invention

[0004] In view of this, the present invention provides a continuous positive airway pressure system and method based on adaptive airflow compensation to solve the technical problems existing in the related art.

[0005] In a first aspect, the present invention provides a continuous positive airway pressure system based on adaptive airflow compensation, including:

[0006] A pressure generation module for generating a continuous positive pressure gas flow;

[0007] A multi-modal sensor group for real-time monitoring of the patient's breathing airflow parameters;

[0008] A data processing module connected to the multi-modal sensor group for performing breathing waveform analysis based on the breathing airflow parameters, calculating the required airflow pressure compensation value according to the analysis result, and generating a corresponding pressure compensation instruction;

[0009] An adaptive pressure regulation module connected to the data processing module and the pressure generation module respectively for dynamically adjusting the pressure of the positive pressure gas flow output by the pressure generation module according to the pressure compensation instruction.

[0010] In an optional embodiment, the multi-modal sensor group includes:

[0011] A pressure sensor for real-time measurement of the pressure value at the patient's airway opening;

[0012] A flow sensor for real-time measurement of the inhaled gas flow and exhaled gas flow on the affected side;

[0013] A physiological sensor for real-time measurement of a patient's blood oxygen saturation and the phase difference of chest and abdomen movement;

[0014] A temperature and humidity sensor for real-time measurement of the temperature and humidity in a patient's airway.

[0015] In an optional embodiment, the data processing module includes:

[0016] A respiratory waveform generation unit, connected to the multi-modal sensor group, for converting the respiratory airflow parameters monitored by the multi-modal sensor group into electrical signals, preprocessing the electrical signals, and drawing a respiratory waveform using the preprocessed electrical signals;

[0017] A characteristic parameter extraction unit, connected to the respiratory waveform generation unit, for extracting key characteristic parameters from the respiratory waveform;

[0018] A waveform analysis unit, connected to the characteristic parameter extraction unit, for evaluating the patient's respiratory state, the working state of the ventilator, and the presence of abnormal waveforms based on the key characteristic parameters;

[0019] An airflow pressure compensation value calculation unit, connected to the characteristic parameter extraction unit, for calculating an airflow pressure compensation value using a control algorithm based on the key characteristic parameters and generating a corresponding pressure compensation instruction.

[0020] In an optional embodiment, the control algorithm uses a hybrid neural network model composed of a convolutional neural network, a Transformer model, and a fully connected layer to calculate the airflow pressure compensation value and generate a corresponding pressure compensation instruction.

[0021] In an optional embodiment, the control algorithm uses a respiratory phase recognition algorithm and an LSTM neural network to calculate the airflow pressure compensation value and generate a corresponding pressure compensation instruction.

[0022] In an optional embodiment, it further includes:

[0023] A pressure release module, disposed at the patient airway interface end and connected to the adaptive pressure regulation module, for automatically releasing pressure when it detects that the patient airway pressure exceeds a preset threshold.

[0024] In an optional embodiment, the multi-modal sensor group further includes:

[0025] An auxiliary pressure sensor, disposed at the proximal end of the patient airway and connected to the data processing module, for compensating for the pressure attenuation caused by the transmission delay of positive pressure gas along the airway.

[0026] In an alternative embodiment, the system is applied to a single-level ventilator, a dual-level ventilator, or other non-invasive ventilators that support the continuous positive airway pressure (CPAP) mode.

[0027] In a second aspect, the present invention provides a method for continuous positive airway pressure based on adaptive airflow compensation, which is applied to a continuous positive airway pressure system based on adaptive airflow compensation. The method includes:

[0028] Real-time monitoring of the patient's respiratory airflow parameters;

[0029] Performing respiratory waveform analysis based on the respiratory airflow parameters, calculating the required airflow pressure compensation value according to the analysis result, and generating a corresponding pressure compensation instruction;

[0030] Dynamically adjusting the pressure of the output positive pressure gas flow according to the pressure compensation instruction.

[0031] In the embodiments of the present invention, the multi-modal sensor group is used to sense the changes in the patient's respiratory airflow in real time, and the machine learning algorithm is used to predict the airflow demand and dynamically adjust the output airflow pressure, thereby optimizing the treatment effect and improving the patient's comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the description of the specific embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0033] Figure 1 is a schematic structural diagram of a continuous positive airway pressure system based on adaptive airflow compensation according to an embodiment of the present invention;

[0034] Figure 2 is a schematic flowchart of a method for continuous positive airway pressure based on adaptive airflow compensation according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0036] Figure 1Schematic diagram of a continuous positive airway pressure system based on adaptive airflow compensation according to an embodiment of the present invention.

[0037] As Figure 1 shown, the present invention provides a continuous positive airway pressure system based on adaptive airflow compensation, including: a pressure generation module, a multimodal sensor group, a data processing module, and an adaptive pressure regulation module.

[0038] Specifically, the pressure generation module is used to generate a continuous positive pressure gas flow; the multimodal sensor group is used to monitor the respiratory airflow parameters of the patient in real time; the data processing module is connected to the multimodal sensor group, and the data processing module is used to perform respiratory waveform analysis based on the respiratory airflow parameters, calculate the required airflow pressure compensation value according to the analysis result, and generate a corresponding pressure compensation instruction; the adaptive pressure regulation module is respectively connected to the data processing module and the pressure generation module, and the adaptive pressure regulation module is used to dynamically adjust the pressure of the positive pressure gas flow output by the pressure generation module according to the pressure compensation instruction.

[0039] Among them, the pressure generation module can provide a stable air pressure source for the entire system to ensure that the patient's airway remains open during sleep.

[0040] The respiratory airflow parameters monitored by the multimodal sensor group can include airflow velocity, flow rate, pressure change, temperature, humidity, blood oxygen saturation, etc., providing an accurate data basis for subsequent adaptive adjustment. It should be noted that the multimodal sensor group includes a variety of sensors and has high accuracy.

[0041] In an optional embodiment, the multimodal sensor group includes: a pressure sensor, a flow sensor, a physiological sensor, and a temperature and humidity sensor.

[0042] Specifically, the pressure sensor is used to measure the pressure value at the airway opening of the patient in real time; the flow sensor is used to measure the inhalation gas flow rate and exhalation gas flow rate of the affected side in real time; the physiological sensor is used to measure the blood oxygen saturation and the phase difference of chest and abdomen movement of the patient in real time; the temperature and humidity sensor is used to measure the temperature and humidity in the patient's airway in real time.

[0043] Among them, the pressure sensor can be set at the pressure sampling port at the patient mask interface to directly monitor the real-time pressure fluctuation at the airway opening of the patient. Preferably, the pressure sensor can be a micro MEMS piezoresistive sensor with a diameter ≤ 5 mm, embedded in the inner layer of the mask silicone pad, which can avoid airflow interference.

[0044] The flow sensor can be set in the airway between the patient airway interface end and the pressure generation module. Preferably, the flow sensor can be a separate two-way flow sensor or two ordinary flow sensors for monitoring the inhalation gas flow rate and exhalation gas flow rate.

[0045] The physiological sensors may include a blood oxygen saturation sensor and a respiratory movement sensor. The blood oxygen saturation sensor is disposed at the forehead bracket of the patient mask to measure the patient's blood oxygen saturation in real time. The respiratory movement sensor is preferably a piezoelectric film sensor, which is embedded in the chest belt and the abdominal belt, and is respectively fixed at a certain intercostal space and a certain distance above the umbilicus to monitor the phase difference of the chest and abdomen movements in real time.

[0046] In an alternative embodiment, the data processing module includes: a respiratory waveform generating unit, a characteristic parameter extracting unit, a waveform analyzing unit, and an air flow pressure compensation value calculating unit.

[0047] Specifically, the respiratory waveform generating unit is connected to the multimodal sensor group. The respiratory waveform generating unit is used to convert the respiratory airflow parameters monitored by the multimodal sensor group into electrical signals, preprocess the electrical signals, and draw a respiratory waveform using the preprocessed electrical signals. The characteristic parameter extracting unit is connected to the respiratory waveform generating unit. The characteristic parameter extracting unit is used to extract key characteristic parameters from the respiratory waveform. The waveform analyzing unit is connected to the characteristic parameter extracting unit. The waveform analyzing unit is used to evaluate the patient's respiratory status, the working status of the ventilator, and whether there are abnormal waveforms according to the key characteristic parameters. The air flow pressure compensation value calculating unit is connected to the characteristic parameter extracting unit. The air flow pressure compensation value calculating unit is used to calculate the air flow pressure compensation value using a control algorithm according to the key characteristic parameters, and generate a corresponding pressure compensation instruction.

[0048] Among them, the multimodal sensor first converts the monitored respiratory airflow parameters into a voltage or current (electrical signal) that changes with time, and then preprocesses the electrical signal, including operations such as filtering and denoising, to improve the accuracy and reliability of the data. Finally, a graphical form of the respiratory waveform is formed by using signal processing techniques on the preprocessed electrical signal.

[0049] In an alternative embodiment, the respiratory waveform generally includes a pressure-time waveform, a flow-time waveform, a volume-time waveform, etc.

[0050] Among them, the pressure-time waveform shows the change of the airway pressure with time, reflecting the pressure output of the ventilator and the patient's respiratory effort.

[0051] The flow-time waveform shows the change of the gas flow with time, reflecting the patient's inhalation and exhalation airflow conditions.

[0052] The volume-time waveform: shows the change of the gas volume inhaled or exhaled by the patient with time, reflecting the patient's tidal volume and respiratory rate.

[0053] The key characteristic parameters extracted by the characteristic parameter extraction unit from the respiratory waveform may include: airflow rate (flow rate changes during inhalation and exhalation phases), tidal volume (gas volume of a single breath), respiratory rate (number of breaths per unit time), inhalation / exhalation time ratio, pressure curve shape (such as airflow limitation during inhalation or pressure oscillation during exhalation), whether there is leakage (abnormal airflow caused by mask leakage), etc.

[0054] The waveform analysis unit can conduct in-depth analysis on the extracted key characteristic parameters, including calculating respiratory parameters such as respiratory rate, tidal volume, inhalation / exhalation ratio, etc., and identifying abnormal characteristics in the respiratory waveform, such as apnea, tachypnea, etc. Specifically, the evaluation indicators can include the following aspects: First, respiratory rate and tidal volume analysis. Through the flow-time waveform and volume-time waveform, the respiratory rate and tidal volume of the patient can be measured to evaluate whether the patient's respiratory status is normal. Second, pressure stability analysis. Through the pressure-time waveform, it can be evaluated whether the pressure provided by the ventilator is stable and whether there are problems such as pressure fluctuations or leakage. Third, human-machine synchronization analysis: By comparing the patient's respiratory effort (such as the negative pressure waveform fluctuation triggered by spontaneous breathing) and the ventilator's response (such as the formation of a pressure plateau), it can be evaluated whether the human-machine synchronization is good. Fourth, treatment effect evaluation: By observing the changes in the respiratory waveform, the treatment effect of the CPAP system can be evaluated, such as whether the airway remains open and whether hypoxemia is improved. Fifth, abnormal waveform identification: Automatically identify abnormal waveforms, such as improper trigger sensitivity setting, large-capacity air leakage, false triggering, human-machine asynchrony, etc., so as to take timely measures for adjustment and intervention.

[0055] In summary, the multi-modal sensor group in the embodiment of the present invention provides accurate data support for the system by real-time monitoring of multiple physiological parameters of the patient. Then, using these data to form a respiratory waveform, and evaluating the patient's respiratory status, treatment effect and potential problems through waveform analysis, the evaluation indicators can be taken into account during subsequent airflow pressure compensation, thereby optimizing the treatment effect.

[0056] The airflow pressure compensation value calculation unit can calculate the airflow pressure compensation value using a control algorithm based on the key characteristic parameters, and generate corresponding pressure compensation instructions. It should be noted that the control algorithm can take the aforementioned multiple evaluation indicators into account.

[0057] In an alternative embodiment, the control algorithm uses a hybrid neural network model composed of a convolutional neural network, a Transformer model, and a fully connected layer to calculate the airflow pressure compensation value and generate corresponding pressure compensation instructions.

[0058] The input data of the hybrid neural network model includes the extracted key feature parameters; in an alternative embodiment, the breathing waveform of the time series, as well as the blood oxygen saturation and the phase difference of the chest and abdomen movement, can also be input simultaneously. The output label includes the target pressure compensation value, which can be labeled by the "ideal pressure response" in the historical data or expert experience. Data augmentation: Add noise to simulate sensor errors (such as flow fluctuations), and generate synthetic data to cover rare events (such as coughing and apnea). Then, the model is pre-trained and optimized. After the pre-training is completed, a small amount of patient individual data is used for model fine-tuning. Preferably, the model parameters can be continuously updated at the device end to adapt to the long-term changes of the patient.

[0059] In an alternative embodiment, the control algorithm uses a breathing phase recognition algorithm and an LSTM neural network to calculate the airflow pressure compensation value and generate the corresponding pressure compensation instruction.

[0060] Specifically, the breathing phase recognition algorithm is first used to decompose the breathing cycle. The preferred breathing phase recognition algorithm can be wavelet transform. The decomposed multi-segment waveforms are input into the pre-trained LSTM neural network, and the predicted voltage fluctuation of the next segment in the future is input, that is, the corresponding pressure compensation instruction is generated.

[0061] Preferably, a feedforward-feedback control method can be used to apply pressure compensation in advance, and at the same time, safety constraint conditions are added. For example, it is specified that the single adjustment amplitude does not exceed 2 cmH2O, and the pressure fluctuation range is controlled within ±15% of the set value.

[0062] In an alternative embodiment, the specific process of the control algorithm is as follows:

[0063] Step a: Perform wavelet transform on the airway pressure signal P(t) to decompose it into multi-scale waveforms P i (t) (such as the inhalation phase and the exhalation phase).

[0064] P i (t) = DWT(P(t), ψ i )

[0065] where ψ i is the wavelet basis function, and i represents different scales.

[0066] Step b: Input the decomposed waveform P i (t) into the pre-trained LSTM model to predict the voltage fluctuation ΔU(t + 1) at the next moment.

[0067] ΔU(t + 1) = LSTM(P(t), θ)

[0068] where θ is the parameter of the LSTM model.

[0069] Step c: According to the calibration curve of the ventilator, convert the predicted voltage fluctuation ΔU(t + 1) into a pressure compensation value ΔP(t + 1):

[0070] ΔP(t + 1) = KΔU(t + 1)

[0071] where K is the calibration coefficient.

[0072] Step d: Feedforward compensation: Apply the predicted pressure compensation value ΔP(t + 1) in advance. Feedback compensation: Calculate the feedback compensation value ΔP set -P meas (t) according to the current pressure error e(t) = P fb (t), where P set is the pressure fluctuation set value, and P meas (t) is the current pressure fluctuation value.

[0073] Total compensation command:

[0074] ΔP total (t) = ΔP(t + 1) + K p e(t)

[0075] where K p is the feedback gain.

[0076] Step e: Limit the single - time pressure adjustment amplitude not to exceed 2 cmH2O:

[0077] ∣ΔP total (t)∣ ≤ 2 cmH2O

[0078] Control the pressure fluctuation range within ±15% of the set value P set :

[0079] 0.85P set ≤ P meas (t) + ΔP total (t) ≤ 1.15P set

[0080] Safety constraint integration:

[0081] If ΔP total (t) exceeds the safety range, then perform clipping:

[0082] ΔP safe (t) = min(max(ΔP tota l(t), - 2), 2)

[0083] where ΔPsafe(t) is the total compensation value after clipping and within the safety range.

[0084] In an alternative embodiment, the system further includes:

[0085] A pressure relief module is disposed at the patient airway interface end and connected to the adaptive pressure regulation module, and is configured to automatically release pressure when it detects that the patient airway pressure exceeds a preset threshold.

[0086] In an alternative embodiment, the multimodal sensor group further includes:

[0087] An auxiliary pressure sensor is disposed at the proximal end of the patient airway and connected to the data processing module, and is configured to compensate for the pressure attenuation caused by the transmission delay of the positive pressure gas flow along the airway.

[0088] In an alternative embodiment, the system is applied to a single-level ventilator, a dual-level ventilator, or other non-invasive ventilators that support the continuous positive airway pressure mode.

[0089] In summary, the embodiments of the present invention use a multimodal sensor group to real-time sense the changes in the patient's respiratory airflow, and use a machine learning algorithm to predict the airflow demand and dynamically adjust the output airflow pressure, thereby optimizing the treatment effect and improving the patient's comfort.

[0090] As Figure 2 shown, the present invention provides a continuous positive airway pressure method based on adaptive airflow compensation, which is applied to a continuous positive airway pressure system based on adaptive airflow compensation. The method includes:

[0091] S1. Real-time monitor the patient's respiratory airflow parameters;

[0092] S2. Perform respiratory waveform analysis based on the respiratory airflow parameters, calculate the required airflow pressure compensation value according to the analysis result, and generate a corresponding pressure compensation instruction;

[0093] S3. Dynamically adjust the pressure of the output positive pressure gas flow according to the pressure compensation instruction.

[0094] Although the embodiments of the present invention are described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations fall within the scope defined by the appended claims.

Claims

1. A continuous positive airway pressure ventilation system based on adaptive airflow compensation, characterized in that, Comprising: A pressure generation module for generating a continuous positive pressure gas flow; A multi-modal sensor group for real-time monitoring of the patient's respiratory airflow parameters; A data processing module connected to the multi-modal sensor group for performing respiratory waveform analysis based on the respiratory airflow parameters, calculating the required airflow pressure compensation value according to the analysis result, and generating a corresponding pressure compensation instruction; An adaptive pressure regulation module connected to the data processing module and the pressure generation module respectively for dynamically adjusting the pressure of the positive pressure gas flow output by the pressure generation module according to the pressure compensation instruction.

2. The system according to claim 1, wherein The multi-modal sensor group includes: A pressure sensor for real-time measurement of the pressure value at the patient's airway opening; A flow sensor for real-time measurement of the inhaled gas flow and exhaled gas flow on the affected side of the patient; A physiological sensor for real-time measurement of the patient's blood oxygen saturation and the phase difference of chest and abdomen movement; A temperature and humidity sensor for real-time measurement of the temperature and humidity in the patient's airway.

3. The system according to claim 2, wherein The data processing module includes: A respiratory waveform generation unit connected to the multi-modal sensor group for converting the respiratory airflow parameters monitored by the multi-modal sensor group into electrical signals, preprocessing the electrical signals, and drawing a respiratory waveform using the preprocessed electrical signals; A characteristic parameter extraction unit connected to the respiratory waveform generation unit for extracting key characteristic parameters from the respiratory waveform; A waveform analysis unit connected to the characteristic parameter extraction unit for evaluating the patient's respiratory state, the working state of the ventilator, and whether there are abnormal waveforms according to the key characteristic parameters; An airflow pressure compensation value calculation unit connected to the characteristic parameter extraction unit for calculating the airflow pressure compensation value using a control algorithm according to the key characteristic parameters and generating a corresponding pressure compensation instruction.

4. The system according to claim 3, characterized in that The control algorithm uses a hybrid neural network model composed of a convolutional neural network, a Transformer model, and a fully connected layer to calculate the airflow pressure compensation value and generate a corresponding pressure compensation instruction.

5. The system according to claim 3, characterized in that, The control algorithm uses a respiratory phase recognition algorithm and an LSTM neural network to calculate the airflow pressure compensation value and generate a corresponding pressure compensation instruction.

6. The system according to claim 1, characterized in that, It further includes: A pressure release module disposed at the patient airway interface end and connected to the adaptive pressure regulation module for automatically relieving pressure when it detects that the patient airway pressure exceeds a preset threshold.

7. The system according to claim 2, wherein The multi-modal sensor group further includes: An auxiliary pressure sensor disposed at the proximal end of the patient airway and connected to the data processing module for compensating the pressure attenuation caused by the transmission delay of the positive pressure gas flow along the airway.

8. The system according to any one of claims 1-7, characterized in that, The system is applied to a single-level ventilator, a double-level ventilator, or other non-invasive ventilators supporting the continuous positive pressure ventilation mode.

9. A continuous positive airway pressure ventilation method based on adaptive airflow compensation, characterized in that, Applied to the system according to any one of claims 1-8, the method includes: Real-time monitoring of the patient's respiratory airflow parameters; Performing respiratory waveform analysis based on the respiratory airflow parameters, calculating the required airflow pressure compensation value according to the analysis result, and generating a corresponding pressure compensation instruction; Dynamically adjusting the pressure of the output positive pressure gas flow according to the pressure compensation instruction.

Images