Mechanical ventilation patient airway acoustic monitoring management optimization method and system based on artificial intelligence, electronic equipment, medium and program product

Abstract

The invention relates to the technical field of biology, and discloses a mechanical ventilation patient airway management optimization method and system based on artificial intelligence, electronic equipment, a medium and a program product, and the method comprises the steps: assembling a sound collection device, collecting airway sound wave data, and screening and classifying the airway sound wave data according to requirements; processing the reserved airway sound wave data, removing redundant information, and separating the airway sound wave data into single respiratory cycle samples by detecting sound loss between breaths; establishing an ANN classification and identification model of the airway secretions, determining parameter setting of the model, and verifying stability and accuracy of the trained ANN classification and identification model of the airway secretions by using the airway sound wave data; and determining an output result of the ANN classification and identification model of the airway secretions, and judging whether the airway secretions of the patient exist or not according to the output result of the ANN classification and identification model of the airway secretions.

Description

Technical Field

[0001] The present invention relates to the field of biotechnology, and particularly to an optimization method, system, electronic device, medium and program product for airway management of mechanically ventilated patients based on artificial intelligence. Background Art

[0002] Currently, mechanical ventilation is a common treatment measure in the ICU, used to support or completely replace a patient's spontaneous breathing, especially in patients with severe respiratory failure or those undergoing surgery requiring general anesthesia. According to data from the Society of Critical Care Medicine (SCCM) in the United States, approximately 790,000 adults receive mechanical ventilation treatment in the United States each year. After mechanical ventilation, it is often difficult to effectively expel respiratory secretions due to the weakened cough reflex. This situation not only hinders normal gas exchange but also increases the risk of complications such as VAP in patients. Airway secretions provide a culture medium for bacteria, enabling bacteria to multiply in the airway and cause pulmonary infections. Especially when airway cleaning and care measures are improper, the incidence of VAP will increase significantly. Relevant data show that the risk of mechanically ventilated patients developing VAP increases by approximately 1% to 3% per day. VAP not only prolongs the duration of mechanical ventilation and hospitalization but is also associated with high mortality. It is estimated that the incidence of VAP in mechanically ventilated patients is approximately 9% to 27%, and the mortality rate of VAP can be as high as 20% to 50%, depending on the patient population and the type of pathogen.

[0003] For patients receiving mechanical ventilation, taking appropriate preventive measures such as regularly changing body positions, maintaining appropriate airway humidity, and promptly removing airway secretions are key strategies to reduce the incidence of VAP. According to the clinical practice guidelines issued by the American Association for Respiratory Care (AARC), implementing effective airway clearance is crucial to maintain airway patency and prevent pulmonary infections. In the ICU, airway clearance techniques are one of the routine respiratory support management measures, and the main purpose is to remove secretions in the patient's airway to maintain airway patency and improve gas exchange. Clinically, the open suction method using a disposable suction catheter is widely adopted. This method involves temporarily disconnecting the patient from the mechanical ventilator to insert the suction catheter for aspirating secretions. Although open suction is widely used due to its simplicity and efficiency, it also has several drawbacks. First, pausing mechanical ventilation during each open suction may cause a temporary deterioration in the patient's oxygenation status, especially for those with severe hypoxemia or prone to respiratory failure. Second, during the open suction process, the patient is directly connected to the external environment, increasing the risk of cross-infection, especially in the ICU environment where strict aseptic operation is highly required. In addition, frequent suction operations may cause damage to the airway mucosa and increase the patient's discomfort. Due to the above drawbacks of frequent suction, it is necessary to identify the presence of airway secretions and perform suction at the appropriate time.

[0004] Currently, the ways to identify airway secretions include auscultation to detect moist rales, chest X-ray showing secretion accumulation, the patient presenting with dyspnea or a decrease in the oxygenation index, etc. However, these methods mainly rely on clinical manifestations and the doctor's experience judgment, such as judging whether suction is needed by auscultating lung sounds, observing the patient's respiratory status, and the characteristics of secretions. However, these methods have limitations. First, they are relatively subjective, and there may be differences in the judgments of different medical staff. Second, some patients may be unable to cough effectively due to low consciousness level or neuromuscular dysfunction, making it difficult for medical staff to accurately assess the secretion situation. Therefore, it is particularly crucial to identify the need and timing of the patient's airway secretions. Summary of the Invention

[0005] This application provides an optimization method, system, electronic device, medium, and program product for airway management of mechanically ventilated patients based on artificial intelligence to solve the problems of how to determine the relationship between airway secretions and airway acoustic wave characteristics in tracheally intubated patients and how to identify airway secretions in tracheally intubated patients.

[0006] To solve the above technical problems, the present application provides an optimization method for airway management of mechanically ventilated patients based on artificial intelligence, including: S1. Assemble a sound collection device, combine an array microphone, a signal acquisition card and acquisition software with the mechanical ventilation device to support real-time acquisition of airway acoustic wave data; S2. When it is determined that there are airway secretions in the trachea, collect the airway acoustic wave data before suctioning, then perform suctioning, collect the airway acoustic wave data after suctioning, and screen and classify the airway acoustic wave data according to requirements; S3. Process the retained airway acoustic wave data, remove redundant information, perform power spectrum and frequency spectrum analysis to obtain the spectral centroid, the maximum power of the power spectrum, the total power of 40 - 500 Hz, the power of each frequency range (40 - 100 Hz, 100 - 200 Hz, 200 - 300 Hz, 300 - 500 Hz), the ratio of each frequency range to the total power, and separate the airway acoustic wave data into single respiratory cycle samples by detecting the absence of sound between breaths; S4. Establish an ANN classification and recognition model for airway secretions, train the ANN classification and recognition model for airway secretions to judge the classification of airway acoustic wave data, determine the parameter settings of the model, and verify the stability and accuracy of the trained ANN classification and recognition model for airway secretions using the airway acoustic wave data; S5. Determine the output result of the ANN classification and recognition model for airway secretions, and judge whether there are airway secretions in the patient's airway according to the output result of the ANN classification and recognition model for airway secretions.

[0007] In some embodiments of the present application, in step S1, the sound collection device includes a T-tube, a Y-tube connected to one end of the T-tube, a gas cannula connected to the other end of the T-tube, the middle side tube of the T-tube is hermetically installed with an array microphone, and the array microphone is electrically connected to a signal acquisition card for collecting and storing airway acoustic wave data.

[0008] In some embodiments of the present application, in step S2, if it is determined that there are airway secretions in the trachea, use the sound collection device to start recording acoustic wave data, perform suctioning after 2 minutes, continue to record acoustic wave data for 2 minutes after suctioning, measure the volume of the aspirated airway secretions using a sputum collection jar, if the airway secretions in the sputum collection jar are not less than 3 ml, the patient's acoustic wave data is retained, if the airway secretions in the sputum collection jar are less than 3 ml, the patient's acoustic wave data is discarded, and the retained patient acoustic wave data is divided into two groups, the acoustic wave data before suctioning is defined as the group with airway secretions, and the acoustic wave data after suctioning is defined as the group without airway secretions.

[0009] In some embodiments of the present application, in step S3, MATLAB R2023a is used for analysis and processing. A low-pass filter is used to remove redundant information, the fast Fourier transform is adopted, and a Hamming window is used for power spectrum and frequency spectrum analysis. The spectral centroid, the maximum power of the power spectrum, the total power of 40 - 500 Hz, the power of each frequency range (40 - 100 Hz, 100 - 200 Hz, 200 - 300 Hz, 300 - 500 Hz), and the ratio of each frequency range to the total power are calculated. The variables obtained from the power spectrum analysis are compared between two groups. According to the characteristic that the airway acoustic wave data shows a parametric distribution, a paired t-test is used to compare the airway acoustic wave data before and after sputum suction. If P is not greater than 0.05, the difference is significant. The wavelet transform is used for time-frequency transformation. By detecting the absence of sound between breaths, the respiratory sound data is separated into individual respiratory cycle samples. Feature extraction is performed on the time-domain features, frequency-domain features, non-linear dynamics features, and entropy features of the signal from each individual respiratory cycle sample.

[0010] In some embodiments of the present application, in step S4, the input layer receives 8 features after the WT transformation of the airway acoustic wave data as input. The first hidden layer includes two parts, "W" and "b", which represent weights and biases respectively. The first hidden layer transforms the input data and outputs it to the second hidden layer. The second hidden layer includes two parts, "W" and "b", which represent weights and biases respectively. It performs a secondary transformation on the input data and outputs the data to the output layer. The output layer converts the output data of the second hidden layer into the final result.

[0011] In some embodiments of the present application, in step S4, after establishing the ANN classification and recognition model for airway secretions, the backpropagation algorithm is used for training to determine the classification of the airway acoustic wave data.

[0012] In some embodiments of the present application, an optimization system for airway management of mechanically ventilated patients based on artificial intelligence includes: a sound collection module for collecting airway acoustic wave data; a data screening module for screening and classifying the collected airway acoustic wave data; and a judgment module for judging whether there are secretions in the airway in the current airway acoustic wave data.

[0013] Compared with the prior art, the present invention has the following characteristics and beneficial effects: The present invention proposes a method for automatically detecting the condition of airway secretions based on acoustic feature recognition. The time-frequency analysis of the airway secretion sound signal is performed using wavelet transform, and the wavelet coefficients are extracted. Finally, an ANN classification model is established and trained using the backpropagation algorithm (BP) to diagnose the classification of the airway secretion sound signal. The method of neural classification of sound using the coefficients of the wavelet-transformed sound wave can help doctors automatically identify the airway secretion conditions of ICU mechanical ventilation patients and help nurses suction sputum in a timely manner. This method has certain practical value in clinically improving the technical level of secretion clearance, saving the lives of ventilated patients, reducing the pain of patients and the workload of medical staff, and good results can be achieved by popularizing its use. Description of the Drawings

[0014] Figure 1 is a schematic flow chart of an embodiment of the present invention; Figure 2 is a schematic diagram of the ANN classification recognition model of an embodiment of the present invention; Figure 3 is a schematic diagram for evaluating the ANN classification recognition model of an embodiment of the present invention; Figure 4 is a waveform and time-frequency diagram of the no-airway-secretion group in an embodiment of the present invention; Figure 5 is a waveform and time-frequency diagram of the airway-secretion group in an embodiment of the present invention; Figure 6 is a waveform diagram of the sound wave of the no-airway-secretion group after FFT transformation in an embodiment of the present invention; Figure 7 is a waveform diagram of the sound wave of the airway-secretion group after FFT transformation in an embodiment of the present invention. Detailed Embodiment

[0015] The following further describes in detail the specific embodiments of the present invention in conjunction with the drawings and embodiments. The following embodiments are used to illustrate the present invention but are not intended to limit the scope of the present invention.

[0016] In the description of the present application, it should be understood that the orientation or positional relationship indicated by the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the present application.

[0017] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of this application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0018] In the description of this application, it should be noted that, unless otherwise clearly specified and defined, the terms "install", "connect", and "couple" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0019] As Figure 1 shown, in some embodiments of this application, an optimization method for airway management of mechanically ventilated patients based on artificial intelligence includes: S1. Assemble a sound collection device, combine an array microphone, a signal acquisition card, and acquisition software with the mechanical ventilation device to support real-time acquisition of airway acoustic wave data; S2. When it is determined that there are airway secretions in the trachea, collect the airway acoustic wave data before suctioning, then perform suctioning, collect the airway acoustic wave data after suctioning, and screen and classify the airway acoustic wave data according to requirements; S3. Process the retained airway acoustic wave data, remove redundant information, perform power spectrum and frequency spectrum analysis to obtain the spectral centroid, the maximum power of the power spectrum, the total power of 40 - 500 Hz, the power of each frequency range (40 - 100 Hz, 100 - 200 Hz, 200 - 300 Hz, 300 - 500 Hz), the ratio of each frequency range to the total power, and separate the airway acoustic wave data into individual respiratory cycle samples by detecting the absence of sound between breaths; S4. Establish an ANN classification and recognition model for airway secretions, train the ANN classification and recognition model for airway secretions to judge the classification of airway acoustic wave data, determine the parameter settings of the model, and verify the stability and accuracy of the trained ANN classification and recognition model for airway secretions using the airway acoustic wave data; S5. Determine the output result of the ANN classification and recognition model for airway secretions, and judge whether there are airway secretions in the patient's airway according to the output result of the ANN classification and recognition model for airway secretions.

[0020] In some embodiments of the present application, in step S1, the sound collection device includes a T-shaped tube, a Y-shaped tube connected to one end of the T-shaped tube, a gas cannula connected to the other end of the T-shaped tube, a middle side tube of the T-shaped tube is hermetically installed with any one of the array microphones, a plurality of array microphones are fixed outside the body, the array microphones are electrically connected to a signal acquisition card, the array microphones are used to collect and save airway acoustic wave data, a plurality of array microphones are placed at the outer contours of the upper respiratory tract and the lungs, and each pickup point of the array microphones collects airway acoustic wave data at the current position.

[0021] In some embodiments of the present application, in step S2, if it is determined that there are airway secretions in the trachea, the sound collection device is used to start recording the acoustic wave data of each pickup point of the array microphones. After 2 minutes, sputum suction is performed. After sputum suction, the acoustic wave data of each pickup point of the array microphones is continuously recorded for another 2 minutes. The volume of the aspirated airway secretions is measured using a sputum collection can. If the airway secretions in the sputum collection can are not less than 3 ml, the acoustic wave data of each pickup point of the array microphones is retained. If the airway secretions in the sputum collection can are less than 3 ml, the acoustic wave data of each pickup point of the array microphones is discarded, and the retained acoustic wave data of each pickup point of the array microphones is divided into two groups. The acoustic wave data before sputum suction is defined as the group with airway secretions, and the acoustic wave data after sputum suction is defined as the group without airway secretions. And each acoustic wave data corresponds to the pickup point at the response position during recording.

[0022] In some embodiments of the present application, in step S3, MATLAB R2023a is used for analysis and processing. A low-pass filter is used to remove redundant information, the fast Fourier transform is adopted, a Hamming window is used for power spectrum and frequency spectrum analysis, the spectral centroid, the maximum power of the power spectrum, the total power from 40 - 500 Hz, the power of each frequency range (40 - 100 Hz, 100 - 200 Hz, 200 - 300 Hz, 300 - 500 Hz), and the ratio of each frequency range to the total power are calculated. The variables obtained from the power spectrum analysis are compared between the two groups. According to the characteristic that the airway acoustic wave data are all parametrically distributed, a paired t-test is used to compare the airway acoustic wave data before and after sputum suction. If P is not greater than 0.05, the difference is significant. A wavelet transform method is used for time-frequency transformation. By detecting the absence of sound between breaths, the respiratory sound data is separated into individual respiratory cycle samples, and the time-domain characteristics, frequency-domain characteristics, non-linear dynamic characteristics, and entropy characteristics of the signal are extracted from each individual respiratory cycle sample.

[0023] As Figure 2As shown, in some embodiments of the present application, in step S4, the input layer receives 8 features after the WT transformation of the airway acoustic wave data as input. The first hidden layer includes two parts, "W" and "b", representing weights and biases respectively. The first hidden layer transforms the input data and outputs it to the second hidden layer. The second hidden layer also includes two parts, "W" and "b", representing weights and biases respectively. It performs a secondary transformation on the input data and outputs the data to the output layer. The output layer converts the output data of the second hidden layer into the final result.

[0024] In some embodiments of the present application, in step S4, after establishing the ANN classification and recognition model for airway secretions, the backpropagation algorithm is used for training to judge the classification of each airway acoustic wave data. Based on the principle that the closer to the secretion, the louder the sound emitted by the airway, multiple airway acoustic wave data are screened out. The corresponding sound pickup point positions are determined using the multiple airway acoustic wave data, and then the secretion accumulation range is determined to indicate the sputum suction position and sputum suction timing.

[0025] As Figure 3 shown, in some embodiments of the present application, R is the correlation coefficient of the fitting line, representing the relationship between the predicted value and the actual value. The value range is from -1 to 1. The closer the value is to 1 or -1, the stronger the correlation. Among them, -1 represents negative correlation and 1 represents positive correlation. The upper left, upper right, lower left, and lower right are the training set, validation set, test set, and all data sets respectively. Generally speaking, the prediction accuracy of all four data sets is > 0.95, that is, the predicted value is very close to the actual value, indicating that the model has high accuracy.

[0026] In some embodiments of the present application, an optimization system for airway management of mechanically ventilated patients based on artificial intelligence includes: a sound collection module for collecting airway acoustic wave data; a data screening module for screening and classifying the collected airway acoustic wave data; and a judgment module for judging whether there are secretions in the airway in the current airway acoustic wave data.

[0027] In some embodiments of the present application, compare the changes in the acoustic wave power within the frequency range of 0 - 100 Hz 2 minutes before and after sputum suction for the same patient. It can be seen that after sputum suction (dashed line), the power at most frequencies decreases compared to before sputum suction (solid line). In the 40 - 100 Hz band, the acoustic wave power peaks before sputum suction, and after sputum suction, the peak value decreases significantly, indicating that sputum suction reduces the energy of the breathing sound at this frequency. Overall, the power spectrum after sputum suction is flatter, showing fewer spikes, which indicates that the energy of the acoustic wave is more evenly distributed at different frequencies rather than concentrated at certain specific frequency points, reflecting a cleaner state inside the airway.

[0028] As shown Figure 4 - in Figure 7, in some embodiments of the present application, the wave crest of the sound wave represents the opening sound of the one-way valve of the ventilator during inhalation, and the wave trough represents the closing sound of the one-way valve at the beginning of exhalation. By comparing the waveforms of the group with airway secretions and the group without airway secretions, it can be found that the waveform of the group without airway secretions is relatively smooth between the two peaks, showing the characteristic of sound wave propagation without obvious interference. When there are airway secretions, the waveform shows obvious irregular changes between the peaks, reflecting the significant interference effect of airway secretions on sound wave propagation; comparison of the time-frequency diagrams of the group with airway secretions and the group without airway secretions after WT transformation. It can be seen that in the low-frequency range of 40 - 100 Hz of the time-frequency diagram of the group without airway secretions, the energy distribution is relatively uniform, and the overall color is more consistent, indicating that the sound is relatively stable in this state. In contrast, the time-frequency diagram of the group with airway secretions shows more regions with concentrated energy in the low-frequency range of 40 - 100 Hz, and these regions are shown in bright colors, indicating that the energy of the sound is higher in this frequency band in the state with airway secretions.

[0029] In summary, the present invention relates to the field of biotechnology, and discloses an optimization method, system, electronic device, medium and program product for airway management of mechanically ventilated patients based on artificial intelligence, including: assembling a sound collection device, collecting airway sound wave data, and screening and classifying the airway sound wave data according to requirements; processing the retained airway sound wave data, removing redundant information, and separating the airway sound wave data into individual respiratory cycle samples by detecting the sound missing between breaths; establishing an ANN classification and recognition model for airway secretions, determining the parameter settings of the model, and verifying the stability and accuracy of the trained ANN classification and recognition model for airway secretions using the airway sound wave data; determining the output result of the ANN classification and recognition model for airway secretions, and judging whether there are airway secretions in the patient's airway according to the output result of the ANN classification and recognition model for airway secretions.

[0030] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art in the technical field, without departing from the technical principle of the present invention, several improvements and substitutions can be made, and these improvements and substitutions should also be regarded as the protection scope of the present invention.

Claims

1. An optimization method for airway management of mechanically ventilated patients based on artificial intelligence, characterized in that, The specific steps of the method are as follows: S1. Assemble the sound collection device by combining the array microphone, signal acquisition card, and acquisition software with the mechanical ventilation device to support real-time acquisition of airway acoustic wave data; S2. When it is judged that there are airway secretions in the trachea, collect the airway acoustic wave data before sputum suction, then perform sputum suction, collect the airway acoustic wave data after sputum suction, and screen and classify the airway acoustic wave data according to requirements; S3. Process the retained airway acoustic wave data, remove redundant information, perform power spectrum and frequency spectrum analysis to obtain the spectral centroid, the maximum power of the power spectrum, the total power of 40 - 500 Hz, the power of each frequency range (40 - 100 Hz, 100 - 200 Hz, 200 - 300 Hz, 300 - 500 Hz), the ratio of each frequency range to the total power, and separate the airway acoustic wave data into single respiratory cycle samples by detecting the absence of sound between breaths; S4. Establish an ANN classification and recognition model for airway secretions, train the ANN classification and recognition model for airway secretions to judge the classification of airway acoustic wave data, determine the parameter settings of the model, and verify the stability and accuracy of the trained ANN classification and recognition model for airway secretions using the airway acoustic wave data; S5. Determine the output result of the ANN classification and recognition model for airway secretions, and judge whether there are airway secretions in the patient's airway according to the output result of the ANN classification and recognition model for airway secretions.

2. The method according to claim 1, characterized in that In the step S1, the sound collection device includes a T-tube, a Y-tube connected to one end of the T-tube, a gas cannula connected to the other end of the T-tube. The middle side tube of the T-tube is sealed and installed with any one of the array microphones. Multiple array microphones are fixed outside the body. The array microphones are electrically connected to the signal acquisition card, and the array microphones are used to collect and save airway acoustic wave data.

3. The method according to claim 1, wherein In the step S2, if it is judged that there are airway secretions in the trachea, use the sound collection device to start recording the acoustic wave data of each pickup point of the array microphone. After 2 minutes, perform sputum suction. After sputum suction, continue to record the acoustic wave data of each pickup point of the array microphone for 2 minutes. Use a sputum collection can to measure the volume of the aspirated airway secretions. If the airway secretions in the sputum collection can are not less than 3 ml, the acoustic wave data of each pickup point of the array microphone are retained. If the airway secretions in the sputum collection can are less than 3 ml, the acoustic wave data of each pickup point of the array microphone are discarded, and the retained acoustic wave data of each pickup point of the array microphone are divided into two groups. The acoustic wave data before sputum suction are defined as the group with airway secretions, and the acoustic wave data after sputum suction are defined as the group without airway secretions.

4. The method according to claim 1, wherein In step S3, MATLAB R2023a is used for analysis and processing. A low-pass filter is used to remove redundant information. The fast Fourier transform is adopted, and the Hamming window is used for power spectrum and frequency spectrum analysis. The spectral centroid, the maximum power of the power spectrum, the total power of 40 - 500 Hz, the power of each frequency range (40 - 100 Hz, 100 - 200 Hz, 200 - 300 Hz, 300 - 500 Hz), and the ratio of each frequency range to the total power are calculated. The variables obtained from the power spectrum analysis are compared between two groups. According to the characteristic that the airway acoustic wave data are all parametrically distributed, a paired t-test is used to compare the airway acoustic wave data before and after sputum suction. If P is not greater than 0.05, the difference is significant. The wavelet transform is used for time-frequency transformation. By detecting the absence of sound between breaths, the respiratory sound data are separated into individual respiratory cycle samples. Feature extraction is performed on the time-domain features, frequency-domain features, non-linear dynamics features, and entropy features of the signal from each individual respiratory cycle sample.

5. The method according to claim 1, characterized in that In step S4, the input layer receives 8 features after the WT transformation of the airway acoustic wave data as input. The first hidden layer consists of two parts, "W" and "b", representing weights and biases respectively. The first hidden layer transforms the input data and outputs it to the second hidden layer. The second hidden layer also consists of two parts, "W" and "b", representing weights and biases respectively. It performs a secondary transformation on the input data and outputs the data to the output layer. The output layer converts the output data of the second hidden layer into the final result.

6. The method according to claim 1, wherein In step S4, after establishing the ANN classification and recognition model for airway secretions, the backpropagation algorithm is used for training to determine the classification of the airway acoustic wave data.

7. An artificial intelligence-based optimized system for airway management of mechanically ventilated patients, characterized in that, Including: A sound collection module for collecting airway acoustic wave data; A data screening module for screening and classifying the collected airway acoustic wave data; A judgment module for judging whether there are secretions in the airway in the current airway acoustic wave data.

8. An electronic device, characterized in that, Including: A processor, a memory, and a computer program stored on the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of an optimization method for airway management of mechanically ventilated patients based on artificial intelligence as described in any one of claims 1 - 6.

9. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium. When the computer program is executed by the processor, it implements the steps of an optimization method for airway management of mechanically ventilated patients based on artificial intelligence as described in any one of claims 1 - 6.

10. A computer program product, comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, it implements the steps of an optimization method for airway management of mechanically ventilated patients based on artificial intelligence as described in any one of claims 1 - 6.

Images