High-frequency whole chest oscillation sputum excretion method and system based on adaptive regulation of clinical data

Abstract

The application discloses a high-frequency full chest oscillation sputum excretion method and system based on clinical data adaptive regulation, and establishes an individualized propagation model of a thoracic cage and lung tissue by collecting individualized clinical data of a patient and short-range responses of chest multi-point acoustics and mechanics. In the model, the current body position of the patient is inputted, a chest drainage path is obtained, and oscillation control parameters are generated under the constraint of the path, which are used to drive the equipment to synthesize a directional propagation pressure traveling wave in multiple regions of the chest wall. In the execution process, safety gate control is performed in combination with electrocardio and respiratory phases, aerosol echoes and high-frequency impedance disturbances are collected as feedback signals, feedback data is formed after the feedback signals are aligned with the execution timing, and the feedback data is used to dynamically update the individualized propagation model and the oscillation control parameters. The application can accurately regulate different patient individual differences and real-time physiological states, effectively enhance the sputum drainage effect, and improve the safety and comfort of treatment, and has high clinical application value.

Description

Technical Field

[0001] This application relates to the field of medical instrument and equipment technology, and in particular to a high-frequency whole-chest oscillation sputum clearance method and system based on clinical data adaptive adjustment. Background Technology

[0002] Currently, high-frequency whole-chest oscillation for sputum clearance is widely used in clinical respiratory rehabilitation and treatment. This method typically applies high-frequency oscillations to the patient's chest wall using an oscillating vest or external mechanical vibration device, loosening and gradually expelling mucus from the airways, thereby improving ventilation and lung function. Existing devices mostly employ preset oscillation frequencies, vibration intensities, and durations. Some products also offer multi-zone vibration units to cover different lung regions, achieving a degree of individualized treatment.

[0003] However, existing technologies have significant limitations. Most sputum clearance devices rely on standardized or empirical parameters, lacking precise adaptation to individual patient differences, leading to significant variations in treatment outcomes across different populations. Furthermore, existing devices typically operate in fixed modes, failing to dynamically adjust based on real-time physiological signals, potentially resulting in insufficient sputum clearance or increased patient discomfort. In addition, the oscillation process lacks effective closed-loop feedback, making it difficult to promptly correct parameters to address changes in the patient's condition during treatment.

[0004] To address the aforementioned shortcomings, there is an urgent need to propose a new method and system for high-frequency whole-chest oscillation sputum clearance that can combine individualized data with real-time feedback to achieve adaptive control. Summary of the Invention

[0005] This application provides a high-frequency whole-chest oscillation sputum drainage method and system based on clinical data adaptive adjustment, in order to enhance the sputum drainage effect and improve the safety and comfort of treatment.

[0006] This application provides a high-frequency whole-chest oscillation sputum clearance method based on clinical data adaptive adjustment, including: Collect individualized clinical data, short-range acoustic response and short-range biomechanical response of the chest at multiple points, and establish an individualized propagation model of the thoracic cage and lung tissue based on the collected data; Input the patient's current position into the individualized propagation model to obtain the chest drainage path; Under the constraints of the thoracic drainage path, the individualized propagation model is invoked to generate oscillation control parameters including frequency, phase and location of action. The oscillation control parameters are used as execution commands, which are used to synthesize directional propagating pressure traveling waves in multiple locations of the chest wall. The execution command is executed and a safety gating output is implemented according to the ECG and respiratory phases, while aerosol echo and high-frequency impedance disturbance are collected as feedback signals. The feedback signal is aligned with the execution timing to form feedback data, and the individualized propagation model and oscillation control parameters of the thoracic cavity and lung tissue are updated based on the feedback data.

[0007] This application provides a high-frequency whole-chest oscillatory sputum clearance system based on clinical data adaptive adjustment, comprising: The acquisition unit is used to collect individualized clinical data, short-range acoustic response and short-range biomechanical response of the chest at multiple points, and to establish an individualized propagation model of the thoracic cage and lung tissue based on the collected data. The input unit is used to input the patient's current position into the individualized propagation model to obtain the chest drainage path; The generation unit is used to call the individualized propagation model under the constraint of the chest drainage path, generate oscillation control parameters including frequency, phase and action area, and use the oscillation control parameters as execution instructions. The execution instructions are used to synthesize directional propagating pressure traveling waves in multiple areas of the chest wall. The execution unit is used to execute the execution instructions and implement safety gating output according to the ECG and respiratory phases, while collecting aerosol echoes and high-frequency impedance disturbances as feedback signals. The alignment unit is used to align the feedback signal with the execution timing to form feedback data, and update the individualized propagation model and oscillation control parameters of the thoracic cavity and lung tissue based on the feedback data.

[0008] This application has the following beneficial technical effects: (1) By establishing an individualized propagation model of the thoracic cavity and lung tissue, the patient's clinical data and acoustic and mechanical responses were comprehensively considered, and the individualized matching of oscillation parameters and drainage path was achieved, which effectively improved the accuracy and adaptability of sputum expectoration treatment. (2) Under the constraint of the thoracic drainage path, the oscillation control parameters of frequency, phase and action area were generated, which can form a directional propagating pressure wave in the chest wall, significantly enhance the migration efficiency of sputum in the airway and improve the sputum expectoration effect. (3) Safety gating control was implemented by combining electrocardiogram and respiratory phase to avoid applying adverse stimulation during the patient's heart or respiratory sensitive period, thereby effectively reducing the risk in the treatment process and improving the patient's safety and comfort. (4) By collecting aerosol echoes and high-frequency impedance disturbances to form feedback data and using it for dynamic updating of the model and parameters, a real-time closed-loop control mechanism was constructed, which enables the treatment process to be automatically optimized according to the patient's condition, ensuring the stability and continuity of the therapeutic effect. Attached Figure Description

[0009] Figure 1 This is a flowchart of a high-frequency whole-chest oscillation sputum clearance method based on clinical data adaptive adjustment provided in the first embodiment of this application.

[0010] Figure 2This is a schematic diagram of a high-frequency whole-chest oscillation sputum clearance system based on clinical data adaptive adjustment, provided in the second embodiment of this application. Detailed Implementation

[0011] Many specific details are set forth in the following description to provide a full understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this application; therefore, this application is not limited to the specific embodiments disclosed below.

[0012] The first embodiment of this application provides a high-frequency whole-chest oscillation sputum clearance method based on adaptive adjustment of clinical data. Please refer to... Figure 1 This figure is a schematic diagram of the first embodiment of this application. The following is in conjunction with... Figure 1 The first embodiment of this application provides a detailed description of a high-frequency whole-chest oscillation sputum clearance method based on adaptive adjustment of clinical data.

[0013] Step S101: Collect individualized clinical data, short-range acoustic response and short-range biomechanical response of the chest at multiple points, and establish an individualized propagation model of the thoracic cage and lung tissue based on the collected data.

[0014] In step S101, comprehensive individualized clinical data of the patient must first be collected. This data includes not only basic anthropometric parameters such as height, weight, chest circumference, age, sex, and body type, but also respiratory-related medical indicators, such as previous pulmonary function test results, types and duration of respiratory diseases, and underlying medical history such as bronchitis, asthma, chronic obstructive pulmonary disease, or cystic fibrosis. This clinical data provides patient-specific boundary conditions and individual variability parameters for model establishment, ensuring that the model can reflect the specific anatomical and pathological characteristics of the patient.

[0015] Next, have the patient rest in a seated or supine position for at least five minutes, maintaining steady breathing. Expose the anterior, lateral, and posterior chests, ensuring the measurement area is dry and clean, and the ambient noise level is as low as possible below normal conversation. Then, place sensors and actuators at anatomical landmarks: on the anterior chest, at the second intercostal space along the bilateral midclavicular line, the midline above the xiphoid process, and the line connecting the two nipples; on the lateral chest, at the fifth intercostal space along the bilateral midaxillary lines; on the posterior back, at the levels of the third, seventh, and tenth thoracic vertebrae along the bilateral scapular lines, and at the corresponding levels along the paravertebral lines. Apply an integrated acoustic sensor patch and a micro-accelerometer sensor patch to each designated location. If necessary, place a reference synchronization sensor patch along the midline of the sternal manubrium. Secure the sensor patches to the skin using medical-grade hypoallergenic adhesive, pressing to release air and prevent loosening; gently touch each patch with your finger to confirm secure adhesion and no slippage.

[0016] After setup, channel calibration is performed. First, a standardized, short acoustic pulse is applied close to the skin at the manubrium position as a reference acoustic excitation. All acoustic channels should simultaneously record clear pulse waveforms, with peak amplitude and arrival time within acceptable tolerances. Then, a tiny mechanical vibration probe is applied at the same location with a minimal displacement, and each acceleration channel is checked to ensure it records the corresponding response peak and aligns with the acoustic pulse on the time marker. If any channel exhibits significant hysteresis, saturation, or noise overload, the sensor patch is re-attached or replaced until all channels meet consistency and signal-to-noise ratio requirements.

[0017] After calibration, short-range acoustic responses at multiple points on the chest were acquired. Two types of excitation were used for short-time measurements: one was narrow-band swept-frequency excitation, covering the low to mid-high frequency bands relevant to expectoration; the other was short-pulse acoustic excitation to obtain transient transmission characteristics. The excitation was emitted close to the skin by an acoustic actuator placed along the midline of the chest, with the intensity controlled below the patient's subjective comfort level. The duration of each excitation was controlled within ten seconds, with sufficient intervals to avoid cumulative effects. The original waveform, energy spectrum, main peak frequency, peak time position, and arrival order relative to the reference channel were recorded for each acoustic channel. The amplitude ratio, energy concentration, and bandwidth of the same excitation in different channels were also recorded. To reduce interference from heart sounds and external noise, DC removal and slow drift correction were performed on the original waveforms after acquisition, abnormal pulses that significantly exceeded the physiological range were removed, bandpass processing was used to retain only the expectoration-related frequency bands, and the time scale of each channel was fine-tuned based on the reference pulse to ensure that all channels were strictly aligned on the same time axis.

[0018] Next, short-range mechanical responses at multiple points on the chest were acquired. In the same position as the acoustic acquisition, a very small, controllable mechanical sweep or short mechanical pulse was applied along the midline of the anterior chest; displacement and acceleration response curves over time were recorded for each acceleration channel. To avoid discomfort, the peak displacement of the mechanical excitation should be below the perceptible threshold, and a soft pad was added to the contact surface to distribute pressure. For each channel, the onset time, time required to rise to the peak value, peak value, and time required to decay to a certain proportion of the baseline were extracted for the instantaneous response, and the stability of these indicators under repeated excitation was calculated. Subsequently, based on the response sequence and peak attenuation degree between adjacent measurement points, the relative directionality of dynamic transmission and coupling strength were summarized. To ensure data quality, a cross-channel consistency check was performed again: when the acoustic and mechanical acquisitions at the same location showed significant inconsistencies in the main response frequency bands (e.g., one side showed strong coupling while the other showed significant attenuation), the acquisition at that location was repeated or the patch position was adjusted until stable and consistent records were obtained.

[0019] After completing the two types of short-range responses, data preprocessing and feature extraction proceed. The timescales of all acoustic and mechanical channels are fine-tuned and aligned against the baseline channel. Slow-changing components of the respiratory cycle are subtracted for background, retaining only the rapid-changing components related to the applied excitation. Repeated trials for each excitation are averaged to improve the signal-to-noise ratio and reduce the impact of occasional noise. A stable set of acoustic and mechanical characteristics is formed for each measurement point. The former includes the frequency range of the main energy concentration, the order of arrival relative to the baseline channel, and the relative amplitude. The latter includes the start-up time, peak size, decay rate, and the transmission sequence between adjacent measurement points for the same excitation. For individualized clinical data, the chest wall surface is divided into a set of discrete locations corresponding to the sensor patch according to height, weight, chest circumference, and body type. Combined with past disease types and baseline pulmonary function indicators, the anatomical correspondence between each discrete location and lung segments and lobes is labeled. This anatomical correspondence is used to map the propagation characteristics measured on the surface to deeper parts of the lung for inference.

[0020] After completing the above preparations, an individualized propagation model of the chest wall and lung tissue is established. Specifically, discrete regions of the chest wall are considered as a set of directed connected areas. Each pair of regions is described by two types of quantitative relationships: one is the acoustic propagation relationship, which reflects the strength and order of energy transfer from the source region to the target region, reflecting the path and attenuation of sound waves between tissue layers; the other is the mechanical coupling relationship, which reflects the order and attenuation rate of acceleration response induced by external forces in adjacent or contralateral regions, reflecting the mechanical connectivity between the chest wall, lung tissue, and airway. These two types of relationships establish a propagation relationship table based on the aforementioned characteristics. The propagation relationship table records the direction, relative intensity, transmission hysteresis, and frequency band sensitivity for each connection, and includes a stability score from multiple repeated measurements. The propagation relationship table and individualized clinical data together constitute the patient's basic propagation information database.

[0021] Based on the database, an individualized propagation model of the thoracic and lung tissues is further developed and can be used in subsequent steps. This model focuses on two core aspects: first, the acoustic transmission and mechanical coupling maps from the chest wall surface to the corresponding anatomical regions within the lungs, which can determine the possible paths and preferred directions of energy propagation between different locations; second, local dynamic properties corrected for individual patient differences, including the effective compliance of each location in the sputum expectoration-related frequency bands, the trend of equivalent impedance changes, and the sensitivity threshold to applied oscillations. The model is stored as a searchable data structure, containing discrete location numbers, bidirectional or unidirectional propagation entries between regions, the strength, hysteresis, frequency band information, and stability score of each entry, and establishing a one-to-one association with the corresponding anatomical regions. To ensure the reliability of subsequent reuse, the model also records the patient's position, excitation type, excitation intensity range, and channel health status at the time of data acquisition, serving as a basis for interpretation and retesting.

[0022] In this embodiment, after the information database is completed, a callable individualized propagation model of the thoracic cavity and lung tissue can be generated as follows: First, using the chest wall coordinate reference and discrete location number as the skeleton, each set of acoustic and mechanical features is written into the propagation entries according to the correspondence between "source location - target location". The strength of the entries is determined by a unified level scale (e.g., divided into extremely strong, strong, medium, weak, and extremely weak) based on the relative amplitude ratio and energy concentration. The hysteresis is determined by a unified time classification based on the arrival order and rise time. The frequency band information is recorded using the dominant frequency band and its adjacent auxiliary frequency bands. The stability score is obtained by weighting the consistency of repeated measurements and cross-batch reproducibility. After the acoustic transmission map and the mechanical coupling map are entered into the database, a cross-check is performed. If there is a significant contradiction between the strength or hysteresis of the two maps in the same pair of locations, the original records are traced back or the entries of adjacent locations are interpolated for verification. If necessary, the entries are marked as low confidence entries for subsequent updates. Subsequently, based on individualized clinical data, a weighting rule table was loaded. For disease-related areas, areas with thicker fat or stiffer chest walls, and areas with lesions indicated by previous imaging, item weight corrections were performed. The corrections followed the principle of "rule traceability and only one weight source added to each item" to avoid excessive stacking. For missing channels or channels with mild artifacts, limited interpolation was performed using the nearest neighbor location and the contralateral mirror location, and such items were automatically marked as low stability to prompt for subsequent priority updates. After completing the item correction, sparse graph structures of acoustic transmission graph and mechanical coupling graph are generated with location as nodes and bidirectional or unidirectional items as edges, respectively. A one-to-one mapping is established with the anatomical correspondence from location to lung lobe and lung segment. Two types of query interfaces are encapsulated in the outer layer of the model. The path derivation interface receives body position and gravity direction, and searches for candidate channels in the intersection of the two graphs according to the criteria of strength priority, minimum hysteresis, and highest consistency with anatomical direction, and outputs the thoracic drainage path and its reachability score. The parameter solving interface receives the thoracic drainage path and selects adjacent locations for the path within the allowable range of frequency band information and strength and hysteresis constraints. It can generate frequency and phase combinations for directional energy transfer and, in conjunction with the location of action, generate directly deployable oscillation control parameters. The model and the database are stored together as a versioned data structure, recording the acquisition position, excitation type and intensity, channel health status, and weight rule source. Each clinical update includes a change summary and rollback point, thus ensuring that the model can be used directly in the first stage to output the thoracic drainage path, and also in the second stage under path constraints to output oscillation control parameters containing frequency, phase, and location of action, ensuring the continuous connection and reproducibility of path calculation and parameter calculation.

[0023] During implementation, the first step is to check whether the relative intensity and hysteresis of each propagation item are consistent with the geometric distance and tissue thickness at the time of acquisition. If an item shows a directionality and hysteresis that is significantly contrary to that of adjacent items, the original record corresponding to that item is backtracked and retested. Secondly, rapid and repeated acquisition is performed on a small number of key areas under the same body position. If the features obtained from the repetition are consistent with existing items in the model or fluctuate slightly within an acceptable range, the model is considered stable. Finally, the model is validated once on another set of short-range excitation records not used for modeling: after triggering with the baseline channel, the relative order of each channel is compared with the directionality shown by the model. If they are consistent, the model is deemed ready for subsequent use.

[0024] At this point, the individualized propagation model of the thoracic cavity and lung tissue is complete. This model is not a static stack of parameters, but a dynamically invoked network structure capable of supporting two different levels of functionality. In the first stage, when the patient's current position is input, the model combines this position with the propagation relationship table, superimposing the influence of gravity direction on energy transfer, thereby outputting the most probable thoracic drainage path—essentially finding the channel in the network where energy flows most easily. In the second stage, when the thoracic drainage path is input, the model selects a set of control schemes that can form a directional traveling wave in the chest wall from existing propagation parameters and actuator mapping rules, outputting oscillation control parameters such as frequency, phase, and location of action—essentially generating specific driving instructions under a given route. Although the inputs and outputs of the two stages differ, they share the same propagation data structure; the former focuses on path calculation, while the latter focuses on parameter generation, logically continuous and interdependent.

[0025] Furthermore, the collection of individualized clinical data, short-range acoustic response and short-range biomechanical response at multiple points in the chest, and the establishment of an individualized propagation model of the thoracic cage and lung tissue based on the collected data, includes: Individualized clinical data were collected from patients at rest and a discrete location mapping table of the thoracic surface was generated. Short-range acoustic and mechanical excitations are applied to the discrete locations, and corresponding signals are collected to form a location signal set. Feature extraction is performed on the location signal set to obtain the propagation parameters between each location and combine them into propagation entries; The propagation entries are weighted and fused with the clinical data to generate a propagation relationship table that includes directionality, relative intensity, hysteresis, and frequency band sensitivity; Based on the propagation relationship table, an individualized propagation model of the thoracic cavity and lung tissue is established. The model is constructed as a searchable data structure, which includes a first calculation module for path deduction, which receives body position parameters and corrects the propagation relationship by combining the direction of gravity to output the thoracic drainage path; at the same time, a second calculation module for parameter back-calculation is provided, which retrieves the corresponding propagation parameters after receiving the thoracic drainage path and generates oscillation control parameters including frequency, phase and location of action, thereby realizing the dual functions of path calculation and parameter generation.

[0026] In this embodiment, comprehensive individualized clinical data must first be collected from the patient at rest. This data includes not only the patient's basic physical parameters, such as height, weight, chest circumference, sex, and age, but also pulmonary function test results and a history of previous respiratory diseases. After standardization, this clinical data is used to generate a discrete location mapping table of the chest wall surface. This table divides the chest wall surface into several anatomically significant discrete regions, each with a number and coordinates, serving as a reference system for subsequent signal acquisition and model building.

[0027] After establishing the mapping table, acoustic and mechanical excitation devices need to be deployed at the corresponding discrete locations, and short-range excitation signals need to be applied. Acoustic excitation typically uses narrowband swept-frequency sound or short pulse sound, while mechanical excitation can be transient small-amplitude mechanical displacement or controlled vibration. Each location will generate a response signal under the excitation. These signals are collected by acoustic and accelerometer sensors placed at the corresponding locations and classified according to the mapping table numbers to form a location signal set. In this way, each discrete location has a corresponding set of original acoustic and mechanical signals, ensuring the consistency of data in space and anatomical location.

[0028] Next, in-depth feature extraction is required from the location signal set. For acoustic signals, the extracted features include relative amplitude, energy concentration, main frequency range, and phase delay; for mechanical signals, the extracted features include start-up delay, peak size, decay rate, and response curve shape. Combining these features forms propagation entries describing the propagation relationship between different locations. Each entry records the strength, delay, and frequency band characteristics of energy propagating between two locations, and is the basic building block of the model.

[0029] After obtaining the propagation entries, they need to be weighted and fused with the patient's individualized clinical data. During the fusion process, the strength and lag of the entries are adjusted based on the patient's chest wall thickness, lung function status, and disease type. For example, for patients with reduced lung compliance, the propagation attenuation weight on the corresponding pathway is increased; for patients with areas of local obstruction or inflammation, the stability score of propagation entries related to that area is reduced. The resulting propagation relationship table, obtained through this adjustment, more accurately reflects the patient's actual acoustic and mechanical propagation characteristics.

[0030] Based on the propagation relationship table, an individualized propagation model for the thoracic cavity and lung tissue is further established. This model is stored in the form of a data structure, possessing searchability and callability. It not only contains a complete set of propagation entries but also records the directionality, relative intensity, hysteresis, and frequency band sensitivity of each propagation path. To support subsequent applications, the model is constructed as a composite structure with two computational modules. The first module is used for path deduction. When the patient's positional parameters are input, this module corrects the propagation relationship by incorporating the direction of gravity. By searching the directed entries in the propagation relationship table, it derives the most probable thoracic drainage path and outputs a path sequence composed of discrete locations and corresponding accessibility scores. The second module is used for parameter backpropagation. Upon receiving the thoracic drainage path, this module retrieves path-related propagation entries from the propagation relationship table. Based on the frequency band information and phase delay of these entries, it calculates the optimal oscillation control strategy, thereby generating oscillation control parameters containing frequency, phase, and location of action.

[0031] Through this structural design, the individualized propagation model can not only output the thoracic drainage path, but also generate oscillation control parameters for the driving device under path constraints, realizing a continuous link from path calculation to control parameter generation. This dual function allows the model to handle both prediction and execution within a single data structure, ensuring the directionality of sputum drainage while providing precise control basis for the device.

[0032] For the first calculation module, upon receiving the patient's position parameters, it first converts the position information into a direction vector in a three-dimensional coordinate system to calibrate the projection of the gravity direction relative to the thoracic coordinate system. Then, in the propagation relationship table, it filters all directed entries close to this projection direction and weights and corrects the relative strength and hysteresis parameters of these entries, increasing the weight of entries in the direction of gravity and decreasing the weight of entries in the direction of gravity against gravity. Next, a graph search algorithm is used to find the optimal path from the chest wall surface to the lung segment region in the corrected directed graph, using the minimum energy attenuation and minimum hysteresis accumulation as selection criteria. Upon output, the module provides the number of each discrete location on the path, forming a complete location sequence, and calculates an accessibility score by accumulating the strength and stability of each entry, ensuring that the result includes not only the path but also a quantitative evaluation of the path's reliability.

[0033] For the second calculation module, upon receiving the thoracic drainage path, it first searches for propagation entries corresponding to adjacent locations along the path, extracting the frequency band sensitivity and phase delay information recorded in these entries. Then, it establishes a candidate frequency set within the frequency band sensitivity range and calculates whether phase combinations at different frequencies can form a continuous traveling wave based on the phase delay matching degree. Through iterative search and matching, the module selects the frequency and phase combinations that produce the most significant coherent superposition effect along the entire path. Next, combining the correspondence between the actuator positions and path locations, it determines the operating area of ​​each actuator and assigns it specific frequency, phase, and output intensity parameters. The final output oscillation control parameters include not only frequency, phase, and operating area, but also the driving sequence and pulse duration of each actuator, thus forming a complete control command that can be directly sent to the device execution end.

[0034] Furthermore, the step of collecting individualized clinical data and generating a discrete location mapping table of the thoracic surface while the patient is at rest includes: Individualized clinical data were collected from patients at rest, including chest circumference, height, weight, vital capacity, one-second rate, and history of respiratory diseases. This individualized clinical data was used as input for subsequent modeling. Based on the individualized clinical data, an initial discrete region division of the thoracic surface is established, which covers the anterior wall, lateral wall and back of the thoracic cavity, forming a set of discrete regions on the thoracic surface. Based on the set of discrete locations on the thoracic surface, the distribution density of the discrete locations is corrected according to the history of respiratory diseases in the individualized clinical data. For patients with chronic obstructive pulmonary disease, the density of discrete locations in the lower lung region is increased, and for patients with cystic fibrosis, the resolution of discrete locations around the central airway is increased, thereby obtaining the corrected set of discrete locations on the thoracic surface. The modified discrete location set of the thoracic surface is correlated with the anatomical associated regions of the lung lobes and bronchi, and a unique number is assigned to each discrete location. Finally, a discrete location mapping table of the thoracic surface is generated. The discrete location mapping table of the thoracic surface serves as a spatial reference for constructing the acoustic and mechanical signal acquisition and propagation relationship.

[0035] In implementing this invention, individualized clinical data needs to be collected from the patient while they are at rest to ensure the accuracy and stability of the input information during subsequent modeling. Individualized clinical data is not limited to basic anthropometric parameters but encompasses a wide range of information regarding the patient's thoracic structure, respiratory function, and disease background. This data includes physical measurements such as chest circumference, height, and weight, as well as functional indicators such as vital capacity and lung rate in one second (LPR). These functional indicators are typically tested using a pulmonary function testing machine and reflect the patient's maximum airflow patency and lung tissue compliance. Furthermore, a history of respiratory diseases, such as chronic obstructive pulmonary disease (COPD), asthma, or cystic fibrosis, must be recorded, as these conditions directly affect the biomechanical and acoustic propagation characteristics of local areas of the thoracic cavity. All the above data, after collection, are compiled into a single individualized clinical dataset for each patient, which will serve as the input for subsequent thoracic modeling.

[0036] After obtaining individualized clinical data, an initial discrete region division of the thoracic surface needs to be established based on this data. The logic behind this division is to view the thoracic surface as a whole composed of multiple smaller regions, each of which is a discrete region. The division must cover the anterior, lateral, and posterior walls of the thoracic cavity, extending from the lower clavicle border to the lower costal margin, and from the mid-axillary line to the scapular line. The number of divisions can be adjusted according to the patient's thoracic circumference and height; for example, for individuals with larger thoracic cavities, the number of discrete regions will be relatively increased to more precisely capture local acoustic and biomechanical properties. In this process, each discrete region forms a set element, resulting in a set of discrete regions on the thoracic surface, which serves as the basis for subsequent refinement and numbering.

[0037] After the initial partitioning, the discrete location set needs to be corrected based on the patient's medical history. This correction is based on increasing the distribution density or resolution of local locations, allowing areas significantly affected by the disease to be represented with higher precision in the model. For example, in patients with chronic obstructive pulmonary disease (COPD), airway obstruction and uneven ventilation are common in the lower lung region; therefore, the number of discrete locations in the lower lung region needs to be increased to more accurately reflect the differences in airflow and tissue response. For patients with cystic fibrosis, the main lesions are usually concentrated in and around the central airway; therefore, increasing the resolution of discrete locations on the thoracic surface corresponding to the central airway makes its representation in the model closer to reality. The corrected discrete location set on the thoracic surface not only reflects the individual's geometric characteristics but also includes the differences in pathological states.

[0038] After correction, it is necessary to establish a correspondence between the corrected set of discrete regions on the thoracic surface and the anatomical regions of the lungs and bronchi. This correspondence is achieved by mapping the position of each discrete region on the thoracic surface to an anatomical atlas. For example, the discrete region in the second intercostal space along the anterior midline can correspond to the anterior segment of the right upper lung, and the discrete region in the fifth intercostal space on the lateral side can correspond to the basal segment of the left lower lung. Through this mapping, each discrete region is not only a geometric location but also carries anatomical significance. To avoid confusion, each discrete region needs to be assigned a unique number, which is bound one-to-one with the spatial location of the discrete region and its corresponding anatomical region, ensuring accurate tracing and retrieval during subsequent signal acquisition and propagation relationship analysis.

[0039] Finally, all discrete locations, their numbers, correspondences, and correction information are compiled to generate a discrete location mapping table for the thoracic surface. This mapping table is a complete reference document that includes information on the geometric division of the thoracic cavity, as well as clinical data and medical history correction factors. It serves as a spatial reference for determining acquisition locations and interpreting propagation paths in subsequent acoustic and mechanical short-range excitation.

[0040] Furthermore, the step of applying acoustic and mechanical short-range excitations to discrete locations on the thoracic surface, acquiring corresponding signals, and forming a location signal set includes: Acoustic short-range excitation is applied to discrete locations on the thoracic surface, the acoustic short-range response of each discrete location is acquired, and a subset of acoustic location signals containing amplitude, energy distribution and phase difference is output. Short-range mechanical excitation is applied to discrete locations on the thoracic surface, the short-range mechanical response of each discrete location is acquired, and a subset of mechanical location signals including start-up delay, peak amplitude, and decay rate is output. The acoustic location signal subset and the mechanical location signal subset are aligned according to the discrete location number, and the trigger time of the reference discrete location is used for unified time-scale correction to output a time-synchronized composite location signal subset. The composite location signal subset is subjected to a consistency check, and responses with amplitude saturation or noise pollution are removed. The remaining responses are then weighted and averaged according to the stability of repeated excitation, and finally the location signal set after noise reduction and stability correction is output.

[0041] During implementation, acoustic and mechanical short-range excitations need to be applied sequentially to discrete locations on the chest wall surface to obtain raw signals that characterize the local tissue propagation properties. First, acoustic short-range excitations are applied to the discrete locations, using either narrowband sweep or short pulses. The excitation signal is applied to the chest wall surface via acoustic actuators, with the amplitude controlled to avoid patient discomfort. Acoustic sensing units positioned at each discrete location synchronously record the acoustic short-range response generated by the excitation. Each response includes signal amplitude, energy distribution across different frequency bands, and phase difference with the reference channel. These recorded results collectively form a subset of the acoustic location signal. The amplitude is obtained by taking the difference between the maximum value of the response waveform and the baseline. The energy distribution is calculated by integrating the energy across each frequency band, with the frequency bands covering the low- to mid-frequency range relevant to sputum expectoration. The phase difference is calculated by comparing the relative time delay between the response waveform and the reference channel waveform at the same frequency; a longer delay indicates a larger phase difference.

[0042] After acoustic excitation, short-range mechanical excitation is applied to discrete locations on the same thoracic surface. This mechanical excitation typically involves a small mechanical oscillator generating a minimal displacement or impact locally and instantaneously, with the amplitude strictly controlled within a safe range that the patient will not perceive or will only perceive slightly. Accelerometers or displacement sensors at each discrete location record the short-range mechanical response. Recorded metrics include onset delay (the time interval between the signal's origin and the response's inception point); peak amplitude (the difference between the maximum value of the response waveform and the baseline); and decay rate (the time required for the peak to decrease to a certain percentage of the baseline). All these quantified results are summarized into a subset of the mechanical location signal.

[0043] The acoustic and mechanical location signal subsets then need to be fused. To ensure that different modal signals can be compared and used under the same time reference, the trigger time of a pre-defined reference discrete location on the thoracic surface must be used as a unified timescale. All signal records in the subset need to be corrected to this trigger point. For example, if an acoustic signal appears two milliseconds earlier than the reference discrete location, its timescale in the record needs to be delayed by two milliseconds; similarly, if a mechanical signal appears three milliseconds later than the reference location, it needs to be aligned three milliseconds earlier. After this timescale unification, the output is a composite location signal subset containing the unified response information of each discrete location under acoustic and mechanical short-range excitation.

[0044] After obtaining the composite location signal subset, a consistency check is required to ensure data reliability. The first step of the check is to eliminate abnormal responses, such as saturation signals with amplitudes exceeding the sensor's normal range, or noise signals that deviate significantly from the normal waveform due to external interference. The second step is to adjust the signal weights based on the stability of repeated excitation. That is, if the response fluctuation of a certain discrete location is small during multiple repeated acquisitions, it indicates high stability and can be given a higher weight; if the fluctuation is large, it indicates poor stability and the corresponding weight should be reduced. Finally, a set of corrected signal results is obtained based on a weighted average. This result can simultaneously eliminate the influence of random noise and maintain a faithful reflection of the main propagation characteristics.

[0045] Through the above steps, the final output location signal set is a dataset that has undergone acoustic and mechanical short-range excitation acquisition, modal alignment, time synchronization, and consistency correction. This dataset fully reflects the propagation characteristics of each discrete location on the thoracic surface under dual-modal excitation and serves as a reliable input for subsequent propagation parameter extraction and propagation entry generation.

[0046] Furthermore, the step of extracting features from the location signal set to obtain propagation parameters between discrete locations on the thoracic surface and combining them into propagation entries includes: Frequency domain analysis is performed on the acoustic short-range response generated by acoustic short-range excitation in the location signal concentration, the dominant frequency, energy concentration and phase delay are extracted, and the results are output as acoustic short-range response characteristics; Time-domain analysis is performed on the short-range mechanical response generated by the short-range mechanical excitation in the location signal set to extract the start-up delay, peak amplitude and decay rate, and the results are output as short-range mechanical response characteristics. The acoustic short-range response features and the mechanical short-range response features are fused according to the discrete location number of the thoracic surface. Features with amplitude differences or inconsistent delays are weighted and corrected, and the fused propagation parameter set is output. Based on the set of propagation parameters, the directionality, relative intensity, hysteresis, and frequency band sensitivity between discrete regions on the thoracic surface are calculated, and the calculation results are combined into propagation entries.

[0047] After acquiring the location signal set, feature extraction is required to ensure that each discrete location on the thoracic surface is not merely a raw signal but forms calculable and comparable propagation parameters. For the acoustic short-range response generated by acoustic short-range excitation, the acquired waveform must first be converted into a frequency domain signal. This process can be achieved using a conventional Fourier transform, the core idea of ​​which is to decompose the original time waveform into several different frequency components. When extracting the dominant frequency, it is necessary to find the frequency point with the highest energy spectral density, that is, the location with the most concentrated energy in the entire frequency distribution. For example, if the energy value at a certain point in the 20 to 200 Hz range is significantly higher than the surrounding frequencies, this frequency is the dominant frequency. The energy concentration is calculated by statistically analyzing the proportion of energy in the narrow band containing the dominant frequency to the total energy. For example, adding the energy within ±5 Hz of the dominant frequency and dividing by the total energy, a higher proportion indicates a more concentrated energy. The extraction of phase delay relies on comparison with a reference location. That is, at the same main frequency, if the signal peak of a certain location lags behind the reference location by 2 milliseconds, the phase delay is equal to the proportion of the time difference within the corresponding period of that frequency, thus quantifying the phase relationship.

[0048] For short-range mechanical responses generated by short-range mechanical excitations, time-domain analysis methods are more suitable. The start-up delay is obtained by measuring the time difference between the trigger point of the reference excitation and the location's first significant deviation from the baseline, typically measured in milliseconds. The peak amplitude is extracted by finding the highest point of the response curve and calculating the difference with the resting baseline; a larger value indicates a higher sensitivity of the location to the excitation. The decay rate requires selecting a fixed threshold after the peak, such as 50% of the peak value, and calculating the time required for the temperature to drop from the peak to this threshold; a shorter time indicates faster decay and weaker coupling. These metrics are combined to output the short-range mechanical response characteristics.

[0049] Next, the acoustic short-range response features and the mechanical short-range response features need to be fused together, ensuring a one-to-one correspondence under the same discrete location number on the thoracic surface. This way, each discrete location has both acoustic and mechanical characteristics. If conflicts are found between the acoustic and mechanical characteristics of the same location during the fusion process—for example, the acoustic features show a small phase delay while the mechanical features show a long start-up delay—then a weighted correction is needed to unify them. The principle of weighted correction is to assign different weights based on signal stability. For example, in repetitive excitation, if the acoustic features fluctuate very little while the mechanical features fluctuate significantly, the weight of the acoustic features can be set to 0.7 and the mechanical features to 0.3 during fusion, resulting in a compromise propagation parameter. This process ensures that the fused parameter set reflects both dual-modal information and avoids a single outlier dominating the result.

[0050] After obtaining the fused propagation parameter set, it is necessary to calculate the specific propagation properties between discrete locations on the thoracic surface. Directionality is calculated based on the order of signal arrival between different locations. For example, in two adjacent locations, if the response at location A always precedes that at location B, the direction of energy propagation can be determined to be from A to B. Relative intensity is calculated by comparing the peak amplitudes of the two locations. For example, if the peak amplitude at location A is twice that at location B, the relative intensity can be expressed as 2:1. Hysteresis is calculated by normalizing the response delay time difference to the excitation period. For example, if the delay is 5 milliseconds and the excitation period is 50 milliseconds, the hysteresis is 0.1. Determining band sensitivity requires comparing the changes in energy concentration at different frequencies. For example, if the energy concentration is significantly higher in the low-frequency band and significantly lower in the high-frequency band, it indicates that the location is more sensitive to low frequencies. Through these calculations, four key indicators—directionality, relative intensity, hysteresis, and band sensitivity—can be formed and combined into propagation entries.

[0051] The essence of a propagation entry is to integrate multiple discrete propagation parameters into a callable data unit. Each entry represents the energy transfer relationship between two or more discrete locations on the thoracic surface.

[0052] Furthermore, the step of weightedly fusing the propagation entries with the individualized clinical data to generate a propagation relationship table including directionality, relative intensity, hysteresis, and frequency band sensitivity includes: Based on the parameters extracted and propagated from the individualized clinical data, the parameters include at least chest circumference, height, weight, vital capacity, one-second rate, and history of respiratory diseases, and the parameters are output as a weighted input set. The propagation entries are categorized according to the anatomically associated regions of the discrete location mapping table of the thoracic surface, and initial values ​​of directionality, relative intensity, hysteresis and frequency band sensitivity are generated for each propagation entry, and the initial propagation entry set is output. The initial propagation entry set is modified based on the weighted input set. For individuals with chronic obstructive pulmonary disease, the hysteresis weight is increased and the relative intensity is decreased in the relevant discrete locations. For individuals with cystic fibrosis, the frequency band sensitivity of the low-frequency band is increased in the relevant discrete locations of the central airway. For individuals whose chest wall thickness is increased due to increased chest circumference or weight, the directionality of the high-frequency band is decreased in the corresponding discrete locations. The modified propagation entry set is then output. The modified propagation entry set is subjected to stability and consistency screening. Based on the stability score of repeated collections and the spatial continuity of the discrete location mapping table of the thoracic surface, propagation entries below the threshold or those that conflict with the directionality of adjacent entries are removed. For propagation entries with opposite directions within the same path, the net dominant direction is retained, and a qualified propagation entry set is output. The qualified propagation entry set is indexed according to the discrete location number of the thoracic surface, and the directionality, relative intensity, hysteresis and frequency band sensitivity of each propagation entry are recorded. At the same time, the entry weight and stability score are retained as auxiliary fields, and finally a propagation relationship table is generated.

[0053] When weighted fusion of propagation entries and individualized clinical data, each entry needs to be linked to the patient's specific physiological characteristics. This ensures that the propagation relationship table reflects not only the objectively collected acoustic and mechanical propagation characteristics but also individual patient differences. First, parameters related to propagation correction need to be extracted from the individualized clinical data, including chest circumference, height, weight, vital capacity, one-second rate, and history of respiratory diseases. These parameters are organized into a weighted input set for subsequent entry correction. Chest circumference, height, and weight primarily affect chest wall thickness and the overall mechanical conduction pathway; vital capacity and one-second rate reflect the patient's basic respiratory function level and affect airway patency and oscillatory transmission efficiency; and medical history determines which areas of pathological change require special weighting.

[0054] After obtaining these parameters, the propagation entries need to be categorized according to the anatomical correspondences indicated by the discrete location mapping table of the thoracic surface. Each propagation entry is essentially an energy propagation relationship between two discrete locations, and therefore can be located to a specific lobe or airway branch through anatomical mapping. Initial values ​​are then generated for each propagation entry, including directionality, relative intensity, hysteresis, and band sensitivity. Directionality is determined by the order in which the signal appears between the two locations; for example, if the signal always appears first in location A and then in location B, the directionality is A to B. Relative intensity is determined by comparing the ratio of the response amplitudes of the two locations; for example, if the peak value of A is twice that of B, the relative intensity is 2:1. Hysteresis is the ratio of the time delay to the excitation period; for example, if the delay is 10 milliseconds and the excitation period is 100 milliseconds, the hysteresis is 0.1. Band sensitivity is determined by comparing the changes in energy concentration at different frequencies; if the energy is significantly concentrated in the low-frequency band, it indicates that the location is sensitive to low frequencies.

[0055] After generating the initial values, they need to be corrected based on the weighted input set. This correction involves adjusting the parameters according to different clinical conditions. For example, if a patient has chronic obstructive pulmonary disease (COPD), airway obstruction is common in the lower lung region, leading to sluggish gas flow; therefore, the hysteresis weight of the relevant entries needs to be increased, and the relative intensity decreased. For individuals with cystic fibrosis, the central airway often contains more viscous secretions, making low-frequency signals easier to capture; therefore, the low-frequency band sensitivity of the relevant entries needs to be increased. For individuals with increased chest circumference or weight, increased chest wall thickness causes high-frequency signals to attenuate more quickly; therefore, the weight of high-frequency directionality needs to be decreased in the corresponding entries. This corrected set of propagation entries more accurately reflects individual differences.

[0056] After correction, stability and consistency screening are required to ensure the items can be reliably used in subsequent models. The screening criteria have two aspects: stability and consistency. Stability refers to whether the response fluctuations of the same location in multiple repeated samplings are within acceptable limits. Excessive fluctuations indicate the item is unreliable and needs to be removed. Consistency refers to whether the directionality of an item forms a continuous spatial transmission relationship with adjacent items. If the directionality of an item is significantly opposite to that of adjacent items and lacks sufficient energy support, it needs to be deleted or adjusted. When directional conflicts exist in the processing path, the net dominance principle is adopted, meaning that only items with greater consistency in directionality and intensity are retained within the same path, while weaker items are removed. Through this process, a set of qualified propagation items is finally formed.

[0057] Finally, the set of qualified transmission entries needs to be indexed according to the discrete location numbers of the thoracic surface, and the directionality, relative intensity, hysteresis, and frequency band sensitivity of each entry need to be recorded. At the same time, to maintain model updability, the weight and stability score of each entry should also be retained as supplementary fields. This information together forms a transmission relationship table, which is not only a static set of connections but also has the ability to be dynamically updated and expanded, continuously correcting itself as patient data increases.

[0058] Furthermore, the first computing module is specifically used for: Receive the body position parameters and combine the body position parameters with the gravity direction to form a set of correction parameters; Based on the set of correction parameters, the propagation entries in the propagation relationship table are weighted and corrected, and the corrected propagation relationship table is output. A directed graph search is performed in the revised propagation table to derive the thoracic drainage path according to the criteria of minimizing energy attenuation and hysteresis accumulation. The chest drainage path is calculated to generate an accessibility score that includes a combination of directionality, relative intensity, and hysteresis, and is output along with the chest drainage path.

[0059] In this embodiment, the first calculation module combines the patient's positional information with the energy propagation characteristics within the thoracic cavity to derive the optimal thoracic drainage path for sputum migration and assigns a quantitative accessibility score to this path. First, it needs to receive positional parameters. These parameters are not simply posture descriptions but include the spatial orientation vector of the patient's chest surface in the current position, the angle between the chest wall and the spine, and the projection of the gravitational direction onto the chest surface coordinate system. For example, if the patient is supine, the gravitational direction is essentially perpendicular to the back plane, and the positional parameters need to quantify this relationship. By combining the positional parameters with the gravitational direction, a set of correction parameters can be formed. This set is used to adjust the priority of propagation between different locations. The correction method involves projecting the gravitational direction onto a discrete location mapping of the chest surface, increasing the weight of propagation in areas aligned with the gravitational direction, and decreasing the weight in areas opposite to the gravitational direction, thus allowing the model to more closely approximate the actual migration trend of sputum under the influence of gravity.

[0060] After obtaining the set of corrected parameters, the propagation relationship table needs to be weighted. The propagation relationship table contains multiple propagation entries, each recording the directionality, relative intensity, hysteresis, and band sensitivity between two discrete locations. During weight correction, the set of corrected parameters needs to be applied to these entries. For example, when an entry represents the propagation direction from the lower lung location to the upper lung location, if the patient's position aligns this direction with the direction of gravity, the weight of this entry needs to be increased. This can be achieved by multiplying the relative intensity coefficient by a correction factor greater than 1, such as 1.2 or 1.3. Conversely, if the direction is opposite to the direction of gravity, the weight needs to be decreased, for example, by multiplying by 0.8 or 0.7. This correction method allows the propagation relationship table to be dynamically adjusted under different body positions, better reflecting the actual physical conduction laws. The corrected propagation relationship table will serve as the basis for the next search step.

[0061] After revising the propagation relationship table, the first calculation module needs to perform a directed graph search within it. The propagation relationship table can be viewed as a directed graph, with discrete locations as nodes and propagation entries as edges, each edge carrying weights such as directionality, relative strength, and hysteresis. The goal of the search is to find a path to the target location from possible starting locations that minimizes energy attenuation and hysteresis accumulation. Minimizing energy attenuation is calculated by multiplying the relative strengths of all edges along the path; the closer the value is to 1, the better the energy retention. Minimizing hysteresis accumulation is calculated by summing the hysteresis values ​​of all edges along the path; the smaller the value, the faster the overall propagation. During the search, an improved Dijkstra's algorithm or the A* algorithm can be used, treating energy attenuation and hysteresis accumulation as dual cost functions, and determining the optimal path by comparing the total costs of different paths. The result of this process is the chest drainage path, representing the main direction of sputum migration under the current body position and propagation characteristics.

[0062] Finally, the chest drainage path needs to be calculated to generate an accessibility score. The accessibility score is a comprehensive evaluation index that combines the consistency of path directionality, the magnitude of relative intensity, and the accumulation of hysteresis. The calculation first confirms whether the path's directionality aligns with the gravity direction in the correction parameter set; higher consistency results in a higher score. Then, the geometric mean of the relative intensity along the path is used as a quantitative indicator of energy transfer capability; a value closer to 1 indicates greater intensity. Next, the average hysteresis along the path is used as a measure of transmission delay; smaller delays result in a higher score. Finally, these three results are combined using a weighted approach, for example, 40% for directionality, 40% for relative intensity, and 20% for hysteresis, to form an accessibility score between 0 and 1. This score is output along with the chest drainage path, providing both the path itself and an evaluation of its reliability, thus providing a solid basis for the subsequent generation of oscillation control.

[0063] Furthermore, the second computing module is specifically used for: Receive the chest drainage path, retrieve the propagation entry corresponding to the chest drainage path in the propagation relationship table, and output a set of path-related propagation entries; Extract the frequency band sensitivity and phase delay information from the path-related propagation entry set, and output the candidate frequency phase combination set; A matching search is performed in the candidate frequency phase combination set to select the optimal combination that can form a continuous traveling wave on the chest drainage path, and the oscillation control parameters are output.

[0064] In this embodiment, the task of the second calculation module is to generate oscillation control parameters that can actually drive the device to perform operations based on the chest drainage path, so that energy forms a continuous traveling wave in the thoracic cavity and lung tissue, thereby effectively promoting the migration of sputum along the predetermined path. Its working logic is not simply calling existing parameters, but through a series of rigorous calculation steps, selecting combinations that meet the drainage requirements from the propagation relationship table layer by layer, and gradually converting them into executable control commands.

[0065] Once the thoracic drainage path has been output by the first calculation module, the second calculation module first receives the path and retrieves the corresponding propagation entry from the propagation relationship table. The propagation relationship table is a directed data structure with discrete regions on the thoracic surface as nodes and propagation entries as edges. Each entry contains parameters such as directionality, relative intensity, hysteresis, and frequency band sensitivity. The propagation entry corresponding to the thoracic drainage path refers to the entry between each discrete region connected by a segment of the path. By retrieving these entries, a set can be obtained that completely records all the local propagation characteristics required to transfer energy along the path; this set is called the path-related propagation entry set.

[0066] After obtaining the set of path-related propagation entries, the second calculation module needs to further extract band sensitivity and phase delay information from it. Band sensitivity reflects the response of the path at different frequencies. For example, an entry may exhibit high energy concentration in the low-frequency band but significant attenuation in the high-frequency band, indicating that the entry is more easily propagated under low-frequency excitation. Phase delay refers to the time delay that occurs when a signal is transmitted from one location to another, which directly affects the phase coordination between multiple locations. If the phase delay is too large, the traveling wave may be disrupted, resulting in local reflections and reducing the expectoration effect. Therefore, during extraction, a corresponding band response curve and a phase delay value need to be recorded for each entry on the path. Integrating these data forms a candidate frequency-phase combination set, which is actually an arrangement of multiple sets of frequencies and phase delays, each representing a possible driving mode.

[0067] After the candidate frequency-phase combination set is generated, the second calculation module needs to perform a matching search to find the optimal combination that can form a continuous traveling wave along the thoracic drainage path. The formation of a continuous traveling wave requires two conditions: first, the frequency must fall within the sensitive range of most entries along the path; second, the phase delays need to be progressively connected, allowing the signal propagation from the start to the end of the path to exhibit a coherent phase progression. In specific calculations, a dominant frequency can be selected from the candidate set, and then the frequency band sensitivity of that frequency for each entry along the path can be checked. If the energy concentration of an entry at that frequency is lower than a set threshold, such as 20% of the total energy, then that frequency is deemed unsuitable. If the sensitivity screening passes, the phase delays are further checked to see if they can form an approximately arithmetic progression relationship. For example, if the phase delays in the three sections of the path are 5 milliseconds, 10 milliseconds, and 15 milliseconds respectively, it indicates that the signal phase is progressively advancing and can form a stable traveling wave. If a combination meets these conditions, it is marked as a candidate optimal combination.

[0068] Among all candidate combinations, the second calculation module also calculates a comprehensive score based on energy retention rate and phase consistency. The energy retention rate is calculated by multiplying the relative intensities of all entries along the path; the closer the result is to 1, the smaller the energy loss. Phase consistency is calculated by checking the deviation of the cumulative phase delay from the ideal arithmetic progression sequence; the smaller the deviation, the better the consistency. The comprehensive score can be calculated using a weighted method, for example, with energy retention rate accounting for 60% and phase consistency accounting for 40%, resulting in a final score. The combination with the highest score is the optimal combination. This combination is directly converted into oscillation control parameters, including three core components: frequency, phase, and location of action. The frequency is determined by the dominant frequency of the optimal combination, the phase is calculated from the delay sequence, and the location of action is determined by key nodes in the thoracic drainage path. Typically, the main oscillation source is set at the beginning and middle of the path, and a phase compensation source is set near the end of the path to ensure smooth energy transfer.

[0069] Ultimately, the oscillation control parameters output by the second calculation module are a set of control commands that can be directly transmitted to the actuator, ensuring that the device generates coordinated directional pressure traveling waves in multiple areas of the chest wall. These parameters not only guarantee matching with the propagation relationship table but also ensure consistency with the thoracic drainage path, thus achieving a closed loop between path calculation and parameter generation.

[0070] Step S102: Input the patient's current position into the individualized propagation model to obtain the chest drainage path.

[0071] In step S102, the patient's current position needs to be passed as input to the individualized propagation model of the thoracic wall and lung tissue established in the previous step. This allows the spatial orientation and gravitational influences to be superimposed on the model, thereby outputting the thoracic drainage path. To ensure the operability of this process, the method of obtaining the current position must first be clarified. Patients may be in common positions such as sitting, supine, prone, or lateral decubitus during treatment. Different positions will cause significant changes in the spatial distribution and stress state of the chest wall and lung tissue. Position information can be collected in two ways: one is by relying on an external sensing system, such as deploying inertial measurement units at key parts of the patient's body to capture three-dimensional posture data in real time; the other is by having the operator directly input the patient's position category through the position selection interface on the treatment device. After obtaining the position information, it needs to be converted into a parameter form that can be combined with the individualized propagation model. For example, Euler angles in a three-dimensional coordinate system can be used to represent the rotation angle of the thoracic wall, or position labels can be mapped to a preset spatial layout of the chest wall area, so that the model can correct the energy propagation direction according to different positions.

[0072] Once the body position parameters are input into the model, the model calls upon the propagation relationship table and anatomical correspondence table established in step S101, superimposing the direction of gravity onto the spatial relationship between the chest wall grid and the lung location. Specifically, the model reorders all possible energy transfer paths, giving higher priority to paths more favorable along the direction of gravity, while reducing the weight of paths in the opposite direction of gravity. At this point, the model comprehensively considers acoustic propagation characteristics and mechanical coupling characteristics, such as the energy transfer intensity, time hysteresis, frequency band response range, and phase sequence between adjacent locations, and combines this with the lung function status and lesion distribution reflected in the patient's individualized clinical data, ultimately deriving the most likely direction and region of sputum migration. This output is the chest drainage path, which is not an abstract curve, but a pathway composed of a series of numbered locations, each with a clear energy transfer intensity, directionality, and reliability score.

[0073] To ensure the practical operability of the results, the model includes a quantitative description of the effectiveness of the output chest drainage pathway. For example, for a drainage pathway, the model calculates the overall attenuation rate of energy transfer along the pathway within the sputum-related frequency band, as well as the expected migration rate of sputum under the combined effects of gravity and oscillation. If the energy attenuation rate of a pathway is too high, or if a key area within it shows a significant increase in mechanical impedance, the model automatically lowers the priority of that pathway, indicating that the feasibility of the pathway may need to be enhanced in subsequent steps through positional adjustments or parameter corrections.

[0074] In clinical applications, the output chest drainage pathways are usually more than one. The model prioritizes several candidate pathways and assigns a score. Some of these pathways may be more suitable for natural drainage in the patient's current position, while others may require subsequent micro-positional adjustments or stronger oscillation control to be fully activated. In either case, the model's output is provided in the form of clear locative sequences and quantitative indicators, enabling the device to use this information to generate specific oscillation control parameters in the next step.

[0075] Therefore, the implementation process of step S102 is to use an individualized transmission model to integrate the patient's current position with the transmission characteristics of the chest and lung tissue, and deduce a chest drainage path that conforms to physical laws and physiological characteristics.

[0076] Step S103: Under the constraint of the thoracic drainage path, the individualized propagation model is invoked to generate oscillation control parameters including frequency, phase and action area. The oscillation control parameters are used as execution instructions, which are used to synthesize directional propagating pressure traveling waves in multiple areas of the chest wall.

[0077] In step S103, given the already obtained chest drainage path, the individualized propagation model of the thoracic cavity and lung tissue needs to be invoked again to generate oscillation control parameters that can drive the device under path constraints. The core of this process is to transform the drainage path into specific execution instructions, enabling the multi-point action units of the chest wall to coordinate their outputs and synthesize directional propagating pressure waves, thereby propelling sputum along the pre-calculated path.

[0078] When the thoracic drainage path is input into the model, the model first filters the regional nodes related to that path in the propagation relationship table and restricts the calculation to only those nodes. This constraint prevents energy from diffusing between locations unrelated to drainage, ensuring the directionality of subsequent control strategies. Subsequently, the model retrieves existing acoustic and mechanical propagation characteristics between the path locations, including parameters such as energy transfer intensity, phase delay, frequency response range, and local impedance. Using this data, the model can calculate whether energy can be effectively transferred between adjacent locations along the path under different frequency and phase combinations.

[0079] During the calculation process, the model prioritizes combinations with the lowest energy attenuation and minimal phase mismatch within the relevant frequency band for sputum clearance. If multiple candidate frequencies exist in a certain path segment, their cumulative effect along the entire path is evaluated through superposition simulation calculations, selecting the frequency range with the highest overall energy transfer efficiency. The phase selection ensures coherent superposition between adjacent locations, allowing the pressure wave to advance continuously rather than cancel each other out. For determining the location of action, the model combines the mapping relationship between each vibration unit on the equipment and discrete locations on the chest wall, prioritizing the activation of units geometrically closest to the path location, while simultaneously activating neighboring units based on propagation strength to create a spatially directional enhancement effect.

[0080] When generating control parameters, the model not only provides the frequency and phase corresponding to each action unit, but also outputs additional parameters such as vibration intensity and pulse width to ensure that the energy transfer is strong enough to push the mucus without exceeding the patient's tolerance range. The final control parameters are a complete set of execution instructions, which clearly list the frequency, phase, intensity, and timing sequence to be applied to each of the multiple points on the chest wall. After receiving these instructions, the device can output them simultaneously at multiple locations according to the settings, thereby synthesizing a directional pressure traveling wave on the chest wall surface.

[0081] The formation mechanism of this pressure traveling wave can be understood as the synergistic superposition effect of multiple oscillating units. When units in different locations output according to the frequency and phase calculated by the model, their vibrations superimpose spatially, forming a wave that propagates from one region to another. Since the position, frequency, and phase of these units have been optimized by the propagation model and are completely consistent with the direction of the chest drainage path, the final synthesized pressure traveling wave will propagate directionally along this path, facilitating the smooth migration of sputum.

[0082] In this way, step S103 realizes the transformation from "path" to "control". The chest drainage path obtained in the previous step only indicates the possible direction of sputum migration, while in this step, the model generates specific control parameters based on this direction, enabling the device to actually synthesize directional pressure traveling waves on the chest wall.

[0083] Step S104: Execute the execution command and implement safety gating output according to the ECG and respiratory phases, while collecting aerosol echoes and high-frequency impedance disturbances as feedback signals.

[0084] In step S104, the execution command generated in the previous step needs to be applied to the patient's chest wall, causing multiple oscillation units to output in tandem according to the set frequency, phase, and action area, forming a directional pressure traveling wave. During this process, it is crucial to ensure that the output of the oscillation signal is strictly aligned with the patient's ECG and respiratory cycles; this is known as safety-gated output. Specifically, the treatment device must monitor the patient's ECG and respiratory signals in real time before outputting the signal and set gating conditions in the algorithm, such as avoiding high-intensity oscillations during the sensitive period of cardiac contraction or at respiratory limit points, thereby preventing potential arrhythmias or respiratory depression. To achieve this, the device typically acquires ECG waveforms through ECG electrodes, monitors respiratory rhythm through a chest strap or airway pressure sensor, and marks each safe time window in the signal processing module. The frequency and phase information in the execution command are triggered within these safe windows, ensuring that the application of the oscillation meets the needs of sputum clearance without interfering with cardiopulmonary function.

[0085] During the oscillation signal output process, multiple units on the chest wall work sequentially or simultaneously according to the model's instructions, forming a pressure traveling wave that propagates from one point to another. To verify the actual effect of oscillation on mucus migration, feedback signals from within the patient's pleural cavity need to be acquired simultaneously. These feedback signals mainly include two types: aerosol echoes and high-frequency impedance perturbations. Aerosol echoes are obtained by detecting the movement of tiny particles and acoustic backscattering signals in the expiratory airflow. They reflect the movement of sputum in the airway; for example, when sputum is pushed, broken, or moved, a brief energy peak appears in the echo. High-frequency impedance perturbations are obtained by applying a weak high-frequency detection signal to the chest wall or airway and measuring changes in impedance in real time. The movement of sputum and changes in airway resistance cause fluctuations in local impedance, which can be used to determine whether the airway has become more open.

[0086] The acquired aerosol echoes and high-frequency impedance disturbances need to be strictly aligned with the timing of the oscillation execution. This means that each feedback signal must correspond to the actual output time point of a certain set of oscillation parameters. To this end, the device adds a precise time stamp when outputting control commands and records the acquired feedback signals using the same time base. This alignment ensures that in subsequent analysis, it is clear which oscillation, which combination of frequency and phase, caused a particular feedback phenomenon, thus providing a precise causal basis for the next step of parameter correction.

[0087] In this way, step S104 not only fulfills the actual function of executing instructions, but also ensures the safety of treatment through the gating mechanism, and provides reliable input for closed-loop control through the feedback acquisition mechanism.

[0088] Step S105: Align the feedback signal with the execution timing to form feedback data, and update the individualized propagation model of the thoracic cavity and lung tissue and the oscillation control parameters based on the feedback data.

[0089] In step S105, the feedback signal acquired in step S104 needs to be precisely aligned with the timing of the execution command to form a feedback dataset that can be analyzed and corrected. To this end, each set of oscillation parameters is first marked with a strict time stamp when the execution command is issued, including the start time, duration of action, and the frequency and phase combinations involved. Simultaneously, time stamps are also written into the data stream to ensure that the two data sources have the same reference. This bidirectional marking method avoids mismatches caused by sampling delays or channel asynchrony, allowing subsequent analysis to clearly identify which set of control parameters corresponds to a particular aerosol echo or a high-frequency impedance disturbance.

[0090] Once the feedback signal is aligned with the execution timing, the system processes and organizes the feedback data. Aerosol echoes typically manifest as a sudden increase in signal intensity or a shift in energy peaks during exhalation, reflecting the process of sputum being pushed, broken, or redistributed. High-frequency impedance perturbations, on the other hand, manifest as instantaneous drops or rises in the impedance curve during the respiratory cycle, indicating whether the airway has become more open due to reduced sputum. After being aligned with time, these raw signals are converted into a set of quantitative indicators, such as estimated sputum migration rates, changes in local airway resistance, and the speed at which impedance recovers to baseline. These indicators objectively assess the sputum expectoration effect of each set of oscillation parameters in the current body position.

[0091] After generating the feedback dataset, it needs to be re-input into the individualized propagation model of the thoracic cage and lung tissue. The model uses these feedback indicators to update the original propagation relationship table and control parameter library. For example, if a certain path consistently exhibits low sputum migration rate or insignificant impedance changes after multiple executions, the model will reduce the priority of that path or even mark it as an inefficient path; conversely, if a combination of frequencies and phases consistently produces significant migration signals and impedance decreases in the feedback, the model will increase the weight of that combination in parameter selection. The update process is not a simple replacement but a gradual adjustment, allowing the model to gradually converge to a propagation structure that better reflects the patient's actual physiological state.

[0092] Simultaneously, the oscillation control parameters are updated. Based on feedback data, the system fine-tunes the frequency range, phase difference, application area, and output intensity based on the original parameters. For example, if feedback shows severe energy attenuation midway through the path, the vibration intensity at the initial location can be appropriately increased, or the phase difference between adjacent units can be adjusted to form a stronger directional traveling wave; if feedback shows insufficient patient tolerance or unstable cardiopulmonary coupling, the oscillation intensity or pulse duration will be reduced. In this way, the updated control parameters can better balance efficacy and safety.

[0093] Ultimately, the introduction of feedback data creates a closed-loop system, where the model is no longer static but continuously optimized based on the patient's real-time responses in each treatment. This means subsequent steps no longer rely on the initial modeling results but instead depend on a continuously updated, individualized propagation model and dynamically adjusted oscillation control parameters, ensuring the method's stability, adaptability, and efficiency in long-term applications.

[0094] Furthermore, the aforementioned high-frequency whole-chest oscillation sputum clearance method based on clinical data adaptive adjustment also includes: During the gating window period, a micro-position adjustment command is generated and executed to align the main direction of oscillation with the direction of gravity to enhance drainage. The process involves repeatedly executing the steps of inputting the patient's current position into the individualized propagation model to obtain the chest drainage path, aligning the feedback signal with the execution timing to form feedback data, and updating the individualized propagation model of the thoracic cavity and lung tissue and the oscillation control parameters based on the feedback data, until the feedback signal indicates that the mucus migration rate and impedance index meet the preset termination conditions.

[0095] In the steps described above in this embodiment, the core idea is to fully utilize the window period formed by safety gating during the oscillation execution process, combined with the dynamic characteristics of real-time feedback signals and the individualized propagation model, to make subtle adjustments to the patient's position, thereby further optimizing the drainage effect of sputum in the pleural cavity. The so-called window period formed by safety gating refers to the prohibited output interval defined by both ECG gating and respiratory gating mechanisms. In ECG gating, the device collects the patient's ECG signal in real time and identifies sensitive periods in the cardiac cycle, such as during cardiac repolarization. To avoid interfering with the heart rhythm, any high-frequency oscillation signal output is prohibited during this period. In respiratory gating, the device synchronously monitors the patient's respiratory cycle, and similarly prohibits the output of high-frequency signals during periods that are prone to causing discomfort or interfering with ventilation, such as the inspiratory peak or expiratory peak. When these two types of gating conditions are superimposed, several intermittent blank windows are formed on the time axis; these windows are the window periods. During the window periods, the device does not apply high-frequency oscillation signals to the chest wall, and the patient's ECG and respiration are in a relatively stable state, thus allowing for safe insertion of micro-positioning adjustments.

[0096] Micro-positioning doesn't involve drastic changes like shifting the patient from a sitting to a supine or vice versa. Instead, it involves using the treatment bed's electric support system or auxiliary positioning devices to tilt or elevate the patient's body at small angles. This might include slightly raising the chest, slightly turning the body, or finely adjusting the height of the shoulder and back support pads. These adjustments are typically between a few and a dozen degrees, ensuring patient comfort without affecting the stability of the sensors and actuators. In this way, the anatomical structures within the chest cavity undergo a slight change relative to gravity, shifting the preferred path for sputum migration. When this adjustment aligns with the predominant oscillation direction derived from the model, the directional traveling wave and gravity create a superposition effect, significantly enhancing the efficiency of sputum migration along the target path.

[0097] After a micro-positioning adjustment, the method enters a cyclic execution mode. This means the patient's new position is input into the individualized propagation model, which recalculates the thoracic drainage path based on the updated spatial orientation. Once the new drainage path is obtained, corresponding oscillation control parameters are generated again to guide the device in outputting new execution commands. Subsequently, oscillation output is executed via safety gating, while simultaneously acquiring feedback signals such as aerosol echoes and high-frequency impedance perturbations. The acquired feedback signals are aligned with the timing of the oscillation execution, forming new feedback data, which is used to update the propagation model and oscillation control parameters. This process forms a continuous closed loop; each positioning adjustment triggers a new round of path calculation, parameter generation, oscillation execution, and feedback acquisition, ensuring the entire treatment process is continuously optimized based on the patient's immediate responses.

[0098] The termination condition for the cycle is given by the feedback signal. When the feedback signal indicates that the sputum migration rate has reached a stable and sufficiently high level, and the airway resistance has significantly decreased and tended to stabilize, it can be considered that sputum expectoration is basically complete, and the treatment process can end. In this way, the treatment no longer relies on a preset fixed time or human judgment, but is based on the patient's own physiological feedback, making the sputum expectoration process adaptive and individualized.

[0099] In the above embodiments, a high-frequency whole-chest oscillation sputum clearance method based on adaptive adjustment of clinical data is described. Correspondingly, this application also provides a high-frequency whole-chest oscillation sputum clearance system based on adaptive adjustment of clinical data. Please refer to... Figure 2 This is a schematic diagram of an embodiment of a high-frequency whole-chest oscillatory sputum clearance system based on adaptive adjustment of clinical data according to this application. Since this embodiment, namely the second embodiment, is basically similar to the method embodiment, it is described simply; relevant details can be found in the description of the method embodiment. The system embodiment described below is merely illustrative.

[0100] The second embodiment of this application provides a high-frequency whole-chest oscillatory sputum clearance system based on clinical data adaptive adjustment, comprising: The acquisition unit 201 is used to acquire individualized clinical data, short-range acoustic response and short-range biomechanical response of the chest at multiple points, and to establish an individualized propagation model of the thoracic cage and lung tissue based on the acquired data. Input unit 202 is used to input the patient's current position into the individualized propagation model to obtain the chest drainage path; The generation unit 203 is used to call the individualized propagation model under the constraint of the chest drainage path, generate oscillation control parameters including frequency, phase and action area, and use the oscillation control parameters as execution instructions. The execution instructions are used to synthesize directional propagating pressure traveling waves in multiple areas of the chest wall. Execution unit 204 is used to execute the execution instructions and implement safety gating output according to the ECG and respiratory phases, while collecting aerosol echoes and high-frequency impedance disturbances as feedback signals; Alignment unit 205 is used to align the feedback signal with the execution timing to form feedback data, and update the individualized propagation model and oscillation control parameters of the thoracic cavity and lung tissue based on the feedback data.

[0101] A third embodiment of this application provides an electronic device, the electronic device comprising: processor; A memory is used to store a program, which, when read and executed by the processor, performs the method provided in the first embodiment of this application.

[0102] The fourth embodiment of this application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, performs the method provided in the first embodiment of this application.

[0103] Although this application discloses preferred embodiments as described above, it is not intended to limit this application. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of this application. Therefore, the scope of protection of this application should be determined by the scope defined in the claims of this application.

Claims

1. An electronic device, characterized in that, The electronic device includes: processor; The memory is used to store programs, which, when read and executed by the processor, perform the following steps: Collect individualized clinical data, short-range acoustic response and short-range biomechanical response of the chest at multiple points, and establish an individualized propagation model of the thoracic cage and lung tissue based on the collected data; Input the patient's current position into the individualized propagation model to obtain the chest drainage path; Under the constraints of the thoracic drainage path, the individualized propagation model is invoked to generate oscillation control parameters including frequency, phase and location of action. The oscillation control parameters are used as execution commands, which are used to synthesize directional propagating pressure traveling waves in multiple locations of the chest wall. The execution command is executed and a safety gating output is implemented according to the ECG and respiratory phases, while aerosol echo and high-frequency impedance disturbance are collected as feedback signals. The feedback signal is aligned with the execution timing to form feedback data, and the individualized propagation model and oscillation control parameters of the thoracic cavity and lung tissue are updated based on the feedback data; The process of collecting individualized clinical data, short-range acoustic response and short-range biomechanical response of the chest at multiple points, and establishing an individualized transmission model of the thoracic cage and lung tissue based on the collected data includes: Individualized clinical data were collected from patients at rest and a discrete location mapping table of the thoracic surface was generated. Short-range acoustic and mechanical excitations are applied to the discrete locations, and corresponding signals are collected to form a location signal set. Feature extraction is performed on the location signal set to obtain the propagation parameters between each location and combine them into propagation entries; The propagation entries are weighted and fused with the clinical data to generate a propagation relationship table that includes directionality, relative intensity, hysteresis, and frequency band sensitivity; Based on the propagation relationship table, an individualized propagation model of the thoracic cavity and lung tissue is established. The model is constructed as a searchable data structure, which includes a first calculation module for path deduction, which receives body position parameters and corrects the propagation relationship by combining the direction of gravity to output the thoracic drainage path; at the same time, a second calculation module for parameter back-calculation is provided, which retrieves the corresponding propagation parameters after receiving the thoracic drainage path and generates oscillation control parameters including frequency, phase and location of action, thereby realizing the dual functions of path calculation and parameter generation.

2. The electronic device according to claim 1, characterized in that, When the program is read and executed by the processor, it also performs the following steps: During the gating window period, a micro-position adjustment command is generated and executed to align the main direction of oscillation with the direction of gravity to enhance drainage. The process involves repeatedly executing the steps of inputting the patient's current position into the individualized propagation model to obtain the chest drainage path, aligning the feedback signal with the execution timing to form feedback data, and updating the individualized propagation model of the thoracic cavity and lung tissue and the oscillation control parameters based on the feedback data, until the feedback signal indicates that the mucus migration rate and impedance index meet the preset termination conditions.

3. The electronic device according to claim 1, characterized in that, When the program is read and executed by the processor, the step of collecting individualized clinical data and generating a discrete location mapping table of the thoracic surface while the patient is at rest includes: Individualized clinical data were collected from patients at rest, including chest circumference, height, weight, vital capacity, one-second rate, and history of respiratory diseases. This individualized clinical data was used as input for subsequent modeling. Based on the individualized clinical data, an initial discrete region division of the thoracic surface is established, which covers the anterior wall, lateral wall and back of the thoracic cavity, forming a set of discrete regions on the thoracic surface. Based on the set of discrete locations on the thoracic surface, the distribution density of the discrete locations is corrected according to the history of respiratory diseases in the individualized clinical data. For patients with chronic obstructive pulmonary disease, the density of discrete locations in the lower lung region is increased, and for patients with cystic fibrosis, the resolution of discrete locations around the central airway is increased, thereby obtaining the corrected set of discrete locations on the thoracic surface. The modified discrete location set of the thoracic surface is correlated with the anatomical associated regions of the lung lobes and bronchi, and a unique number is assigned to each discrete location. Finally, a discrete location mapping table of the thoracic surface is generated. The discrete location mapping table of the thoracic surface serves as a spatial reference for constructing the acoustic and mechanical signal acquisition and propagation relationship.

4. The electronic device according to claim 1, characterized in that, When the program is read and executed by the processor, the step of applying acoustic and mechanical short-range excitations to discrete regions on the thoracic surface, acquiring corresponding signals, and forming a region signal set includes: Acoustic short-range excitation is applied to discrete locations on the thoracic surface, the acoustic short-range response of each discrete location is acquired, and a subset of acoustic location signals containing amplitude, energy distribution and phase difference is output. Short-range mechanical excitation is applied to discrete locations on the thoracic surface, the short-range mechanical response of each discrete location is acquired, and a subset of mechanical location signals including start-up delay, peak amplitude, and decay rate is output. The acoustic location signal subset and the mechanical location signal subset are aligned according to the discrete location number, and the trigger time of the reference discrete location is used for unified time-scale correction to output a time-synchronized composite location signal subset. The composite location signal subset is subjected to a consistency check, and responses with amplitude saturation or noise pollution are removed. The remaining responses are then weighted and averaged according to the stability of repeated excitation, and finally the location signal set after noise reduction and stability correction is output.

5. The electronic device according to claim 1, characterized in that, When the program is read and executed by the processor, the step of extracting features from the location signal set to obtain propagation parameters between discrete locations on the thoracic surface and combining them into propagation entries includes: Frequency domain analysis is performed on the acoustic short-range response generated by acoustic short-range excitation in the location signal concentration, the dominant frequency, energy concentration and phase delay are extracted, and the results are output as acoustic short-range response characteristics; Time-domain analysis is performed on the short-range mechanical response generated by the short-range mechanical excitation in the location signal set to extract the start-up delay, peak amplitude and decay rate, and the results are output as short-range mechanical response characteristics. The acoustic short-range response features and the mechanical short-range response features are fused according to the discrete location number of the thoracic surface. Features with amplitude differences or inconsistent delays are weighted and corrected, and the fused propagation parameter set is output. Based on the set of propagation parameters, the directionality, relative intensity, hysteresis, and frequency band sensitivity between discrete regions on the thoracic surface are calculated, and the calculation results are combined into propagation entries.

6. The electronic device according to claim 1, characterized in that, When the program is read and executed by the processor, the step of weightedly fusing the propagation entries with the individualized clinical data to generate a propagation relationship table including directionality, relative intensity, hysteresis, and frequency band sensitivity includes: Based on the parameters extracted and propagated from the individualized clinical data, the parameters include at least chest circumference, height, weight, vital capacity, one-second rate, and history of respiratory diseases, and the parameters are output as a weighted input set. The propagation entries are categorized according to the anatomically associated regions of the discrete location mapping table of the thoracic surface, and initial values ​​of directionality, relative intensity, hysteresis and frequency band sensitivity are generated for each propagation entry, and the initial propagation entry set is output. The initial propagation entry set is modified based on the weighted input set. For individuals with chronic obstructive pulmonary disease, the hysteresis weight is increased and the relative intensity is decreased in the relevant discrete locations. For individuals with cystic fibrosis, the frequency band sensitivity of the low-frequency band is increased in the relevant discrete locations of the central airway. For individuals whose chest wall thickness is increased due to increased chest circumference or weight, the directionality of the high-frequency band is decreased in the corresponding discrete locations. The modified propagation entry set is then output. The modified propagation entry set is subjected to stability and consistency screening. Based on the stability score of repeated collections and the spatial continuity of the discrete location mapping table of the thoracic surface, propagation entries below the threshold or those that conflict with the directionality of adjacent entries are removed. For propagation entries with opposite directions within the same path, the net dominant direction is retained, and a qualified propagation entry set is output. The qualified propagation entry set is indexed according to the discrete location number of the thoracic surface, and the directionality, relative intensity, hysteresis and frequency band sensitivity of each propagation entry are recorded. At the same time, the entry weight and stability score are retained as auxiliary fields, and finally a propagation relationship table is generated.

7. The electronic device according to claim 1, characterized in that, When the program is read and executed by the processor, the first computing module is specifically used for the following steps: Receive the body position parameters and combine the body position parameters with the gravity direction to form a set of correction parameters; Based on the set of correction parameters, the propagation entries in the propagation relationship table are weighted and corrected, and the corrected propagation relationship table is output. A directed graph search is performed in the revised propagation table to derive the thoracic drainage path according to the criteria of minimizing energy attenuation and hysteresis accumulation. The chest drainage path is calculated to generate an accessibility score that includes a combination of directionality, relative intensity, and hysteresis, and is output along with the chest drainage path.

8. The electronic device according to claim 1, characterized in that, When the program is read and executed by the processor, the second computing module is specifically used for the following steps: Receive the chest drainage path, retrieve the propagation entry corresponding to the chest drainage path in the propagation relationship table, and output a set of path-related propagation entries; Extract the frequency band sensitivity and phase delay information from the path-related propagation entry set, and output the candidate frequency phase combination set; A matching search is performed in the candidate frequency phase combination set to select the optimal combination that can form a continuous traveling wave on the chest drainage path, and the optimal oscillation control parameters are output.

Images