Personalized rehabilitation intervention management system for post-thoracic surgery patients

Abstract

This invention discloses a personalized rehabilitation intervention management system for post-thoracic surgery patients, belonging to the field of medical rehabilitation technology. The invention includes a dynamic perception module, an evolution prediction module, a strategy generation module, an interactive output module, and a correction module. The dynamic perception module collects respiratory waveforms and sputum position data to construct a sputum movement characteristic model representing the airway resistance response characteristics and sputum inertial response characteristics. The evolution prediction module simulates the sputum movement path based on the current model, predicts risks, and triggers signals. The strategy generation module performs safety and efficiency screening within a variable parameter space, selecting the target respiratory guidance parameters that maximize sputum displacement. The correction module uses the deviation between the actual trajectory and the predicted trajectory to adjust the model parameters in reverse, forming a closed-loop adaptive mechanism. This invention resolves the contradiction between sputum expectoration and pain in post-thoracic surgery rehabilitation, achieving a dual gain in safety and efficiency.

Description

Technical Field

[0001] This invention relates to the field of medical rehabilitation technology, and in particular to a personalized rehabilitation intervention management system for post-thoracic surgery patients. Background Technology

[0002] Thoracic surgery is an important means of treating chest diseases. The core rehabilitation challenge faced by patients after surgery lies in the contradiction between clearing airway secretions and incision pain. Due to surgical trauma and the effects of anesthesia, a large amount of sputum accumulates in the patient's airway. If it is not cleared in time, it can lead to serious complications such as atelectasis or lung infection. Therefore, coughing forcefully to expel sputum is a key aspect of postoperative rehabilitation after thoracic surgery. However, coughing causes chest cavity contraction and increased abdominal pressure, which directly pulls on the surgical incision that has not yet fully healed, causing severe pain. Some patients find that rapid and shallow breathing can effectively reduce the urge to cough, thereby avoiding pain caused by incision traction due to coughing. Ultimately, this creates a vicious cycle where patients suppress their cough due to fear of pain, leading to sputum accumulation and complications.

[0003] Current technologies use flow or pressure sensors to collect patients' respiratory waveforms and monitor parameters such as respiratory rate and tidal volume, issuing alarms when abnormal breathing patterns are detected. Alternatively, physical vibration expectoration devices or nebulized inhalation therapy are used to passively or actively assist in loosening and expelling airway secretions for rehabilitation intervention. Some studies have also attempted to combine acoustic sensor arrays to locate sputum positions, providing reference information for clinical assessment. However, current technologies can only identify abnormal breathing patterns but cannot distinguish the causes of shallow and rapid breathing. Furthermore, they cannot quantitatively weigh and optimize decisions between the two mutually restrictive goals of avoiding pain and effective expectoration. This can easily lead to misjudgment and inappropriate interventions, exacerbating patient discomfort and making it difficult to achieve truly personalized rehabilitation management. Summary of the Invention

[0004] The purpose of this invention is to provide a personalized rehabilitation intervention management system for post-thoracic surgery patients, which solves the problem that existing technologies cannot identify the true cause of shallow and rapid breathing, and easily confuse and misjudge anxiety-related shallow and rapid breathing, pain-related shallow and rapid breathing, and pathological shallow and rapid breathing, thereby leading to inappropriate rehabilitation intervention processes that aggravate patient discomfort.

[0005] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: This invention is a personalized rehabilitation intervention management system for post-thoracic surgery patients, comprising: The dynamic sensing module continuously acquires postoperative temporal respiratory waveform data and temporal position data of sputum within the airway. Based on the temporal respiratory waveform data, it extracts the instantaneous airflow velocity, respiratory phase, and velocity change rate at the current moment. Simultaneously, it calculates the instantaneous movement velocity and acceleration amplitude of the sputum based on the temporal position data. The velocity change rate represents the airway's resistance response characteristics, and the acceleration amplitude represents the sputum's inertial response characteristics. It constructs and dynamically updates a sputum movement characteristic model that represents the current physical state of the patient's airway. A set of parameters for the current respiratory state is formed based on the instantaneous airflow velocity, respiratory phase, resistance response characteristics, and inertial response characteristics. The evolution prediction module inputs the current respiratory state parameter set and the updated sputum movement feature model into the simulation engine. Taking the current moment as the initial state, it presets a fixed time period and simulates the continuous movement path of sputum with respiratory airflow along the time axis in the forward direction within the fixed time period. It also checks the direction of the movement path point by point. If any displacement vector of any future time step is detected pointing to the preset danger area, a risk trigger signal is immediately generated. The strategy generation module, upon receiving a risk trigger signal, locks the current sputum motion feature model and respiratory state parameter set. It constructs a variable parameter space using respiratory rate, inspiratory-to-expiratory ratio, and peak flow rate. Each parameter value in this space is considered a candidate parameter combination, and safety and efficiency screening are performed within this space. Based on the locked model and parameter set, it calculates the rate of change curve of intrathoracic pressure under each candidate parameter combination, eliminating parameter combinations whose pressure change rate exceeds a preset pain threshold, thus forming a safe parameter subset. Within this safe parameter subset, it calculates the expected sputum displacement vector magnitude corresponding to each parameter combination, selecting the parameter combination that maximizes this magnitude as the target respiratory guidance parameter. The interactive output module is used to convert the target breathing guidance parameters into breathing guidance signals and output them to the human-computer interaction interface; The correction module, after the respiratory guidance signal is issued and a complete respiratory cycle has elapsed, calls the dynamic sensing module again to obtain the actual movement trajectory of the sputum during that cycle, calculates the positional deviation between the actual movement trajectory and the previously generated predicted movement trajectory by the evolution prediction module, and uses the positional deviation as feedback error to adjust the model parameters in the sputum motion characteristic model that represent inertial response and resistance response characteristics, so that the model's predicted output converges to the actual observed value. Subsequently, the corrected sputum motion characteristic model is re-inputted into the evolution prediction module to trigger the next round of trajectory extrapolation and risk prediction process, and so on, until the actual movement trajectory of the sputum converges to the preset target airway position.

[0006] Furthermore, in the dynamic perception module, the specific method for constructing and dynamically updating the sputum movement characteristic model is as follows: the rate of change of flow velocity is used as a quantitative indicator of airway resistance response characteristics, and the acceleration amplitude is used as a quantitative indicator of sputum inertial response characteristics; using a preset physical mapping relationship, the resistance response characteristics and inertial response characteristics are mapped to a set of feature parameters that characterize the relationship between the aerodynamic force and motion state of sputum, thus forming a sputum movement characteristic model.

[0007] Furthermore, in the dynamic sensing module, flow sensors and pressure sensors installed on the patient's respiratory pathway continuously collect the patient's respiratory flow and respiratory pressure signals at a preset sampling frequency, and generate time-series respiratory waveform data after analog-to-digital conversion; through an acoustic sensor array installed on the patient's airway surface projection area, respiratory sound signals are continuously collected, and time-frequency analysis and sound source localization processing are performed on the respiratory sound signals to calculate the real-time position of sputum in the airway and generate time-series position data.

[0008] Furthermore, the steps for extracting the respiratory phase and flow rate change rate at the current moment include: performing zero-crossing detection on the time-series respiratory waveform data, determining the respiratory phase at the current moment based on the zero-crossing points where the waveform amplitude changes from negative to positive and from positive to negative, where the inspiratory phase corresponds to the positive interval of the waveform and the expiratory phase corresponds to the negative interval of the waveform; performing differential operations on the continuous sampling points in the time-series respiratory waveform data, calculating the difference in airflow velocity between the current moment and the previous moment, and dividing the difference by the sampling time interval to obtain the flow rate change rate.

[0009] Furthermore, in the evolution prediction module, a fixed time period is preset as the length of a complete respiratory cycle. During the process of checking the direction of movement path point by point, the airway space is discretized into multiple continuous segments, and several of these segments are marked as danger zones in advance. When the endpoint of the displacement vector at any time step falls into the danger zone, a risk trigger signal is generated.

[0010] Furthermore, in the safety and efficiency screening, the pleural pressure change rate curve was calculated based on the locked sputum movement characteristic model, respiratory state parameter set and preset pleural mechanics model, and the pain threshold was the preset upper limit of the pressure change rate. Let the candidate parameter combination be ,in, Respiratory rate, For the breathing ratio, For peak flow rate, the corresponding curve for rate of change of intrathoracic pressure is as follows: The pain threshold is The security screening criteria are shown in equation (1): (1) The parameter combinations that satisfy the above conditions constitute a subset of safety parameters. .

[0011] Furthermore, for the parameter combinations in the safety parameter subset, based on the locked sputum motion feature model, the cumulative displacement vector of sputum within a preset time period is simulated under the respiratory waveform driven by the parameter combination, and the magnitude is taken as the expected sputum displacement vector magnitude. Define a subset of safety parameters Given any combination of parameters c, under the influence of the respiratory waveform driven by c, the cumulative displacement vector of sputum within time period T is: The expected magnitude of the sputum displacement vector is shown in equation (2): (2) Target breathing guidance parameters As shown in equation (3): (3).

[0012] Furthermore, in the correction module, when adjusting the model parameters in reverse using the position deviation, a closed-loop adaptive correction strategy is adopted: the position deviation is input into the parameter optimizer, with the goal of minimizing the deviation between the predicted movement trajectory and the actual movement trajectory, and the model parameters of the inertial response characteristics and resistance response characteristics in the sputum motion feature model are iteratively adjusted until the deviation converges to within the preset threshold.

[0013] Furthermore, the breathing guidance signals output by the interactive output module include graphical breathing guidance waveforms, voice prompts, and tactile feedback signals.

[0014] Furthermore, when performing safety and efficiency screening, the strategy generation module first sorts the candidate parameter combinations in the safety parameter subset according to the peak value of the rate of change of intrathoracic pressure from smallest to largest; then, from the sorting results, it selects the parameter combinations ranked in the top N to form the preferred subset, where N is a preset positive integer; finally, from the preferred subset, it selects the parameter combination with the largest expected sputum displacement vector magnitude as the target respiratory guidance parameter.

[0015] The present invention has the following beneficial effects: This invention inputs the current sputum movement characteristic model and respiratory state parameter set into the simulation engine, simulates the continuous movement path of sputum in a future respiratory cycle along the time axis, and checks the direction of displacement vectors point by point. Once any displacement in the future time step is detected pointing to a preset danger area, a risk trigger signal is immediately generated, thereby issuing an early warning before the sputum enters the danger area, which significantly improves the initiative and safety of rehabilitation intervention.

[0016] Upon receiving a risk trigger signal, this invention constructs a variable parameter space using respiratory rate, inspiratory-to-expiratory ratio, and peak flow rate. Within this space, it performs dual screening for safety and efficiency. The safety screening eliminates parameter combinations that would cause the rate of change in intrathoracic pressure to exceed a preset pain threshold, ensuring that the intervention process does not cause severe pain to the patient. The efficiency screening selects the parameter combination that maximizes the magnitude of the sputum displacement vector from the subset of safety parameters as the target respiratory guidance parameters, thereby resolving the contradiction between sputum expectoration and pain in postoperative rehabilitation of thoracic surgery and achieving a dual gain in safety and efficiency.

[0017] This invention compares the positional deviation between the actual and predicted sputum movement trajectories after the respiratory guidance signal is issued and a complete respiratory cycle has elapsed. This deviation is used as feedback error to adjust the model parameters representing inertial and resistance response characteristics in the sputum movement feature model, so that the model's predicted output converges to the actual observed value. This allows the system to learn from each intervention, and as the number of uses increases, the model's fitting accuracy to the individual physiological characteristics of the patient continues to improve.

[0018] In the process of safety and efficiency screening, this invention first ensures that the parameter combinations entering the preferred subset all have a low rate of pressure change, avoiding the selection of a critical safety scheme close to the pain threshold due to excessive pursuit of sputum expectoration efficiency; secondly, it provides a flexible adjustment capability for the weights of safety and efficiency, which can be dynamically adjusted according to individual differences such as the patient's pain tolerance and recovery stage. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0021] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0022] See Figure 1 A personalized rehabilitation intervention management system for post-thoracic surgery patients, including: The dynamic sensing module continuously acquires postoperative temporal respiratory waveform data and temporal position data of sputum within the airway. Based on the temporal respiratory waveform data, it extracts the instantaneous airflow velocity, respiratory phase, and velocity change rate at the current moment. Simultaneously, it calculates the instantaneous movement velocity and acceleration amplitude of the sputum based on the temporal position data. The velocity change rate represents the airway's resistance response characteristics, and the acceleration amplitude represents the sputum's inertial response characteristics. It constructs and dynamically updates a sputum movement characteristic model that represents the current physical state of the patient's airway. A set of parameters for the current respiratory state is formed based on the instantaneous airflow velocity, respiratory phase, resistance response characteristics, and inertial response characteristics. It should be further explained that the specific method for constructing and dynamically updating the sputum movement feature model in the dynamic perception module is as follows: the rate of change of flow velocity is used as a quantitative indicator of airway resistance response characteristics, and the acceleration amplitude is used as a quantitative indicator of sputum inertial response characteristics; using a preset physical mapping relationship, the resistance response characteristics and inertial response characteristics are mapped to a set of feature parameters that characterize the relationship between the aerodynamic force and motion state of sputum, thus forming a sputum movement feature model.

[0023] In the dynamic sensing module, flow sensors and pressure sensors installed on the patient's respiratory pathway continuously collect the patient's respiratory flow and respiratory pressure signals at a preset sampling frequency, and generate time-series respiratory waveform data after analog-to-digital conversion; through an acoustic sensor array installed on the patient's airway surface projection area, respiratory sound signals are continuously collected, and time-frequency analysis and sound source localization processing are performed on the respiratory sound signals to calculate the real-time position of sputum in the airway and generate time-series position data.

[0024] The steps for extracting the respiratory phase and rate of change of flow velocity at the current moment include: performing zero-crossing detection on the time-series respiratory waveform data, determining the respiratory phase at the current moment based on the zero-crossing points where the waveform amplitude changes from negative to positive and from positive to negative, where the inspiratory phase corresponds to the positive interval of the waveform and the expiratory phase corresponds to the negative interval of the waveform; performing differential operation on the continuous sampling points in the time-series respiratory waveform data, calculating the difference in airflow velocity between the current moment and the previous moment, and dividing the difference by the sampling time interval to obtain the rate of change of flow velocity.

[0025] The construction and parameter updating of the sputum motion feature model are as follows: This invention employs a respiratory mechanics model oriented towards airway secretions to characterize sputum movement characteristics. Here, airway secretions are sputum, and the sputum movement characteristic model structure consists of two core parameters: the sputum inertia coefficient M and the airway viscous resistance coefficient R. Its dynamic description is shown in equation (4): (4) in, The inertial response characteristic parameter of sputum during the acceleration phase, namely the sputum inertia coefficient, is expressed in kg and characterizes the inherent inertia of sputum in resisting changes in motion state. The airway viscous resistance coefficient is a parameter representing the resistance response of the airway to the movement of sputum. Its unit is N·s / m, and it characterizes the sliding friction and viscous effect between sputum and the airway wall. The instantaneous acceleration of sputum is calculated by the dynamic sensing module based on the temporal location data of sputum at the current moment. The unit is m / s². 2 ; The velocity of sputum movement at the current moment is calculated by the dynamic sensing module based on the temporal location data of sputum, i.e., the instantaneous velocity of sputum, in m / s; The aerodynamic force exerted by airflow on sputum, i.e., the airflow driving force, is expressed in N and is calculated based on the instantaneous airflow velocity and airway cross-sectional area in the respiratory waveform data. The minimum driving force required to move sputum from a stationary state is the static friction threshold for sputum movement, measured in N. When the airflow driving force is below this threshold, the sputum will not move.

[0026] According to the mechanical framework of Duncan's second law, the product of the net driving force on sputum and the sputum inertia coefficient M is positively correlated with the sputum acceleration. Here, the static driving force is the airflow driving force minus the static friction threshold. At the same time, the product of the airway viscous drag coefficient R and the sputum velocity represents the resistance component continuously consumed during the movement. A quantitative mapping relationship between respiratory mechanical parameters and sputum movement state is established through a first-order linear ordinary differential equation.

[0027] The dynamic identification method for sputum movement feature model parameters in this invention is as follows: At each sampling time t during system operation, the parameters provided by the dynamic sensing module include: instantaneous acceleration of sputum. Instantaneous velocity of sputum Current displacement length of sputum Instantaneous airflow velocity in the breathing waveform and airway pressure signal Substitute the above parameters into the equation of motion shown in equation (5) and solve simultaneously: (5) In this context, subscripts 1 and 2 represent the observations at two adjacent sampling times, respectively. Based on instantaneous airflow velocity Estimation based on airway cross-sectional area A: in, For gas density, A is preset based on the patient's preoperative or postoperative CT impact data; and The results were obtained by using the first and second derivatives of the sputum temporal location data with respect to time.

[0028] After each respiratory cycle, the correction module inputs the positional deviation between the actual movement trajectory and the predicted movement trajectory into the parameter optimizer. With the goal of minimizing the prediction deviation, it iteratively adjusts the values ​​of M and R to achieve closed-loop adaptive update of the model parameters.

[0029] The evolution prediction module inputs the current respiratory state parameter set and the updated sputum movement feature model into the simulation engine. Taking the current moment as the initial state, it presets a fixed time period and simulates the continuous movement path of sputum with respiratory airflow along the time axis in the forward direction within the fixed time period. It also checks the direction of the movement path point by point. If any displacement vector of any future time step is detected pointing to the preset danger area, a risk trigger signal is immediately generated. It should be added that in the evolution prediction module, the preset fixed time period is the length of a complete respiratory cycle. During the process of checking the direction of movement path point by point, the airway space is discretized into multiple continuous segments, and several of these segments are marked as danger areas in advance. When the endpoint of the displacement vector at any time step falls into the danger area, a risk trigger signal is generated.

[0030] The strategy generation module, upon receiving a risk trigger signal, locks the current sputum motion feature model and respiratory state parameter set. It constructs a variable parameter space using respiratory rate, inspiratory-to-expiratory ratio, and peak flow rate. Each parameter value in this space is considered a candidate parameter combination, and safety and efficiency screening are performed within this space. Based on the locked model and parameter set, it calculates the rate of change curve of intrathoracic pressure under each candidate parameter combination, eliminating parameter combinations whose pressure change rate exceeds a preset pain threshold, thus forming a safe parameter subset. Within this safe parameter subset, it calculates the expected sputum displacement vector magnitude corresponding to each parameter combination, selecting the parameter combination that maximizes this magnitude as the target respiratory guidance parameter. It should be noted that in the safety and efficiency screening, the pleural pressure change rate curve was calculated based on the locked sputum movement characteristic model, respiratory state parameter set and preset pleural mechanics model, and the pain threshold was the preset upper limit of the pressure change rate. Let the candidate parameter combination be ,in, Respiratory rate, For the breathing ratio, For peak flow rate, the corresponding curve for rate of change of intrathoracic pressure is as follows: The pain threshold is The security screening criteria are shown in equation (1): (1) The parameter combinations that satisfy the above conditions constitute a subset of safety parameters. .

[0031] For parameter combinations in the safety parameter subset, based on the locked sputum motion feature model, the cumulative displacement vector of sputum within a preset time period is simulated under the respiratory waveform driven by the parameter combination, and the magnitude is taken as the expected sputum displacement vector magnitude. Define a subset of safety parameters Given any combination of parameters c, under the influence of the respiratory waveform driven by c, the cumulative displacement vector of sputum within time period T is: The expected magnitude of the sputum displacement vector is shown in equation (2): (2) Target breathing guidance parameters As shown in equation (3): (3).

[0032] When performing safety and efficiency screening, the strategy generation module first sorts the candidate parameter combinations in the safety parameter subset according to the peak value of the rate of change of intrathoracic pressure from smallest to largest; then, from the sorting results, it selects the parameter combinations ranked in the top N to form the preferred subset, where N is a preset positive integer; finally, from the preferred subset, it selects the parameter combination with the largest expected sputum displacement vector magnitude as the target breathing guidance parameter. The interactive output module is used to convert the target breathing guidance parameters into breathing guidance signals and output them to the human-computer interaction interface; It should be noted that the breathing guidance signals output by the interactive output module include graphical breathing guidance waveforms, voice prompts, and tactile feedback signals.

[0033] The correction module, after the respiratory guidance signal is issued and a complete respiratory cycle has elapsed, calls the dynamic sensing module again to obtain the actual movement trajectory of the sputum during that cycle, calculates the positional deviation between the actual movement trajectory and the previously generated predicted movement trajectory by the evolution prediction module, and uses the positional deviation as feedback error to adjust the model parameters in the sputum motion characteristic model that represent inertial response and resistance response characteristics, so that the model's predicted output converges to the actual observed value. Subsequently, the corrected sputum motion characteristic model is re-inputted into the evolution prediction module to trigger the next round of trajectory extrapolation and risk prediction process, and so on, until the actual movement trajectory of the sputum converges to the preset target airway position.

[0034] It should be noted that in the correction module, when adjusting the model parameters in reverse using the position deviation, a closed-loop adaptive correction strategy is adopted: the position deviation is input into the parameter optimizer, and the model parameters of the inertial response characteristics and resistance response characteristics in the sputum motion characteristic model are iteratively adjusted with the goal of minimizing the deviation between the predicted movement trajectory and the actual movement trajectory, until the deviation converges to within the preset threshold.

[0035] In this invention, a dynamic sensing module first continuously collects the patient's temporal respiratory waveform data and sputum temporal location data, extracting instantaneous airflow velocity, respiratory phase, velocity change rate, and the instantaneous movement velocity and acceleration amplitude of the sputum. The velocity change rate and acceleration amplitude are quantified as airway resistance response characteristics and sputum inertial response characteristics, respectively, constructing and dynamically updating a sputum movement characteristic model representing the patient's current airway physical state. Subsequently, the evolution prediction module, using the current moment as the initial state, inputs the sputum movement characteristic model and respiratory state parameter set into the simulation engine, simulating the continuous movement path of sputum within a preset fixed time period along the time axis, and checking the displacement vector direction point by point. If any displacement in any future time step is detected pointing to a preset danger area, a risk trigger signal is immediately generated. Upon receiving the risk trigger signal, the strategy generation module uses the respiratory frequency... A variable parameter space is constructed using parameters such as rate of response, inspiratory-to-expiratory ratio, and peak flow rate. Within this space, parameter combinations that exceed a preset pain threshold are eliminated through a safety screening process to form a safe parameter subset. Then, the parameter combination with the largest expected sputum displacement vector magnitude is selected from this subset as the target breathing guidance parameter. The interactive output module converts this target breathing guidance parameter into multimodal breathing guidance signals, including graphical waveforms, voice prompts, and tactile feedback, and outputs them to the human-computer interaction interface to guide the patient to execute the optimal breathing mode. Finally, after each complete breathing cycle, the correction module compares the positional deviation between the actual and predicted sputum movement trajectories. This deviation is used to adjust the model parameters representing inertial and resistance response characteristics in the sputum motion feature model, causing the model prediction to converge to the actual observation value. This process is repeated to achieve closed-loop adaptive personalized rehabilitation intervention.

[0036] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A personalized rehabilitation intervention management system for post-thoracic surgery patients, characterized in that: include: The dynamic sensing module is used to continuously acquire the patient's postoperative temporal respiratory waveform data and the temporal position data of sputum in the airway, extract feature parameters to construct and dynamically update the sputum movement feature model that represents the current physical state of the patient's airway, and form the current respiratory state parameter set. The evolution prediction module inputs the current respiratory state parameter set and sputum movement feature model into the simulation engine. Taking the current moment as the initial state, it simulates the movement path of sputum with the respiratory airflow in a preset fixed time period along the time axis in the forward direction. It checks the direction of the movement path point by point. If any displacement vector of any future time step is detected to point to the preset danger area, a risk trigger signal is immediately generated. The strategy generation module receives a risk trigger signal, locks the current sputum movement feature model and respiratory state parameter set, constructs a variable parameter space using respiratory rate, inspiratory-to-expiratory ratio, and peak flow rate, uses each set of parameter values ​​as candidate parameter combinations, performs safety and efficiency screening, calculates the intrathoracic pressure change rate curve under each candidate parameter combination, eliminates parameter combinations with pressure change rates exceeding a preset pain threshold, forming a safe parameter subset; calculates the expected sputum displacement vector magnitude corresponding to each parameter combination in the safe parameter subset, and selects the parameter combination with the largest magnitude as the target respiratory guidance parameter output. The correction module, after the respiratory guidance parameters are released and a complete respiratory cycle has passed, obtains the actual movement trajectory of sputum during that cycle, calculates the positional deviation between the actual movement trajectory and the predicted movement trajectory, and uses the positional deviation to adjust the model parameters in the sputum motion characteristic model that represent inertial response characteristics and resistance response characteristics, so that the prediction of the sputum motion characteristic model converges to the actual observation value, until the actual movement trajectory of sputum converges to the preset target airway position.

2. The personalized rehabilitation intervention management system for post-thoracic surgery patients according to claim 1, characterized in that, In the dynamic sensing module, based on the time-series respiratory waveform data, the instantaneous airflow velocity, respiratory phase, and velocity change rate at the current moment are extracted in real time. Simultaneously, based on the time-series location data, the instantaneous movement velocity and acceleration amplitude of the sputum are calculated. The velocity change rate represents the airway resistance response characteristics, and the acceleration amplitude represents the sputum inertial response characteristics. The current respiratory state parameter set includes instantaneous airflow velocity, respiratory phase, resistance response characteristics, and inertial response characteristics. The specific method for constructing and dynamically updating the sputum movement feature model is as follows: the velocity change rate is used as a quantitative index of airway resistance response characteristics, and the acceleration amplitude is used as a quantitative index of sputum inertial response characteristics. Using a preset physical mapping relationship, the resistance response characteristics and inertial response characteristics are mapped to a feature parameter set that characterizes the relationship between the aerodynamic force and the motion state of the sputum, thus forming the sputum movement feature model.

3. The personalized rehabilitation intervention management system for post-thoracic surgery patients according to claim 1, characterized in that, In the dynamic sensing module, flow sensors and pressure sensors installed on the patient's respiratory pathway continuously collect the patient's respiratory flow and respiratory pressure signals at a preset sampling frequency. After analog-to-digital conversion, time-series respiratory waveform data is generated. An acoustic sensor array installed on the patient's airway surface projection area continuously collects respiratory sound signals, and performs time-frequency analysis and sound source localization processing on the respiratory sound signals to calculate the real-time position of sputum in the airway and generate the time-series position data.

4. The personalized rehabilitation intervention management system for post-thoracic surgery patients according to claim 1, characterized in that, The steps for extracting the respiratory phase and flow rate change rate at the current moment include: performing zero-crossing detection on the time-series respiratory waveform data; determining the respiratory phase at the current moment based on the zero-crossing points where the waveform amplitude changes from negative to positive and from positive to negative, wherein the inspiratory phase corresponds to the positive interval of the waveform and the expiratory phase corresponds to the negative interval of the waveform; performing differential operation on the continuous sampling points in the time-series respiratory waveform data to calculate the airflow velocity difference between the current moment and the previous moment; and dividing the difference by the sampling time interval to obtain the flow rate change rate.

5. The personalized rehabilitation intervention management system for post-thoracic surgery patients according to claim 1, characterized in that, In the evolution prediction module, a fixed time period is preset as the length of a complete respiratory cycle. During the process of checking the direction of movement path point by point, the airway space is discretized into multiple continuous segments. Several of these segments are marked as danger zones in advance. When the endpoint of the displacement vector at any time step falls into a danger zone, a risk trigger signal is generated.

6. The personalized rehabilitation intervention management system for post-thoracic surgery patients according to claim 1, characterized in that, In the safety and efficiency screening, the pleural pressure change rate curve is calculated based on a locked sputum movement characteristic model, a respiratory state parameter set, and a preset pleural mechanics model, and the pain threshold is a preset upper limit value of the pressure change rate. Let the candidate parameter combination be ,in, Respiratory rate, For the breathing ratio, For peak flow rate, the corresponding curve for rate of change of intrathoracic pressure is as follows: The pain threshold is The security screening criteria are shown in equation (1): （1） The parameter combinations that satisfy the above conditions constitute a subset of safety parameters. .

7. The personalized rehabilitation intervention management system for post-thoracic surgery patients according to claim 6, characterized in that, For the parameter combination in the safety parameter subset, based on the locked sputum motion feature model, the cumulative displacement vector of sputum within a preset time period is simulated under the respiratory waveform driven by the parameter combination, and the modulus is taken as the expected sputum displacement vector modulus. Define a subset of safety parameters Given any combination of parameters c, under the influence of the respiratory waveform driven by c, the cumulative displacement vector of sputum within time period T is: The expected magnitude of the sputum displacement vector is shown in equation (2): （2） Target breathing guidance parameters As shown in equation (3): （3）。 8. The personalized rehabilitation intervention management system for post-thoracic surgery patients according to claim 1, characterized in that, In the correction module, when adjusting the parameters of the sputum motion feature model in reverse using the position deviation, the position deviation is input into the parameter optimizer. With the goal of minimizing the deviation between the predicted movement trajectory and the actual movement trajectory, the parameters of the sputum motion feature model that represent the inertial response characteristics and the resistance response characteristics in the sputum motion feature model are iteratively adjusted until the deviation converges to within a preset threshold.

9. The personalized rehabilitation intervention management system for post-thoracic surgery patients according to claim 1, characterized in that, The breathing guidance signals output by the interactive output module include graphical breathing guidance waveforms, voice prompts, and tactile feedback signals.

10. The personalized rehabilitation intervention management system for post-thoracic surgery patients according to claim 1, characterized in that, When performing safety and efficiency screening, the strategy generation module first sorts the candidate parameter combinations in the safety parameter subset according to the peak value of the rate of change of intrathoracic pressure from smallest to largest. Then, it selects the parameter combinations ranked in the top N from the sorting results to form an optimal subset, where N is a preset positive integer. Finally, in the optimal subset, the parameter combination with the largest expected sputum displacement vector magnitude is selected as the target respiratory guidance parameter.

Images