Simulation teaching method and device for emergency rescue of laryngeal obstruction

Abstract

The invention relates to a simulation teaching method and device for emergency rescue of laryngeal obstruction. The method comprises the following steps: constructing a replaceable airway physical model integrated with a pressure sensor, a position sensor and an airflow sensor; the airway resistance-displacement dynamic model is used for controlling the airway resistance change of the replaceable airway physical model in real time; constructing a case library according to the collected real first-aid cases; when the student operates the replaceable airway physical model, synchronously calling the corresponding teaching guide content in the case library; and comparing the student operation data with a preset standard operation threshold, performing uncertainty reasoning on the pressure curve stability, the operation time and the airway smoothness by adopting a Dempster-Shafer evidence algorithm, outputting an evaluation result and an operation error report, and automatically generating a visual training report. According to the invention, through the replaceable airway physical model and the Dempster-Shafer evidence algorithm, pathophysiological changes of laryngeal obstruction can be accurately simulated, tiny flaws in operation of trainees can be identified, and the training accuracy and efficiency are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of medical simulation education technology, specifically to a simulation teaching method and device for emergency rescue of laryngeal obstruction, and computer equipment. Background Technology

[0002] Laryngeal obstruction is a common and rapidly progressing emergency; if not treated promptly, patients can die from asphyxiation within a short period. The core of clinical emergency care lies in rapidly establishing an artificial airway (such as endotracheal intubation or cricothyroidotomy), requiring medical personnel to possess extremely high proficiency, accurate judgment, and stable psychological qualities. However, conducting basic training on real patients carries extremely high risks; therefore, highly realistic simulation training has become an irreplaceable teaching method.

[0003] Current laryngeal obstruction emergency simulation training mainly includes static anatomical models, computer simulation systems with simple sensors, and highly realistic physiologically driven mannequins. Static anatomical models, typically made of plastic or silicone, can demonstrate the anatomical structure of the airway, but their limitation lies in their inability to simulate the dynamic pathophysiological changes of laryngeal obstruction (such as the gradual increase in airway resistance and changes in blood oxygen saturation). Trainees can only practice the operational steps but cannot experience the time pressure and physiological feedback of real resuscitation, resulting in limited training effectiveness. Computer simulation systems with simple sensors integrate a few sensors (such as position sensors), providing prompts when tools (such as laryngoscopes or intubations) reach specific positions. However, the evaluation dimensions are singular, lacking quantitative analysis of key quality indicators such as operational force, stability, and smoothness. For example, it only determines whether the intubation tube has passed the glottis but cannot assess the potential pressure damage to the airway mucosa during intubation. Highly realistic physiologically driven mannequins are currently a more advanced solution, capable of simulating various vital signs and generating corresponding physiological responses. However, their core technology is closed, preventing teaching institutions from customizing and adjusting them according to specific clinical cases or teaching needs. Furthermore, the assessment of trainees' operations relies heavily on the instructors' subjective experience and on-site observation, making it difficult to achieve standardized scoring, especially when assessing subtle differences in operational techniques. In addition, the purchase and maintenance costs of the entire system are extremely high, making it difficult to widely implement in primary healthcare institutions.

[0004] To address the aforementioned issues, this invention proposes a simulation-based teaching method for emergency rescue of laryngeal obstruction. By employing a replaceable airway physical model and the Dempster-Shafer evidence algorithm, it accurately simulates the pathophysiological changes of laryngeal obstruction and identifies subtle flaws in trainees' operations, thereby improving the accuracy and efficiency of training. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a simulation teaching method and device for emergency rescue of laryngeal obstruction, as well as computer equipment, in order to overcome the shortcomings of the prior art.

[0006] According to one aspect of the present invention, a simulation teaching method for emergency rescue of laryngeal obstruction is provided, comprising:

[0007] A replaceable airway physical model integrating pressure, position, and airflow sensors is constructed to simulate laryngeal obstruction from normal to Class I and Class II. The parameters of the airway resistance-displacement dynamic model corresponding to the replaceable airway physical model are dynamically learned and adaptively adjusted using a BP neural network. The airway resistance-displacement dynamic model is used to control the airway resistance changes of the replaceable airway physical model in real time to fit the dynamic resistance curve and ensure model fit, thereby improving the physiological realism of the simulation under different obstruction levels.

[0008] A case database is constructed based on collected real emergency cases, including medical history, vital signs, auscultation sounds, and laryngoscopy views. When trainees operate the replaceable airway physical model, the corresponding standard operating procedure video, 3D anatomical model, and real-time auscultation audio from the case database are simultaneously called as teaching guidance content.

[0009] Real-time acquisition of student operation data from sensors in the replaceable airway physical model and simultaneous recording of student operation videos, wherein the student operation data includes intubation depth, airway pressure and airflow rate;

[0010] The student's operation data is compared with the preset standard operation threshold. The Dempster-Shafer evidence algorithm is used to perform uncertainty reasoning on the stability of the pressure curve, operation time and airway patency. The output includes an evaluation result including operation standardization judgment and comprehensive score and an operation error report.

[0011] Based on the evaluation results, the operation error report, the student's operation data, and the student's operation video, a visual training report is automatically generated. The visual training report includes an operation trajectory diagram, a score distribution diagram, a comparative analysis of the teaching guidance content and the student's operation, and related case content and operation video clips, which are used for post-training review and analysis.

[0012] In one alternative approach, the expression for the total drag functional of the airway drag-displacement dynamic model is:

[0013]

[0014] in, For laminar flow resistance operators; This is the turbulent drag tensor; This is the viscoelastic dissipation integral; For inertial drag functional; This represents the interaction potential between the tube walls; For the velocity field vector, pressure field, density field, and temperature field; For Reynolds number tensors; For strain tensor, For stress tensor, It is a complex modulus tensor; For acceleration field, It is a function of cross-sectional area; Let be the displacement field vector. These are the force field acting on the pipe wall and the wall stress tensor, respectively.

[0015] In one alternative approach, the expression for the laminar flow resistance operator is:

[0016]

[0017] in, For the velocity field vector, pressure field, density field, and temperature field; ; For density-temperature dependent apparent activation energy tensor; Here is the Arrhenius relation for the temperature field; For a velocity gradient-dependent Cross model; It is a time-periodic pulsation modulation function; It is the divergence modulation function; For curvature correction operators; This is a function of the instantaneous radius of the airway; This refers to the three-dimensional spatial region of the airway; The reference viscosity coefficient;

[0018] The expression for the turbulent drag tensor is:

[0019]

[0020] in, For Reynolds number tensors; It is a function of cross-sectional area; This is the turbulence intensity coefficient; The Reynolds number exponent; This is the acceleration response function; The airflow acceleration vector;

[0021] The expression for the viscoelastic dissipation integral is:

[0022]

[0023] in, For stress tensor; It is a complex modulus tensor; The strain function; For the matter derivative operator; Angular frequency; It is a spatial position vector; for; Let be the gradient damping function; , For the strain-dependent spatial dissipative modulus field, For nonlinear coupling coefficient tensors, It is a fractional strain rate sensitivity index. This is the strain correction factor;

[0024] The expression for the inertial drag functional is:

[0025]

[0026] in, For acceleration fields; The mass derivative of the airflow vector; The square norm of the airflow rate; This is the local velocity field vector;

[0027] The expression for the interaction potential between the tube walls is:

[0028]

[0029] in, It is a curvature weighting function; This is the time decay function.

[0030] In an alternative approach, the generalized basic trust assignment functional of the Dempster-Shafer evidence algorithm is:

[0031]

[0032] in, , where is the Mahalanobis distance of the i-th evidence source; for The kernel function; Let be the weight tensor of the i-th evidence source; A mixed weighting for quantum evidence and fuzzy evidence; Let f(x) be the probability amplitude function of the quantum field. For fuzzy measure integral function; Normalization factor; ; The weighting coefficients are for hyperbolic secant. It is a hyperbolic secant function; This is the error function.

[0033] In one alternative approach, the quantum field probability amplitude function is:

[0034]

[0035] in, The quantum probability amplitude; For time evolution interval; This is the decoherence time constant; For quantum oscillation amplitude coefficient; This refers to the quantum oscillation angular frequency; Evolutionary time; For quantum phase angle;

[0036] The fuzzy measure integral function is:

[0037]

[0038] in, The cutoff level parameter; For fuzzy membership functions; It is a fuzzy measure function; This is the Gaussian smoothing kernel function; For bandwidth parameters; For supremum operators; To select the smaller operator; For function composition operators;

[0039] The weight tensor of the evidence source is:

[0040]

[0041] in, Let be the quantum von Neumann entropy of the i-th source of evidence; Jensen-Shannon divergence; These are the divergence weighting coefficients; , Let be the probability distribution vectors of the i-th and j-th evidence sources, respectively; As a measure of the certainty of evidence;

[0042] The quantum von Neumann entropy is:

[0043]

[0044] in, , where is the density operator for the i-th evidence source; For trace operators; Let be the quantum state vector of the i-th evidence source.

[0045] In one alternative approach, the generalized synthesis rule of the Dempster-Shafer evidence algorithm is:

[0046]

[0047] in:

[0048] ;

[0049] ;

[0050] ;

[0051] in, For global conflict measurement; The similarity weight of propositions; This is the confidence level difference function; This is the conflict attenuation coefficient; These are the eigenvectors of propositions B and C; It is the covariance matrix; For Bhattacharyya distance weights; Distance to Bhattacharyya; It is the hyperbolic tangent function.

[0052] In one alternative approach, the replaceable airway physical model includes seven functional levels, wherein the innermost functional level is a medical-grade silicone matrix with antibacterial properties, and the surface is modified with a nanoscale rough structure to simulate the frictional characteristics of a real airway mucosa.

[0053] The second functional layer is a shape memory alloy woven mesh, and the transformation temperature is controlled within the physiological range of 35-37℃ by the ratio of alloy components.

[0054] The third functional layer is a piezoelectric ceramic sensor array, which uses interdigitated electrodes to realize gradient detection of pressure distribution, with a spatial resolution of 0.5mm×0.5mm and a sampling frequency of 1.13~1.25kHz.

[0055] The fourth functional layer is an electroactive hydrogel response layer, whose volume phase change characteristics enable a rapid response to external electric field stimulation by adjusting the crosslinking density and monomer concentration.

[0056] The fifth functional layer is the microfluidic network layer, which contains 30-32 independently controlled microvalves and 58-64 pressure sensing nodes;

[0057] The sixth functional layer is the temperature control layer, which uses Peltier elements and thermistors to achieve closed-loop control of the model surface temperature.

[0058] The outermost functional layer is the optical marking layer, which contains 250-256 infrared reflective markers to provide position reference.

[0059] In one alternative approach, the airway drag-displacement dynamic model includes a spatiotemporal graph convolutional-recurrent hybrid network module consisting of a graph convolutional coding submodule, a gated cyclic evolution submodule, and a physical information decoding submodule connected in sequence.

[0060] The graph convolutional coding submodule constructs a graph structure of a dynamic adjacency matrix based on prior knowledge of airway anatomy, and uses Chebyshev polynomial approximation graph convolution kernels to extract the spatial correlation of resistance features in each region of the airway.

[0061] The internal state update equation of the gated cyclic evolution submodule is filtered based on the cross-time step information of the attention weights;

[0062] The physical information decoding submodule maps the encoded and evolved high-dimensional features back to the physical space to ensure the consistency between the network prediction results and the underlying fluid dynamics principles.

[0063] According to another aspect of the present invention, a simulation teaching device for emergency rescue of laryngeal obstruction is provided, comprising:

[0064] The physiological simulation and resistance control module is used to construct a replaceable airway physical model integrating pressure sensors, position sensors, and airflow sensors to simulate laryngeal obstruction states from normal to Class I and Class II. Specifically, it uses a BP neural network to dynamically learn and adaptively adjust the parameters of the airway resistance-displacement dynamic model corresponding to the replaceable airway physical model. The airway resistance-displacement dynamic model is used to control the airway resistance changes of the replaceable airway physical model in real time to fit the dynamic resistance curve and ensure model fit, thereby improving the physiological realism of the simulation under different obstruction levels.

[0065] The multimodal teaching guidance module is used to build a case library based on collected real emergency cases, including medical history, vital signs, auscultation sounds, and laryngoscopy views. When trainees operate the replaceable airway physical model, the corresponding standard operating procedure video, 3D anatomical model, and real-time auscultation audio from the case library are called up simultaneously as teaching guidance content.

[0066] The operation data synchronous acquisition module is used to collect student operation data from the sensors in the replaceable airway physical model in real time and synchronously record the student's operation video. The student operation data includes intubation depth, airway pressure and airflow rate.

[0067] The intelligent assessment and error analysis module is used to compare the trainee's operation data with the preset standard operation threshold, and uses the Dempster-Shafer evidence algorithm to perform uncertainty reasoning on the stability of the pressure curve, operation time and airway patency, and outputs assessment results including operation standardization judgment and comprehensive score and operation error report.

[0068] The comprehensive report generation module is used to automatically generate a visual training report based on the evaluation results, the operation error report, the student operation data, and the student operation video. The visual training report includes an operation trajectory diagram, a score distribution diagram, a comparative analysis of the teaching guidance content and the student operation, and related case content and operation video clips for post-training review and analysis.

[0069] According to another aspect of the present invention, a computer device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus;

[0070] The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described simulation teaching method for emergency rescue of laryngeal obstruction.

[0071] According to the solution provided by this invention, a replaceable airway physical model integrating pressure sensors, position sensors, and airflow sensors is constructed to simulate laryngeal obstruction states from normal to Class I and Class II. Specifically, a BP neural network dynamically learns and adaptively adjusts the parameters of an airway resistance-displacement dynamic model corresponding to the replaceable airway physical model. The airway resistance change of the replaceable airway physical model is controlled in real time using the airway resistance-displacement dynamic model to fit the dynamic resistance curve and ensure model fit, thereby improving the physiological realism of the simulation under different obstruction levels. A case database including medical history, vital signs, auscultation sounds, and laryngoscopy views is constructed based on collected real emergency cases. When trainees operate the replaceable airway physical model, corresponding standard operating procedure videos, 3D anatomical models, and real-time auscultation audio from the case database are simultaneously invoked as teaching guidance content. The system collects real-time operator data from sensors in the replaceable airway physical model and simultaneously records operator videos. The operator data includes intubation depth, airway pressure, and airflow rate. The operator data is compared with preset standardized operation thresholds. The Dempster-Shafer evidence algorithm is used to perform uncertainty inference on pressure curve stability, operation time, and airway patency, outputting an evaluation result including an assessment of operational standardization and a comprehensive score, as well as an operation error report. Based on the evaluation result, the operation error report, and the operator data and video, a visual training report is automatically generated. This report includes an operation trajectory diagram, a score distribution diagram, a comparative analysis of the teaching guidance content and operator actions, and related case content and video clips for post-training review and analysis. This invention, through a replaceable airway physical model and the Dempster-Shafer evidence algorithm, can accurately simulate the pathophysiological changes of laryngeal obstruction and identify subtle flaws in operator actions, significantly improving the accuracy and efficiency of training.

[0072] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0073] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0074] Figure 1A flowchart illustrating the simulated teaching method for emergency rescue of laryngeal obstruction according to an embodiment of the present invention is shown.

[0075] Figure 2 A schematic diagram of the framework of the throat obstruction emergency rescue simulation teaching device according to an embodiment of the present invention is shown;

[0076] Figure 3 A schematic diagram of the structure of a computer device according to an embodiment of the present invention is shown. Detailed Implementation

[0077] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0078] Figure 1 A flowchart illustrating the simulated teaching method for emergency rescue of laryngeal obstruction according to an embodiment of the present invention is shown. Specifically, as... Figure 1 As shown, it includes the following steps:

[0079] Step S101: Construct a replaceable airway physical model integrating pressure sensors, position sensors, and airflow sensors to simulate laryngeal obstruction states from normal to Class I and Class II. The model dynamically learns and adaptively adjusts the parameters of the airway resistance-displacement dynamic model corresponding to the replaceable airway physical model using a BP neural network. The airway resistance-displacement dynamic model is used to control the airway resistance changes of the replaceable airway physical model in real time to fit the dynamic resistance curve and ensure model fit, thereby improving the physiological realism of the simulation under different obstruction levels.

[0080] In this embodiment, the continuous and dynamic changes in airway resistance can be simulated, accurately fitting the dynamic resistance curve under real physiological conditions. This allows trainees to experience not just "open" or "closed," but a more realistic progression of airway obstruction from normal to Class I and Class II. The replaceable airway physical model is made of biocompatible materials with appropriate mechanical properties, and its internal structure accurately replicates the anatomical morphology from the oral cavity to the trachea. Pressure sensors are distributed at key airway locations (such as the glottis and near the obstruction point) to measure the pressure of the tube against the airway wall and changes in airway pressure during procedures (such as intubation). Position sensors, using electromagnetic or optical trackers, are mounted on the simulated laryngoscope or endotracheal tube to accurately track the position and depth of the instruments in real time. Airflow sensors are installed at the airway outlet of the model to measure the airflow rate during simulated breathing. Computer-controlled mechanical actuators (such as expandable rings and movable sliders) are designed inside the model to dynamically change the airway diameter, thereby altering airflow resistance. The airway drag-displacement dynamic model is based on principles of fluid mechanics, solid mechanics, and physiology. It takes displacement commands from the actuator as input and outputs predicted airway drag, including parameters related to the physical model's characteristics. When using the model for the first time or replacing it with a new physical model, initial parameters are set for the dynamic model through a standardized testing procedure (e.g., sending a series of known displacement commands and recording the resulting actual drag). In actual teaching use, the system runs continuously. Given a displacement command, the model predicts the drag value, while sensors simultaneously measure the actual drag value. The error between the predicted and actual values ​​is used as a training signal for the BP network, propagating back to fine-tune the network weights. This dynamically and adaptively corrects the parameters of the drag-displacement dynamic model, making its predictions increasingly accurate.

[0081] For example, in training in intubation and resuscitation for Class II laryngeal obstruction, a Class II laryngeal obstruction case is selected. A preset Class II obstruction dynamic resistance curve (showing significantly increased airway resistance fluctuating dramatically with the respiratory cycle) is invoked. In the initial state, the replaceable airway physical model, under the action of the resistance adjustment mechanism, has a reduced inner diameter, and the trainee can feel obvious airflow obstruction and high respiratory resistance at the model inlet. The trainee inserts a laryngoscope, and the position sensor tracks the position and angle of the laryngoscope blade in real time. When the blade tip touches and slightly lifts the epiglottis, it has a minor effect on the obstruction site. The pressure sensor detects local pressure changes, and the airflow sensor detects a momentary improvement in airflow. The BP neural network determines whether there is a deviation between the actual response of the current physical model and the prediction of the resistance-displacement dynamic model. Due to slight model aging, its resistance response is slightly slower than predicted. The BP neural network adaptively adjusts key parameters of the dynamic model (such as the viscoelastic coefficient). The adjusted model recalculates and sends fine-tuning instructions to the resistance adjustment mechanism, making the resistance feedback of the model to the trainee's operations (such as further lifting the laryngoscope) more consistent with the physiological state of a real Class II obstruction. This means that not only is the resistance high, but it is also accompanied by a soft and unstable feeling due to tissue edema. Throughout the operation, the trainee does not feel a static, narrow passage, but rather a dynamic, responsive tissue with a sense of life. When the operation is correct, the resistance is moderately relieved in a physiological manner; when the operation is improper (such as excessive force), it simulates a sudden increase in resistance due to spasm or further edema. Thus, the highly realistic experience significantly enhances the immersion in training and the teaching effect.

[0082] In one alternative approach, the expression for the total drag functional of the airway drag-displacement dynamic model is:

[0083]

[0084] in, For laminar flow resistance operators; This is the turbulent drag tensor; This is the viscoelastic dissipation integral; For inertial drag functional; This represents the interaction potential between the tube walls; For the velocity field vector, pressure field, density field, and temperature field; For Reynolds number tensors; For strain tensor, For stress tensor, It is a complex modulus tensor; For acceleration field, It is a function of cross-sectional area; Let be the displacement field vector. These are the force field acting on the pipe wall and the wall stress tensor, respectively.

[0085] In this embodiment, the functional decomposes the total drag into five interrelated physical components (laminar flow, turbulent flow, tissue viscoelasticity, fluid inertia, and tube wall mechanics), enabling the model to simulate the full spectrum of respiratory states, from steady breathing (laminar flow dominant) to rapid panting (turbulent and inertial flow dominant), as well as the viscoelastic response of airway soft tissue under stress. Each operator is dynamic (time / velocity related) and nonlinear. For example, It describes the stress relaxation and creep characteristics of airway tissue, that is, when compressed by instruments, the tissue does not deform or recover instantly, but has a time-dependent, lag process. This nonlinear dynamic behavior is a core feature of real physiological tissue and enhances the physiological realism of the simulation.

[0086] For example, the simulation of airway resistance changes during inspiration in a patient with grade II laryngeal obstruction is shown. At this point, the airway is significantly narrowed, and the airflow accelerates, causing swollen soft tissues (such as the epiglottis and aryepiglottic folds) to vibrate and further deform under the suction force of the airflow. Five operators work together to accurately describe this complex physiological process. The laminar flow resistance operator increases significantly at the narrowing point due to the sharp decrease in radius, forming the basic resistance. The turbulent flow resistance operator increases the Reynolds number when the airflow velocity exceeds a critical value at the narrowing point, simulating the additional energy loss caused by turbulence, manifested as audible wheezing and a nonlinear spike in resistance. The viscoelastic dissipation operator reflects physiological realism (the swollen soft tissue is modeled as a viscoelastic material). When a strong inspiration begins, the airflow exerts a step-like pull on the soft tissue. Due to viscoelasticity, the tissue does not immediately reach maximum deformation (creep) but has a delay. The viscoelastic dissipation integral calculates the energy consumed due to the delayed deformation, manifested as a non-transient increase in resistance over time at the beginning of inspiration, perfectly simulating the mechanical behavior of real biological tissues. The inertial drag functional simulates the inertial force required to accelerate the airflow itself. When a patient attempts to inhale rapidly, even without viscous loss, some inertial drag must be overcome, making the model more realistic in simulating the effort required for inhalation during acute respiratory distress. The airway wall interaction potential operator simulates the feedback effect of deformation and stress on the total resistance caused by internal pressure and external contact (such as compression by the duct). When the laryngoscope is raised above the epiglottis, it not only changes the geometry but also alters the local stress, thus affecting the resistance.

[0087] In one alternative approach, the expression for the laminar flow resistance operator is:

[0088]

[0089] in, For the velocity field vector, pressure field, density field, and temperature field; ; For density-temperature dependent apparent activation energy tensor; Here is the Arrhenius relation for the temperature field; For a velocity gradient-dependent Cross model; It is a time-periodic pulsation modulation function; It is the divergence modulation function; For curvature correction operators; This is a function of the instantaneous radius of the airway; This refers to the three-dimensional spatial region of the airway; The reference viscosity coefficient;

[0090] The expression for the turbulent drag tensor is:

[0091]

[0092] in, For Reynolds number tensors; It is a function of cross-sectional area; This is the turbulence intensity coefficient; The Reynolds number exponent; This is the acceleration response function; The airflow acceleration vector;

[0093] The expression for the viscoelastic dissipation integral is:

[0094]

[0095] in, For stress tensor; It is a complex modulus tensor; The strain function; For the matter derivative operator; Angular frequency; It is a spatial position vector; for; Let be the gradient damping function; , For the strain-dependent spatial dissipative modulus field, For nonlinear coupling coefficient tensors, It is a fractional strain rate sensitivity index. This is the strain correction factor;

[0096] The expression for the inertial drag functional is:

[0097]

[0098] in, For acceleration fields; The mass derivative of the airflow vector; The square norm of the airflow rate; This is the local velocity field vector;

[0099] The expression for the interaction potential between the tube walls is:

[0100]

[0101] in, It is a curvature weighting function; This is the time decay function.

[0102] In this embodiment, dynamic viscosity defines fluid viscosity as a function of temperature, density, and shear rate, accurately simulating the non-Newtonian fluid properties of respiratory mucus (such as shear thinning) and the effects of body temperature and respiratory cycle on mucus flowability. Compressibility correction considers the compressibility of gases under high-speed flow or pressure changes, making the model more accurate when simulating forceful breathing or coughing. Geometric curvature correction quantifies geometric effects such as secondary flow in airway bends using the Laplace operator, addressing the insufficient applicability of straight-tube models in real-world curved airways.

[0103] In one alternative approach, the replaceable airway physical model includes seven functional levels, wherein the innermost functional level is a medical-grade silicone matrix with antibacterial properties, and the surface is modified with a nanoscale rough structure to simulate the frictional characteristics of a real airway mucosa.

[0104] The second functional layer is a shape memory alloy woven mesh, and the transformation temperature is controlled within the physiological range of 35-37℃ by the ratio of alloy components.

[0105] The third functional layer is a piezoelectric ceramic sensor array, which uses interdigitated electrodes to realize gradient detection of pressure distribution, with a spatial resolution of 0.5mm×0.5mm and a sampling frequency of 1.13~1.25kHz.

[0106] The fourth functional layer is an electroactive hydrogel response layer, whose volume phase change characteristics enable a rapid response to external electric field stimulation by adjusting the crosslinking density and monomer concentration.

[0107] The fifth functional layer is the microfluidic network layer, which contains 30-32 independently controlled microvalves and 58-64 pressure sensing nodes;

[0108] The sixth functional layer is the temperature control layer, which uses Peltier elements and thermistors to achieve closed-loop control of the model surface temperature.

[0109] The outermost functional layer is the optical marking layer, which contains 250-256 infrared reflective markers to provide position reference.

[0110] In this embodiment, a high-density distributed tactile sensing network is constructed through the third layer (piezoelectric sensor array) and the fifth layer (microfluidic pressure nodes). For example, a spatial resolution of 0.5 mm and a sampling frequency of >1 kHz can capture the microscopic pressure distribution and dynamic changes in the contact between instruments (such as laryngeal blades and catheters) and the airway wall. The second layer (shape memory alloy mesh) endows the model with shape memory and self-recovery capabilities, automatically restoring its initial anatomical shape after deformation due to manipulation, thereby improving the model's durability and training efficiency. The fourth layer (electroactive hydrogel) and the fifth layer (microfluidic network) receive electrical or fluid signals from the airway resistance-displacement dynamic model. Through the expansion / contraction of the hydrogel or the opening and closing of microvalves, the airway inner diameter and flexibility are changed in real time, thus accurately simulating the dynamic physiological and pathological changes from normal to Class I and II obstruction. The innermost nanoscale rough surface simulates the sensation of mucosal friction, and the temperature control of the sixth layer (maintained at 35-37°C) allows trainees to feel the real human body temperature when in contact, avoiding the discomfort caused by a cold model. The outermost optical markers, combined with an external motion capture system, can track the three-dimensional spatial trajectory of instruments (such as laryngoscope handles) without contact, for operation path analysis.

[0111] In one alternative approach, the airway drag-displacement dynamic model includes a spatiotemporal graph convolutional-recurrent hybrid network module consisting of a graph convolutional coding submodule, a gated cyclic evolution submodule, and a physical information decoding submodule connected in sequence.

[0112] The graph convolutional coding submodule constructs a graph structure of a dynamic adjacency matrix based on prior knowledge of airway anatomy, and uses Chebyshev polynomial approximation graph convolution kernels to extract the spatial correlation of resistance features in each region of the airway.

[0113] The internal state update equation of the gated cyclic evolution submodule is filtered based on the cross-time step information of the attention weights;

[0114] The physical information decoding submodule maps the encoded and evolved high-dimensional features back to the physical space to ensure the consistency between the network prediction results and the underlying fluid dynamics principles.

[0115] In this embodiment, the graph convolutional encoding submodule abstracts the airway into a graph structure (nodes represent different regions of the airway, and edges represent their anatomical connections), enabling the network to possess prior anatomical knowledge. Using Chebyshev polynomials for graph convolution allows for efficient and stable capture of spatial dependencies between nodes. For example, it understands that narrowing of the glottis inevitably affects the airflow pattern in the downstream trachea, thus achieving more accurate global spatial drag feature extraction. The gated cyclic evolution submodule (such as a variant of LSTM or GRU) learns the temporal patterns of drag changes. The physical information decoding submodule, acting as a physical constraint layer, forces the decoding and mapping of the high-dimensional abstract features learned by the network back to physical quantities (such as pressure, velocity, and drag) that conform to the basic principles of fluid mechanics (such as the Navier-Stokes equations), ensuring that each predicted output of the network is not only mathematically fitted to the data but also physically reliable and interpretable.

[0116] Step S102: Construct a case database based on collected real emergency cases, including medical history, vital signs, auscultation sounds, and laryngoscopy views; when trainees operate the replaceable airway physical model, synchronously call up the corresponding standard operating procedure video, 3D anatomical model, and real-time auscultation audio from the case database as teaching guidance content.

[0117] In this embodiment, a case database (including medical history and vital signs) forces trainees to assess and make decisions about patients' conditions like real doctors before and during procedures. For example, when faced with cases of anaphylactic shock with laryngeal edema and cases of laryngeal tumors, although both may present as Grade II obstruction, their medical history, rate of progression, and focus of rescue strategies differ, thereby training trainees' clinical differential diagnosis and response capabilities. Simultaneously, multiple stimuli are provided: auditory (auscultation sounds), visual (laryngoscopy view, 3D model, operation video), and tactile (physical model resistance). When trainees experience high resistance on the physical model, they simultaneously hear the characteristic high-pitched inspiratory stridor of the case and see the swollen glottis in the laryngoscopy view. Through strong multi-sensory association, trainees deepen their three-dimensional understanding and memory of specific conditions.

[0118] Step S103: Real-time acquisition of student operation data from sensors in the replaceable airway physical model and simultaneous recording of student operation video, wherein the student operation data includes intubation depth, airway pressure and airflow rate.

[0119] In this embodiment, multiple sensors are used to precisely quantify the intubation depth, airway pressure, and airflow rate at the millisecond level, so that the trainee's description is no longer a vague expression such as feeling that the intubation was too deep or that too much force was used.

[0120] Step S104: The student's operation data is compared with the preset standard operation threshold. The Dempster-Shafer evidence algorithm is used to perform uncertainty reasoning on the stability of the pressure curve, operation time and airway patency, and outputs the evaluation results including operation standardization judgment and comprehensive score and operation error report.

[0121] In this embodiment, the stability of the stress curve lies between stable and unstable. The Dempster-Shafer evidence algorithm expresses the degree of trust in stable, unstable, and unknown (i.e., undeterminable) conditions through a basic trust assignment function, rather than forcing a binary judgment. This aligns better with the thinking patterns in actual assessments and can handle uncertainties caused by sensor data noise and individual differences. Different assessment metrics can sometimes conflict. For example, a trainee might have a short operation time (evidence supporting 'excellent'), but their stress curve might fluctuate greatly (evidence supporting 'unsatisfactory'). Simple average scores would be "averaged out." The Dempster-Shafer evidence algorithm's synthesis rules quantify the degree of conflict and reasonably discount the influence of conflicting evidence during synthesis, thus arriving at a more robust comprehensive judgment rather than giving misleading high or low scores. When the evidence is very weak or completely conflicting, the Dempster-Shafer evidence algorithm can honestly output "undeterminable" instead of forcibly giving a potentially erroneous conclusion.

[0122] In an alternative approach, the generalized basic trust assignment functional of the Dempster-Shafer evidence algorithm is:

[0123]

[0124] in, , where is the Mahalanobis distance of the i-th evidence source; for The kernel function; Let be the weight tensor of the i-th evidence source; A mixed weighting for quantum evidence and fuzzy evidence; Let f(x) be the probability amplitude function of the quantum field. For fuzzy measure integral function; Normalization factor; ; The weighting coefficients are for hyperbolic secant. It is a hyperbolic secant function; This is the error function.

[0125] In this embodiment, during operational evaluation, different sensor data (such as pressure and airflow) exhibit varying fluctuation ranges and importance. Mahalanobis distance considers the correlation (through the covariance matrix) and dimensional differences of each data dimension, automatically adjusting the contribution of each dimension to the distance. This makes the evaluation results insensitive to data scale, better reflecting the distribution characteristics of real data, and thus more accurately measuring the deviation of the current operational data from the standard template. The quantum probability amplitude function, based on quantum mechanics principles, can express non-classical correlations such as interference and superposition between evidence. For example, a trainee's psychological state of tension may simultaneously affect their operational time and pressure stability; the two pieces of evidence are not independent within the quantum framework but exhibit entanglement or interference effects. The fuzzy measure integral function uses fuzzy membership functions and fuzzy measures to accurately characterize fuzzy states such as "relatively stable" and "very unstable," making the evaluation more consistent with human thinking.

[0126] In one alternative approach, the quantum field probability amplitude function is:

[0127]

[0128] in, The quantum probability amplitude; For time evolution interval; This is the decoherence time constant; For quantum oscillation amplitude coefficient; This refers to the quantum oscillation angular frequency; Evolutionary time; For quantum phase angle;

[0129] The fuzzy measure integral function is:

[0130]

[0131] in, The cutoff level parameter; For fuzzy membership functions; It is a fuzzy measure function; This is the Gaussian smoothing kernel function; For bandwidth parameters; For supremum operators; To select the smaller operator; For function composition operators;

[0132] The weight tensor of the evidence source is:

[0133]

[0134] in, Let be the quantum von Neumann entropy of the i-th source of evidence; Jensen-Shannon divergence; These are the divergence weighting coefficients; , Let be the probability distribution vectors of the i-th and j-th evidence sources, respectively; As a measure of the certainty of evidence;

[0135] The quantum von Neumann entropy is:

[0136]

[0137] in, , where is the density operator for the i-th evidence source; For trace operators; Let be the quantum state vector of the i-th evidence source.

[0138] In this embodiment, The interference effect describing the existence of evidence, in operational assessment, means that two seemingly independent pieces of evidence (such as "hand tremor" and "stress instability") may not be independent, but rather superimposed or canceled out like waves. The quantum field probability amplitude function can capture this deep correlation. The memory decay effect simulates the memory effect of evidence, where errors occurring early in the operation have a diminishing impact over time (decoherence), while recent operational performance has a greater impact on the current assessment, making the assessment focus more on the overall dynamic process of the operation rather than isolated moments.

[0139] In one alternative approach, the generalized synthesis rule of the Dempster-Shafer evidence algorithm is:

[0140]

[0141] in:

[0142] ;

[0143] ;

[0144] ;

[0145] in, For global conflict measurement; The similarity weight of propositions; This is the confidence level difference function; This is the conflict attenuation coefficient; These are the eigenvectors of propositions B and C; It is the covariance matrix; For Bhattacharyya distance weights; Distance to Bhattacharyya; It is the hyperbolic tangent function.

[0146] In this embodiment, the impact of brief conflicts occurring early in the operation rapidly diminishes as more consistent evidence emerges. By treating conflicts as a matter that fades over time, the system is prevented from crashing due to temporary data fluctuations or noise. Propositional similarity weights are determined through... By fusion with high weights, a comprehensive evaluation that is both fast and smooth is obtained, enabling intelligent evidence synthesis based on semantic similarity. Confidence difference calibration mitigates the veto effect of low-quality evidence on high-quality evidence, making the synthesized result more reflective of the quality differences among the evidence.

[0147] Step S105: Based on the evaluation results, the operation error report, the college's operation data, and the student's operation video, an automatic visualization training report is generated. The visualization training report includes an operation trajectory diagram, a score distribution diagram, a comparative analysis of the teaching guidance content and the student's operation, and related case content and operation video clips, which are used for post-training review and analysis.

[0148] In this embodiment, quantitative data (operational data), qualitative assessment (error report), spatial information (trajectory graph), standard reference (teaching guidance), and time series (operational video) are integrated together. Instructors and trainees can simultaneously see "what was done" (video), "how well it was done" (score / error), "why this is said" (data trajectory), and "how it should have been done" (standard comparison) in a single report, improving the efficiency and depth of the debriefing. Abstract concepts such as the stability and trustworthiness of the pressure curve output by the evidence algorithm are obscure for trainees, but the operation trajectory graph and synchronized video clips can concretize abstract assessments. For example, if the assessment report points to a deviation in the laryngoscope insertion trajectory, trainees can directly see on the trajectory graph how their path deviated from the standard path and review the specific actions that caused the deviation in the adjacent video window, making the error immediately clear and accelerating the understanding process.

[0149] For example, after a trainee completes a "Level II laryngeal obstruction" intubation training session, a personal visual training report is automatically generated, with a comprehensive score of 78 points.

[0150] The scoring distribution chart (radar chart) shows that the trainee scored very high (near full marks) in operation speed and anatomical exposure, but extremely low in gentleness of operation, forming a significant weakness. During the laryngoscope insertion and glottic exposure phases, the operation trajectory diagram shows that the trainee's laryngoscope trajectory (red line) exhibited a sudden, rightward lateral swing as it approached the epiglottis, while the standard trajectory (green line) was smooth and straight upward. The left side of the split-screen video shows the trainee's operation, displaying a violent movement of using the laryngoscope blade to pry open the epiglottis at this moment; the right side shows the standard video, where the expert smoothly lifts it upward. Synchronous data curves show a sharp spike in the airway pressure curve at the moment the trainee "pried" the epiglottis, far exceeding the safety threshold.

[0151] According to the solution provided by this invention, a replaceable airway physical model integrating pressure sensors, position sensors, and airflow sensors is constructed to simulate laryngeal obstruction states from normal to Class I and Class II. Specifically, a BP neural network dynamically learns and adaptively adjusts the parameters of an airway resistance-displacement dynamic model corresponding to the replaceable airway physical model. The airway resistance change of the replaceable airway physical model is controlled in real time using the airway resistance-displacement dynamic model to fit the dynamic resistance curve and ensure model fit, thereby improving the physiological realism of the simulation under different obstruction levels. A case database including medical history, vital signs, auscultation sounds, and laryngoscopy views is constructed based on collected real emergency cases. When trainees operate the replaceable airway physical model, corresponding standard operating procedure videos, 3D anatomical models, and real-time auscultation audio from the case database are simultaneously invoked as teaching guidance content. The system collects real-time operator data from sensors in the replaceable airway physical model and simultaneously records operator videos. The operator data includes intubation depth, airway pressure, and airflow rate. The operator data is compared with preset standardized operation thresholds. The Dempster-Shafer evidence algorithm is used to perform uncertainty inference on pressure curve stability, operation time, and airway patency, outputting an evaluation result including an assessment of operational standardization and a comprehensive score, as well as an operation error report. Based on the evaluation result, the operation error report, and the operator data and video, a visual training report is automatically generated. This report includes an operation trajectory diagram, a score distribution diagram, a comparative analysis of the teaching guidance content and operator actions, and related case content and video clips for post-training review and analysis. This invention, through a replaceable airway physical model and the Dempster-Shafer evidence algorithm, can accurately simulate the pathophysiological changes of laryngeal obstruction and identify subtle flaws in operator actions, significantly improving the accuracy and efficiency of training.

[0152] Figure 2 A schematic diagram of the framework of a throat obstruction emergency rescue simulation teaching device according to an embodiment of the present invention is shown. The throat obstruction emergency rescue simulation teaching device includes:

[0153] The physiological simulation and resistance control module 210 is used to construct a replaceable airway physical model integrating pressure sensors, position sensors, and airflow sensors to simulate laryngeal obstruction states from normal to Class I and Class II. Specifically, it uses a BP neural network to dynamically learn and adaptively adjust the parameters of the airway resistance-displacement dynamic model corresponding to the replaceable airway physical model. The airway resistance-displacement dynamic model is used to control the airway resistance changes of the replaceable airway physical model in real time to fit the dynamic resistance curve and ensure model fit, thereby improving the physiological realism of the simulation under different obstruction levels.

[0154] The multimodal teaching guidance module 220 is used to construct a case library including medical history, vital signs, auscultation sounds and laryngoscopy views based on collected real emergency cases; when the trainee operates the replaceable airway physical model, the corresponding standard operation procedure video, 3D anatomical model and real-time auscultation audio in the case library are called up simultaneously as teaching guidance content;

[0155] The operation data synchronous acquisition module 230 is used to collect student operation data from the sensors in the replaceable airway physical model in real time and synchronously record the student's operation video. The student operation data includes intubation depth, airway pressure and airflow rate.

[0156] The intelligent assessment and error analysis module 240 is used to compare the student's operation data with the preset standard operation threshold, and use the Dempster-Shafer evidence algorithm to perform uncertainty reasoning on the stability of the pressure curve, operation time and airway patency, and output the assessment results including operation standardization judgment and comprehensive score and operation error report.

[0157] The comprehensive report generation module 250 is used to automatically generate a visual training report based on the evaluation results, the operation error report, the student operation data, and the student operation video. The visual training report includes an operation trajectory diagram, a score distribution diagram, a comparative analysis of the teaching guidance content and the student's operation, and related case content and operation video clips for post-training review and analysis.

[0158] Figure 3 The diagram shows a structural schematic of an embodiment of the computer device of the present invention. The specific embodiments of the present invention do not limit the specific implementation of the computer device.

[0159] like Figure 3 As shown, the computer device may include: a processor 302, a communications interface 304, a memory 306, and a communications bus 308.

[0160] The processor 302, communication interface 304, and memory 306 communicate with each other via communication bus 308. Communication interface 304 is used to communicate with other network elements such as clients or other servers. The processor 302 executes program 310, specifically performing the relevant steps in the above-described embodiment of the throat obstruction emergency rescue simulation teaching method.

[0161] Specifically, program 310 may include program code that includes computer operation instructions.

[0162] Processor 302 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The computer device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0163] Memory 306 is used to store program 310. Memory 306 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0164] According to the solution provided by this invention, a replaceable airway physical model integrating pressure sensors, position sensors, and airflow sensors is constructed to simulate laryngeal obstruction states from normal to Class I and Class II. Specifically, a BP neural network dynamically learns and adaptively adjusts the parameters of an airway resistance-displacement dynamic model corresponding to the replaceable airway physical model. The airway resistance change of the replaceable airway physical model is controlled in real time using the airway resistance-displacement dynamic model to fit the dynamic resistance curve and ensure model fit, thereby improving the physiological realism of the simulation under different obstruction levels. A case database including medical history, vital signs, auscultation sounds, and laryngoscopy views is constructed based on collected real emergency cases. When trainees operate the replaceable airway physical model, corresponding standard operating procedure videos, 3D anatomical models, and real-time auscultation audio from the case database are simultaneously invoked as teaching guidance content. The system collects real-time operator data from sensors in the replaceable airway physical model and simultaneously records operator videos. The operator data includes intubation depth, airway pressure, and airflow rate. The operator data is compared with preset standardized operation thresholds. The Dempster-Shafer evidence algorithm is used to perform uncertainty inference on pressure curve stability, operation time, and airway patency, outputting an evaluation result including an assessment of operational standardization and a comprehensive score, as well as an operation error report. Based on the evaluation result, the operation error report, and the operator data and video, a visual training report is automatically generated. This report includes an operation trajectory diagram, a score distribution diagram, a comparative analysis of the teaching guidance content and operator actions, and related case content and video clips for post-training review and analysis. This invention, through a replaceable airway physical model and the Dempster-Shafer evidence algorithm, can accurately simulate the pathophysiological changes of laryngeal obstruction and identify subtle flaws in operator actions, significantly improving the accuracy and efficiency of training.

[0165] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination of all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed can be employed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose. Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination. The invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims listing several devices, several of these devices may be embodied by the same hardware item. Unless otherwise specified, the steps in the above embodiments should not be construed as limiting the order of execution.

Claims

1. A simulation teaching method for emergency rescue of laryngeal obstruction, characterized in that, include: A replaceable airway physical model integrating pressure, position, and airflow sensors is constructed to simulate laryngeal obstruction from normal to Class I and Class II. The parameters of the airway resistance-displacement dynamic model corresponding to the replaceable airway physical model are dynamically learned and adaptively adjusted using a BP neural network. The airway resistance-displacement dynamic model is used to control the airway resistance changes of the replaceable airway physical model in real time to fit the dynamic resistance curve and ensure model fit, thereby improving the physiological realism of the simulation under different obstruction levels. A case database is constructed based on collected real emergency cases, including medical history, vital signs, auscultation sounds, and laryngoscopy views. When trainees operate the replaceable airway physical model, the corresponding standard operating procedure video, 3D anatomical model, and real-time auscultation audio from the case database are simultaneously called as teaching guidance content. Real-time acquisition of student operation data from sensors in the replaceable airway physical model and simultaneous recording of student operation videos, wherein the student operation data includes intubation depth, airway pressure and airflow rate; The student's operation data is compared with the preset standard operation threshold. The Dempster-Shafer evidence algorithm is used to perform uncertainty reasoning on the stability of the pressure curve, operation time and airway patency. The output includes an evaluation result including operation standardization judgment and comprehensive score and an operation error report. Based on the evaluation results, the operation error report, the student's operation data, and the student's operation video, a visual training report is automatically generated. The visual training report includes an operation trajectory diagram, a score distribution diagram, a comparative analysis of the teaching guidance content and the student's operation, and related case content and operation video clips, which are used for post-training review and analysis.

2. The simulated teaching method for emergency rescue of laryngeal obstruction according to claim 1, characterized in that, The expression for the total resistance functional of the airway resistance-displacement dynamic model is as follows: ； in, For laminar flow resistance operators; This is the turbulent drag tensor; This is the viscoelastic dissipation integral; For inertial drag functional; This represents the interaction potential between the tube walls; For the velocity field vector, pressure field, density field, and temperature field; For Reynolds number tensors; For strain tensor, For stress tensor, It is a complex modulus tensor; For acceleration field, It is a function of cross-sectional area; Let be the displacement field vector. These are the force field acting on the pipe wall and the wall stress tensor, respectively.

3. The simulated teaching method for emergency rescue of laryngeal obstruction according to claim 2, characterized in that, The expression for the laminar flow resistance operator is: ； in, For the velocity field vector, pressure field, density field, and temperature field; ; For density-temperature dependent apparent activation energy tensor; Here is the Arrhenius relation for the temperature field; For a velocity gradient-dependent Cross model; It is a time-periodic pulsation modulation function; It is the divergence modulation function; For curvature correction operators; This is a function of the instantaneous radius of the airway; This refers to the three-dimensional spatial region of the airway; The reference viscosity coefficient; The expression for the turbulent drag tensor is: ； in, For Reynolds number tensors; It is a function of cross-sectional area; This is the turbulence intensity coefficient; The Reynolds number exponent; This is the acceleration response function; The airflow acceleration vector; The expression for the viscoelastic dissipation integral is: ； in, For stress tensor; It is a complex modulus tensor; The strain function; For the matter derivative operator; Angular frequency; It is a spatial position vector; for; Let be the gradient damping function; , For the strain-dependent spatial dissipative modulus field, For nonlinear coupling coefficient tensors, It is a fractional strain rate sensitivity index. This is the strain correction factor; The expression for the inertial drag functional is: ； in, For acceleration fields; The mass derivative of the airflow vector; The square norm of the airflow rate; This is the local velocity field vector; The expression for the interaction potential between the tube walls is: ； in, It is a curvature weighting function; This is the time decay function.

4. The simulated teaching method for emergency rescue of laryngeal obstruction according to claim 1, characterized in that, The generalized basic trust assignment functional of the Dempster-Shafer evidence algorithm is: ； in, , where is the Mahalanobis distance of the i-th evidence source; for The kernel function; Let be the weight tensor of the i-th evidence source; A mixed weighting for quantum evidence and fuzzy evidence; Let f(x) be the probability amplitude function of the quantum field. For fuzzy measure integral function; Normalization factor; ; The weighting coefficients are for hyperbolic secant. It is a hyperbolic secant function; This is the error function.

5. The simulation teaching method for emergency rescue of laryngeal obstruction according to claim 4, characterized in that, The quantum field probability amplitude function is: ； in, The quantum probability amplitude; For time evolution interval; This is the decoherence time constant; For quantum oscillation amplitude coefficient; This refers to the quantum oscillation angular frequency; Evolutionary time; For quantum phase angle; The fuzzy measure integral function is: ； in, The cutoff level parameter; For fuzzy membership functions; It is a fuzzy measure function; This is the Gaussian smoothing kernel function; For bandwidth parameters; For supremum operators; To select the smaller operator; For function composition operators; The weight tensor of the evidence source is: ； in, Let be the quantum von Neumann entropy of the i-th source of evidence; Jensen-Shannon divergence; These are the divergence weighting coefficients; , Let be the probability distribution vectors of the i-th and j-th evidence sources, respectively; As a measure of the certainty of evidence; The quantum von Neumann entropy is: ； in, , where is the density operator for the i-th evidence source; For trace operators; Let be the quantum state vector of the i-th evidence source.

6. The simulated teaching method for emergency rescue of laryngeal obstruction according to claim 1, characterized in that, The generalized synthesis rule of the Dempster-Shafer evidence algorithm is as follows: ； in: ； ； ； in, For global conflict measurement; The similarity weight of propositions; This is the confidence level difference function; This is the conflict attenuation coefficient; These are the eigenvectors of propositions B and C; It is the covariance matrix; For Bhattacharyya distance weights; Distance to Bhattacharyya; It is the hyperbolic tangent function.

7. The simulated teaching method for emergency rescue of laryngeal obstruction according to claim 1, characterized in that, The replaceable airway physical model includes seven functional levels, of which the innermost functional level is a medical-grade silicone matrix with antibacterial properties, and the surface is modified with a nanoscale rough structure to simulate the frictional characteristics of the real airway mucosa. The second functional layer is a shape memory alloy woven mesh, and the transformation temperature is controlled within the physiological range of 35-37℃ by the ratio of alloy components. The third functional layer is a piezoelectric ceramic sensor array, which uses interdigitated electrodes to realize gradient detection of pressure distribution, with a spatial resolution of 0.5mm×0.5mm and a sampling frequency of 1.13~1.25kHz. The fourth functional layer is an electroactive hydrogel response layer, whose volume phase change characteristics enable a rapid response to external electric field stimulation by adjusting the crosslinking density and monomer concentration. The fifth functional layer is the microfluidic network layer, which contains 30-32 independently controlled microvalves and 58-64 pressure sensing nodes; The sixth functional layer is the temperature control layer, which uses Peltier elements and thermistors to achieve closed-loop control of the model surface temperature. The outermost functional layer is the optical marking layer, which contains 250-256 infrared reflective markers to provide position reference.

8. The simulated teaching method for emergency rescue of laryngeal obstruction according to claim 1, characterized in that, The airway resistance-displacement dynamic model includes a spatiotemporal graph convolutional-cyclic hybrid network module consisting of a graph convolutional coding submodule, a gated cyclic evolution submodule, and a physical information decoding submodule connected in sequence. The graph convolutional coding submodule constructs a graph structure of a dynamic adjacency matrix based on prior knowledge of airway anatomy, and uses Chebyshev polynomial approximation graph convolution kernels to extract the spatial correlation of resistance features in each region of the airway. The internal state update equation of the gated cyclic evolution submodule is filtered based on the cross-time step information of the attention weights; The physical information decoding submodule maps the encoded and evolved high-dimensional features back to the physical space to ensure the consistency between the network prediction results and the underlying fluid dynamics principles.

9. A simulation teaching device for emergency rescue of laryngeal obstruction, characterized in that, include: The physiological simulation and resistance control module is used to construct a replaceable airway physical model integrating pressure sensors, position sensors, and airflow sensors to simulate laryngeal obstruction states from normal to Class I and Class II. Specifically, it uses a BP neural network to dynamically learn and adaptively adjust the parameters of the airway resistance-displacement dynamic model corresponding to the replaceable airway physical model. The airway resistance-displacement dynamic model is used to control the airway resistance changes of the replaceable airway physical model in real time to fit the dynamic resistance curve and ensure model fit, thereby improving the physiological realism of the simulation under different obstruction levels. The multimodal teaching guidance module is used to build a case library based on collected real emergency cases, including medical history, vital signs, auscultation sounds, and laryngoscopy views. When trainees operate the replaceable airway physical model, the corresponding standard operating procedure video, 3D anatomical model, and real-time auscultation audio from the case library are called up simultaneously as teaching guidance content. The operation data synchronous acquisition module is used to collect student operation data from the sensors in the replaceable airway physical model in real time and synchronously record the student's operation video. The student operation data includes intubation depth, airway pressure and airflow rate. The intelligent assessment and error analysis module is used to compare the trainee's operation data with the preset standard operation threshold, and uses the Dempster-Shafer evidence algorithm to perform uncertainty reasoning on the stability of the pressure curve, operation time and airway patency, and outputs assessment results including operation standardization judgment and comprehensive score and operation error report. The comprehensive report generation module is used to automatically generate a visual training report based on the evaluation results, the operation error report, the student operation data, and the student operation video. The visual training report includes an operation trajectory diagram, a score distribution diagram, a comparative analysis of the teaching guidance content and the student operation, and related case content and operation video clips for post-training review and analysis.

10. A computer device, comprising: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described simulation teaching method for emergency rescue of laryngeal obstruction.

Images