Intelligent teaching system for traditional Chinese medicine clinical skill cultivation

Abstract

The invention relates to the technical field of intelligent teaching, in particular to an intelligent teaching system for traditional Chinese medicine clinical skill cultivation. The system comprises a virtual reality simulation module, an operation interaction module, an operation data analysis module, an adaptive learning feedback module and a teaching evaluation module. The virtual reality simulation module constructs a vivid traditional Chinese medicine skill training scene through three-dimensional modeling, ray tracing and physical simulation technologies, and the vivid traditional Chinese medicine skill training scene covers pulse feeling, acupuncture and moxibustion, massage and other operations. The operation interaction module collects user operation data and extracts features in combination with a variational auto-encoder. The operation data analysis module analyzes user operation deviation by using a graph neural network and a generative adversarial network, the adaptive learning feedback module generates a personalized learning strategy, and the teaching evaluation module evaluates a learning effect through fuzzy evaluation and analytic hierarchy process. According to the invention, intelligence, individuation and high efficiency of traditional Chinese medicine clinical skill cultivation are realized, the fidelity of teaching and the accuracy of feedback are improved, and the skill level and the training effect of learners are also improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent teaching technology, and in particular to an intelligent teaching system for cultivating clinical skills of traditional Chinese medicine. Background Art

[0002] The inheritance and development of TCM clinical skills are of great significance to the sustainable development of TCM. In the process of cultivating TCM clinical skills, the traditional teaching method is mainly for students to gradually master TCM clinical skills through observation, imitation and practice of the teacher's operation. However, there are many shortcomings in the traditional teaching model, and there are also the following problems: the TCM skill training scenes in the existing technology lack authenticity and cannot effectively simulate actual operations, which affects the learners' immersion and training effects; the existing system supports limited interactive devices and cannot accurately and comprehensively capture and analyze the user's operation characteristics, resulting in unsatisfactory training effects; the traditional system does not conduct in-depth data analysis on user operations, lacks intelligent identification of key nodes and patterns in operations, and the feedback mechanism is also imperfect. It is also impossible to dynamically adjust the learning strategy according to the learners' operation performance, resulting in large differences in individual learning effects. Summary of the invention

[0003] To solve the above problems, the present invention provides an intelligent teaching system for cultivating clinical skills in traditional Chinese medicine, which solves the problems of lack of authenticity in existing traditional Chinese medicine skills training scenarios, incomplete capture of operation features, shallow data analysis, and insufficient intelligence in feedback and learning strategy adjustment, improves the realism of teaching and the accuracy of feedback, and also enhances learners' skill levels and training effects.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] An intelligent teaching system for cultivating clinical skills in traditional Chinese medicine, comprising a virtual reality simulation module, an operation interaction module, an operation data analysis module, an adaptive learning feedback module and a teaching evaluation module which are sequentially connected in communication;

[0006] The virtual reality simulation module is used to construct a simulation operation scene for TCM clinical skills training based on three-dimensional modeling and real-time rendering technology in combination with virtual reality technology, and utilizes ray tracing algorithm and physical simulation algorithm to improve the realism of the simulation operation scene; the simulation operation scene includes pulse diagnosis, acupuncture and massage;

[0007] The operation interaction module is used to perform real-time interaction through user input devices, collect operation data, and use variational autoencoders to extract key features of user operations; the user input devices include keyboards, handles, touch pads, joysticks, and force feedback devices;

[0008] The operation data analysis module is used to analyze the relationship and time evolution between actions based on the operation data and key features using a graph neural network, capture key nodes and patterns in the operation, and use a generative adversarial network to generate a comparison model between standard operations and user operations, automatically generate deviation indicators from actual operations, and generate an operation data analysis report;

[0009] The adaptive learning feedback module is used to generate a dynamically adjusted personalized learning strategy based on the operation data analysis report using a learning path optimization algorithm;

[0010] The teaching evaluation module is used to evaluate the user's performance in various traditional Chinese medicine clinical skills based on the operation data analysis report and personalized learning strategy, using fuzzy evaluation algorithm and hierarchical analysis algorithm to generate a teaching evaluation report.

[0011] Furthermore, the operation data includes operation position information, force information, time series, operation frequency and operation area positioning information input by the user.

[0012] Furthermore, the operation process of the virtual reality simulation module includes the following steps:

[0013] Based on the preset TCM clinical skills training requirements, a virtual operation scene model is created through 3D modeling technology;

[0014] Using a physical simulation algorithm based on finite element analysis to perform mechanical simulation on the operation behavior of the virtual operation scene model;

[0015] Use ray tracing algorithms to render virtual scenes in real time, including the light scattering effect of the skin, skin deformation during acupuncture needle insertion, and reflection phenomena;

[0016] The physical simulation engine automatically adjusts the response state of the virtual operation scene model, combines the dynamic feedback of pulse changes, acupuncture depth and massage force, and updates the operation difficulty and response speed in the simulation scene in real time.

[0017] Furthermore, the virtual operation scene model includes human skeleton, muscles, anatomical structure, meridian information, acupoint distribution and virtual medical instruments.

[0018] Furthermore, the operation process of the operation interaction module includes the following steps:

[0019] Receive real-time operation instructions input by the user through the user input device and obtain operation data;

[0020] Using a variational autoencoder to extract key features from the collected operation data, the key features include operation path, force change, angle change and time continuity;

[0021] Based on the key features, dynamically adjusting scene responses in the virtual reality simulation module;

[0022] The collected operation data and extracted key features are stored in a database.

[0023] Furthermore, the operation process of the operation data analysis module includes the following steps:

[0024] According to the operation data and key features, an operation analysis model based on a graph neural network is constructed to analyze the mutual relationship and time evolution between actions and capture the dynamic evolution law in the operation; the nodes represent the user's operation actions, operation features and operation time points; the edges represent the user's action sequence, the correlation between actions and the time sequence correlation;

[0025] Combined with the generative adversarial network algorithm, a standard operation model is established based on the expert system, and the standard operation model is compared with the operation analysis model to generate error analysis, including deviations in operation accuracy, strength and sequence;

[0026] Based on the above analysis results, an operation data analysis report is generated, including operation evaluation, difference analysis and improvement suggestions.

[0027] Furthermore, the formula of the operation analysis model is as follows:

[0028]

[0029] in, represents the new features of the user operation after being processed in the model, W (k) Represents the weight matrix of the kth layer, which is used to map and adjust the operation features; and Represents the node feature vector of the Kth layer; Represents the set of neighbor nodes of node v, that is, the time and order relationship between operation actions; d v and d u represents the degree of nodes v and u; σ represents the activation function; λ represents the weighting function; D(x r ) represents the discriminator, which is used to evaluate the real user operation data; G(z g ) indicates that the generator passes the noise z g The generated standard operation data simulates the standard operation of experts.

[0030] Furthermore, the operation process of the adaptive learning feedback module includes the following steps:

[0031] receiving the operation data analysis report of the operation data analysis module, and identifying the operation mode and learning progress of the user;

[0032] Based on the user's operation performance and learning history, the learning path optimization algorithm is used to analyze the user's mastery of various skill points in clinical skills and generate dynamically adjusted personalized learning strategies;

[0033] Provide instant feedback to remind users of problems or deficiencies in the operation process, and give specific improvement suggestions based on deviations in the operation data;

[0034] Based on personalized learning strategies, phased learning goals are generated, and the learning path is continuously updated through algorithms.

[0035] Furthermore, the formula of the learning path optimization algorithm is as follows:

[0036]

[0037] Among them, L opt represents the optimal learning path obtained by the optimization algorithm; w i represents the weight coefficient, that is, the importance of the i-th skill point; represents the user operation error of the i-th skill point at the current time point t; represents the user learning efficiency of the i-th skill point at the current time point t; η represents the difficulty influence coefficient; D i Indicates the difficulty of the task currently faced by the user; Indicates the user's mastery of this skill point; P target represents the target progress; β represents the historical learning adaptability coefficient; Indicates the user's learning adaptability to the system in the previous learning process.

[0038] The beneficial effects of the present invention are:

[0039] The virtual reality simulation module of the present invention uses three-dimensional modeling and real-time rendering technology, combined with ray tracing and physical simulation algorithms, to make traditional Chinese medicine skill training scenes (such as pulse diagnosis, acupuncture, and massage) more realistic and vivid, enhance the learner's sense of immersion and experience, and thus improve learning efficiency. The operation interaction module supports a variety of input devices, and uses a variational autoencoder to extract key features of the operation data, which can accurately capture the user's operation details. This real-time interactive mechanism helps to improve the learner's feedback speed in practice and enhance the accuracy of skill mastering. The operation data analysis module analyzes the action relationship and time evolution through a graph neural network, and uses a generative adversarial network to compare standard operations and user operations, and automatically generates deviation instructions. This automated data analysis greatly improves the ability to accurately evaluate user operations, helps learners better understand the deficiencies in their own operations, and provides clear directions for improvement. The adaptive learning feedback module uses a learning path optimization algorithm based on the operation data analysis report to formulate a personalized learning plan for each learner, and dynamically adjusts the learning path according to the learning progress to ensure that learners can perform skill training at the rhythm that best suits them and improve learning effects. The teaching evaluation module uses fuzzy evaluation algorithm and hierarchical analysis method to comprehensively evaluate learners' operations and generate personalized teaching evaluation reports. This precise evaluation system enables clear feedback at every stage of the teaching process, further optimizes the teaching effect, and provides a scientific basis for subsequent teaching planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a module schematic diagram of an intelligent teaching system for cultivating clinical skills of traditional Chinese medicine according to the present invention.

[0041] Figure 2 It is a flowchart of the operation process of the virtual reality simulation module provided by an embodiment of the present invention.

[0042] Figure 3 It is a flowchart of the operation data analysis module operation process provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] See also Figure 1-3 As shown, the present invention relates to an intelligent teaching system for cultivating clinical skills in traditional Chinese medicine.

[0044] An intelligent teaching system for cultivating clinical skills in traditional Chinese medicine, comprising a virtual reality simulation module, an operation interaction module, an operation data analysis module, an adaptive learning feedback module and a teaching evaluation module which are sequentially connected in communication;

[0045] The virtual reality simulation module is used to construct a simulation operation scene for TCM clinical skills training based on three-dimensional modeling and real-time rendering technology in combination with virtual reality technology, and utilizes ray tracing algorithm and physical simulation algorithm to improve the realism of the simulation operation scene; the simulation operation scene includes pulse diagnosis, acupuncture and massage;

[0046] The operation process of the virtual reality simulation module includes the following steps:

[0047] Based on the preset TCM clinical skill training requirements, a virtual operation scene model is created through three-dimensional modeling technology; the virtual operation scene model includes human skeleton, muscles, anatomical structure, meridian information, acupoint distribution and virtual medical instruments;

[0048] Specifically, human skeleton and muscle modeling: 3D modeling technology is used to finely model the human skeleton, ensuring that the structure and joint movement of each bone conform to the actual physiological situation. At the same time, the muscle model is built based on anatomical data to ensure that the contraction, tension and relaxation of muscles can be realistically simulated during pulse diagnosis, acupuncture, massage and other operations.

[0049] Meridian and acupoint distribution modeling: Based on traditional Chinese medicine theory, the system constructs an accurate meridian and acupoint distribution model. The meridian model covers the whole body and corresponds to different internal organs in the human body. The acupoint distribution is closely integrated with the human anatomical structure, ensuring that users can accurately locate the acupoints through VR devices during acupuncture operations.

[0050] Virtual medical equipment modeling: Acupuncture needles, massage instruments and other medical equipment are also constructed using 3D modeling technology to ensure that the appearance, weight and operating feel are consistent with the actual equipment. In the virtual scene, users can simulate the operation of these virtual instruments through interactive devices to complete the selection and use of medical instruments.

[0051] Using a physical simulation algorithm based on finite element analysis to perform mechanical simulation on the operation behavior of the virtual operation scene model;

[0052] It should be noted that the system uses finite element analysis technology to simulate the mechanical behavior in the virtual operation scene. Finite element analysis can segment the complex structure of the virtual model, decompose it into a limited number of small units, and calculate the physical properties of each unit such as stress and strain under the action of external forces, as follows:

[0053] Mechanical simulation of pulse diagnosis: During the pulse diagnosis process, the system simulates the force of the user pressing the wrist. Finite element analysis calculates the pressure distribution of the finger on the skin and blood vessels and the deformation of the skin when pressing. This mechanical simulation ensures a realistic touch when diagnosing the pulse. At the same time, according to the change in applied pressure, the system can provide corresponding pulse feedback (such as floating pulse, sinking pulse, etc.).

[0054] Mechanical simulation of acupuncture operation: Acupuncture operation involves the process of needle piercing the skin. Through finite element analysis, the system calculates the force on the skin and muscles during acupuncture, and simulates the entire process of the needle piercing the skin and passing through the muscle tissue. Mechanical simulation can accurately reflect the impact of the angle, depth, and force of needle insertion on the needle sensation and treatment effect.

[0055] Mechanical simulation of massage operation: Massage operation needs to simulate the effects of different massage techniques on the skin, muscles, and bones. Finite element analysis can simulate how the pressure applied by massage techniques acts on muscles and bones, and calculate the stress distribution of muscle tissues by techniques of different strengths, directions, and frequencies. Through mechanical simulation, users can obtain real feedback on massage operations.

[0056] Use ray tracing algorithms to render virtual scenes in real time, including the light scattering effect of the skin, skin deformation during acupuncture needle insertion, and reflection phenomena;

[0057] It should be noted that in order to improve the realism of the virtual operation scene, the system uses the ray tracing algorithm to render the scene in real time, especially the light effect of the skin, as follows:

[0058] Skin light scattering effect: The ray tracing algorithm can accurately simulate the effect of light penetrating the surface of the skin and scattering under the skin. This light scattering phenomenon makes the skin look more natural and soft. Especially during the operation, when the fingers or needles touch the skin, the skin's luster and tone will change slightly, and the system can dynamically present these details through the ray tracing algorithm.

[0059] Skin deformation and reflection during acupuncture: When a user performs acupuncture, the needle will cause slight skin deformation after entering the skin. The ray tracing algorithm renders these deformation processes in real time, including the shadow changes and light reflection effects caused by the needle on the skin surface. When the needle is inserted at different angles, the light reflection effect of the skin will also change accordingly, further enhancing the visual realism.

[0060] The physical simulation engine automatically adjusts the response state of the virtual operation scene model, combines the dynamic feedback of pulse changes, acupuncture depth and massage force, and updates the operation difficulty and response speed in the simulation scene in real time.

[0061] It should be noted that the system uses a physical simulation engine to adjust the feedback of user operations in real time to ensure that the virtual operation scene can respond according to the operation situation. The physical simulation engine adjusts the physical state of the virtual human model in real time according to the user's operation data, as follows:

[0062] Dynamic feedback of pulse changes: During the pulse diagnosis operation, the system will simulate subtle changes in the pulse according to the pressure, position and duration applied by the user. For example, when the user presses lightly, the system will feedback a floating pulse, and when the user presses deeply, the system will feedback a sinking pulse. Real-time feedback of the pulse will be presented through tactile devices and visual effects.

[0063] Dynamic feedback of acupuncture depth and force: The physical simulation engine calculates the depth and force of acupuncture in real time, and adjusts the difficulty of acupuncture according to different depths and angles. For example, when acupuncture is deeper, the force feedback of the needle increases, providing the user with a more difficult operation experience. At the same time, the system simulates different needle sensations (such as soreness, swelling, numbness, etc.) according to the depth of acupuncture, and transmits it to the user through the force feedback device.

[0064] Dynamic feedback of massage force: During massage operation, the user's technique, force and massage rhythm will be simulated in real time through the physical engine. The system can adjust the tension of the muscles and the elasticity of the skin to ensure that the response of massage techniques under different forces is real and natural. During the operation, as the user's technique changes, the system will immediately feedback the impact of the massage force on different parts of the body, increase the difficulty of the operation or give prompts.

[0065] The operation interaction module is used to perform real-time interaction through user input devices, collect operation data, and use variational autoencoders to extract key features of user operations; the user input devices include keyboards, handles, touch pads, joysticks, and force feedback devices; the operation data includes user input operation position information, force information, time series, operation frequency, and operation area positioning information;

[0066] Specifically, the system supports a variety of input devices, including:

[0067] Keyboard / Mouse: Users can input diagnosis and treatment steps and related information through the keyboard, and the mouse is used for detailed operations and selections in the scene.

[0068] Handle: The handle is used for acupuncture and massage operations. Combined with the force feedback device, it can simulate real mechanical effects.

[0069] Touchpad: The touchpad is used for pulse diagnosis. Users can simulate different pulse conditions by pressing different pressures. The multi-touch function of the touchpad allows users to simulate multi-finger presses at the same time.

[0070] Joystick: used to move the position in the virtual scene, making it easier for users to switch between different perspectives.

[0071] Force feedback device: used to simulate the force effect of the user, especially in acupuncture and massage operations, the user can feel the applied force and the body's reaction force.

[0072] The operation process of the operation interaction module includes the following steps:

[0073] Receive real-time operation instructions input by the user through the user input device and obtain operation data;

[0074] Using a variational autoencoder to extract key features from the collected operation data, the key features include operation path, force change, angle change and time continuity;

[0075] Based on the key features, dynamically adjusting scene responses in the virtual reality simulation module;

[0076] The collected operation data and extracted key features are stored in a database.

[0077] The operation data analysis module is used to analyze the relationship and time evolution between actions based on the operation data and key features using a graph neural network, capture key nodes and patterns in the operation, and use a generative adversarial network to generate a comparison model between standard operations and user operations, automatically generate deviation indicators from actual operations, and generate an operation data analysis report;

[0078] The operation process of the operation data analysis module includes the following steps:

[0079] According to the operation data and key features, an operation analysis model based on a graph neural network is constructed to analyze the mutual relationship and time evolution between actions and capture the dynamic evolution law in the operation; the nodes represent the user's operation actions, operation features and operation time points; the edges represent the user's action sequence, the correlation between actions and the time sequence correlation;

[0080] Specifically, node construction: Each node represents the specific action or operation feature performed by the user during the operation. For example, during acupuncture, the node can represent features such as "needling depth", "needling angle", and "force size" at a specific time point; during massage operations, the node can represent information such as "manipulation changes" and "force application area". The node contains a timestamp to record the time point when the operation occurred. This can capture the timing information of the user's operation, which is convenient for subsequent analysis of the dynamic changes of the operation.

[0081] Edge construction: Edges represent the correlation between nodes and are used to describe the order and timing of operation actions. Through the weight value of the edge, the model can identify the strong correlation between adjacent actions and the smoothness of action connection. For example, during acupuncture, the system will analyze the correlation between acupuncture depth, angle and force. The larger the weight value of the edge, the stronger the correlation between these actions. At the same time, the edge is also used to analyze the evolution of the action over time, and the model will automatically identify the temporal order of the action. For example, during massage operation, the connection order and force changes of different techniques will be modeled and evaluated through the edge.

[0082] Combined with the generative adversarial network algorithm, a standard operation model is established based on the expert system, and the standard operation model is compared with the operation analysis model to generate error analysis, including deviations in operation accuracy, strength and sequence;

[0083] Specifically, the establishment of a standard operation model: the system constructs a standard operation model through a large amount of operation data of Chinese medicine experts. These data sets may include: the precise operation steps of experts in different Chinese medicine skills (such as acupuncture, massage, and pulse diagnosis). Each key step of the operation and the corresponding time series data, such as the angle, depth, strength and accuracy of each acupuncture of the expert in acupuncture. For each operation step, the expert's standard operation not only includes a description of the action itself, but also an ideal mechanical model. For example, in a massage operation, the expert's force should be uniform and consistent with the physiological response of human muscles. : Construct a standard model of expert operation through a graph neural network. Each node represents the operation performed by the expert at a specific moment and its mechanical characteristics, and the edge represents the timing and connection of the expert's operation. This model is used to compare with the user's operation model to ensure that the user's actions meet the standards.

[0084] The detailed process of generative adversarial network algorithm comparison is as follows:

[0085] Generator: The generator attempts to simulate virtual operation data similar to the user's operation. By observing the user's actual operation, the generator will generate a set of virtual operation feature data that will be as close to the user's operation behavior as possible.

[0086] Discriminator: The task of the discriminator is to distinguish between the generated virtual operation data and the standard operation data. It will gradually improve the accuracy of the comparison and finally conclude the difference between the user operation and the standard operation, as follows:

[0087] Operation accuracy deviation: for example, whether the deviation between the angle of acupuncture and the standard operation is within a reasonable range, or whether the force applied during massage is consistent with the standard operation.

[0088] Operation sequence deviation: The system will check whether the user's operations are performed in the standard order, whether there are skipped steps, disordered sequences, or missing actions.

[0089] Force deviation: During the massage operation, the generator simulates the user's force application behavior, and the discriminator compares its deviation from the expert's standard force, for example, whether the force change is too large or too small.

[0090] Time deviation: By comparing the time nodes of user operations with the timing of standard operations, GAN can detect whether the timing of user actions is reasonable, such as the duration of each technique in massage, the time interval between finger presses when taking a pulse, etc.

[0091] During the training process, the generative adversarial network algorithm will continuously optimize the confrontation between the generator and the discriminator, so as to accurately identify the deviation between the user operation and the standard operation. The generator will continuously generate data close to the user's operation, while the discriminator will continuously improve its ability to distinguish between real and generated data, so that the final generated operation analysis can highly match the actual operation scenario. When the operation deviation is identified, the system will automatically generate an error analysis report, indicating the specific steps in which the user deviates from the standard operation, such as insufficient force during acupuncture or too fast movements during massage.

[0092] Based on the above analysis results, generate an operation data analysis report, including operation evaluation, difference analysis and improvement suggestions;

[0093] Furthermore, the formula of the operation analysis model is as follows:

[0094]

[0095] in, represents the new features of the user operation after being processed in the model, W (k) Represents the weight matrix of the kth layer, which is used to map and adjust the operation features, mapping the features related to TCM operations, such as the force changes during pulse diagnosis, the operation depth during acupuncture, the force and angle changes during massage, etc.; and Represents the node feature vector of the Kth layer; Represents the set of neighbor nodes of node v, that is, the time and order relationship between the operation actions. For example, in acupuncture operation, a node may represent the action of needle piercing the skin, and the neighbor node represents the subsequent adjustment action (such as deepening the needle or rotating the needle), thus reflecting the continuity of the operation action; d v and d u represents the degree of nodes v and u, reflecting the connectivity and complexity between the operation actions. The higher the degree, the stronger the association between the operation action and other actions, which is suitable for analyzing the complex relationship between the change of finger position and the change of force during pulse diagnosis; σ represents the activation function; λ represents the weighting function; D(x r ) represents the discriminator, which is used to evaluate the real user operation data; G(z g ) indicates that the generator passes the noise z g The generated standard operation data simulates the standard operation of experts.

[0096] The adaptive learning feedback module is used to generate a dynamically adjusted personalized learning strategy based on the operation data analysis report using a learning path optimization algorithm;

[0097] The operation process of the adaptive learning feedback module includes the following steps:

[0098] receiving the operation data analysis report of the operation data analysis module, and identifying the operation mode and learning progress of the user;

[0099] Based on the user's operation performance and learning history, the learning path optimization algorithm is used to analyze the user's mastery of various skill points in clinical skills and generate dynamically adjusted personalized learning strategies;

[0100] It should be noted that the learning path optimization algorithm dynamically adjusts the user's learning path based on the user's current skill mastery. The learning path is a personalized learning plan that prioritizes the skills that need to be improved based on the user's performance at different skill points. For example, if the user is weak in acupuncture point positioning, the system will prioritize more acupuncture point positioning exercises and reduce the training time for already mastered skills (such as massage force control). The learning path optimization algorithm not only adjusts the order of training, but also dynamically adjusts the difficulty of training based on the user's performance. For example, if the user's performance in a certain skill point has been significantly improved, the system will automatically increase the difficulty of training, such as requiring deeper acupuncture or more accurate angles, and more uniform force during massage. Each user's learning path is generated based on their unique operational performance, and the system will select specific training content based on the user's error type and progress. For example, the system may arrange more force control exercises for massage operations and more complex pulse training for pulse diagnosis operations.

[0101] Provide instant feedback to remind users of problems or deficiencies in the operation process, and give specific improvement suggestions based on deviations in the operation data;

[0102] Generate phased learning goals based on personalized learning strategies, and continuously update learning paths through algorithms;

[0103] Furthermore, the formula of the learning path optimization algorithm is as follows:

[0104]

[0105] Among them, L opt represents the optimal learning path obtained by the optimization algorithm; w i Represents the weight coefficient, that is, the importance of the i-th skill point, which is assigned according to the TCM clinical skill requirements set by the system; Indicates the user operation error of the i-th skill point at the current time point t, which comes from the error analysis report of the operation data analysis module; represents the user learning efficiency of the i-th skill point at the current time point t, and represents the efficiency of the user practicing on this skill point. The higher the value, the faster the operation error decreases; η represents the difficulty influence coefficient; D iIndicates the difficulty of the task currently faced by the user. The greater the difficulty, the higher the optimization cost. Indicates the user's mastery of this skill point, ranging from 0 to 1; P target It represents the target progress and is usually set to 1, indicating that a skill has been fully mastered; β represents the historical learning adaptability coefficient; Indicates the user's learning adaptability to the system in the previous learning process.

[0106] The teaching evaluation module is used to evaluate the user's performance in various traditional Chinese medicine clinical skills based on the operation data analysis report and personalized learning strategy, using fuzzy evaluation algorithm and hierarchical analysis algorithm to generate a teaching evaluation report.

[0107] Specifically, the core of the fuzzy evaluation algorithm is to classify the multidimensional data of user operations (such as strength, angle, time, etc.) into different fuzzy sets. Each fuzzy set represents a different level of operation, such as "very standard", "basic standard", "slight deviation", "serious deviation", etc. The operation standard corresponding to each level may have a certain range of fuzziness. The system defines a corresponding fuzzy set for each skill action to ensure that the subtle differences in skill operations can be covered. The system defines the degree of membership of the user's current operation with these strength standards through a fuzzy membership function. The degree of membership ranges from 0 to 1, representing the degree of compliance of the current operation with a certain set of standards. For example, if the user's pressing force is in the "moderate force" range, its membership may be 0.8, indicating that it is relatively close to the ideal state.

[0108] Each fuzzy set corresponds to a membership function, which is used to determine the extent to which the user's operation meets a certain standard. The design of the membership function determines the precision of the fuzzy evaluation. The system can select the appropriate function according to the characteristics of different operation behaviors, as follows:

[0109] Triangular membership function: It is suitable for relatively linear operation behaviors with clear boundaries. For example, when the acupuncture angle is too large or too small, the corresponding membership will drop rapidly. In this scenario, the triangular membership function can clearly define the situation where an operation is within the ideal range.

[0110] Trapezoidal membership function: Suitable for behaviors with large tolerances in the operating standards. For example, in a massage operation, there may be a certain range of reasonable changes in the strength of the manipulation, so the trapezoidal membership function can define a wider "standard" area to accommodate this natural operating fluctuation.

[0111] Gaussian membership function: Applicable to operation behaviors that require smoother transitions, especially in scenarios where the operation evaluation has continuous changes. For example, the pressure during pulse diagnosis does not change much with the slight movement of the fingertips, and the Gaussian membership function can smoothly capture this slight change.

[0112] The system converts the fuzzy evaluation value into a specific evaluation result through a defuzzification algorithm, including the maximum membership method and the centroid method:

[0113] Maximum membership method: select the fuzzy set with the largest membership value as the final evaluation result. For example, the user's pulse diagnosis operation belongs to the "moderate force" set with a membership of 0.75, and belongs to the "insufficient force" set with a membership of 0.25. The system selects "moderate force" as the final evaluation result.

[0114] Center of gravity method: weighted calculation of the membership degree is performed to obtain the final score. For example, the score of pulse diagnosis strength can be the weighted average of different membership values ​​to obtain a specific value (such as 85 points), and the operation feedback is generated based on this.

[0115] The analytic hierarchy process (AHP) conducts a multi-dimensional hierarchical evaluation of user operations. The specific hierarchical division is as follows:

[0116] First level: Skill categories

[0117] Based on the different categories of TCM clinical skills, the system first divides the assessment into three main operation categories: pulse diagnosis, acupuncture, and massage. Each type of operation has its own independent scoring criteria and assessment model.

[0118] Second level: Skill subdivision

[0119] Each major skill category is further divided into multiple specific assessment items. For example, in acupuncture, the main assessment items include: the angle of the needle (correctness of acupuncture), the depth of needle insertion (force control), the stability of the needle holding, the standardization of the technique during acupuncture, the speed and accuracy of needle removal, the assessment of pulse diagnosis skills, the intensity of pressing, the accuracy of the pressing position, the duration of pressing, the stability of the fingering and the rhythm control during pulse diagnosis

[0120] The third level: key feature dimensions

[0121] For each skill sub-item, the system further evaluates the operation data collected by the operation interaction module according to different dimensions. Each operation data dimension is linked to a specific operation performance, such as angle error during acupuncture, force deviation during pulse diagnosis, and rhythm uniformity during massage.

[0122] Each key dimension is assigned a different weight to reflect its importance to the overall skill performance. The system calculates the importance weight of each dimension through hierarchical analysis and generates the final comprehensive score.

[0123] After completing the fuzzy evaluation and hierarchical analysis, the teaching evaluation module generates a comprehensive evaluation result. The result is presented in the following ways:

[0124] Quantitative scoring and grading: Based on the above analysis, the system calculates the user's total score in each operation category. The scoring adopts a 0-100 point system and is divided into different levels such as "excellent", "good", "qualified", and "needs improvement". Each level corresponds to different operation performance, and users can clearly understand in which aspects they have outstanding performance and which aspects still need further improvement.

[0125] Detailed skill performance report: In addition to the total score, the system also generates a detailed evaluation report showing the user's performance in each operation sub-item. For example, in acupuncture operations, the system will provide a detailed analysis of the user's acupuncture angle, needle insertion depth, and stability, pointing out the specific values ​​of deviations and possible operational errors. The evaluation report will also present the comparison between the user's operation and the standard operation in the form of a chart or operation process animation. For example, through the data of the force feedback device, the system can show the change curve of the user's manipulation force during the massage process, and overlap it with the force curve of the standard massage technique to intuitively show the difference.

[0126] Personalized improvement suggestions: Based on the evaluation results, the system generates personalized improvement suggestions. For example, for users with large deviations in acupuncture angles, the system will recommend specific angle exercises; for users who have poor control over pulse diagnosis strength, the system may recommend appropriate increase in strength feedback training. These suggestions are not only based on the current evaluation results, but also provide long-term skill improvement strategies based on the user's historical operation data and learning progress, helping users to improve their own operation level more specifically in the subsequent learning process.

[0127] In summary, the present invention constructs a highly realistic TCM operation scene through a virtual reality simulation module, combined with three-dimensional modeling, ray tracing and physical simulation algorithms. The simulation effects of pulse diagnosis, acupuncture, massage and other skills are extremely realistic, especially the operation mechanics simulation realized by finite element analysis and the skin light effect of the ray tracing algorithm, which can effectively improve the students' operation experience and help them understand the operation details more intuitively and deeply. Through the physical simulation engine and the adaptive learning feedback module, the system can capture the key operation parameters such as the strength and angle of the students in the operation in real time, and provide dynamic feedback according to the operation situation. Combined with the operation deviation analysis report generated by the operation data analysis module, the system can generate personalized learning strategies to help students carry out targeted intensive training for weak links.

[0128] The present invention uses graph neural networks, generative adversarial networks and other technologies to accurately analyze the key features of user operations, the relationship between actions and time evolution. The system can not only capture the errors of user operations, but also provide specific improvement suggestions, providing students with a systematic skill improvement path. Through fuzzy evaluation algorithm and hierarchical analysis method, the system can evaluate the students' operation performance from multiple angles. The system provides operation evaluation, difference analysis and personalized improvement suggestions to ensure that students can fully and accurately understand their skill mastery, and continuously optimize the learning path and improve learning efficiency. With the support of the operation interaction module and the data analysis module, the system can accurately record every operation detail of the students, including operation path, force change, time continuity, etc., to further improve the students' fine mastery of clinical operations of traditional Chinese medicine and ensure the improvement of teaching quality.

[0129] The above implementation modes are merely descriptions of the preferred implementation modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering and technical personnel in the field shall fall within the protection scope determined by the claims of the present invention.

Claims

1. An intelligent teaching system for cultivating clinical skills in traditional Chinese medicine, characterized in that: It includes a virtual reality simulation module, an operation interaction module, an operation data analysis module, an adaptive learning feedback module and a teaching evaluation module which are sequentially connected in communication; The virtual reality simulation module is used to construct a simulation operation scene for TCM clinical skills training based on three-dimensional modeling and real-time rendering technology in combination with virtual reality technology, and utilizes ray tracing algorithm and physical simulation algorithm to improve the realism of the simulation operation scene; the simulation operation scene includes pulse diagnosis, acupuncture and massage; The operation interaction module is used to perform real-time interaction through user input devices, collect operation data, and use variational autoencoders to extract key features of user operations; the user input devices include keyboards, handles, touch pads, joysticks, and force feedback devices; The operation data analysis module is used to analyze the relationship and time evolution between actions based on the operation data and key features using a graph neural network, capture key nodes and patterns in the operation, and use a generative adversarial network to generate a comparison model between standard operations and user operations, automatically generate deviation indicators from actual operations, and generate an operation data analysis report; The adaptive learning feedback module is used to generate a dynamically adjusted personalized learning strategy based on the operation data analysis report using a learning path optimization algorithm; The teaching evaluation module is used to evaluate the user's performance in various traditional Chinese medicine clinical skills based on the operation data analysis report and personalized learning strategy, using fuzzy evaluation algorithm and hierarchical analysis algorithm to generate a teaching evaluation report.

2. According to claim 1, the intelligent teaching system for cultivating clinical skills in traditional Chinese medicine is characterized in that: The operation data includes operation position information, force information, time series, operation frequency and operation area positioning information input by the user.

3. The intelligent teaching system for cultivating clinical skills of traditional Chinese medicine according to claim 1, characterized in that: The operation process of the virtual reality simulation module includes the following steps: Based on the preset TCM clinical skills training requirements, a virtual operation scene model is created through 3D modeling technology; Using a physical simulation algorithm based on finite element analysis to perform mechanical simulation on the operation behavior of the virtual operation scene model; Use ray tracing algorithms to render virtual scenes in real time, including the light scattering effect of the skin, skin deformation during acupuncture needle insertion, and reflection phenomena; The physical simulation engine automatically adjusts the response state of the virtual operation scene model, combines the dynamic feedback of pulse changes, acupuncture depth and massage force, and updates the operation difficulty and response speed in the simulation scene in real time.

4. The intelligent teaching system for cultivating clinical skills of traditional Chinese medicine according to claim 3, characterized in that: The virtual operation scene model includes human skeleton, muscles, anatomical structure, meridian information, acupoint distribution and virtual medical instruments.

5. The intelligent teaching system for cultivating clinical skills of traditional Chinese medicine according to claim 1, characterized in that: The operation process of the operation interaction module includes the following steps: Receive real-time operation instructions input by the user through the user input device and obtain operation data; Using a variational autoencoder to extract key features from the collected operation data, the key features include operation path, force change, angle change and time continuity; Based on the key features, dynamically adjusting scene responses in the virtual reality simulation module; The collected operation data and extracted key features are stored in a database.

6. The intelligent teaching system for cultivating clinical skills of traditional Chinese medicine according to claim 1, characterized in that: The operation process of the operation data analysis module includes the following steps: According to the operation data and key features, an operation analysis model based on a graph neural network is constructed to analyze the mutual relationship and time evolution between actions and capture the dynamic evolution law in the operation; the nodes represent the user's operation actions, operation features and operation time points; the edges represent the user's action sequence, the correlation between actions and the time sequence correlation; Combined with the generative adversarial network algorithm, a standard operation model is established based on the expert system, and the standard operation model is compared with the operation analysis model to generate error analysis, including deviations in operation accuracy, strength and sequence; Based on the above analysis results, an operation data analysis report is generated, including operation evaluation, difference analysis and improvement suggestions.

7. The intelligent teaching system for cultivating clinical skills of traditional Chinese medicine according to claim 6, characterized in that: The formula of the operation analysis model is as follows: in, represents the new features of the user operation after being processed in the model, W (k) Represents the weight matrix of the kth layer, which is used to map and adjust the operation features; and represents the node feature vector of the Kth layer; N(v) represents the set of neighboring nodes of node v, that is, the time and order relationship between operation actions; d v and d u represents the degree of nodes v and u; σ represents the activation function; λ represents the weighting function; D(x r ) represents the discriminator, which is used to evaluate the real user operation data; G(z g ) indicates that the generator passes the noise z g The generated standard operation data simulates the standard operation of experts.

8. The intelligent teaching system for cultivating clinical skills of traditional Chinese medicine according to claim 1, characterized in that: The operation process of the adaptive learning feedback module includes the following steps: receiving the operation data analysis report of the operation data analysis module, and identifying the operation mode and learning progress of the user; Based on the user's operation performance and learning history, the learning path optimization algorithm is used to analyze the user's mastery of various skill points in clinical skills and generate dynamically adjusted personalized learning strategies; Provide instant feedback to remind users of problems or deficiencies in the operation process, and give specific improvement suggestions based on deviations in the operation data; Based on personalized learning strategies, phased learning goals are generated, and the learning path is continuously updated through algorithms.

9. The intelligent teaching system for cultivating clinical skills of traditional Chinese medicine according to claim 8, characterized in that: The formula of the learning path optimization algorithm is as follows: Among them, L opt represents the optimal learning path obtained by the optimization algorithm; w i represents the weight coefficient, that is, the importance of the i-th skill point; represents the user operation error of the i-th skill point at the current time point t; represents the user learning efficiency of the i-th skill point at the current time point t; η represents the difficulty influence coefficient; D i Indicates the difficulty of the task currently faced by the user; Indicates the user's mastery of this skill point; P target represents the target progress; β represents the historical learning adaptability coefficient; Indicates the user's learning adaptability to the system in the previous learning process.

Images