Sensor-based physical education virtual teaching simulation system and method

Abstract

The invention discloses a sports virtual teaching simulation system and method based on a sensor, and belongs to the technical field of sports virtual teaching, and the method comprises the following steps: collecting a large number of physical characteristics and motion data of different athletes through the sensor and a camera, setting the classification range of the physical characteristics of the athletes, and setting the classification range of the physical characteristics; classifying the acquired athlete data according to the range, and synchronously classifying the corresponding sensor data into corresponding categories; through the combination of the steps of multi-source data acquisition and preprocessing of a sensor and a camera, classification modeling according to project and physical features, student feature matching and personalized standard generation, real-time data comparison and feedback and the like, the action posture and the core force generation state of a student can be accurately captured; according to the physical education virtual teaching system and method, the teaching standard and 1: 1 virtual coach guidance matched with individuals are provided, the professionality, pertinence and error correction efficiency of physical education are improved, then efficient, accurate and personalized physical education virtual teaching is achieved, and the limitation that traditional teaching is only heavy in body and single in standard is broken through.

Description

Technical Field

[0001] This invention relates to the field of virtual sports teaching technology, and more specifically, to a sensor-based virtual sports teaching simulation system and method. Background Technology

[0002] Traditional physical education teaching has significant limitations in terms of precision, personalization, and feedback effectiveness, making it difficult to meet the actual needs of high-quality physical education. At the level of movement instruction, traditional teaching relies heavily on video equipment to standardize students' external postures, failing to capture the core muscle activation. For sports like tennis and golf, which require precise control of muscle activation, mere physical standardization is insufficient to address the core of instruction, leading to frequent instances where posture is correct but force application is incorrect, significantly reducing teaching effectiveness and hindering precise guidance from the surface to the essence. At the level of evaluation standards, traditional teaching uses a uniform standard to judge the movement quality of different students, ignoring individual differences in physical characteristics such as height, arm length, and leg length, as well as individual sports styles. This results in evaluations lacking specificity and accuracy, failing to adapt to the different physiological structures and sports characteristics of different students, and restricting the achievement of students' optimal athletic performance. Therefore, there is an urgent need for a physical education teaching technology solution that can break through the limitations of traditional teaching and achieve precise, personalized, and efficient guidance.

[0003] No effective solutions have yet been proposed to address the problems in the relevant technologies. Summary of the Invention

[0004] To address the problems in related technologies, this invention proposes a sensor-based virtual sports teaching simulation system and method to overcome the technical problems of insufficient force capture and single standard in existing related technologies.

[0005] Therefore, the specific technical solution adopted by the present invention is as follows: A sensor-based virtual teaching simulation method for physical education, comprising the following steps: S1. Collect a large amount of physical characteristics and motion data of different athletes through sensors and cameras, set the classification range of athletes' physical characteristics, classify the collected athlete data according to the range, synchronize the corresponding sensor data into the corresponding category, and build a standard motion model for each category. S2. When students use the system, they first select the sports program and teaching category to learn, then the system identifies the students' physical characteristics data through the camera, and then the students wear skin sensors. At the same time, the system matches the students' physical characteristics with the standard movement model of the previously preset classification range. S3. Generate a 1:1 virtual coach based on the trainee's physical characteristics to guide the trainee's training. At the same time, collect the trainee's learning data and compare it with the virtual coach's standard data to analyze the trainee and generate specific improvement suggestions. The virtual coach then demonstrates the correct movements based on the analysis results.

[0006] In a preferred embodiment, the steps of collecting a large amount of physical characteristics and motion data of different athletes through sensors and cameras, setting a classification range for the athletes' physical characteristics, classifying the collected athlete data according to the range, synchronously assigning the corresponding sensor data to the corresponding category, and constructing a standardized motion model for each category include the following steps: S11. By deploying thin-film epidermal sensors on the surface of key muscle groups of athletes, high-precision electromyographic signals and triboelectric signals are collected simultaneously to obtain multi-dimensional data such as muscle activation timing, force intensity and movement status in real time; at the same time, in conjunction with a multi-angle high-speed camera system, the overall movement trajectory of athletes when performing sports events is captured, and their key physical characteristics data, including height, wingspan and leg length, are recorded simultaneously. S12. The collected data is structured and archived according to the sports and their different teaching categories to ensure that the data is clearly associated with specific sports techniques and styles. At the same time, based on the principles of anthropometry, the classification interval thresholds of key physical characteristics are set for each sport. S13. Compare the static physical feature data of each athlete with the above-mentioned preset interval, and classify the overall data, including physical features and synchronously collected dynamic sensor data, into the corresponding physical feature categories. Then, clean, align and standardize the massive data after classification, and further create independent datasets for each combination of sports category and physical feature category. S14. Based on each independent dataset, machine learning and biomechanical modeling techniques are used to analyze and extract the optimal movement trajectory pattern and muscle coordination pattern of this type of person in this action, and a standardized movement model is constructed for each defined category.

[0007] This model defines the standard spatial path, rhythm, and standard force curves for each major muscle group from the start to the end of a movement, and also serves as the gold standard for subsequent teaching.

[0008] As a preferred implementation, the step of analyzing and extracting the optimal movement trajectory pattern and muscle coordination pattern of the population in this action based on each independent dataset using machine learning and biomechanical modeling techniques, and constructing a standardized movement model for each defined category, includes the following steps: S141. Perform precise temporal synchronization and spatial alignment of the electromyographic and triboelectric signals from the epidermal sensors from each independent dataset with the 3D motion capture data from the camera to form a complete data sample of temporal, spatial posture, and muscle activation. S142. Using a dynamic time warping time alignment algorithm, the data samples of all athletes in the same category performing the same action are normalized and aligned on the time axis to ensure that the start point, key phase point and end point of the action can match each other, so that actions of different durations can be compared. S143. Using K-means clustering in cluster analysis, standard motion trajectories are extracted from the motion trajectory data of numerous samples. At the same time, through biomechanical modeling technology of muscle synergy analysis, the core muscle synergy patterns and their activation time curves corresponding to the above standard motion trajectories are decomposed and extracted to form standard muscle force exertion patterns. S144. Calculate the confidence interval of the spatial position of each joint point on the standard movement trajectory at each moment, and the standard deviation range of each muscle signal in the standard muscle force pattern, quantify the reasonable individual difference boundary of the standard, and form a movement template that includes standard values ​​and tolerance range. S145. Integrate the standard movement trajectory, standard muscle exertion pattern and its corresponding spatiotemporal tolerance range, key phase points of the movement into a standardized motion model.

[0009] In a preferred embodiment, the standard movement trajectory, standard muscle exertion pattern and its corresponding spatiotemporal tolerance range, and key phase points of the movement are integrated into a standardized motion model. The specific model formula is as follows: ; in, Represents a complete canonical motion model. Let be a vector function that defines the standard motion trajectory. It represents the sequence of standard positions or angles of key human joints in three-dimensional space over a normalized motion time t. It is a vector function that defines the standard muscle activation pattern. It represents the standardized muscle activation level of each major target muscle group at the same time t. It is a set that defines the key phase points of the action. It is a set of tolerance parameters, which includes and The allowable deviation range at each point in time.

[0010] In a preferred embodiment, when the student uses the system, they first select the sport and teaching category to learn, then the system identifies the student's physical characteristic data through a camera, and then the student wears a skin sensor. Simultaneously, the system matches the student's physical characteristics with a pre-set, standardized movement model within a pre-defined classification range, including the following steps: S21. After the student starts the system, he / she first selects the target sports to learn in the interactive interface. Then the system determines whether there are different teaching categories for the selected sports. If there are different teaching categories, the interface will present these subcategories for the student to further select and determine the specific skills or styles to learn. S22. The system guides the trainee to stand in the designated area, scans them with a camera, and automatically identifies and measures the trainee's key physical characteristics, including height, arm length, and leg length. S23. Following the system's instructions, the trainee correctly attaches the epidermal sensor to the skin surface of the relevant major muscle groups in preparation for collecting electromyographic biological signals during movement. At the same time, the system compares the trainee's physical feature data with the preset category ranges and automatically matches it to the most suitable physical feature classification. S24. Based on the sports program, teaching category, and matched physical feature classification selected by the student, the system calls the corresponding standard motion model from the pre-built model library, and then scales the standard spatial trajectory of the general model according to the student's personal height and limb length data.

[0011] The standard spatial trajectory of the general model is scaled up to calculate and generate a set of customized standard movement paths and muscle exertion curves for each trainee, which can serve as a personalized standard for their subsequent training.

[0012] As a preferred implementation, the step of calling the corresponding standard motion model from the pre-built model library based on the sports project, teaching category, and matched physical feature classification selected by the student, and then scaling the standard spatial trajectory of the general model according to the student's personal height and limb length data includes the following steps: S241. After calling the corresponding standard motion model, first extract the standard motion trajectory data based on standard human body dimensions from the model. S242. Calculate the scaling factor between the student's individual physical data and the model baseline data. S243. The system applies these scaling factors to geometrically scale the standard spatial trajectory. Then, based on the personalized movement path, it combines the scaled limb length, the estimated limb mass distribution parameters, and the biomechanical calculation model to fine-tune the standard muscle force curve in the original standard model. This matches the theoretically required muscle strength level when the student completes the new size movement path and generates a personalized muscle force target curve that fits their physiological structure. The biomechanical computational model was trained by analyzing a large amount of kinematic, dynamic, and electromyographic data collected synchronously from top athletes and performing inverse dynamic calculations based on physical laws. S244. Integrate and synchronize the personalized standard movement path obtained through geometric scaling with the personalized muscle force curve adjusted through biomechanics for subsequent real-time comparison and teaching guidance.

[0013] In a preferred embodiment, the process of generating a 1:1 virtual coach based on the trainee's physical characteristics to guide the trainee's training, simultaneously collecting data on the trainee's learning process, comparing it one-to-one with the virtual coach's standard data, analyzing the trainee, generating specific improvement suggestions, and having the virtual coach demonstrate the correct movements based on the analysis results includes the following steps: S31. The system drives a 1:1 three-dimensional virtual coach model with the same body size in real time based on the trainee's height, arm length, and leg length. S32. When the trainee begins training, the skin sensors on his body and the system camera are activated simultaneously to continuously collect the trainee's real muscle electrical signals, body posture and movement trajectory data, forming a real-time data stream of the trainee. S33. Align the timeline of the student's real-time data stream with the timeline of the personalized standard model, and compare the spatial position of the student's joints with the corresponding joint position of the virtual coach one by one. Calculate the spatial coordinate deviation, quantify the static deviation and dynamic trajectory error of the posture, and then compare the activation sequence, intensity and coordination mode between different muscles of the student's muscle signals with the standard force curve to identify force errors. S34. Based on the analysis results of the combined spatial trajectory deviation and muscle force deviation, the fundamental error is located through a pre-set rule base and expert knowledge model, and improvement suggestions are generated. S35. Based on the main errors identified in the analysis, the virtual coach highlights the correct trajectory or force application point at key points and breaks down and demonstrates the standard movements, forming a real-time teaching loop of assessment, diagnosis, and guidance.

[0014] The sensor-based virtual sports teaching simulation system adopts any of the above-mentioned sensor-based virtual sports teaching simulation methods, including a multi-source data acquisition and preprocessing module, a standardized movement model construction module, a student feature matching and personalized standard generation module, and a real-time teaching interaction and feedback module. The multi-source data acquisition and preprocessing module collects electromyography and triboelectric signals through sensors, and captures movement trajectories and physical features such as height, arm length, and leg length with a high-speed camera. It archives data according to sports projects and teaching categories, sets physical feature classification thresholds, and completes data cleaning and standardization preprocessing to create independent datasets for combinations of project categories and physical feature categories. The standardized motion model construction module, based on the preprocessed independent dataset, performs data synchronization and alignment, time-series normalization, extracts standard motion trajectories through K-means clustering, obtains standard force exertion patterns through muscle synergy analysis, quantifies the tolerance range, integrates information such as trajectory and force exertion patterns, and constructs a standardized motion model that defines the standard of motion. The student feature matching and personalized standard generation module allows the system to identify the student's physical characteristics through a camera after the student selects a sports activity and teaching category. It then matches the corresponding standard movement model, calculates the scaling factor to adjust the trajectory, and combines the biomechanical model to fine-tune the force curve, generating a personalized standard that adapts to the student's physiological structure. The real-time teaching interaction and feedback module generates a 1:1 virtual coach, synchronously collects student training data and compares it with the personalized model, quantifies posture and force deviations, locates core errors and generates improvement suggestions.

[0015] The beneficial effects of this invention are as follows: 1. Traditional physical education teaching often relies on cameras to constrain students' external postures, but struggles to capture core muscle activation, leading to poor teaching results for some sports requiring precise force application. This invention fuses motion trajectory data collected by cameras with electromyographic and triboelectric signals captured by skin sensors. It precisely records students' body postures and movement paths using multi-angle high-speed cameras, while simultaneously acquiring core data such as muscle activation timing and force intensity in real time using sensors. This enables a comprehensive analysis of both external posture and internal force application. This dual data fusion method can accurately identify whether students have correct postures but incorrect force application, overcoming the shortcomings of traditional teaching that only focuses on appearances and neglects the essence. It allows teaching guidance to delve from physical form to core force application, significantly improving the professionalism and effectiveness of teaching. 2. Traditional physical education teaching uses a single standard to judge the quality of different students' movements, ignoring individual differences in physical characteristics, including height, arm length, leg length, and athletic style, resulting in a lack of specificity and accuracy in the evaluation results. This invention collects a large amount of multidimensional data from different athletes, classifies and models them according to sports, teaching categories, and physical characteristics, and constructs a standardized movement model library covering multiple scenarios. When students use it, they first match the corresponding basic model according to their selected sports, teaching categories, and identified physical characteristics. Then, by calculating personalized scaling factors and combining them with biomechanical models, the standard trajectory and force curve are fine-tuned to generate a unique standard adapted to the student's physiological structure. This classification modeling and individual adaptation mode can ensure the accuracy and adaptability of movement evaluation, enabling students to achieve optimal athletic performance based on their own conditions. 3. This invention generates a 1:1 virtual coach based on the student's height, arm length, and leg length. This provides students with an intuitive reference for movement and enables precise guidance through real-time data comparison. During training, the system simultaneously collects the student's movement trajectory and muscle exertion data. A dynamic time warping algorithm aligns the student's data with the virtual coach's standard data, comparing the spatial position of joints, muscle activation sequence, and exertion intensity to quantify posture deviations and exertion errors. Subsequently, the system uses a pre-built rule base and expert knowledge model to pinpoint the root cause of the problem and generate specific improvement suggestions. Simultaneously, the virtual coach visually presents the correction direction by highlighting key trajectories and demonstrating standard movements in detail, forming a closed-loop teaching process of real-time data collection, precise comparison, error diagnosis, and demonstration guidance. This immersive and interactive guidance mode not only allows students to immediately recognize their shortcomings but also enables them to quickly receive targeted correction solutions, improving correction efficiency and learning effectiveness. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a sensor-based virtual sports teaching simulation method according to an embodiment of the present invention. Detailed Implementation

[0018] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention. The components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0019] According to embodiments of the present invention, a sensor-based virtual teaching simulation system and method for physical education are provided.

[0020] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 As shown, the sensor-based virtual sports teaching simulation method according to an embodiment of the present invention includes the following steps: S1. Collect a large amount of physical characteristics and motion data of different athletes through sensors and cameras, set the classification range of athletes' physical characteristics, classify the collected athlete data according to the range, synchronize the corresponding sensor data into the corresponding category, and build a standard motion model for each category. Furthermore, by collecting a large amount of physical characteristics and motion data of different athletes through sensors and cameras, setting a classification range for the athletes' physical characteristics, classifying the collected athlete data according to this range, and synchronously assigning the corresponding sensor data to the corresponding category, and constructing a standardized motion model for each category, the following steps are included: S11. By deploying thin-film epidermal sensors on the surface of key muscle groups of athletes, high-precision electromyographic signals and triboelectric signals are collected simultaneously to obtain multi-dimensional data such as muscle activation timing, force intensity and movement status in real time; at the same time, in conjunction with a multi-angle high-speed camera system, the overall movement trajectory of athletes when performing sports events is captured, and their key physical characteristics data, including height, wingspan and leg length, are recorded simultaneously. S12. The collected data is structured and archived according to the sports and their different teaching categories to ensure that the data is clearly associated with specific sports techniques and styles. At the same time, based on the principles of anthropometry, the classification interval thresholds of key physical characteristics are set for each sport. S13. Compare the static physical feature data of each athlete with the above-mentioned preset interval, and classify the overall data, including physical features and synchronously collected dynamic sensor data, into the corresponding physical feature categories. Then, clean, align and standardize the massive data after classification, and further create independent datasets for each combination of sports category and physical feature category. S14. Based on each independent dataset, using machine learning and biomechanical modeling techniques, analyze and extract the optimal movement trajectory pattern and muscle coordination pattern of this group of people in this movement, and construct a standardized movement model for each defined category. Furthermore, based on each independent dataset, machine learning and biomechanical modeling techniques were used to analyze and extract the optimal movement trajectory pattern and muscle coordination pattern of this group of people in this action, and a standardized movement model was constructed for each defined category, including the following steps: S141. Perform precise temporal synchronization and spatial alignment of the electromyographic and triboelectric signals from the epidermal sensors from each independent dataset with the 3D motion capture data from the camera to form a complete data sample of temporal, spatial posture, and muscle activation. S142. Using a dynamic time warping time alignment algorithm, the data samples of all athletes in the same category performing the same action are normalized and aligned on the time axis to ensure that the start point, key phase point and end point of the action can match each other, so that actions of different durations can be compared. S143. Using K-means clustering in cluster analysis, standard motion trajectories are extracted from the motion trajectory data of numerous samples. At the same time, through biomechanical modeling technology of muscle synergy analysis, the core muscle synergy patterns and their activation time curves corresponding to the above standard motion trajectories are decomposed and extracted to form standard muscle force exertion patterns. S144. Calculate the confidence interval of the spatial position of each joint point on the standard movement trajectory at each moment, and the standard deviation range of each muscle signal in the standard muscle force pattern, quantify the reasonable individual difference boundary of the standard, and form a movement template that includes standard values ​​and tolerance range. S145. Integrate the standard movement trajectory, standard muscle exertion pattern and its corresponding spatiotemporal tolerance range, key phase points of the movement into a standardized movement model. Furthermore, the standard movement trajectory, standard muscle exertion pattern, and their corresponding spatiotemporal tolerance range, as well as the key phase points of the movement, are integrated into a standardized motion model. The specific model formula is as follows: ; in, Represents a complete canonical motion model. Let be a vector function that defines the standard motion trajectory. It represents the sequence of standard positions or angles of key human joints in three-dimensional space over a normalized motion time t. It is a vector function that defines the standard muscle activation pattern. It represents the standardized muscle activation level of each major target muscle group at the same time t. It is a set that defines the key phase points of the action. It is a set of tolerance parameters, which includes and The allowable deviation range at each point in time.

[0021] It should be noted that, in application, the system will use dynamic time warping to align the trainee's real-time data sequence with the model's time axis, then calculate the differences between the trainee's actual trajectory and actual muscle signals and the standard model at each point, and compare them with the tolerance set, thereby quantifying the spatial posture error and muscle exertion error to achieve accurate assessment. S2. When students use the system, they first select the sports program and teaching category to learn, then the system identifies the students' physical characteristics data through the camera, and then the students wear skin sensors. At the same time, the system matches the students' physical characteristics with the standard movement model of the previously preset classification range. Furthermore, when using the system, students first select the sport and teaching category they want to learn. Then, the system identifies the student's physical characteristics through a camera. Next, the student wears a skin sensor, and the system matches the student's physical characteristics with a pre-set, standardized movement model within a pre-defined category, including the following steps: S21. After the student starts the system, he / she first selects the target sports to learn in the interactive interface. Then the system determines whether there are different teaching categories for the selected sports. If there are different teaching categories, the interface will present these subcategories for the student to further select and determine the specific skills or styles to learn. It should be noted that, for example, tennis has teaching categories such as defensive counter-attack, baseline attack, all-court, and serve-and-volley; badminton includes teaching categories with different tactical styles such as serve-and-attack, attack the backcourt, drop shot and net play, and defense followed by attack; and table tennis has teaching categories based on grip and technical characteristics such as penhold close-to-the-table fast attack, shakehand topspin combined with fast attack, and shakehand chop. These subcategories all correspond to different technical movements, power generation patterns, and tactical logics, allowing students to choose to learn according to their own needs. S22. The system guides the trainee to stand in the designated area, scans them with a camera, and automatically identifies and measures the trainee's key physical characteristics, including height, arm length, and leg length. S23. Following the system's instructions, the trainee correctly attaches the epidermal sensor to the skin surface of the relevant major muscle groups in preparation for collecting electromyographic biological signals during movement. At the same time, the system compares the trainee's physical feature data with the preset category ranges and automatically matches it to the most suitable physical feature classification. S24. Based on the sports program, teaching category, and matched physical feature classification selected by the student, the system calls the corresponding standard motion model from the pre-built model library, and then scales the standard spatial trajectory of the general model according to the student's personal height and limb length data.

[0022] Furthermore, based on the sports program, teaching category, and matched physical feature classification selected by the student, the system retrieves the corresponding standard motion model from the pre-built model library. Then, based on the student's individual height and limb length data, the system scales the standard spatial trajectory of this general model, including the following steps: S241. After calling the corresponding standard motion model, first extract the standard motion trajectory data based on standard human body dimensions from the model. S242. Calculate the scaling factor between the student's individual physical data and the model baseline data. It should be noted that in practical applications, the actual arm length of the trainee can be divided by the standard arm length of the model to obtain the arm length scaling factor. The same calculation can be performed on key dimensions such as leg length and torso length to obtain a set of personalized scaling coefficients for different body segments. S243. The system applies these scaling factors to geometrically scale the standard spatial trajectory. Then, based on the personalized movement path, it combines the scaled limb length, the estimated limb mass distribution parameters, and the biomechanical calculation model to fine-tune the standard muscle force curve in the original standard model. This matches the theoretically required muscle strength level when the student completes the new size movement path and generates a personalized muscle force target curve that fits their physiological structure. It should be noted that for joint coordinate data, the system will scale the spatial vector of the corresponding limb in the trajectory according to the length ratio of each limb; for skeletal angle data, the system will keep the angle unchanged, and recalculate the absolute spatial position of the hands and feet according to the new limb length, thereby generating a personalized standard movement path with the correct spatial scale that adapts to the individual body shape of the trainee. S244. Integrate and synchronize the personalized standard movement path obtained through geometric scaling with the personalized muscle exertion curve adjusted through biomechanics for subsequent real-time comparison and teaching guidance. S3. Generate a 1:1 virtual coach based on the trainee's physical characteristics to guide the trainee's training. At the same time, collect the trainee's learning data and compare it with the virtual coach's standard data to analyze the trainee and generate specific improvement suggestions. The virtual coach then demonstrates the correct movements based on the analysis results.

[0023] Furthermore, a 1:1 virtual coach is generated based on the trainee's physical characteristics to guide the training. Simultaneously, data from the trainee's learning process is collected and compared one-to-one with the virtual coach's standard data. The data is analyzed to generate specific improvement suggestions. Based on the analysis results, the virtual coach demonstrates the correct movements, including the following steps: S31. The system drives a 1:1 three-dimensional virtual coach model with the same body size in real time based on the trainee's height, arm length, and leg length. S32. When the trainee begins training, the skin sensors on his body and the system camera are activated simultaneously to continuously collect the trainee's real muscle electrical signals, body posture and movement trajectory data, forming a real-time data stream of the trainee. S33. Align the timeline of the student's real-time data stream with the timeline of the personalized standard model, and compare the spatial position of the student's joints with the corresponding joint position of the virtual coach one by one. Calculate the spatial coordinate deviation, quantify the static deviation and dynamic trajectory error of the posture, and then compare the activation sequence, intensity and coordination mode between different muscles of the student's muscle signals with the standard force curve to identify force errors. S34. Based on the analysis results of the combined spatial trajectory deviation and muscle force deviation, the fundamental error is located through a pre-set rule base and expert knowledge model, and improvement suggestions are generated. S35. Based on the main errors identified in the analysis, the virtual coach highlights the correct trajectory or force application point at key points and breaks down and demonstrates the standard movements, forming a real-time teaching loop of assessment, diagnosis, and guidance.

[0024] The sensor-based virtual sports teaching simulation system adopts any of the above sensor-based virtual sports teaching simulation methods, including a multi-source data acquisition and preprocessing module, a standardized movement model construction module, a student feature matching and personalized standard generation module, and a real-time teaching interaction and feedback module. The multi-source data acquisition and preprocessing module collects electromyography and triboelectric signals through sensors, and captures movement trajectories and physical features such as height, arm length, and leg length with a high-speed camera. It archives data according to sports projects and teaching categories, sets physical feature classification thresholds, and completes data cleaning and standardization preprocessing to create independent datasets for combinations of project categories and physical feature categories. The standardized motion model construction module, based on the preprocessed independent dataset, performs data synchronization and alignment, temporal normalization, extracts standard motion trajectories through K-means clustering, obtains standard force exertion patterns through muscle synergy analysis, quantifies the tolerance range, integrates the trajectory and force exertion pattern information, and constructs a standardized motion model that defines the standard of motion. The student feature matching and personalized standard generation module allows the system to identify the physical characteristics of students after they select a sports program and teaching category, match them with the corresponding standard movement model, calculate the scaling factor to adjust the trajectory, and fine-tune the force curve in combination with the biomechanical model to generate personalized standards that are adapted to the student's physiological structure. The real-time teaching interaction and feedback module generates a 1:1 virtual coach, synchronously collects student training data and compares it with the personalized model, quantifies posture and force deviations, locates core errors and generates improvement suggestions.

[0025] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A sensor-based virtual teaching simulation method for physical education, characterized in that, The method includes the following steps: S1. Collect a large amount of physical characteristics and motion data of different athletes through sensors and cameras, set the classification range of athletes' physical characteristics, classify the collected athlete data according to the range, synchronize the corresponding sensor data into the corresponding category, and build a standard motion model for each category. S2. When students use the system, they first select the sports program and teaching category to learn, then the system identifies the students' physical characteristics data through the camera, and then the students wear skin sensors. At the same time, the system matches the students' physical characteristics with the standard movement model of the previously preset classification range. S3. Generate a 1:1 virtual coach based on the trainee's physical characteristics to guide the trainee's training. At the same time, collect the trainee's learning data and compare it with the virtual coach's standard data to analyze the trainee and generate specific improvement suggestions. The virtual coach then demonstrates the correct movements based on the analysis results.

2. The sensor-based virtual sports teaching simulation method according to claim 1, characterized in that, The process of collecting a large amount of physical characteristics and motion data from different athletes using sensors and cameras, setting a classification range for the athletes' physical characteristics, classifying the collected athlete data according to this range, synchronously assigning the corresponding sensor data to the corresponding category, and constructing a standardized motion model for each category includes the following steps: S11. By deploying thin-film epidermal sensors on the surface of key muscle groups of athletes, high-precision electromyographic signals and triboelectric signals are collected simultaneously to obtain multi-dimensional data such as muscle activation timing, force intensity and movement status in real time; at the same time, in conjunction with a multi-angle high-speed camera system, the overall movement trajectory of athletes when performing sports events is captured, and their key physical characteristics data, including height, wingspan and leg length, are recorded simultaneously. S12. The collected data is structured and archived according to the sports and their different teaching categories to ensure that the data is clearly associated with specific sports techniques and styles. At the same time, based on the principles of anthropometry, the classification interval thresholds of key physical characteristics are set for each sport. S13. Compare the static physical feature data of each athlete with the above-mentioned preset interval, and classify the overall data, including physical features and synchronously collected dynamic sensor data, into the corresponding physical feature categories. Then, clean, align and standardize the massive data after classification, and further create independent datasets for each combination of sports category and physical feature category. S14. Based on each independent dataset, machine learning and biomechanical modeling techniques are used to analyze and extract the optimal movement trajectory pattern and muscle coordination pattern of this type of person in this action, and a standardized movement model is constructed for each defined category.

3. The sensor-based virtual sports teaching simulation method according to claim 2, characterized in that, Based on independent datasets, machine learning and biomechanical modeling techniques are used to analyze and extract the optimal movement trajectory pattern and muscle coordination pattern of this group of people in this action, and a standardized movement model is constructed for each defined category, including the following steps: S141. Perform precise temporal synchronization and spatial alignment of the electromyographic and triboelectric signals from the epidermal sensors from each independent dataset with the 3D motion capture data from the camera to form a complete data sample of temporal, spatial posture, and muscle activation. S142. Using a dynamic time warping time alignment algorithm, the data samples of all athletes in the same category performing the same action are normalized and aligned on the time axis to ensure that the start point, key phase point and end point of the action can match each other, so that actions of different durations can be compared. S143. Using K-means clustering in cluster analysis, standard motion trajectories are extracted from the motion trajectory data of numerous samples. At the same time, through biomechanical modeling technology of muscle synergy analysis, the core muscle synergy patterns and their activation time curves corresponding to the above standard motion trajectories are decomposed and extracted to form standard muscle force exertion patterns. S144. Calculate the confidence interval of the spatial position of each joint point on the standard movement trajectory at each moment, and the standard deviation range of each muscle signal in the standard muscle force pattern, quantify the reasonable individual difference boundary of the standard, and form a movement template that includes standard values ​​and tolerance range. S145. Integrate the standard movement trajectory, standard muscle exertion pattern and its corresponding spatiotemporal tolerance range, key phase points of the movement into a standardized motion model.

4. The sensor-based virtual sports teaching simulation method according to claim 3, characterized in that, The standard movement trajectory, standard muscle exertion pattern and its corresponding spatiotemporal tolerance range, and key phase points of the movement are integrated into a standardized motion model. The specific model formula is as follows: ； in, Represents a complete canonical motion model. Let be a vector function that defines the standard motion trajectory. It represents the sequence of standard positions or angles of key human joints in three-dimensional space over a normalized motion time t. It is a vector function that defines the standard muscle activation pattern. It represents the standardized muscle activation level of each major target muscle group at the same time t. It is a set that defines the key phase points of the action. It is a set of tolerance parameters, which includes and The allowable deviation range at each point in time.

5. The sensor-based virtual sports teaching simulation method according to claim 1, characterized in that, When a student uses the system, they first select the sport and teaching category they want to learn. Then, the system identifies the student's physical characteristics through a camera. Next, the student wears a skin sensor, and the system matches the student's physical characteristics with a pre-defined, standardized movement model within a pre-defined category. This process includes the following steps: S21. After the student starts the system, he / she first selects the target sports to learn in the interactive interface. Then the system determines whether there are different teaching categories for the selected sports. If there are different teaching categories, the interface will present these subcategories for the student to further select and determine the specific skills or styles to learn. S22. The system guides the trainee to stand in the designated area, scans them with a camera, and automatically identifies and measures the trainee's key physical characteristics, including height, arm length, and leg length. S23. Following the system's instructions, the trainee correctly attaches the epidermal sensor to the skin surface of the relevant major muscle groups in preparation for collecting electromyographic biological signals during movement. At the same time, the system compares the trainee's physical feature data with the preset category ranges and automatically matches it to the most suitable physical feature classification. S24. Based on the sports program, teaching category, and matched physical feature classification selected by the student, the system calls the corresponding standard motion model from the pre-built model library, and then scales the standard spatial trajectory of the general model according to the student's personal height and limb length data.

6. The sensor-based virtual sports teaching simulation method according to claim 5, characterized in that, The process of classifying a student's chosen sport, teaching category, and matched physical features, then calling the corresponding standard motion model from a pre-built model library, and scaling the standard spatial trajectory of this general model according to the student's individual height and limb length data, includes the following steps: S241. After calling the corresponding standard motion model, first extract the standard motion trajectory data based on standard human body dimensions from the model. S242. Calculate the scaling factor between the student's individual physical data and the model baseline data. S243. The system applies these scaling factors to geometrically scale the standard spatial trajectory. Then, based on the personalized movement path, it combines the scaled limb length, the estimated limb mass distribution parameters, and the biomechanical calculation model to fine-tune the standard muscle force curve in the original standard model. This matches the theoretically required muscle strength level when the student completes the new size movement path and generates a personalized muscle force target curve that fits their physiological structure. S244. Integrate and synchronize the personalized standard movement path obtained through geometric scaling with the personalized muscle force curve adjusted through biomechanics for subsequent real-time comparison and teaching guidance.

7. The sensor-based virtual sports teaching simulation method according to claim 1, characterized in that, The process involves generating a 1:1 virtual coach based on the trainee's physical characteristics to guide their training. Simultaneously, data from the trainee's learning process is collected and compared one-to-one with the virtual coach's standard data. The analysis of the trainee's data yields specific improvement suggestions, and the virtual coach demonstrates the correct movements based on the analysis results. This includes the following steps: S31. The system drives a 1:1 three-dimensional virtual coach model with the same body size in real time based on the trainee's height, arm length, and leg length. S32. When the trainee begins training, the skin sensors on his body and the system camera are activated simultaneously to continuously collect the trainee's real muscle electrical signals, body posture and movement trajectory data, forming a real-time data stream of the trainee. S33. Align the timeline of the student's real-time data stream with the timeline of the personalized standard model, and compare the spatial position of the student's joints with the corresponding joint position of the virtual coach one by one. Calculate the spatial coordinate deviation, quantify the static deviation and dynamic trajectory error of the posture, and then compare the activation sequence, intensity and coordination mode between different muscles of the student's muscle signals with the standard force curve to identify force errors. S34. Based on the analysis results of the combined spatial trajectory deviation and muscle force deviation, the fundamental error is located through a pre-set rule base and expert knowledge model, and improvement suggestions are generated. S35. Based on the main errors identified in the analysis, the virtual coach highlights the correct trajectory or force application point at key points and breaks down and demonstrates the standard movements, forming a real-time teaching loop of assessment, diagnosis, and guidance.

8. A sensor-based virtual teaching simulation system for physical education, characterized in that: The system adopts the sensor-based virtual sports teaching simulation method as described in any one of claims 1-7, including a multi-source data acquisition and preprocessing module, a standardized movement model construction module, a student feature matching and personalized standard generation module, and a real-time teaching interaction and feedback module; The multi-source data acquisition and preprocessing module collects electromyography and triboelectric signals through sensors, and captures movement trajectories and physical features such as height, arm length, and leg length with a high-speed camera. It archives data according to sports projects and teaching categories, sets physical feature classification thresholds, and completes data cleaning and standardization preprocessing to create independent datasets for combinations of project categories and physical feature categories. The standardized motion model construction module, based on the preprocessed independent dataset, performs data synchronization and alignment, time-series normalization, extracts standard motion trajectories through K-means clustering, obtains standard force exertion patterns through muscle synergy analysis, quantifies the tolerance range, integrates information such as trajectory and force exertion patterns, and constructs a standardized motion model that defines the standard of motion. The student feature matching and personalized standard generation module allows the system to identify a student's physical characteristics via camera after the student selects their sport and teaching category. Match the corresponding standard motion model, calculate the scaling factor to adjust the trajectory, combine the biomechanical model to fine-tune the force curve, and generate a personalized standard that adapts to the student's physiological structure. The real-time teaching interaction and feedback module generates a 1:1 virtual coach, synchronously collects student training data and compares it with the personalized model, quantifies posture and force deviations, locates core errors and generates improvement suggestions.

Images