Old people real-time interactive exercise training teaching system combining virtual reality and sensing technology

Abstract

The invention provides an old people real-time interactive exercise training teaching system combining virtual reality and a sensing technology, and the system comprises a VR display module which is used for constructing a human-computer interaction interface for a user to start training; the VR function service module is used for capturing action data in a user training process based on a sensing technology, analyzing user action accuracy based on a deep learning algorithm, and generating a real-time feedback instruction to dynamically correct user actions; and the data collection module is used for being connected with the VR function service module and storing the user action data and the training historical data for the VR display module to call so as to generate a user exercise performance summary report. According to the technical scheme, the friendliness of the exercise training system in the interactive experience of the elderly user can be improved, personalized adaptation is carried out, and real-time feedback of exercise training of the elderly user is achieved.

Description

Technical Field

[0001] The present invention belongs to the field of virtual reality technology and human factors engineering technology, and in particular relates to a real-time interactive sports training and teaching system for the elderly that combines virtual reality and sensing technology. Background Art

[0002] In recent years, the application of VR technology in elderly health management has made significant progress. Through virtual environments, seniors can engage in various sports training without having to leave their homes or go to high-risk locations, significantly increasing their participation in exercise. Numerous studies have demonstrated that VR sports training can effectively improve balance, muscle strength, and cognitive function in older adults, particularly in preventing falls and improving exercise endurance. For example, some VR systems use virtual coaches to guide users through simple physical exercises or promote mobility and social interaction among seniors through simulated environments.

[0003] Among existing VR sports training systems, some solutions have been attempted to apply to physical exercise for the elderly. Many VR applications help seniors complete simple exercises through the guidance of virtual coaches. For example, Watson and others use virtual coaches to help seniors with gait training, balance training, and other exercises, helping them maintain good exercise habits. Furthermore, some VR systems incorporate wearable sensors to record users' movement data and help evaluate exercise effectiveness. For example, the Max Planck Institute for Informatics invented Captury live motion capture technology. Its main markets are the healthcare and entertainment industries, ranging from virtual reality experiences to the recording of professional dance performances, to biometric gait analysis and functional movement presentation for kindergarten children.

[0004] However, these systems still have some insurmountable limitations, particularly in terms of real-time feedback, personalized adaptation, and ease of use, failing to fully meet the actual needs of the elderly. First, while some existing exercise training systems can record user exercise data, most systems lack real-time feedback and dynamic adjustment and personalized training content. This prevents users from flexibly adjusting their training based on their individual needs, hindering their ability to maximize training effectiveness. Second, existing technologies generally fail to provide personalized, progressive training programs for different elderly populations (such as those with mild cognitive impairment), lacking sufficient adaptability. While some systems use virtual coaches or motion capture technology to provide feedback, they lack mechanisms to adjust exercise intensity and content in real time based on the user's health status, thus failing to truly address the personalized training challenges faced by the elderly. Finally, most existing VR training systems are targeted at middle-aged or young adults who enjoy exercise, with few systems specifically designed for the elderly. Consequently, existing systems often lack concise user interfaces and interaction designs, lacking specific adaptation for elderly users, increasing their difficulty in use and impacting both the user experience and the effectiveness of the systems.

[0005] In summary, existing technologies have significant limitations in terms of real-time feedback, personalized adaptation, and aging-friendly design. Existing VR sports training systems lack dynamic feedback mechanisms, fail to effectively adjust training content based on individual differences, and fail to fully address operational difficulties faced by elderly users. These deficiencies directly impact the elderly population's acceptance and effectiveness of related technologies, highlighting the innovative and practical nature of this invention in terms of aging-friendly design, personalized training programs, and comprehensive assessment capabilities. Summary of the Invention

[0006] The present invention provides a real-time interactive sports training teaching system for the elderly that combines virtual reality and sensing technology to solve the technical problems of existing sports training systems in terms of unfriendly interactive experience for elderly users, insufficient personalized adaptation, and insufficient real-time feedback.

[0007] To achieve the above object, the present invention provides the following solutions:

[0008] A real-time interactive sports training and teaching system for the elderly that combines virtual reality and sensing technology, including:

[0009] VR display module, used to build a human-computer interaction interface for users to start training;

[0010] The VR function service module is used to capture the user's motion data during training based on sensor technology, analyze the user's motion accuracy based on deep learning algorithms, and generate real-time feedback instructions to dynamically correct the user's motion;

[0011] The data collection module is used to connect with the VR function service module to store user motion data and training history data for the VR display module to call to generate a user sports performance summary report.

[0012] Preferably, the VR display module includes:

[0013] An introduction unit for allowing the user to input a user ID, select the gender of the virtual user avatar, and the virtual environment;

[0014] a warm-up training unit, configured to introduce three-dimensional character models of a virtual coach and a virtual user avatar, and to recommend corresponding learning modes to the user by having the virtual user avatar follow the warm-up training of the virtual coach;

[0015] A formal training unit, configured to independently select a learning mode for formal training based on the learning mode recommendation result of the warm-up training unit;

[0016] The action coaching unit is used to provide users with corresponding error correction guidance based on the formal training learning model.

[0017] Preferably, the VR function service module includes:

[0018] A motion capture unit is used to monitor the user's whole body skeletal joints during warm-up exercises to obtain the user's warm-up motion data, and use a motion matching algorithm to calculate the difference between the user's warm-up motion data and the virtual coach's standard motion data to obtain the user's motion accuracy, thereby supporting the recommendation of learning modes in the warm-up training unit;

[0019] A feedback unit is used to monitor and record the user's movement accuracy, and compare the user's formal training movement data with the warm-up movement data, so as to support the movement coaching unit in providing the user with corresponding error correction guidance;

[0020] The adaptive user avatar unit is used for allowing the user to adjust the viewing angle of the virtual user avatar, wherein the viewing angle includes a first viewing angle or a third viewing angle.

[0021] Preferably, in the motion capture unit, the process of obtaining the user's warm-up motion data includes:

[0022] Collect thermal infrared images and depth images of the user through the Azure Kinect RGB camera and depth sensor;

[0023] Based on the thermal infrared image and the depth image, using a deep learning model, obtaining the user's 3D joint coordinates, 3D joint directions and time feature information;

[0024] The user's warm-up movement data is obtained based on the user's 3D joint coordinates, 3D joint directions and time feature information.

[0025] Preferably, the error correction guidance includes voice instructions, video demonstrations and bone color prompts.

[0026] Preferably, the user motion data and training history data collected by the data collection module include: skeletal coordinate position, total motion duration, number of correct motions, number of incorrect motions, motion accuracy and frequency of repeated instructions.

[0027] Preferably, the VR display module also includes a summary reporting unit for constructing a user motion report and tracking training progress based on the user motion data and training history data collected by the data collection module.

[0028] Preferably, the user movement report includes the performance level, movement accuracy and duration of each action of the user.

[0029] Compared to existing technologies, the present invention offers the following benefits: It integrates an "adaptive user avatar" function. This technology automatically adjusts the avatar's interaction style and feedback intensity based on the elderly user's athletic ability, physical condition, and cognitive level. Beginners receive real-time, personalized corrective feedback, while more experienced users receive basic guidance and simple movement feedback. This avatar technology, which dynamically adjusts based on user performance, ensures that each elderly user receives personalized training tailored to their abilities, thereby improving training effectiveness and reducing the risk of sports injuries caused by improper operation. This invention utilizes a Microsoft Kinect device for precise motion capture and combines it with a deep learning algorithm to analyze the user's motion data in real time, providing instant movement feedback. Furthermore, through in-depth interviews and the collection of movement data from sports coaches, standard movement data specifications have been calibrated. Furthermore, the system can adjust the content and intensity of training sessions in real time based on the analysis results, ensuring that elderly users receive accurate guidance during their workouts. This real-time feedback mechanism dynamically adjusts training based on each user's movement performance, improving training effectiveness and optimizing the user experience.

[0030] Special consideration was given to the unique needs of elderly users. By surveying seniors and human-computer interaction experts, we clarified the principles of age-friendly design and designed a simple and intuitive user interface to reduce the user's cognitive burden. The interface utilizes large fonts, simple icons, and clear operation instructions to ensure that seniors can easily understand and operate the system. This age-friendly design makes the system more compatible with the cognitive and physiological characteristics of elderly users, improving user acceptance and effectiveness.

[0031] Integrated multiple feedback methods (voice commands, video demonstrations, or bone color cues): These adaptively adjust to the user's training progress and performance, providing varying intensities and forms of feedback to ensure the user fully understands and executes the exercise task. This diverse feedback mechanism not only improves training effectiveness but also provides elderly users with comprehensive perceptual guidance during exercise. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0033] Figure 1 This is a schematic diagram of the operation of a real-time interactive sports training and teaching system for the elderly that combines virtual reality and sensing technology according to an embodiment of the present invention;

[0034] Figure 2 A recommended flow chart for the learning mode of an embodiment of the present invention;

[0035] Figure 3 This is a flow chart of a real-time interactive sports training and teaching system for the elderly that combines virtual reality and sensing technology according to an embodiment of the present invention;

[0036] Figure 4 This is a schematic diagram of the structure of a real-time interactive sports training and teaching system for the elderly that combines virtual reality and sensing technology according to an embodiment of the present invention;

[0037] Figure 5 A flowchart of an algorithm for constructing an adaptive user avatar according to an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0039] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0040] Example 1

[0041] like Figure 4As shown, a real-time interactive sports training and teaching system for the elderly that combines virtual reality and sensing technology includes: a VR display module, a VR function service module, and a data collection module.

[0042] The VR display module is used to build a human-computer interaction interface for users to begin training. Specifically, this module utilizes the Unity platform (a real-time 3D interactive content creation and operation platform) to build a user interface that interacts with the VR functional service module. This module includes five key units: an introduction unit, a warm-up training unit, a formal training unit, a movement coaching unit, and a summary report unit.

[0043] In a further embodiment, the VR display module includes:

[0044] The introduction unit is used for the user to input the user number, select the gender of the virtual user avatar and the virtual environment; in this embodiment, the user enters the virtual environment (such as Figure 1 As shown). The virtual environment includes a three-dimensional virtual city park environment. Specifically, the process of constructing the three-dimensional virtual city park environment includes:

[0045] Real-life photos were collected to capture the layout of a realistic urban park, including key elements such as trails, lakes, greenery, and buildings. The terrain was then manually sculpted using 3ds Max and Maya 3D software, and finally, high-resolution textures were created in Substance Painter to match the actual soil, grass, and stones.

[0046] The warm-up training unit is used to introduce a three-dimensional character model of a virtual coach (i.e., a coach in the virtual world, responsible for demonstrating the correct exercise guidance movements to the user) and a virtual user avatar (i.e., the image of the user in the virtual world, which synchronously reflects the user's real-time movements in reality to the virtual world). The virtual user avatar follows the virtual coach's warm-up training. The user follows the virtual coach for 1 minute of warm-up activities. The system evaluates the user's exercise level based on the user's exercise performance and recommends the corresponding learning mode to the user. Specifically, in the system's warm-up training unit, the user follows the coach for warm-up training. Among them, the Eight Section Brocade moves guided by the coach are warm-up moves, that is, each move is simplified, and each move is only performed once (the complete move is a move movement in three cycles). Therefore, the system recommends a formal training learning mode suitable for the user based on the user's movement proficiency in the warm-up stage.

[0047] The formal training unit is used to select a learning mode for formal training based on the learning mode recommendation results of the warm-up training unit. In this embodiment, there are three learning modes, and users can choose different modes for formal training. The learning modes include novice mode (i.e., the user's movement accuracy is 0%-70%), experienced mode (i.e., the user's movement accuracy is 70%-75%), and expert mode (i.e., the user's movement accuracy is 75%-100%). The algorithm steps are as follows: Figure 2 shown.

[0048] The movement coaching unit (an error correction and coaching unit) is used to provide users with corresponding error correction guidance based on the learning mode of formal training. In this embodiment, different levels of error correction guidance are provided according to the learning mode. Novice mode provides full coaching and user avatar functions and comprehensive feedback, namely guidance through voice commands, video demonstrations, and bone color cues, allowing users to gradually fine-tune their performance and improve their movement precision. Experienced mode introduces a virtual coach and self-avatar, providing voice commands and video demonstrations. Expert mode only uses virtual coach instructions, reducing visual complexity and reducing feedback that affects the flow experience of expert users.

[0049] The VR function service module is used to capture the user's motion data during training based on sensing technology, analyze the user's motion accuracy based on deep learning algorithms, and generate real-time feedback instructions to dynamically correct user movements.

[0050] A further implementation method is that the VR function service module includes: a motion capture unit, a feedback unit and an adaptive user avatar unit.

[0051] The motion capture unit is used to monitor the user's 32 skeletal joints throughout the body (including key parts such as shoulders, elbows, wrists and knees) during warm-up exercises to obtain the user's warm-up movement data, and use the motion matching algorithm to calculate the difference between the user's warm-up movement data and the virtual coach's standard movement data. This difference is used to establish a correction threshold for the user's formal training unit, obtain the user's movement accuracy, and support the recommendation of learning modes in the warm-up training unit.

[0052] The action matching accuracy formula is as follows:

[0053] , , ,

[0054] In the above formula, represents the joint coordinates of the user or trainer, and Indicates the number of joints selected for calculation. is the normalized vector calculated from the two joint coordinates. Calculate the average cosine similarity between the user's motion vector sequence and the standard motion vector sequence. Note that the present invention sets the maximum allowable angle between matching vectors to 90°. If the angle exceeds 90°, the matching accuracy is determined to be 0.

[0055] A further embodiment is that, in the motion capture unit, the process of obtaining the user's warm-up motion data includes:

[0056] Collect thermal infrared images and depth images of the user through the Azure Kinect RGB camera and depth sensor;

[0057] Based on thermal infrared images and depth images, the user's 3D joint coordinates, 3D joint direction and time feature information are obtained using a deep learning model;

[0058] The user's warm-up movement data is obtained based on the user's 3D joint coordinates, 3D joint directions and time feature information.

[0059] The feedback unit is configured to monitor and record the user's movement accuracy and compare the user's formal training movement data with the warm-up movement data. If the error between the two sets of data exceeds a correction threshold set by the warm-up training unit, the movement coaching unit is enabled to provide the user with corresponding error correction guidance. In a further embodiment, the error correction guidance includes voice commands, video demonstrations, and bone color cues.

[0060] In this embodiment, the feedback unit is implemented using various functional systems in the Unity Editor. Specifically, language commands are implemented using the Audio Source class of Unity's audio system; video demonstrations are implemented using the VideoPlayer class of Unity's video system; and bone colors are implemented using Unity's UGUI system.

[0061] (1) The steps for calling language instructions are as follows: the system references the animation event in the coach animation. When the coach animation plays a specific action, it calls the play function of the Audio Source and plays the corresponding language. In addition, the system calculates the matching degree between the user and the coach's movements (the calculation formula is the same as the above movement matching accuracy formula) and outputs the user's joint information. The language instruction code determines whether the user's movement matching accuracy meets the conditions for triggering the language instruction. If so, it plays the specified feedback audio.

[0062] (2) The steps of video demonstration are described as follows: the system calculates the matching degree between the user and the coach (calculation formula is the same as the above-mentioned matching accuracy formula) and outputs the user's joint information. The video playback code determines whether the user's matching accuracy meets the demonstration video conditions. If so, it activates the object containing the Video Player component and demonstrates the correct movement video of the move.

[0063] (3) The steps for improving bone color are as follows: The system calculates the degree of motion matching between the user and the coach (the calculation formula is the same as the motion matching accuracy formula above) and outputs the user's joint information. The video playback code determines whether the user's motion matching accuracy meets the conditions for displaying bone color. If so, it replaces the object containing the Image component in the parent Canvas object and changes the color of the corresponding bone position of the user.

[0064] The adaptive user avatar unit is used to allow users to adjust the viewing angle of the virtual user avatar, where the viewing angle includes the first perspective (i.e., only part of the virtual body can be seen) or the third perspective (i.e., the entire virtual body can be seen), helping users complete training at the most comfortable observation angle.

[0065] In this embodiment, the process of constructing an adaptive user avatar includes:

[0066] The user's cognitive level is assessed using the MoCA questionnaire (MOCA questionnaire source: https: / / mocacognition.com / ) and divided into different cognitive levels according to different score segments, which is recorded as parameter C n Physical health status was collected through the EQ-5D-5L self-assessment questionnaire (EQ-5D-5L questionnaire source: https: / / euroqol.org / information-and-support / euroqol-instruments / eq-5d-5l / ), and different score segments were used to represent different health status, which was recorded as parameter P n The athletic ability is evaluated by the accuracy of the warm-up phase, which is recorded as parameter A. n The present invention adopts the dynamic weight algorithm Comscore n =0.4C n +0.3P n +0.3A nTo comprehensively evaluate the user's overall ability, n is the number of each user (e.g., 001, 002). In the formal practice scenario, the system will adaptively create different self-avatar interaction modes based on Comscore, including adaptive UI decision mode, adaptive audio decision mode, and adaptive error correction decision mode. This function ensures that each elderly user can receive personalized training based on their own abilities, thereby effectively improving the training effect. The specific algorithm structure is as follows: Figure 5 shown.

[0067] The data collection module (database module) is used to connect with the VR function service module to store user motion data and training history data for the VR display module to call to generate a user sports performance summary report.

[0068] A further implementation method is that the user motion data and training history data collected by the data collection module include: skeletal coordinate position, total motion duration, number of correct motions, number of incorrect motions, motion accuracy and frequency of repeated instructions.

[0069] A further implementation method is that the VR display module also includes a summary reporting unit for constructing a user motion report and tracking training progress based on the user motion data and training history data collected by the data collection module.

[0070] A further embodiment is that the user's motion report includes the user's performance level, motion accuracy, and duration of each action.

[0071] In this embodiment, an electronic device supporting system functions includes a Meta Quest 2 VR headset, Azure Kinect 3.0, and a Windows computer. The Azure Kinect is connected to the computer via a wired connection for skeletal tracking. The Meta Quest 2 VR device is wirelessly connected to the computer's projection display to create an immersive VR experience. The Windows computer is placed directly in front of the participant. The computer transmits resources to the VR headset and is responsible for supporting scene calculations and execution.

[0072] The interaction logic of the present invention is divided into the following five main stages:

[0073] Introduction: Users perform virtual workouts directly by wearing a VR headset, without the need for handheld controllers. Click the Login button and select the gender of their avatar.

[0074] Warm-up training session: The user faces the camera and follows the instructions of the virtual coach to perform warm-up training.

[0075] Formal training session: Users can freely choose any learning mode to start formal sports training.

[0076] Real-time error correction and action coaching: During formal training, users will receive customized error reminder feedback (such as voice instructions, video demonstrations, or bone color prompts) and can freely choose the perspective of the user's avatar (first-person perspective or third-person perspective).

[0077] Ending stage: After watching the training summary report, the user can take off the VR headset to end the experience.

[0078] In summary, the present invention has the following advantages over the prior art:

[0079] Adaptive Avatar Technology: One of the core innovations of this invention is the integrated "adaptive avatar" feature. This technology automatically adjusts the avatar's interaction style and feedback intensity based on the elderly user's motor ability, physical condition, and cognitive level. Beginners receive real-time, personalized corrective feedback, while more experienced users receive basic guidance and simple movement feedback. This avatar technology, which dynamically adjusts based on user performance, ensures that each elderly user receives personalized training tailored to their abilities, thereby improving training effectiveness and reducing the risk of sports injuries caused by improper operation.

[0080] Real-time Motion Feedback and Dynamic Adjustment: This system uses a Microsoft Kinect device for precise motion capture and combines it with deep learning algorithms to analyze the user's motion data in real time, providing immediate feedback. Furthermore, through in-depth interviews and the collection of motion data from sports coaches, standard motion data specifications have been calibrated. This system can then adjust the content and intensity of training sessions in real time based on the analysis results, ensuring that elderly users receive accurate guidance during their workouts. This real-time feedback mechanism dynamically adjusts training based on each user's performance, improving training effectiveness and optimizing the user experience.

[0081] Elderly-Friendly Design: This system specifically addresses the specific needs of senior users. By surveying seniors and human-computer interaction experts, we identified principles for elderly-friendly design and designed a simple, intuitive user interface to reduce the user's cognitive burden. The interface utilizes large fonts, simple icons, and clear operation instructions, ensuring that seniors can easily understand and operate the system. This elderly-friendly design makes the system more compatible with the cognitive and physiological characteristics of senior users, improving user acceptance and user experience.

[0082] Integrated multiple feedback methods (voice commands, video demonstrations, or bone color cues): These adaptively adjust to the user's training progress and performance, providing varying intensities and forms of feedback to ensure the user fully understands and executes the exercise task. This diverse feedback mechanism not only improves training effectiveness but also provides elderly users with comprehensive perceptual guidance during exercise.

[0083] Example 2

[0084] like Figure 3 As shown, this embodiment provides a specific application process of a real-time interactive sports training and teaching system for the elderly that combines virtual reality and sensing technology:

[0085] Step 1: System initialization and introduction unit settings

[0086] This implementation uses an Oculus Quest2 VR headset. Participants will wear a head-mounted display (HMD) and experience VR without controllers. After understanding the training process and entering a participant ID, the experiment officially begins. This information will be used for subsequent performance analysis and health profile tracking.

[0087] Step 2: User avatar gender

[0088] The user will select the gender of the user's avatar, press the OK button, and enter the warm-up training unit.

[0089] Step 3: Warm-up training unit configuration

[0090] In the warm-up training section of the real-time interactive sports training teaching system, participants follow the virtual coach to warm up for one minute. After the warm-up training, participants can choose their preferred learning mode based on the system's recommendations.

[0091] Step 4: Configuration of formal training unit, error correction and coaching unit

[0092] Participants will follow a virtual instructor through a full 12-minute Ba Duan Jin exercise session. They will receive customized error alerts (e.g., voice commands, video demonstrations, or bone color cues) and can freely choose their avatar's perspective (first-person or third-person) during the training.

[0093] Step 5: Summary Report Unit Configuration

[0094] In the summary report section of the real-time interactive sports training teaching system, participants can review the level and accuracy of their sports performance through the practice summary report. Removing the VR headset will automatically exit and end the sports training.

[0095] Through the above-mentioned implementation examples, this invention overcomes the limitations of existing technologies in terms of personalized adaptation, real-time feedback, and elderly-friendly design, providing a more efficient, safe, and adaptable virtual reality sports training system for the elderly. This system not only improves the physical health and cognitive function of the elderly, but also effectively reduces the stress of family caregivers, providing important support for the further development of elderly health interventions.

[0096] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. A real-time interactive sports training and teaching system for the elderly that combines virtual reality and sensing technology, characterized by: include: VR display module, used to build a human-computer interaction interface for users to start training; The VR function service module is used to capture the user's motion data during training based on sensor technology, analyze the user's motion accuracy based on deep learning algorithms, and generate real-time feedback instructions to dynamically correct the user's motion; The data collection module is used to connect with the VR function service module to store user motion data and training history data for the VR display module to call to generate a user sports performance summary report.

2. The system according to claim 1, wherein: The VR display module includes: An introduction unit for allowing the user to input a user ID, select the gender of the virtual user avatar, and the virtual environment; a warm-up training unit, configured to introduce three-dimensional character models of a virtual coach and a virtual user avatar, and to recommend corresponding learning modes to the user by having the virtual user avatar follow the warm-up training of the virtual coach; A formal training unit, configured to independently select a learning mode for formal training based on the learning mode recommendation result of the warm-up training unit; The action coaching unit is used to provide users with corresponding error correction guidance based on the formal training learning model.

3. The system according to claim 2, characterized in that The VR function service module includes: A motion capture unit is used to monitor the user's whole body skeletal joints during warm-up exercises to obtain the user's warm-up motion data, and use a motion matching algorithm to calculate the difference between the user's warm-up motion data and the virtual coach's standard motion data to obtain the user's motion accuracy, thereby supporting the recommendation of learning modes in the warm-up training unit; A feedback unit is used to monitor and record the user's movement accuracy, and compare the user's formal training movement data with the warm-up movement data, so as to support the movement coaching unit in providing the user with corresponding error correction guidance; The adaptive user avatar unit is used for allowing the user to adjust the viewing angle of the virtual user avatar, wherein the viewing angle includes a first viewing angle or a third viewing angle.

4. The system according to claim 3, characterized in that In the motion capture unit, the process of obtaining the user's warm-up motion data includes: Collect thermal infrared images and depth images of the user through the Azure Kinect RGB camera and depth sensor; Based on the thermal infrared image and the depth image, using a deep learning model, obtaining the user's 3D joint coordinates, 3D joint directions and time feature information; The user's warm-up movement data is obtained based on the user's 3D joint coordinates, 3D joint directions and time feature information.

5. The system according to claim 3, wherein: The error correction tutorial includes voice instructions, video demonstrations, and bone color prompts.

6. The system according to claim 5, characterized in that The user motion data and training history data collected by the data collection module include: skeletal coordinate position, total motion duration, number of correct motions, number of incorrect motions, motion accuracy, and frequency of repeated instructions.

7. The system according to claim 5, characterized in that The VR display module also includes a summary reporting unit for constructing a user motion report based on the user motion data and training history data collected by the data collection module, and tracking training progress.

8. The system according to claim 7, characterized in that The user's motion report includes the user's performance level, motion accuracy, and duration of each action.

Images