Wushu action optimization training method combined with virtual reality technology

Abstract

The invention discloses a martial arts motion optimization training method combined with a virtual reality technology, and relates to the technical field of motion image analysis and processing. The martial arts motion optimization training method combined with the virtual reality technology is characterized in that complex martial arts motions are disassembled into mechanical fractal units which can be recursively combined through fractal mechanical decomposition training; a trainer performs targeted intensive training on each fractal unit with the help of the VR tactile glove, deeply understands the action mechanics principle, greatly improves the control ability of action details, generates a personalized optimal action path light track for the trainer, guides the trainer to complete actions with the minimum joint load, not only reduces action errors, but also improves the training efficiency. In addition, the martial arts action entropy evaluation system quantifies the action deviation in real time, through dynamic adjustment of virtual fluid resistance, a trainer is prompted to correct wrong actions in time, and a set of complete and efficient training system is formed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of analysis and processing of action images, in particular to a martial arts action optimization training method combined with virtual reality technology. BACKGROUND

[0002] Martial arts not only has the practical value of strengthening the body, self-defense and self-protection, but also plays an important role in cultivating the will and spiritual outlook of people. However, the traditional martial arts training has long relied on the mode of oral instruction, which requires high professional level and teaching experience of the coach, and is limited by factors such as venue and time, making it difficult to meet the growing training needs of the public. With the rapid development of science and technology, virtual reality (VR) technology has been gradually applied to the field of sports training, bringing new opportunities for the innovative development of martial arts training.

[0003] At present, the existing martial arts training method based on VR technology still has many deficiencies in action optimization. In the application of biomechanics, it can only simply simulate martial arts actions, and cannot construct personalized biomechanical redundancy models according to the individual bone structure and movement characteristics of the trainee, which makes it difficult for the trainee to complete the action in the most scientific way and increases the risk of joint injury. In the action decomposition training, the linear decomposition method is generally used, which lacks in-depth analysis and flexible recombination of mechanical subunits, limiting the trainee's understanding and mastery of complex actions. In addition, the existing training method lacks quantitative evaluation of action specification, making it difficult to accurately feedback the trainee's action deviation in real time, and unable to guide the trainee to correct the wrong action through effective intervention mechanism.

[0004] In summary, there is an urgent need for an innovative training method in the current martial arts training field to overcome the deficiencies of existing VR martial arts training and achieve scientific, personalized and efficient martial arts training to meet the diverse needs of martial arts training in modern society. SUMMARY

[0005] The purpose of the present application is to overcome the deficiencies of the prior art and provide a martial arts action optimization training method combined with virtual reality technology. It can decompose complex martial arts actions into recursively combinable mechanical subunits through fractal mechanics decomposition training, and the trainee can conduct targeted intensive training on each subunit with the help of VR haptic gloves, thereby deeply understanding the action mechanics principle and greatly improving the control ability of action details. At the same time, the virtual skeleton mapping technology based on biomechanical redundancy optimization generates personalized optimal action path tracks for the trainee, guiding them to complete the action with the minimum joint load, which not only reduces the action error, but also improves the fluency and specification of the action. In addition, the martial arts action entropy evaluation system quantifies the action deviation in real time, and through the dynamic adjustment of virtual fluid resistance, it prompts the trainee to correct the wrong action in time, forming a complete and efficient training system, and comprehensively improving the quality and effect of martial arts training.

[0006] The application provides the following technical solutions to solve the above technical problems: a martial arts action optimization training method combined with virtual reality technology, and the specific steps of the method are as follows:

[0007] S100, a trainer wears a VR device and a VR haptic glove, enters a VR training environment, constructs a personalized biomechanics redundancy model by analyzing user skeletal motion data, calculates the distribution of joint torque in real time and generates a double-color light rail path, and triggers a cutting signal of a fractal unit when a joint torque sudden increase point is detected;

[0008] S200, positioning a martial arts action fractal boundary according to the torque sudden increase point output by S100, cutting a complete action into a mechanically fractal unit library that can be recursively combined, and recording fingertip pressure distribution data by using the haptic glove for correction of the biomechanics redundancy model;

[0009] S300, calculating an action entropy value E, and activating fluid resistance in a virtual environment when the entropy value is greater than an entropy threshold value, wherein the resistance strength of the fluid resistance dynamically decays with the training progress of the fractal unit;

[0010] S400, updating the fractal unit library with the entropy value evaluation result of S300 as haptic feedback data to support nonlinear action recombination;

[0011] S500, automatically adjusting the recursive depth of the updated fractal unit according to the entropy value change rate when the optimized martial arts action sequence is executed in the VR scene.

[0012] Further, the construction of the personalized biomechanics redundancy model in S100 includes:

[0013] After the trainer wears the VR device and the VR haptic glove and enters the VR training environment, real-time skeletal motion data of the trainer is collected, the skeletal motion data contains the position, angle, and motion speed information of each joint, and the initial joint torque theoretical value τ th is obtained.

[0014] The joint fault tolerance threshold τ tol = μ·τ max +(1-μ)·τ th , wherein μ∈[0.4, 0.7] is a risk coefficient, τ max is the maximum bearing torque of the joint.

[0015] In the VR environment, the double-color light rail path is distinguished by different color light rails.

[0016] Further, the double-color light rail path in S100 includes:

[0017] Red warning area: Joint torque surge point, that is, the joint torque is greater than the joint fault tolerance threshold τ tol , triggering the fractal unit cutting signal;

[0018] Blue guidance zone: The movement guidance beam is activated when the joint load is lower than the safety threshold of 50%. The joint load is the comprehensive force that the trainee is subjected to during the training process.

[0019] Furthermore, in the process of the trainee completing each fractal unit movement, the tactile glove synchronously records the fingertip pressure distribution data, organizes the recorded fingertip pressure data according to the time series and fingertip position, constructs the fingertip pressure distribution matrix P, and calculates the mean value of each fingertip pressure distribution. Among them, P ij Represents the pressure value at the jth fingertip position at the i-th time point, where n is the total number of time points and j represents the fingertip position index, reflecting the average force of each fingertip during the entire movement and calculating the variance of the pressure Variance is used to reflect the stability of fingertip force and introduces a risk adjustment factor where m is the total number of fingertip positions, ω j is the weight of each fingertip position, which is used to indicate the importance of different fingertips in affecting joint load. is the overall mean of the average pressures across all fingertips and m is the total number of fingertip positions, and the risk coefficient μ is adjusted according to the risk adjustment factor λ, that is, μ′=μ+k0(λ-λ0), where μ ′ is the adjusted risk coefficient, k0 is the adjustment amplitude coefficient, which is used to control the adjustment amplitude and k0∈(0,1), λ0 is the fingertip pressure reference value, which indicates the discrete degree of standard fingertip pressure distribution and λ0=0.5. By adjusting the risk coefficient μ, the joint fault tolerance threshold τ tol The optimization is carried out, and the fingertip pressure distribution data is also used to update the fractal unit library. According to the trainee's force habits and movement characteristics, the division of fractal units and training methods are optimized.

[0020] Furthermore, the S300 collects the trainee's motion data in real time, including multiple dimensional data of joint angle, velocity, and acceleration, and divides the time point i of each martial arts action fractal unit into multiple time segments, each segment lasting Δt. For each time segment, the motion feature vector X of each joint is extracted. t =[x 1,t ,x 2,t ,…,x n,t ], n is the total number of time points, t represents the time segment number of the training cycle, and the action entropy value of each fractal unit is calculated Among them, T ais the total number of time segments contained in the ath fractal unit, p(x t ) represents the action feature vector X corresponding to the time segment t in the ath fractal unit t The probability of occurrence, set an entropy threshold E for each martial arts action th , when the calculated fractal unit action entropy value E k >E th When the trainee is in a state of shock, the fluid resistance in the virtual environment is immediately activated, prompting the trainee to correct the movement and complete the fractal movement with minimal joint load.

[0021] Furthermore, the S300 applies fluid resistance to the trainee's hand through the VR tactile glove for each fractal unit. The direction of the fluid resistance is opposite to the direction of movement of the hand in the fractal unit. The resistance F is dynamically adjusted according to the difficulty of the fractal unit and the trainee's current joint load. Among them, τ act is the actual joint torque detected during the training of the current fractal unit, τ tol is the joint fault tolerance threshold, F max is the maximum resistance that the tactile glove can exert, k is the adjustment coefficient, and its value range is [0.5, 1.5]. As the trainee gradually reaches the standard for the training of the fractal unit, the fluid resistance gradually decreases. Each time a fractal unit training is successfully completed, and the action entropy value is reduced to the threshold E th Below this point, the fluid resistance decays at a fixed ratio, i.e., the resistance is reduced by 20% for every unit that meets the standard.

[0022] Furthermore, in the fractal unit library updated in S400, the mechanical fractal units are updated to L1-L3 level fractal mechanical units according to the target martial arts movements, wherein:

[0023] The L1 level unit corresponds to the motion trajectory of a single joint and is the most basic action unit;

[0024] L2 level units are recursively combined from 2-3 L1 units;

[0025] Level L3 units include compound movement patterns that are coordinated across limbs;

[0026] The nonlinear motion reorganization selects a combination with the smallest joint load according to real-time load data, and generates a two-color light track path matching the selected combination in a VR environment.

[0027] Furthermore, the S500 is based on the entropy value change rate of the training cycle Where ΔE is the change in action entropy during the training cycle, and Δt is the time elapsed during the training cycle. The logic for adjusting the recursive depth based on the entropy change rate is:

[0028] when When , higher-order fractal units are automatically unlocked, that is, from L1 to L2 and L3;

[0029] when When training, it is forced to downgrade to the basic unit, that is, downgrade from L3 to L2 and L1;

[0030] when When the entropy value reaches the standard, the nonlinear reorganization permission is opened to the fractal units to generate a new combination library of moves.

[0031] Compared with the existing technology, this martial arts movement optimization training method combined with virtual reality technology has the following beneficial effects:

[0032] 1. The present invention uses fractal mechanics decomposition training to break down complex martial arts movements into recursively combinable mechanical fractal units. Trainers use VR tactile gloves to conduct targeted intensive training on each fractal unit, gaining a deep understanding of the principles of movement mechanics and greatly improving their ability to control movement details. At the same time, it generates personalized optimal movement path light tracks for trainers, guiding them to complete movements with minimal joint load. This not only reduces movement errors but also improves the smoothness and standardization of movements. In addition, the martial arts movement entropy value evaluation system quantifies movement deviations in real time and, through dynamic adjustment of virtual fluid resistance, encourages trainers to correct incorrect movements in a timely manner, forming a complete and efficient training system that comprehensively improves the quality and effectiveness of martial arts training.

[0033] 2. The personalized biomechanical redundancy model constructed by the present invention calculates the torque distribution of each joint in real time, clarifies the joint load warning area and the safety guidance area, and triggers the fractal unit cutting signal in time when a sudden increase in joint torque is detected, exceeding the joint fault tolerance threshold, to prevent the trainee from being injured due to continuous excessive joint load. At the same time, by monitoring and analyzing the trainee's fingertip pressure distribution data, the biomechanical model is further optimized to ensure that the training movements are in line with the trainee's physical characteristics and force habits, providing multiple safety guarantees for the trainee and effectively reducing the risk of injury during training.

[0034] Other advantages, objects and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be learned from the practice of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0036] Figure 1 An operational flow chart for a martial arts movement optimization training method combined with virtual reality technology;

[0037] Figure 2 A step-by-step diagram of a martial arts movement optimization training method that incorporates virtual reality technology. DETAILED DESCRIPTION

[0038] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0039] Example 1

[0040] like Figure 2 As shown, this embodiment details the specific implementation steps of a martial arts movement optimization training method combined with virtual reality technology. The method integrates biomechanics, fractal mechanics and entropy value evaluation, and aims to provide martial arts trainees with a personalized, efficient and safe training experience. By constructing a personalized biomechanical redundancy model, fractal mechanics decomposition training, movement entropy value evaluation and dynamic adjustment mechanism, it can achieve precise optimization of martial arts movements, improve training effects and reduce the risk of injury.

[0041] The trainee wears VR equipment and VR tactile gloves and enters the VR training environment. The motion capture system of the VR equipment starts working, which can collect the trainee's skeletal movement data with high precision. This data contains rich information such as the position, angle, and movement speed of each joint. This data is the basis for the subsequent construction of a personalized biomechanical redundancy model.

[0042] S100, build a personalized biomechanical redundant model and generate light tracks: Based on the collected skeletal motion data, calculate the force and torque of each connecting rod, and then obtain the theoretical value of each joint torque τ th In order to ensure training safety, a joint fault tolerance threshold τ is established tol =μ·τ max +(1-μ)·τ th , where μ is the risk coefficient, ranging from [0.4, 0.7], and is flexibly adjusted according to the trainee's fingertip pressure distribution data and the current training goal, τ maxis the maximum load moment of the joint. In the VR environment, a two-color light track path is generated according to the calculation results. The red warning area represents the point where the joint torque suddenly increases. When the joint torque is greater than the joint fault tolerance threshold τ tol When the joint load is lower than the safety threshold of 50%, the blue guiding beam activates the action guiding beam. The joint load here is the comprehensive force exerted during the training process, which comprehensively considers factors such as joint torque, muscle strength and external resistance. The blue guiding beam provides guidance for further optimization of the movement when the trainee's movement is relatively easy and the joint load is low.

[0043] S200, fractal unit cutting and data recording: When a sudden increase in joint torque is detected in S100, triggering the fractal unit cutting signal, the fractal boundary of the martial arts action is accurately located according to the point of torque sudden increase, and the complex complete action is reasonably cut into multiple mechanical fractal units that can be recursively combined. When the trainee completes each fractal unit action, the VR tactile gloves synchronously record the fingertip pressure distribution data. These data are sorted according to time series and fingertip position, and constructed into a fingertip pressure distribution matrix P. By calculating the mean value of each fingertip pressure distribution, the VR tactile gloves can automatically record the fingertip pressure distribution data. Among them, P ij Represents the pressure value at the jth fingertip position at the i-th time point, where n is the total number of time points and j represents the fingertip position index, reflecting the average force of each fingertip during the entire movement and calculating the variance of the pressure Variance is used to reflect the stability of fingertip force and introduces a risk adjustment factor where m is the total number of fingertip positions, ω j is the weight of each fingertip position, which is used to indicate the importance of different fingertips on joint load. The risk coefficient μ is adjusted according to the risk adjustment factor λ, i.e., μ′=μ+k0(λ-λ0), where μ′ is the adjusted risk coefficient, k0 is the adjustment amplitude coefficient, which is used to control the adjustment amplitude and k0∈(0,1), λ0 is the fingertip pressure reference value, which indicates the discrete degree of standard fingertip pressure distribution and λ0=0.5. By adjusting the risk coefficient μ, the joint fault tolerance threshold τ tol The optimization is carried out, and the fingertip pressure distribution data is also used to update the fractal unit library. According to the trainee's force habits and movement characteristics, the division of fractal units and training methods are optimized.

[0044] S300, Action Entropy Calculation and Fluid Resistance Adjustment: Real-time collection of the trainee's action data, including multi-dimensional information such as joint angle, velocity, acceleration, etc., divides each martial arts action fractal unit into multiple time segments, each segment lasting Δt, and extracts the action feature vector X of each joint for each time segment. t =[x1,t ,x 2,t ,…,x n,t ], where n is the number of features, t represents the time segment number, and based on the information entropy principle, the action entropy value of each fractal unit is calculated T a is the total number of time segments contained in the ath fractal unit, p(x t ) is the action feature vector X corresponding to the time segment t in the ath fractal unit t The probability of occurrence reflects the degree of deviation between the current action and the standard action. The entropy threshold E is set for each martial arts action. th , when the fractal unit action entropy value E k >E th When , it means that the trainee's movement deviates from the standard, and the fluid resistance in the virtual environment is activated. For each fractal unit, the VR tactile glove applies fluid resistance to the trainee's hand. The direction of resistance is opposite to the direction of hand movement. The resistance size F is adjusted according to the movement difficulty of the fractal unit and the trainee's current joint load. The calculation formula is: Among them, τ act is the joint torque actually detected during the current fractal unit training, T tol is the joint fault tolerance threshold, F max is the maximum resistance that the tactile glove can exert, and k is the adjustment coefficient, ranging from [0.5 to 1.5]. As the trainee reaches the target of fractal unit training, that is, the action entropy value drops below the threshold, the fluid resistance decays at a fixed ratio. For each successful completion of fractal unit training, the resistance decreases by 20%. This dynamic adjustment mechanism encourages the trainee to continuously correct their movements and complete the training with minimal joint load.

[0045] S400, fractal unit library update and nonlinear motion reorganization: The entropy evaluation results of S300 are used as tactile feedback data to update the fractal unit library. In the updated fractal unit library, the mechanical fractal units are divided into L1-L3 levels. The L1-level unit corresponds to the motion trajectory of a single joint and is the most basic motion unit. The L2-level unit is composed of 2-3 L1 units recursively combined, involving the coordinated movement of multiple joints. The L3-level unit contains a compound motion pattern of cross-limb coordination. Nonlinear motion reorganization selects the combination with the smallest joint load based on real-time load data. In the VR environment, a matching two-color light track path is generated according to the selected combination, providing trainees with new movement practice guidance and exploring movement combinations that are more in line with their own biomechanical characteristics.

[0046] S500, recursive depth adjustment: real-time calculation of the entropy change rate of the training cycle This parameter reflects the changing trend of the trainee's movement standardization over a period of time. When , it indicates that the trainee's action standardization is improving rapidly, and higher-order fractal units are automatically unlocked, from L1 to L2 and L3, allowing the trainee to challenge more complex action combinations and further improve the training effect; when When , it means that the trainee's action standardization has fluctuated greatly or declined, and the trainee is forced to downgrade to the basic unit training, from L3 to L2 and L1, to help the trainee consolidate the basic action; when When the entropy value of the fractal unit reaches the standard, the nonlinear reorganization permission is opened. The trainer can use these fractal units to generate a new combination library of moves, which will stimulate the trainer's creativity and training enthusiasm.

[0047] In summary, this embodiment demonstrates in detail the complete implementation process of the martial arts movement optimization training method combined with virtual reality technology. By constructing a personalized biomechanical redundancy model, performing fractal unit cutting and training, calculating the movement entropy value and adjusting the fluid resistance, updating the fractal unit library and reorganizing the movement, adjusting the recursive depth and other steps, the optimized training of martial arts movements is achieved. This method can accurately analyze the joint load of the trainee, provide personalized training plans, effectively improve the training effect and reduce the risk of injury, and meet the needs of scientific, personalized and efficient martial arts training.

[0048] Example 2

[0049] like Figure 1 As shown, this embodiment provides a trainer with an operation process for martial arts movement training and optimization through a martial arts movement optimization training method combined with virtual reality technology, specifically:

[0050] Equipment preparation and environment entry: The trainee puts on the VR equipment and VR tactile gloves and enters the VR martial arts training environment;

[0051] Initial data collection and model building: After entering the environment, the VR device's motion capture system begins to work, automatically collecting the trainee's skeletal motion data, including information such as the position, angle, and movement speed of each joint. Based on this data, a personalized biomechanical redundancy model is constructed, and the torque distribution of each joint is calculated.

[0052] Two-color light track viewing and guidance: In the VR environment, trainees can see the generated two-color light track path. During training, if the joint load is normal and in the blue guidance zone, trainees can optimize their movements according to the prompts of the activated action guidance beam. If the joint torque increases and enters the red warning zone, trainees should pay attention to adjust their movements to avoid excessive joint load.

[0053] Fractal unit training: When a sudden increase in joint torque is detected, triggering a fractal unit cutting signal, the complete martial arts movement is divided into multiple mechanical fractal units. The trainer uses VR tactile gloves to conduct targeted training on each fractal unit. During the training process, the tactile gloves will record the fingertip pressure distribution data;

[0054] Action entropy feedback and adjustment: The trainee's action data is collected in real time and the action entropy is calculated. If the entropy is greater than the set entropy threshold, the fluid resistance in the virtual environment will be activated, and the trainee will feel resistance in the opposite direction of the hand movement. At this time, the trainee needs to adjust their movements according to the resistance feedback to reduce the action entropy. As the training progresses, the fluid resistance will gradually decay after each successful completion of a fractal unit training and the action entropy drops below the threshold.

[0055] Fractal unit library update and movement reorganization: During training, the fractal unit library is updated based on entropy evaluation results and fingertip pressure distribution data. The updated fractal unit library contains fractal mechanics units of different levels. Trainees can use these units to reorganize nonlinear movements. During reorganization, the optimal movement combination with the least joint load is selected for the trainee based on real-time load data, and a corresponding two-color light track path is generated. Trainees then practice according to the new light track path.

[0056] Recursive depth adjustment: The entropy change rate of the optimized action combination training cycle is continuously calculated to adjust the recursive depth. When the entropy change rate is less than 0.15, the trainee can unlock higher-order fractal units, such as unlocking from level L1 to levels L2 and L3 for training; when the entropy change rate is greater than 0.25, the trainee needs to be forced to downgrade to basic unit training, that is, downgrade from level L3 to levels L2 and L1; when the entropy change rate is between 0.15 and 0.25, the trainee has the right to nonlinearly reorganize the fractal units that meet the entropy value requirements, and can generate a new move combination library and practice it;

[0057] Continuous training and optimization: The trainees will repeat the training according to the above process. During the training process, the training plan will be continuously optimized and adjusted according to various data. The trainees will also continuously improve their martial arts movements based on system feedback to achieve better training results.

[0058] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as above in terms of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can, without departing from the scope of the technical solution of the present invention, make some changes or modifications to equivalent embodiments using the technical contents disclosed above. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A martial arts action optimization training method combined with virtual reality technology, characterized in that: The specific steps of this method are: S100: The trainee wears VR equipment and VR tactile gloves and enters the VR training environment. By analyzing the user's skeletal motion data, a personalized biomechanical redundancy model is constructed, the torque distribution of each joint is calculated in real time, and a two-color light track path is generated. When a sudden increase in joint torque is detected, the fractal unit cutting signal is triggered; S200: Locate the fractal boundary of the martial arts action based on the torque surge point output by S100, divide the complete action into recursively combinable mechanical fractal units to form a fractal unit library, and record the fingertip pressure distribution data through the tactile glove for use in correcting the biomechanical redundancy model; S300, calculating an action entropy value E, and when the entropy value is greater than an entropy value threshold, activating a fluid resistance in the virtual environment, wherein the resistance strength of the fluid resistance dynamically decays with the progress of fractal unit training; S400, uses the entropy evaluation result of S300 as tactile feedback data to update the fractal unit library and support nonlinear action reorganization; S500: When executing the optimized martial arts action sequence in the VR scene, the recursive depth of the updated fractal unit is automatically adjusted according to the rate of change of the entropy value.

2. The martial arts action optimization training method combined with virtual reality technology according to claim 1 is characterized in that: The construction of the personalized biomechanical redundancy model in S100 includes: After the trainee wears VR equipment and VR tactile gloves and enters the VR training environment, the trainee's skeletal motion data is collected in real time. The skeletal motion data includes information on the position, angle, and movement speed of each joint, and the initial theoretical value of each joint torque τ is obtained. th ; Establish joint fault tolerance threshold τ tol =μ·τ max +(1-μ)·τ th , where μ∈[0.4, 0.7] is the risk coefficient, τ max is the maximum load moment of the joint; Use different colored light tracks to distinguish the dual-color light track paths in the VR environment.

3. The martial arts action optimization training method combined with virtual reality technology according to claim 1 is characterized in that: The S100 dual-color optical track path includes: Red warning area: Joint torque surge point, that is, the joint torque is greater than the joint fault tolerance threshold τ tol , triggering the fractal unit cutting signal; Blue guidance zone: The movement guidance beam is activated when the joint load is lower than the safety threshold of 50%. The joint load is the comprehensive force that the trainee is subjected to during the training process.

4. The martial arts action optimization training method combined with virtual reality technology according to claim 1 is characterized in that: In the process of the trainee completing each fractal unit movement, the tactile glove synchronously records the fingertip pressure distribution data, organizes the recorded fingertip pressure data according to time series and fingertip position, constructs the fingertip pressure distribution matrix P, and calculates the mean value of each fingertip pressure distribution. Among them, P ij Represents the pressure value at the jth fingertip position at the i-th time point, where n is the total number of time points and j represents the fingertip position index, reflecting the average force of each fingertip during the entire movement and calculating the variance of the pressure Variance is used to reflect the stability of fingertip force and introduces a risk adjustment factor where m is the total number of fingertip positions, ω j is the weight of each fingertip position, which is used to indicate the importance of different fingertips in affecting joint load. is the overall mean of the average pressures across all fingertips and m is the total number of fingertip positions, and the risk coefficient μ is adjusted according to the risk adjustment factor λ, that is, μ′=μ+k0(λ-λ0), where μ′ is the adjusted risk coefficient, k0 is the adjustment amplitude coefficient, which is used to control the adjustment amplitude and k0∈(0,1), λ0 is the fingertip pressure reference value, which represents the degree of dispersion of the standard fingertip pressure distribution and λ0=0.

5. By adjusting the risk coefficient μ, the joint fault tolerance threshold τ tol The optimization is carried out, and the fingertip pressure distribution data is also used to update the fractal unit library. According to the trainee's force habits and movement characteristics, the division of fractal units and training methods are optimized.

5. The martial arts action optimization training method combined with virtual reality technology according to claim 1 is characterized in that: The S300 collects the trainee's motion data in real time, including multiple dimensions of joint angle, velocity, and acceleration, and divides the time point i of each martial arts action fractal unit into multiple time segments, each segment lasting Δt. For each time segment, the motion feature vector X of each joint is extracted. t =[x 1,t ,x 2,t ,…,x n,t ], n is the total number of time points, t represents the time segment number of the training cycle, and the action entropy value of each fractal unit is calculated Among them, T a is the total number of time segments contained in the ath fractal unit, p(x t ) represents the action feature vector X corresponding to the time segment t in the ath fractal unit t The probability of occurrence, set an entropy threshold E for each martial arts action th , when the calculated fractal unit action entropy value E k >E th When the trainee is in a state of shock, the fluid resistance in the virtual environment is immediately activated, prompting the trainee to correct the movement and complete the fractal movement with minimal joint load.

6. The martial arts action optimization training method combined with virtual reality technology according to claim 1 is characterized in that: The S300 applies fluid resistance to the trainee's hand through the VR tactile glove for each fractal unit. The direction of the fluid resistance is opposite to the direction of movement of the hand in the fractal unit. The resistance F is dynamically adjusted according to the difficulty of the fractal unit and the trainee's current joint load. Among them, τ act is the actual joint torque detected during the training of the current fractal unit, τ tol is the joint fault tolerance threshold, F max is the maximum resistance that the tactile glove can exert, k is the adjustment coefficient, and its value range is [0.5, 1.5]. As the trainee gradually reaches the standard for the training of the fractal unit, the fluid resistance gradually decreases. Each time a fractal unit training is successfully completed, and the action entropy value is reduced to the threshold E th Below this point, the fluid resistance decays at a fixed ratio, i.e., the resistance is reduced by 20% for every unit that meets the standard.

7. The martial arts action optimization training method combined with virtual reality technology according to claim 1 is characterized in that: In the fractal unit library updated in S400, the mechanical fractal units are updated to L1-L3 level fractal mechanical units according to the target martial arts movements, wherein: The L1 level unit corresponds to the motion trajectory of a single joint and is the most basic action unit; L2 level units are recursively combined from 2-3 L1 units; Level L3 units include compound movement patterns that are coordinated across limbs; The nonlinear motion reorganization selects a combination with the smallest joint load according to real-time load data, and generates a two-color light track path matching the selected combination in a VR environment.

8. The martial arts action optimization training method combined with virtual reality technology according to claim 1 is characterized in that: The S500 is based on the entropy value change rate of the training cycle Where ΔE is the change in action entropy during the training cycle, and Δt is the time elapsed during the training cycle. The logic for adjusting the recursive depth based on the entropy change rate is: when When , higher-order fractal units are automatically unlocked, that is, from L1 to L2 and L3; when When training, it is forced to downgrade to the basic unit, that is, downgrade from L3 to L2 and L1; when When the entropy value reaches the standard, the nonlinear reorganization permission is opened to the fractal units to generate a new combination library of moves.

Images