Exercise intervention method, apparatus and system based on virtual reality and deep reinforcement learning technologies

Abstract

The present application relates to the field of artificial intelligence, and provides an exercise intervention method based on virtual reality and deep reinforcement learning technologies. The method comprises: acquiring an exercise knowledge template matching a virtual exercise scene associated with a virtual reality device; on the basis of the exercise knowledge template, performing reinforcement learning training on an exercise guidance model to be trained, and obtaining a pre-trained general exercise guidance model, the pre-trained general exercise guidance model being used to provide general standardized exercise guidance prompts to a target subject wearing the virtual reality device; acquiring virtual exercise interaction data of the target subject in the virtual exercise scene; and on the basis of the virtual exercise interaction data, performing reinforcement learning-based fine-tuning training on the pre-trained general exercise guidance model, and obtaining a personalized exercise guidance model for the target subject, the personalized exercise guidance model being used to provide personalized exercise guidance prompts to the target subject.

Description

Motion intervention methods, devices, and systems based on virtual reality and deep reinforcement learning technologies

[0001] Cross-references to related applications

[0002] This application claims priority to Chinese Patent Application No. 2024115551299, filed on November 4, 2024, entitled "Motion Intervention Method, Device and System Based on Virtual Reality and Deep Reinforcement Learning Technology", the entire contents of which are incorporated herein by reference. Technical Field

[0003] This application relates to the field of artificial intelligence technology, and in particular to a motion intervention method, device, and system based on virtual reality and deep reinforcement learning technology. Background Technology

[0004] Sports science and technology refers to a discipline that uses scientific principles and technological means to research, design, develop, and apply sports activities. It encompasses multiple disciplines including exercise physiology, exercise psychology, exercise biomechanics, and sports training, aiming to improve people's physical fitness and promote the development and innovation of sports techniques.

[0005] However, in modern society, people face problems such as increased screen time and reduced outdoor activities, leading to insufficient physical activity and consequently affecting their physical and mental health. The inventors realized that in related technologies, real-world exercise interventions typically rely on traditional teacher-led instruction and are limited by venue, failing to effectively stimulate people's interest in exercise and sustained participation, and unable to provide scientific exercise feedback and intelligent personalized exercise guidance.

[0006] Therefore, the related technologies suffer from low intelligence in motion guidance. Summary of the Invention

[0007] According to various embodiments disclosed in this application, a motion intervention method, apparatus, and system based on virtual reality and deep reinforcement learning technology are provided.

[0008] A motion intervention method based on virtual reality and deep reinforcement learning technology includes:

[0009] Obtain a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content;

[0010] The motion guidance model to be trained is subjected to reinforcement learning training based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device;

[0011] Acquire virtual motion interaction data of the target object in the virtual motion scene; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene; and

[0012] The pre-trained general motion guidance model is fine-tuned based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the target object's motion state.

[0013] A motion intervention device based on virtual reality and deep reinforcement learning technology includes:

[0014] The first acquisition module is used to acquire a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content.

[0015] The training module is used to perform reinforcement learning training on the motion guidance model to be trained based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device.

[0016] The second acquisition module is used to acquire virtual motion interaction data of the target object in the virtual motion scene; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene; and

[0017] The fine-tuning module is used to perform reinforcement learning-based fine-tuning training on the pre-trained general motion guidance model based on the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the motion state of the target object.

[0018] A motion intervention system based on virtual reality and deep reinforcement learning technology includes a memory and one or more processors. The memory stores computer-readable instructions, which, when executed by the processors, cause the one or more processors to perform the following steps:

[0019] Obtain a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content;

[0020] The motion guidance model to be trained is subjected to reinforcement learning training based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device;

[0021] Acquire virtual motion interaction data of the target object in the virtual motion scene; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene; and

[0022] The pre-trained general motion guidance model is fine-tuned based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the target object's motion state.

[0023] One or more computer-readable storage media storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the following steps:

[0024] Obtain a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content;

[0025] The motion guidance model to be trained is subjected to reinforcement learning training based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device;

[0026] Acquire virtual motion interaction data of the target object in the virtual motion scene; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene; and

[0027] The pre-trained general motion guidance model is fine-tuned based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the target object's motion state.

[0028] The aforementioned motion intervention method, device, system, and storage medium based on virtual reality and deep reinforcement learning technologies acquire a motion knowledge template matching a virtual motion scene associated with a virtual reality device. The virtual motion scene is obtained by digital twin modeling based on a real motion scene. The motion knowledge template includes standardized motion knowledge content. A pre-trained general motion guidance model is obtained by reinforcement learning training based on the motion knowledge template. The pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to a target object wearing a virtual reality device. Virtual motion interaction data of the target object in the virtual motion scene is acquired. The virtual motion interaction data is obtained based on the actions performed by the target object in the virtual motion scene. The pre-trained general motion guidance model is fine-tuned based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object. The personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on its motion state.

[0029] Thus, a two-stage deep reinforcement learning method is used to train a motion guidance model capable of providing both standardized and personalized motion guidance. In the digital twin stage, the motion guidance model is trained based on a motion knowledge template driven by standardized motion knowledge content to learn standardized motion guidance, resulting in a pre-trained general motion guidance model. In the human-computer interaction feedback stage, the pre-trained general motion guidance model is fine-tuned using virtual motion interaction data of the target object in a virtual motion scene, resulting in a personalized motion guidance model for the target object. This personalized model can adjust the general standardized motion guidance prompts according to the target object's motion state, ensuring the adjusted standardized prompts match the target object's current motion state, thus providing personalized motion guidance on top of providing standardized motion guidance, effectively improving the intelligence of motion guidance.

[0030] Details of one or more embodiments of this application are set forth in the following drawings and description. Other features and advantages of this application will become apparent from the specification, drawings, and claims. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 is a flowchart illustrating a motion intervention method based on virtual reality and deep reinforcement learning technology according to one or more embodiments;

[0033] Figure 2 is a flowchart illustrating the steps of obtaining a motion knowledge template that matches a virtual motion scene associated with a virtual reality device according to one or more embodiments.

[0034] Figure 3 is a flowchart illustrating a physical simulation algorithm for parametric modeling according to one or more embodiments;

[0035] Figure 4 is a schematic diagram of the framework of a deep reinforcement learning algorithm based on motion knowledge templates and human-computer interaction according to one or more embodiments;

[0036] Figure 5 is a flowchart illustrating a motion intervention method based on virtual reality and deep reinforcement learning technology in another embodiment;

[0037] Figure 6 is a block diagram of a motion intervention device based on virtual reality and deep reinforcement learning technology according to one or more embodiments;

[0038] Figure 7 is a block diagram of a computer device according to one or more embodiments. Detailed Implementation

[0039] To make the technical solutions and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0040] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0041] In one embodiment, as shown in Figure 1, a motion intervention method based on virtual reality and deep reinforcement learning technology is provided. This embodiment illustrates the application of this method to a computer device. It is understood that the computer device can be a terminal device, including but not limited to smartphones, tablets, portable personal computers, etc. This application does not limit the type of terminal device. Optionally, the computer device can be a server, which can be an independent physical server, a server cluster consisting of multiple physical servers, or a distributed system. The computer device can also be a system including terminals and servers, and the motion intervention method based on virtual reality and deep reinforcement learning technology is implemented through the interaction between the terminals and servers. In this embodiment, the method includes the following steps:

[0042] Step S110: Obtain a motion knowledge template that matches the virtual motion scene associated with the virtual reality device.

[0043] The virtual reality (VR) devices can include VR headsets and motion controllers. The virtual sports scene is obtained through digital twin modeling of a real-world sports scene. The virtual sports scene can be associated with the motion controller. Different motion controllers can be provided depending on the type of sport. For example, in table tennis, a racket-like controller is provided to offer a more realistic experience of receiving and serving the ball. In football, a leg controller is provided to offer a more realistic kicking experience. The sports knowledge template includes standardized sports knowledge content. This standardized sports knowledge content can be standardized sports knowledge content related to the type of sport corresponding to the virtual sports scene.

[0044] In practice, the computer device can obtain a motion knowledge template matching the virtual motion scene associated with the motion controller in the virtual reality device. This motion knowledge template includes standardized motion knowledge content related to the type of motion corresponding to the virtual motion scene. For example, if the motion controller is a racket-like controller and the associated virtual motion scene is a virtual table tennis scene, the matching motion knowledge template can include standardized motion knowledge content related to table tennis; if the motion controller is a leg controller and the associated virtual motion scene is a virtual soccer scene, the matching motion knowledge template can include standardized motion knowledge content related to soccer.

[0045] Step S120: Based on the motion knowledge template, perform reinforcement learning training on the motion guidance model to be trained to obtain a pre-trained general motion guidance model.

[0046] Among them, the pre-trained general motion guidance model is used to provide general standardized motion guidance cues to target objects wearing virtual reality devices.

[0047] In practice, computer devices can perform reinforcement learning training on the motion guidance model to be trained based on the motion knowledge template to obtain a pre-trained general motion guidance model. The pre-trained general motion guidance model can be used to provide general standardized motion guidance prompts to target objects wearing virtual reality devices.

[0048] The standardized, universal exercise guidance prompts provide general posture guidance and skill tips applicable to a wide range of people during exercise, ensuring the safety and effectiveness of the exercise. After wearing a virtual reality headset and motion controllers, the target user can access the intervention software to participate in the exercise. The target user can choose different difficulty levels and types of exercises based on their own circumstances. Different difficulty levels and types of exercises correspond to different pre-trained universal exercise guidance models. This application's embodiments explain the exercise guidance model corresponding to a specific exercise; it is understood that the method provided in this application can be applied to adjust the exercise guidance models for exercises of different difficulty levels and types.

[0049] Step S130: Obtain virtual motion interaction data of the target object in the virtual motion scene.

[0050] Among them, virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in a virtual motion scene.

[0051] In practice, computer devices can acquire virtual motion interaction data of a target object within a virtual motion scene based on the actions performed by that object. Specifically, the target object has a corresponding virtual character within the virtual motion scene, and the computer device can obtain virtual motion interaction data based on the actions performed by the corresponding virtual character.

[0052] In some embodiments, the actions performed by the target object in the real world can be synchronously mapped to the virtual character corresponding to the target object, thereby obtaining the actions performed by the virtual character corresponding to the target object in a virtual motion scene.

[0053] For example, a real-world gesture might correspond to a specific action or effect in a virtual environment. For instance:

[0054] Example 1: Jump

[0055] Real World: The target object performs a jump in reality, and sensors detect its vertical displacement and time information. Mapping Process: By analyzing the height and duration of the jump, computer equipment maps the jump event to a corresponding effect in a virtual motion scene. Virtual Motion Scene: The virtual character corresponding to the target object jumps in the virtual motion scene, which may trigger a reaction from a virtual prop, such as opening a door or starting a task.

[0056] Example 2: Walking and moving

[0057] Real world: The target object walks in the real world, and sensors capture its position changes, speed, and direction. Mapping process: Computer devices map the target object's walking motion data to its position coordinates in a virtual motion scene. Virtual motion scene: The virtual character corresponding to the target object moves accordingly in the virtual motion scene, perhaps walking in a virtual city, following the same path as in reality.

[0058] Step S140: Based on the virtual motion interaction data, the pre-trained general motion guidance model is fine-tuned using reinforcement learning to obtain a personalized motion guidance model for the target object.

[0059] The personalized exercise guidance model is used to provide personalized exercise guidance prompts to the target object based on the target object's exercise state. The target object's exercise state may include at least one of the target object's body data (such as heart rate, posture, speed, range of motion, etc.) and current virtual exercise interaction data.

[0060] In practice, computer devices can fine-tune a pre-trained general motion guidance model based on reinforcement learning using virtual motion interaction data to obtain a personalized motion guidance model for the target object. This personalized model can adjust the general standardized motion guidance prompts according to the target object's motion state, providing personalized motion guidance prompts. Specifically, personalized motion guidance prompts can provide posture guidance and skill suggestions tailored to the individual during exercise, meeting the target object's unique circumstances and goals, achieving precise motion guidance for the target object, and enhancing the intelligence of motion guidance.

[0061] In the aforementioned motion intervention method based on virtual reality and deep reinforcement learning technologies, the following steps are taken: A motion knowledge template matching a virtual motion scene associated with a virtual reality device is acquired. The virtual motion scene is obtained through digital twin modeling based on a real motion scene. The motion knowledge template includes standardized motion knowledge content. A pre-trained general motion guidance model is trained using reinforcement learning based on the motion knowledge template. This pre-trained general motion guidance model provides standardized motion guidance prompts to a target object wearing a virtual reality device. Virtual motion interaction data of the target object in the virtual motion scene is acquired. This virtual motion interaction data is obtained based on the actions performed by the target object in the virtual motion scene. The pre-trained general motion guidance model is fine-tuned using reinforcement learning based on the virtual motion interaction data to obtain a personalized motion guidance model for the target object. This personalized motion guidance model provides personalized motion guidance prompts to the target object based on its motion state.

[0062] Thus, a two-stage deep reinforcement learning method is used to train a motion guidance model capable of providing both standardized and personalized motion guidance. In the digital twin stage, the motion guidance model is trained based on a motion knowledge template driven by standardized motion knowledge content to learn standardized motion guidance, resulting in a pre-trained general motion guidance model. In the human-computer interaction feedback stage, the pre-trained general motion guidance model is fine-tuned using virtual motion interaction data of the target object in a virtual motion scene, resulting in a personalized motion guidance model for the target object. This personalized model can adjust the general standardized motion guidance prompts according to the target object's motion state, ensuring the adjusted standardized prompts match the target object's current motion state, thus providing personalized motion guidance on top of providing standardized motion guidance, effectively improving the intelligence of motion guidance.

[0063] In one embodiment, as shown in FIG2, step S110, obtaining a motion knowledge template matching the virtual motion scene associated with the virtual reality device, includes the following steps:

[0064] Step S210: Obtain the parametric static modeling results of the virtual objects associated with the virtual motion scene.

[0065] The virtual objects associated with the virtual motion scene can be used to represent real objects in the actual motion scene, i.e., to represent real-world objects. For example, in a table tennis scene, the virtual objects associated with the virtual motion scene can include virtual objects corresponding to real objects such as the table tennis ball and the table tennis racket. In a soccer scene, the virtual objects associated with the virtual motion scene can include virtual objects corresponding to real objects such as the soccer field and the soccer ball. The parametric static modeling result is obtained by modeling the static characteristics of the real objects corresponding to the virtual objects when they are in a static state. These static characteristics can include static position information and static orientation information when in a static state.

[0066] In practice, computer devices can obtain parametric static modeling results of virtual objects associated with virtual motion scenes. These parametric static modeling results are obtained by modeling based on the static characteristics of the real objects corresponding to the virtual objects when they are in a static state.

[0067] Step S220: Obtain the parametric dynamic modeling results of the virtual objects associated with the virtual motion scene.

[0068] The parametric dynamic modeling result is obtained by modeling the dynamic characteristics of real-world objects in motion. These dynamic characteristics can include changes in the object's position, collisions, capture, scaling, surface deformation, etc.

[0069] In practice, computer devices can obtain the parametric dynamic modeling results of virtual objects associated with virtual motion scenes. These parametric dynamic modeling results are obtained by modeling the dynamic characteristics of real objects when they are in motion.

[0070] Step S230: Construct a motion knowledge template based on the parametric static modeling results and the parametric dynamic modeling results.

[0071] In practice, computer devices can construct motion knowledge templates based on parametric static modeling results and parametric dynamic modeling results.

[0072] The technical solution of this embodiment obtains parametric static modeling results of virtual objects associated with a virtual motion scene; these parametric static modeling results characterize the static characteristics of the virtual objects when they are stationary; it also obtains parametric dynamic modeling results of virtual objects associated with the virtual motion scene; these parametric dynamic modeling results characterize the dynamic characteristics of the virtual objects when they are in motion; and a motion knowledge template is constructed based on the parametric static and dynamic modeling results. Thus, by obtaining parametric static and dynamic modeling results obtained from modeling based on the static and dynamic characteristics of the corresponding real-world objects when they are stationary, a motion knowledge template containing standardized motion knowledge content can be constructed, enabling more accurate motion simulation and analysis.

[0073] In one embodiment, obtaining the parametric static modeling result of a virtual object associated with a virtual motion scene includes: obtaining physical characteristic data corresponding to the real object; constructing a virtual model corresponding to the real object based on the physical characteristic data; assigning target physical attributes and target model textures to the virtual model corresponding to the real object to obtain a target virtual model corresponding to the real object; and obtaining the parametric static modeling result based on the target virtual model corresponding to the real object.

[0074] The target physical properties are physical properties that match the real-world object. The target model texture is a model texture that matches the real-world object. Physical property data can include dimensions, shape, weight, material, etc. Physical attributes can include density, coefficient of friction, elastic modulus, etc.

[0075] In practice, during the process of acquiring the parametric static modeling results of virtual objects associated with a virtual motion scene, the computer device can obtain the physical property data of the real-world object and construct a virtual model corresponding to the real-world object based on this data. For example, by collecting the physical property data of the real-world object, using design software to construct a 3D model that conforms to the actual shape, and setting the size and shape as adjustable parameters, a virtual model is obtained. Then, target physical attributes and target model textures can be assigned to the virtual model corresponding to the real-world object to obtain the target virtual model corresponding to the real-world object; where the target physical attributes are physical attributes that match the real-world object, and the target model textures are model textures that match the real-world object. Thus, by assigning target physical attributes, the movement and interaction behavior of the real-world object during motion can be accurately simulated; by assigning target model textures, the visual effect can be enhanced. Furthermore, material settings (such as reflection, gloss, etc.) can be applied to the virtual model corresponding to the real-world object to improve realism. Therefore, the computer device can obtain the parametric static modeling results of the virtual object based on the target virtual model corresponding to the real-world object.

[0076] The technical solution of this embodiment involves acquiring physical characteristic data corresponding to a real-world object, constructing a virtual model of the real-world object based on the physical characteristic data, assigning target physical attributes and target model textures to the virtual model of the real-world object, thereby obtaining a target virtual model of the real-world object. The target physical attributes are physical attributes that match the real-world object; the target model textures are model textures that match the real-world object; and a parametric static modeling result is obtained based on the target virtual model of the real-world object. In this way, a virtual model conforming to the actual shape can be constructed using physical characteristic data, the movement and interaction behavior of the real-world object during motion can be accurately simulated by assigning target physical attributes, and the visual effect can be enhanced by assigning target model textures, making the obtained parametric static modeling result more accurately simulate the static characteristics of the real-world object when it is at rest.

[0077] In one embodiment, obtaining the parametric dynamic modeling result of a virtual object associated with a virtual motion scene includes: obtaining the dynamic equation of the real object when it is in motion; the dynamic equation is used to characterize the dynamic characteristics of the real object when it is in motion; and parametrically modeling the dynamic equation to obtain the parametric dynamic modeling result.

[0078] In practice, during the process of acquiring the parametric dynamic modeling results of virtual objects associated with virtual motion scenes, the computer device can acquire the dynamic equations of real objects in motion. The dynamic equations are used to characterize the dynamic characteristics of real objects in motion. The dynamic equations are then parametrically modeled to obtain the parametric dynamic modeling results.

[0079] The technical solution of this embodiment obtains the dynamic equations of a real-world object in motion; these dynamic equations characterize the dynamic properties of the object in motion; and parametric modeling of these dynamic equations yields parametric dynamic modeling results. Thus, by parametrically modeling the dynamic equations of a real-world object in motion, the dynamic properties of the object in motion can be simulated more accurately.

[0080] To facilitate understanding by those skilled in the art, Figure 3 provides a flowchart of a physical simulation algorithm for parametric modeling. Specifically, Figure 3 illustrates a flowchart of a human-in-the-loop (HIL) physical simulation algorithm. To enable the motion guidance model to master standardized motion knowledge content during the digital twin stage, a HIL-based physical simulation algorithm was designed to construct the VR environment and preset motion knowledge templates. For real objects, such as sports fields, ping-pong paddles, and soccer balls, computer-aided methods are used for parametric modeling, and high-resolution textures are applied to obtain parametric static modeling results. For dynamic modeling, such as the interaction force between the paddle and the ball, and the ball's rotation, computer-aided methods are used to parametrically model the dynamic equations to obtain parametric dynamic modeling results. After the virtual sports scene (virtual sports world) is created, professional coaches and physical education teachers for teenagers (professional users) test it and provide feedback. Based on the feedback, the parametric modeling method is adjusted to achieve high realism, high immersion, and a safe level for the user group.

[0081] In one embodiment, acquiring virtual motion interaction data of a target object in a virtual motion scene includes: acquiring current virtual environment information of the virtual motion scene, and inputting the current virtual environment information into a pre-trained general motion guidance model; the policy function of the pre-trained general motion guidance model is used to generate motion guidance action prompts in response to the current virtual environment information; the motion guidance action prompts are used to instruct the target object to perform corresponding actions in the virtual motion scene to obtain virtual motion interaction data.

[0082] The current virtual environment information can include the current scene, the ball's position, the ball's rotation, and the ball's speed. The policy function of the pre-trained general motion guidance model can be a neural network.

[0083] In practice, during the process of acquiring virtual motion interaction data of the target object in a virtual motion scene, the computer device can obtain the current virtual environment information of the virtual motion scene and input the current virtual environment information into a pre-trained general motion guidance model. The policy function of the pre-trained general motion guidance model is used to generate motion guidance action prompts in response to the current virtual environment information. In this way, the target object can perform corresponding actions according to the motion guidance action prompts. By synchronously mapping the action to the virtual character corresponding to the target object in the virtual motion scene, the virtual character corresponding to the target object can perform corresponding actions in the virtual motion scene to obtain virtual motion interaction data.

[0084] Specifically, the motion guidance cue can refer to the motion guidance cue determined by a pre-trained general motion guidance model from among various candidate motion guidance cues, based on the current virtual environment information, that maximizes the predicted probability of the target object continuing to participate in motion interaction. In this way, motion guidance can be accurately provided to the target object based on the current virtual environment information.

[0085] In one embodiment, a pre-trained general motion guidance model is fine-tuned based on reinforcement learning using virtual motion interaction data to obtain a personalized motion guidance model for the target object. This includes: acquiring virtual environment information after the virtual motion scene responds to the virtual motion interaction data, obtaining new virtual environment information, and acquiring reward data after the pre-trained general motion guidance model outputs motion guidance action prompts; updating the policy function of the pre-trained general motion guidance model based on the new virtual environment information and reward data until the policy function converges to obtain the personalized motion guidance model.

[0086] The reward data is determined based on the predicted probability that the target object will continue to participate in the sports interaction.

[0087] In practice, during the process of fine-tuning a pre-trained general motion guidance model based on reinforcement learning using virtual motion interaction data to obtain a personalized motion guidance model for the target object, the computer device can acquire virtual environment information after the virtual motion scene responds to the virtual motion interaction data, obtain new virtual environment information, and acquire reward data after the pre-trained general motion guidance model outputs motion guidance action prompts. Based on the new virtual environment information and reward data, the policy function of the pre-trained general motion guidance model is updated until the policy function converges, thus obtaining the personalized motion guidance model.

[0088] The technical solution of this embodiment obtains new virtual environment information by acquiring virtual environment information after the virtual motion scene responds to virtual motion interaction data, and obtains reward data after the pre-trained general motion guidance model outputs motion guidance prompts. Based on the new virtual environment information and reward data, the policy function of the pre-trained general motion guidance model is updated until the policy function converges, resulting in a personalized motion guidance model. Thus, by continuously acquiring new virtual environment information and reward data after the virtual motion scene responds to virtual motion interaction data, the pre-trained general motion guidance model can adaptively adjust its policy function to better provide personalized training plans and motion guidance for the target object, making it more in line with individual needs and goals, and improving the accuracy and safety of motion guidance for the target object.

[0089] To facilitate understanding by those skilled in the art, Figure 4 provides a framework for a deep reinforcement learning algorithm based on motion knowledge templates and human-computer interaction. The motion guidance model can be understood as a virtual intelligent coach. As shown in Figure 4, a virtual intelligent coach (REVERIE Coach) is constructed using a transformer network to learn motion knowledge content. REVERIE Coach utilizes a multimodal encoder to acquire current virtual environment information from the virtual motion scene, including the current scene, ball position, ball rotation, and ball speed. In the digital twin stage, REVERIE Coach is trained on motion knowledge templates in the virtual motion scene, mastering standardized motion knowledge content for standardized motion guidance. In the human-computer interaction stage, REVERIE Coach fine-tunes itself based on virtual motion interaction data fed back by the target object to develop personalized motion techniques suitable for each target object. After two stages of deep reinforcement learning training, REVERIE Coach can provide standardized and personalized motion guidance to the target object.

[0090] In this framework, the server uses virtual environment information as the state and motion guidance prompts as actions to establish a Reinforcement Learning (RL) framework to maximize long-term benefits (such as user retention). The RL framework comprises two key components: the state of the environment and the interaction between the environment and the decision-maker. In a changing environment, at a given moment, the environment is in a certain state, and the decision-maker needs to make a decision based on this state, taking a certain action. Each action changes the environment, leading to a new state, which in turn affects the decision-maker's action at that new moment. Within the RL framework, different states and different actions generate rewards (which can be negative). Through the continuous interaction and development between the environment and the decision-maker, RL aims to find a policy—how to determine actions for different states—that maximizes the long-term cumulative reward. In motion guidance, the virtual environment information is the state influencing action guidance decisions, while the output motion guidance prompts (actions) affect subsequent states. The policy is how to provide motion guidance prompts based on the virtual environment information.

[0091] In one embodiment, as shown in Figure 5, a motion intervention method based on virtual reality and deep reinforcement learning technology is provided. Taking the application of this method to a computer device as an example, the method includes the following steps:

[0092] Step S502: Obtain the physical property data corresponding to the real-world object, and construct a virtual model corresponding to the real-world object based on the physical property data. Step S504: Assign target physical attributes and target model textures to the virtual model corresponding to the real-world object, obtaining the target virtual model corresponding to the real-world object. Step S506: Obtain the parametric static modeling result based on the target virtual model corresponding to the real-world object. Step S508: Obtain the dynamic equations of the real-world object when it is in motion.

[0093] Step S510: Parametrically model the dynamic equations to obtain parametric dynamic modeling results. Step S512: Construct a motion knowledge template based on the parametric static and dynamic modeling results. Step S514: Train the motion guidance model to be trained using reinforcement learning based on the motion knowledge template to obtain a pre-trained general motion guidance model. Step S516: Obtain the current virtual environment information of the virtual motion scene and input it into the pre-trained general motion guidance model. Step S518: Obtain the virtual environment information after the virtual motion scene responds to virtual motion interaction data to obtain new virtual environment information, and obtain the reward data obtained after the pre-trained general motion guidance model outputs motion guidance prompts. Step S520: Update the policy function of the pre-trained general motion guidance model based on the new virtual environment information and reward data until the policy function converges to obtain a personalized motion guidance model. It should be noted that the specific limitations of the above steps can be found in the specific limitations of a motion intervention method based on virtual reality and deep reinforcement learning technology described above.

[0094] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0095] Based on the same inventive concept, this application also provides a motion intervention device based on virtual reality and deep reinforcement learning technology for implementing the motion intervention method based on virtual reality and deep reinforcement learning technology described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of the motion intervention device based on virtual reality and deep reinforcement learning technology provided below can be found in the limitations of the motion intervention method based on virtual reality and deep reinforcement learning technology described above, and will not be repeated here.

[0096] In one exemplary embodiment, as shown in FIG6, a motion intervention device based on virtual reality and deep reinforcement learning technology is provided, comprising: a first acquisition module 610, a training module 620, a second acquisition module 630, and a fine-tuning module 640, wherein:

[0097] The first acquisition module 610 is used to acquire a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content.

[0098] The training module 620 is used to perform reinforcement learning training on the motion guidance model to be trained based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device.

[0099] The second acquisition module 630 is used to acquire virtual motion interaction data of the target object in the virtual motion scene; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene.

[0100] The fine-tuning module 640 is used to perform reinforcement learning-based fine-tuning training on the pre-trained general motion guidance model based on the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the motion state of the target object.

[0101] In one embodiment, the first acquisition module 610 is specifically used to acquire the parametric static modeling result of the virtual object associated with the virtual motion scene; the parametric static modeling result is obtained by modeling based on the static characteristics of the real object corresponding to the virtual object when it is in a static state; acquire the parametric dynamic modeling result of the virtual object associated with the virtual motion scene; the parametric dynamic modeling result is obtained by modeling based on the dynamic characteristics of the real object when it is in a moving state; and construct the motion knowledge template based on the parametric static modeling result and the parametric dynamic modeling result.

[0102] In one embodiment, the first acquisition module 610 is specifically used to acquire physical characteristic data corresponding to the real object, construct a virtual model corresponding to the real object based on the physical characteristic data, assign target physical attributes and target model textures to the virtual model corresponding to the real object, and obtain a target virtual model corresponding to the real object; the target physical attributes are physical attributes that match the real object; the target model textures are model textures that match the real object; and obtain the parametric static modeling result based on the target virtual model corresponding to the real object.

[0103] In one embodiment, the first acquisition module 610 is specifically used to acquire the dynamic equation of the real object when it is in motion; the dynamic equation is used to characterize the dynamic characteristics of the real object when it is in motion; and the dynamic equation is parametrically modeled to obtain the parametric dynamic modeling result.

[0104] In one embodiment, the second acquisition module 630 is specifically used to acquire the current virtual environment information of the virtual motion scene and input the current virtual environment information into the pre-trained general motion guidance model; the policy function of the pre-trained general motion guidance model is used to generate motion guidance action prompts in response to the current virtual environment information; the motion guidance action prompts are used to instruct the target object to perform corresponding actions in the virtual motion scene to obtain the virtual motion interaction data.

[0105] In one embodiment, the fine-tuning module 640 is specifically used to acquire virtual environment information after the virtual motion scene responds to the virtual motion interaction data, to obtain new virtual environment information, and to acquire reward data obtained after the pre-trained general motion guidance model outputs the motion guidance action prompt; based on the new virtual environment information and the reward data, to update the policy function of the pre-trained general motion guidance model until the policy function converges, to obtain the personalized motion guidance model.

[0106] The various modules in the aforementioned motion intervention device based on virtual reality and deep reinforcement learning technologies can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0107] In an exemplary embodiment, a computer device is provided, which can be a motion intervention system based on virtual reality and deep reinforcement learning technology. Its internal structure is shown in Figure 7. The computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system, computer-readable instructions, and a database. The internal memory provides an environment for the operation of the operating system and computer-readable instructions stored in the non-volatile or volatile storage media. The database stores model parameter data. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer-readable instructions are executed by the processor, a motion intervention method based on virtual reality and deep reinforcement learning technology is implemented.

[0108] Those skilled in the art will understand that the structure shown in Figure 7 is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0109] A motion intervention system based on virtual reality and deep reinforcement learning technology includes a memory and one or more processors. The memory stores computer-readable instructions, which, when executed by the processors, cause the one or more processors to perform the following steps:

[0110] Obtain a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content;

[0111] The motion guidance model to be trained is subjected to reinforcement learning training based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device;

[0112] Acquire virtual motion interaction data of the target object in the virtual motion scene; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene; and

[0113] The pre-trained general motion guidance model is fine-tuned based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the target object's motion state.

[0114] One or more computer-readable storage media storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the following steps:

[0115] Obtain a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content;

[0116] The motion guidance model to be trained is subjected to reinforcement learning training based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device;

[0117] Acquire virtual motion interaction data of the target object in the virtual motion scene; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene; and

[0118] The pre-trained general motion guidance model is fine-tuned based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the target object's motion state.

[0119] The computer-readable storage medium may be non-volatile or volatile.

[0120] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0121] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware through computer-readable instructions. These computer-readable instructions can be stored in a computer-readable storage medium, and when executed, they can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0122] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0123] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A motion intervention method based on virtual reality and deep reinforcement learning technology, comprising: Obtain a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content; The motion guidance model to be trained is subjected to reinforcement learning training based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device; The virtual motion interaction data of the target object in the virtual motion scene is obtained; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene. and The pre-trained general motion guidance model is fine-tuned based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the target object's motion state.

2. The method according to claim 1, wherein, The acquisition of motion knowledge templates that match virtual motion scenes associated with virtual reality devices includes: Obtain the parametric static modeling results of the virtual objects associated with the virtual motion scene; the parametric static modeling results are obtained by modeling based on the static characteristics of the real objects corresponding to the virtual objects when they are in a static state; Obtain the parametric dynamic modeling results of the virtual objects associated with the virtual motion scene; the parametric dynamic modeling results are obtained by modeling the dynamic characteristics of the real objects when they are in motion; and Based on the parametric static modeling results and the parametric dynamic modeling results, the motion knowledge template is constructed.

3. The method according to claim 2, wherein, The step of obtaining the parametric static modeling results of the virtual objects associated with the virtual motion scene includes: Obtain the physical characteristic data corresponding to the real object, and construct a virtual model corresponding to the real object based on the physical characteristic data; Assigning target physical attributes and target model textures to the virtual model corresponding to the real-world object yields the target virtual model corresponding to the real-world object; the target physical attributes are physical attributes that match the real-world object; the target model texture is a model texture that matches the real-world object; and The parametric static modeling result is obtained based on the target virtual model corresponding to the real object.

4. The method according to claim 2, wherein, The step of obtaining the parametric dynamic modeling results of the virtual objects associated with the virtual motion scene includes: Obtain the dynamic equations of the real-world object when it is in motion; the dynamic equations characterize the dynamic properties of the real-world object when it is in motion; and The dynamic equations are parametrically modeled to obtain the parametric dynamic modeling results.

5. The method according to claim 1, wherein, The step of acquiring the virtual motion interaction data of the target object in the virtual motion scene includes: The current virtual environment information of the virtual motion scene is obtained, and the current virtual environment information is input into the pre-trained general motion guidance model; the policy function of the pre-trained general motion guidance model is used to generate motion guidance action prompts in response to the current virtual environment information; the motion guidance action prompts are used to instruct the target object to perform corresponding actions in the virtual motion scene to obtain the virtual motion interaction data.

6. The method according to claim 5, wherein, The step of fine-tuning the pre-trained general motion guidance model based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object includes: The system acquires virtual environment information after the virtual motion scene responds to the virtual motion interaction data, obtains new virtual environment information, and acquires reward data obtained after the pre-trained general motion guidance model outputs the motion guidance action prompt; and Based on the new virtual environment information and the reward data, the policy function of the pre-trained general motion guidance model is updated until the policy function converges, thus obtaining the personalized motion guidance model.

7. A motion intervention device based on virtual reality and deep reinforcement learning technology, comprising: The first acquisition module is used to acquire motion knowledge templates that match the virtual motion scenes associated with the virtual reality device; The virtual sports scene is obtained by digital twin modeling based on the real sports scene; the sports knowledge template includes standardized sports knowledge content; The training module is used to perform reinforcement learning training on the motion guidance model to be trained based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device. The second acquisition module is used to acquire virtual motion interaction data of the target object in the virtual motion scene; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene; and The fine-tuning module is used to perform reinforcement learning-based fine-tuning training on the pre-trained general motion guidance model based on the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the motion state of the target object.

8. A motion intervention system based on virtual reality and deep reinforcement learning technology, comprising a memory and one or more processors, wherein the memory stores computer-readable instructions, and when executed by the one or more processors, the one or more processors cause the one or more processors to perform the following steps: Obtain a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content; The motion guidance model to be trained is subjected to reinforcement learning training based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device; The virtual motion interaction data of the target object in the virtual motion scene is obtained; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene. and The pre-trained general motion guidance model is fine-tuned based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the target object's motion state.

9. The system according to claim 8, wherein, When the processor executes the computer-readable instructions, it also performs the following steps: Obtain the parametric static modeling results of the virtual objects associated with the virtual motion scene; the parametric static modeling results are obtained by modeling based on the static characteristics of the real objects corresponding to the virtual objects when they are in a static state; Obtain the parametric dynamic modeling results of the virtual objects associated with the virtual motion scene; the parametric dynamic modeling results are obtained by modeling the dynamic characteristics of the real objects when they are in motion. and Based on the parametric static modeling results and the parametric dynamic modeling results, the motion knowledge template is constructed.

10. The system according to claim 9, wherein, When the processor executes the computer-readable instructions, it also performs the following steps: Obtain the physical characteristic data corresponding to the real object, and construct a virtual model corresponding to the real object based on the physical characteristic data; Assigning target physical attributes and target model textures to the virtual model corresponding to the real-world object yields the target virtual model corresponding to the real-world object; the target physical attributes are physical attributes that match the real-world object; the target model texture is a model texture that matches the real-world object; and The parametric static modeling result is obtained based on the target virtual model corresponding to the real object.

11. The system according to claim 9, wherein, When the processor executes the computer-readable instructions, it also performs the following steps: Obtain the dynamic equations of the real-world object when it is in motion; the dynamic equations characterize the dynamic properties of the real-world object when it is in motion; and The dynamic equations are parametrically modeled to obtain the parametric dynamic modeling results.

12. The system according to claim 8, wherein, When the processor executes the computer-readable instructions, it also performs the following steps: The current virtual environment information of the virtual motion scene is obtained, and the current virtual environment information is input into the pre-trained general motion guidance model; the policy function of the pre-trained general motion guidance model is used to generate motion guidance action prompts in response to the current virtual environment information; the motion guidance action prompts are used to instruct the target object to perform corresponding actions in the virtual motion scene to obtain the virtual motion interaction data.

13. The system according to claim 12, wherein, When the processor executes the computer-readable instructions, it also performs the following steps: The virtual environment information of the virtual motion scene after responding to the virtual motion interaction data is obtained, and new virtual environment information is obtained, as well as reward data obtained after the pre-trained general motion guidance model outputs the motion guidance action prompt; and Based on the new virtual environment information and the reward data, the policy function of the pre-trained general motion guidance model is updated until the policy function converges, thus obtaining the personalized motion guidance model.

14. One or more computer-readable storage media storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the following steps: Obtain a motion knowledge template that matches a virtual motion scene associated with a virtual reality device; the virtual motion scene is obtained by digital twin modeling based on a real motion scene; the motion knowledge template includes standardized motion knowledge content; The motion guidance model to be trained is subjected to reinforcement learning training based on the motion knowledge template to obtain a pre-trained general motion guidance model; the pre-trained general motion guidance model is used to provide general standardized motion guidance prompts to target objects wearing the virtual reality device; The virtual motion interaction data of the target object in the virtual motion scene is obtained; the virtual motion interaction data is motion interaction data obtained based on the actions performed by the target object in the virtual motion scene. and The pre-trained general motion guidance model is fine-tuned based on reinforcement learning using the virtual motion interaction data to obtain a personalized motion guidance model for the target object; the personalized motion guidance model is used to provide personalized motion guidance prompts to the target object based on the target object's motion state.

15. The storage medium according to claim 14, wherein, When the computer-readable instructions are executed by the processor, the following steps are also performed: Obtain the parametric static modeling results of the virtual objects associated with the virtual motion scene; the parametric static modeling results are obtained by modeling based on the static characteristics of the real objects corresponding to the virtual objects when they are in a static state; Obtain the parametric dynamic modeling results of the virtual objects associated with the virtual motion scene; the parametric dynamic modeling results are obtained by modeling the dynamic characteristics of the real objects when they are in motion. and Based on the parametric static modeling results and the parametric dynamic modeling results, the motion knowledge template is constructed.

16. The storage medium according to claim 15, wherein, When the computer-readable instructions are executed by the processor, the following steps are also performed: Obtain the physical characteristic data corresponding to the real object, and construct a virtual model corresponding to the real object based on the physical characteristic data; Assigning target physical attributes and target model textures to the virtual model corresponding to the real-world object yields the target virtual model corresponding to the real-world object; the target physical attributes are physical attributes that match the real-world object; the target model texture is a model texture that matches the real-world object; and The parametric static modeling result is obtained based on the target virtual model corresponding to the real object.

17. The storage medium according to claim 15, wherein, When the computer-readable instructions are executed by the processor, the following steps are also performed: Obtain the dynamic equations of the real-world object when it is in motion; the dynamic equations characterize the dynamic properties of the real-world object when it is in motion; and The dynamic equations are parametrically modeled to obtain the parametric dynamic modeling results.

18. The storage medium according to claim 14, wherein, When the computer-readable instructions are executed by the processor, the following steps are also performed: The current virtual environment information of the virtual motion scene is obtained, and the current virtual environment information is input into the pre-trained general motion guidance model; the policy function of the pre-trained general motion guidance model is used to generate motion guidance action prompts in response to the current virtual environment information; the motion guidance action prompts are used to instruct the target object to perform corresponding actions in the virtual motion scene to obtain the virtual motion interaction data.

19. The storage medium according to claim 18, wherein, When the computer-readable instructions are executed by the processor, the following steps are also performed: The virtual environment information of the virtual motion scene after responding to the virtual motion interaction data is obtained, and new virtual environment information is obtained, as well as reward data obtained after the pre-trained general motion guidance model outputs the motion guidance action prompt; and Based on the new virtual environment information and the reward data, the policy function of the pre-trained general motion guidance model is updated until the policy function converges, thus obtaining the personalized motion guidance model.

Images