Virtual reality rehabilitation training self-adaptive regulation and control method and system based on heart flow double-mapping model

Abstract

The invention provides a virtual reality rehabilitation training self-adaptive regulation and control method and system based on a heart flow double-mapping model, and the method comprises the steps: constructing the heart flow double-mapping model in advance, which comprises a patient data and heart flow state input mapping model and a task parameter and heart flow state output mapping model; collecting multi-dimensional data of the patient in virtual reality rehabilitation training in real time; the current cardiac flow state is calculated through the patient data and cardiac flow state input mapping model, if the current cardiac flow state does not reach the target cardiac flow state, the optimal task parameter combination is solved through the task parameter and cardiac flow state output mapping model under the medical rule constraint condition, virtual reality training task parameters are dynamically adjusted, and the target cardiac flow state is obtained. The heart flow state of the patient reaches the preset target. According to the method, through multi-dimensional data fusion and cardiac flow double-mapping model collaborative optimization, a high cardiac flow state of a patient can be accurately induced and maintained, and the participation degree and the curative effect of rehabilitation training are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical technology, and in particular to a method and system for adaptively controlling virtual reality rehabilitation training based on a heart flow dual mapping model. Background Art

[0002] In recent years, virtual reality (VR) technology has demonstrated tremendous potential in the field of rehabilitation medicine. Its immersive environment can effectively enhance patient engagement and rehabilitation outcomes. However, rehabilitation training typically requires long-term commitment, and patients can easily experience fatigue from repetitive training, leading to decreased compliance. Psychological research has shown that when individuals are in a "flow" state, they exhibit a high degree of focus, a sense of control, and intrinsic motivation. This optimal experience can significantly improve training outcomes and patient adherence. Therefore, inducing and maintaining a patient's flow state during VR rehabilitation training has become a key issue in improving rehabilitation efficacy.

[0003] Currently, some research has attempted to apply flow theory to VR rehabilitation. One existing approach involves collecting the user's EEG signals, analyzing the relationship between specific frequency band characteristics and flow states, and adjusting the rhythmic stimulation of the virtual scene accordingly. Another existing approach utilizes user performance data, such as movement accuracy or reaction time, to establish a correlation model between task difficulty and flow experience, thereby enabling dynamic adjustment of task difficulty. These methods have improved the user training experience to a certain extent.

[0004] However, existing technologies still have obvious shortcomings. First, existing technologies usually rely on only a single type of data input, such as using only physiological signals or only analyzing behavioral performance, which cannot fully reflect the patient's comprehensive state. Secondly, the control models are mostly one-way mapping relationships, which fail to fully consider the synergy between patient characteristics and task parameters. Furthermore, existing solutions lack consideration of the special needs of rehabilitation scenarios, such as the balance between medical safety constraints and personalized rehabilitation goals. These limitations make it difficult for existing methods to achieve accurate and stable flow state control in actual rehabilitation applications. Summary of the Invention

[0005] In view of this, the embodiments of the present invention provide a method and system for adaptive control of virtual reality rehabilitation training based on the flow dual mapping model to eliminate or improve one or more defects existing in the existing technology, and solve the problems of single data dimension, simple control logic, and insufficient rehabilitation adaptability existing in the existing technology.

[0006] In one aspect, the present invention provides a method for adaptively controlling virtual reality rehabilitation training based on a flow dual-map model, the method comprising the following steps: Real-time collection of multi-dimensional data of patients during virtual reality rehabilitation training tasks; Obtaining a pre-built dual-flow mapping model, the dual-flow mapping model comprising a patient data-flow state input mapping model and a task parameter-flow state output mapping model; wherein the patient data-flow state input mapping model is used to characterize the nonlinear mapping relationship between the patient's multidimensional data and the flow state, the patient's multidimensional data comprising at least basic patient information, clinical assessment data, and physiological signals, behavioral performance data, and subjective experience data in each virtual reality rehabilitation training task; and the task parameter-flow state output mapping model is provided with medical rule constraints and is used to characterize the mapping relationship between the task parameter combination of virtual reality rehabilitation training and the flow state; In one control, the multidimensional data is input into the patient data and flow state input mapping model to calculate the current flow state; If the current flow state does not reach the preset target flow state, the task parameters and the flow state output mapping model are used to solve the optimal parameter combination that meets the medical rule constraint conditions; the task parameters in the virtual reality rehabilitation training task are adjusted according to the optimal parameter combination; and the adjustment is repeated until the current flow state reaches the target flow state; If the current flow state reaches the preset target flow state, the current flow state is maintained to continue the virtual reality rehabilitation training task.

[0007] In some embodiments of the present invention, the method further includes constructing a mapping model between the patient data and the flow state input, wherein the mapping model between the patient data and the flow state input is obtained based on the following steps: Collecting and preprocessing multidimensional data of patients to construct a training set; labeling the training set with a label of a true flow state; Constructing an initial model, wherein the initial model adopts a multi-output machine learning model; the initial model takes the patient data in the training set as input and outputs a predicted flow state; The initial model is trained using the training set, and the initial model is optimized with the goal of minimizing the error between the predicted flow state and the actual flow state, ultimately obtaining a mapping model between the patient data and the flow state input.

[0008] In some embodiments of the present invention, the actual flow state is quantified based on the following steps: A preset flow state scale is used to perform multi-dimensional scoring of the flow experience in each virtual reality rehabilitation training task to form a flow state vector as the true flow state; wherein the flow state scale includes the following dimensions: challenging skill balance, behavioral awareness integration, clear goals, clear feedback, concentration, sense of control, weakening of self-awareness, distorted time perception, and intrinsic motivation perception.

[0009] In some embodiments of the present invention, the method further includes constructing a mapping model between the task parameters and the flow state output, wherein the mapping model between the task parameters and the flow state output is obtained based on the following steps: Constructing a task parameter system, and using a preset sampling method for each task parameter in the task parameter system to generate multiple task parameter combinations; Constructing a state space using the task parameter combination, the patient's real-time state data, and the flow state, constructing an action space using the adjustment amount of the task parameters, and constructing a reward function based on the flow state improvement amount and the medical rule constraint condition; An Actor-Critic network is used to construct a reinforcement learning model. The state vector of the state space is input into the Actor network to generate an action probability distribution, and the state vector is input into the Critic network to generate a state value estimate. A mean square error loss is constructed based on the state value estimate. With the goal of minimizing the mean square error loss, the parameters of the Actor network and the Critic network are updated, and finally a mapping model between the task parameters and the flow state output is obtained.

[0010] In some embodiments of the present invention, the task parameter system includes a plurality of adjustable task parameters, and the task parameters include at least avatar parameters, narrative parameters, task mechanism parameters, aesthetic parameters, music parameters and multimodal stimulation parameters.

[0011] In some embodiments of the present invention, the medical rule constraints include but are not limited to motor function constraints, training dose constraints and safety threshold constraints; wherein the motor function constraints are used to limit the control range of the task parameters to not exceed the tolerable threshold of the patient's current physiological function; the training dose constraints are used to control the intensity of the virtual reality rehabilitation training to comply with clinical treatment standards; and the safety threshold constraints are used to ensure that the intensity of sensory stimulation does not exceed the patient's individualized tolerance limit.

[0012] In some embodiments of the present invention, using the task parameters and flow state output mapping model to solve the optimal parameter combination that meets the medical rule constraints includes: Calculating a difference between the current flow state and the target flow state; Based on the back propagation gradient of the task parameters and the flow state output mapping model, the contribution of each task parameter to the improvement of the flow state is calculated, and an adjustment priority sequence is generated; Filtering the first preset number of task parameters with the greatest contribution as optimization variables according to the adjustment priority sequence; In a feasible solution space that satisfies the constraints of the medical rule, an iterative optimization algorithm is used to optimize the optimization variables with the goal of minimizing the difference to obtain the optimal parameter combination; Among them, the iterative optimization algorithm performs the following operations in one iteration: generates a candidate parameter combination according to the current gradient direction; inputs the candidate parameter combination that passes the medical rule constraint condition into the task parameter and flow state output mapping model to obtain a predicted flow state; calculates a new difference between the predicted flow state and the target flow state; if the new difference is less than the original difference, updates the optimal parameter combination; iterates until a preset number of iterations is reached or the new difference converges.

[0013] In some embodiments of the present invention, regulating the task parameters in the virtual reality rehabilitation training task according to the optimal parameter combination includes collaboratively adjusting at least two types of task parameters, wherein the adjustment method of each task parameter includes: Adjusting the avatar parameters, including adjusting one or more of the avatar's limb proportions, range of motion, and action guide markers; Adjusting the narrative parameters, including adjusting one or more of task progress, task difficulty, and task environment story elements; Adjusting the task mechanism parameters, including adjusting one or more of virtual obstacle attributes, task time limit, and interaction mechanism; Adjusting the aesthetic parameters, including adjusting one or more of the color contrast of the virtual scene and the visual prominence of the interface elements; Adjusting the music parameters, including adjusting one or more of the background music type, rhythm, and prompt audio frequency; Adjusting the multimodal stimulation parameters includes adjusting one or more of multisensory feedback intensity and environmental interference elements.

[0014] On the other hand, the present invention also provides a virtual reality rehabilitation training adaptive control system based on the flow dual-map model, which, when executed, implements the steps of any of the methods mentioned above, comprising: Model building module, used to build and maintain the flow dual-map model; A closed-loop control module is used to control the parameters of the virtual reality rehabilitation training task in real time based on the heart flow dual mapping model so that the patient's heart flow state reaches a preset target heart flow state; The virtual reality training module includes a task scene construction unit and a human-computer interaction interface, which is used to execute and present virtual reality rehabilitation training tasks.

[0015] On the other hand, the present invention further provides a computer-readable storage medium having a computer program / instruction stored thereon, which implements the steps of any of the methods mentioned above when executed by a processor.

[0016] The present invention provides a method and system for adaptively controlling virtual reality rehabilitation training based on a flow dual-mapping model. The method comprises: a pre-constructed flow dual-mapping model comprising a patient data-flow state input mapping model and a task parameter-flow state output mapping model; real-time acquisition of multidimensional data from patients during virtual reality rehabilitation training; calculation of the current flow state using the patient data and the flow state input mapping model; and, if the current flow state does not reach the target flow state, the task parameter-flow state output mapping model is used to determine the optimal task parameter combination under medical rule constraints. The virtual reality training task parameters are then dynamically adjusted until the patient's flow state reaches the preset target. Through the synergistic optimization of multi-dimensional data fusion and the flow dual-mapping model, the present invention accurately induces and maintains a high flow state in patients, significantly improving the engagement and efficacy of rehabilitation training. Furthermore, the method supports flexible configuration of data collection methods (such as wearable devices or VR interactive data), reducing reliance on a single device. The method is suitable for a variety of scenarios, including motor function recovery and cognitive training, and offers both personalized adaptability and medical safety.

[0017] Additional advantages, objects, and features of the present invention will be set forth in part in the following description and will become apparent to those skilled in the art upon examination of the following or may be learned from practice of the present invention. The objects and other advantages of the present invention may be realized and obtained by the structures particularly pointed out in the description and drawings.

[0018] Those skilled in the art will understand that the purposes and advantages that can be achieved by the present invention are not limited to the above specific descriptions, and the above and other purposes that can be achieved by the present invention will be more clearly understood based on the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of this application, and do not constitute a limitation of the present invention. In the drawings: Figure 1 Schematic diagram of the steps of a method for adaptively controlling virtual reality rehabilitation training based on a flow dual-map model in one embodiment of the present invention.

[0020] Figure 2 Schematic diagram of the construction of the dual-map model of flow in one embodiment of the present invention.

[0021] Figure 3 Flowchart of a method for adaptively controlling virtual reality rehabilitation training according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0023] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, the accompanying drawings only show structures and / or processing steps closely related to the solutions according to the present invention, while other details that are not closely related to the present invention are omitted.

[0024] It should be emphasized that the term "include / comprises" when used herein refers to the existence of features, elements, steps or components, but does not exclude the existence or addition of one or more other features, elements, steps or components.

[0025] It should also be noted that, unless otherwise specified, the term "connection" herein may refer not only to a direct connection but also to an indirect connection involving an intermediate.

[0026] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0027] It should be emphasized here that the step marks mentioned below do not limit the order of the steps, but it should be understood that the steps can be executed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be executed simultaneously.

[0028] In order to solve the problems of single data dimension, simple control logic and insufficient rehabilitation adaptability in the existing technology, the present invention provides a virtual reality rehabilitation training adaptive control method based on the flow dual mapping model, such as Figure 1 As shown, the method includes the following steps S101 to S105: Step S101: Real-time collection of multi-dimensional data of the patient during the virtual reality rehabilitation training task.

[0029] Step S102: Obtain a pre-built dual-flow mapping model. The dual-flow mapping model includes a patient data-to-flow state input mapping model and a task parameter-to-flow state output mapping model. The patient data-to-flow state input mapping model is used to characterize the nonlinear mapping relationship between a patient's multidimensional data and flow state. The multidimensional patient data includes at least basic patient information, clinical assessment data, and physiological signals, behavioral performance data, and subjective experience data during each virtual reality rehabilitation training task. The task parameter-to-flow state output mapping model, which is constrained by medical rules, characterizes the mapping relationship between task parameter combinations and flow state during virtual reality rehabilitation training.

[0030] Step S103: In one control, multi-dimensional data is input into the patient data and flow state input mapping model to calculate the current flow state.

[0031] Step S104: If the current flow state does not reach the preset target flow state, the task parameters and flow state output mapping model are used to solve the optimal parameter combination that meets the medical rule constraints; the task parameters in the virtual reality rehabilitation training task are adjusted according to the optimal parameter combination; and the adjustment is repeated until the current flow state reaches the target flow state.

[0032] Step S105: If the current flow state reaches the preset target flow state, the current flow state is maintained and the virtual reality rehabilitation training task is continued.

[0033] The pre-built flow dual map model is required in the adaptive control method of virtual reality rehabilitation training based on the flow dual map model. Therefore, the flow dual map model is first explained for easier understanding.

[0034] like Figure 2 As shown in Figure 2, this is a schematic diagram of the construction of the flow dual-mapping model.

[0035] The dual-flow mapping model consists of two models: a patient data-to-flow state input mapping model and a task parameter-to-flow state output mapping model. The patient data-to-flow state input mapping model characterizes the nonlinear mapping relationship between multidimensional patient data and flow state, while the task parameter-to-flow state output mapping model characterizes the mapping relationship between task parameter combinations and flow state during virtual reality rehabilitation training.

[0036] Among them, the flow state is a comprehensive experience that includes nine dimensions: challenge-skill balance, behavioral awareness integration, clear goals, clear feedback, concentration, sense of control, weakened self-awareness, distorted time perception, and intrinsic motivation perception. It can be obtained based on the scoring results of flow state scales such as the Flow State Scale-2.

[0037] It should be noted that regulating a single task parameter is unlikely to fully address the multidimensional optimization requirements of the flow state. For example, the "challenge-skill balance" dimension primarily relies on the adjustment of task mechanism parameters (such as obstacle difficulty); the "concentration" dimension is more susceptible to the influence of music parameters (such as music rhythm); and the "sense of control" dimension needs to be enhanced through avatar parameters (such as action-guiding markers). Therefore, in the present invention, a mapping model between task parameters and flow state output is constructed. By mapping multidimensional task parameters to multidimensional flow states, the multi-task parameters are jointly adjusted to synchronously optimize interrelated flow dimensions. For example, increasing obstacle height while adding auxiliary markers enhances both the challenge and the sense of control. At the same time, the negative interference between dimensions caused by traditional single-parameter regulation is avoided, such as the problem that simply increasing task difficulty may lead to a decrease in sense of control.

[0038] Collect multi-dimensional data of patients, including basic information of patients, clinical assessment data, as well as physiological signals, behavioral performance data and subjective experience data in various virtual reality rehabilitation training tasks. Specifically: Basic information includes: patient's age, gender, medical history and other basic information.

[0039] Clinical assessment data refers to the patient's clinical function assessment data, including motor function assessment data, daily activities function assessment data, cognitive function assessment data, etc. Motor function assessment can be assessed using scales such as the Fugl-Meyer Scale and the Berg Balance Scale; cognitive function assessment can be assessed using methods such as the MoCA (Montreal Cognitive Assessment) and the MMSE (Mini-Mental State Examination).

[0040] Physiological signals include: heart rate, skin conductivity, respiratory rate, facial expression, blood oxygen saturation, EEG signals, eye movement data, etc. collected in real time during virtual reality rehabilitation training tasks. Physiological signals are potentially important physiological characteristics that can reflect the flow state.

[0041] Behavioral performance data includes real-time data collected during VR rehabilitation training, including the patient's movement completion (e.g., joint mobility, movement standardization, etc.), reaction time, and task completion duration. This data can reflect the patient's skill performance and adaptation to the task during training.

[0042] Subjective experience data includes the patient's subjective rating of fatigue, focus, and sense of control during the current training task. For example, a flow experience questionnaire can be presented to the patient through an interactive interface within a virtual reality scene for rating. The flow experience questionnaire can use the Flow State Scale-2 to obtain multi-dimensional subjective flow experience data.

[0043] The collected multidimensional patient data are cleaned and normalized to remove noise and outliers to ensure the accuracy and usability of the data.

[0044] Based on the patient's multidimensional data after the above processing, a patient data and flow state input mapping model and a task parameter and flow state output mapping model are constructed.

[0045] In some embodiments, a method for constructing (training) a model mapping patient data to flow state input includes: Constructing a training set involves collecting and preprocessing multidimensional patient data. For example, timestamp synchronization technology is used to unify data with different sampling frequencies to the same time base (e.g., 100 Hz for physiological signals and 30 Hz for behavioral data), achieve alignment through interpolation or downsampling, and synchronously control errors.

[0046] Each sample in the training set is labeled with a true flow state. A pre-defined flow state scale is used to assess the multi-dimensional flow experience during each VR rehabilitation training task, generating a flow state vector as the true flow state. For example, the pre-defined flow state scale is the Flow State Scale-2, which includes the following nine dimensions: challenging skill balance, behavioral awareness integration, clear goals, clear feedback, focus, sense of control, weakened self-awareness, distorted time perception, and perceived intrinsic motivation.

[0047] Build an initial model using a multi-output machine learning model, such as a tensor regression network or graph neural network. This model takes the patient's multidimensional data as input and outputs the corresponding predicted flow state.

[0048] The initial model is trained using the training set to construct a weighted mean square loss between the predicted flow state and the actual flow state. The initial model is optimized with the goal of minimizing this loss, and finally a mapping model between patient data and flow state input is obtained.

[0049] It's important to note that in the model mapping patient data to flow state input, the patient's physiological signals (heart rate, EEG, etc.), behavioral performance data (movement accuracy, reaction time, etc.), and subjective feelings (fatigue, interest, etc.) are all multidimensional data, and their associations with flow state are specific. For example, changes in heart rate primarily affect the "sense of control" dimension, while movement accuracy is strongly correlated with the "challenge-skill balance" dimension. Therefore, using only a single-dimensional flow metric cannot accurately map to the patient's multidimensional data. Furthermore, some patient data types may only be strongly correlated with certain dimensions of flow state and have no correlation with overall flow. For example, EEG signals may only be strongly correlated with the "concentration" dimension and have no correlation with the overall flow score. However, this does not mean that EEG signals are unrelated to flow state; rather, it indicates that EEG signals can reflect certain dimensional characteristics of flow state. Therefore, simply mapping the overall flow score to EEG signals loses the mapping relationship between EEG signals and the "concentration" dimension, making it impossible to accurately map to the patient's multidimensional data.

[0050] Similarly, based on the collected multi-dimensional data of patients, especially the physiological signals, behavioral performance data and subjective experience data in various virtual reality rehabilitation training tasks, a back-propagation neural network or reinforcement learning model is used to construct a mapping model between task parameters and flow state output.

[0051] In some embodiments, a method for constructing (training) a task parameter-flow state output mapping model includes: Construct a task parameter system, including multiple adjustable task parameters: avatar parameters, narrative parameters, task mechanism parameters, aesthetic parameters, music parameters, and multimodal stimulation parameters. Avatar parameters include body proportions and range of motion, narrative parameters include task difficulty and story progression, task mechanism parameters include obstacle attributes and task time limit, aesthetic parameters include color contrast, music parameters include rhythm and volume, and multimodal stimulation parameters include tactile feedback intensity.

[0052] Intelligent sampling methods such as Latin hypercube sampling method are used for each task parameter to generate multiple task parameter combinations.

[0053] A reinforcement learning model was constructed using an actor-critic network. The state space was constructed using the current task parameters, the patient's real-time status data (including physiological signals and behavioral performance data), and the current flow state. The action space was constructed using adjustments to the task parameters. A reward function was constructed based on the improvement in flow and the constraints of medical rules. The improvement in flow was calculated as a weighted sum of nine dimensions.

[0054] The state vector of the state space is input into the Actor network to generate an action probability distribution; the state vector is then input into the Critic network to generate a state value estimate. Based on the state value estimate, a mean squared error loss is constructed. With the goal of minimizing the mean squared error loss, the parameters of the Actor and Critic networks are updated. This allows the model to learn the mapping relationship between task parameter combinations and flow states, outputting the optimal task parameter adjustment plan, and ultimately obtaining a mapping model between task parameters and flow states.

[0055] Different patients, due to differences in basic information such as age, disease type, and gender, have varying tolerance and sensitivity to task parameters. For example, elderly patients have a lower tolerance for complex tasks than younger patients, and elderly patients are less sensitive to musical rhythm than younger patients. By mapping multidimensional task parameters to flow states, we can match task parameter combinations to the patient's flow shortcomings. For example, when the "clear feedback" dimension is insufficient, the prompt audio frequency and interface feedback intensity can be adjusted simultaneously. Alternatively, for example, the target threshold for the "challenge-skill balance" dimension for early-stage stroke patients is lower than that for patients in recovery, achieving personalized adaptation.

[0056] In some embodiments, medical rule constraints include: Motor function constraints are used to limit the control range of the task parameters to not exceed the patient's current physiological function tolerance threshold, such as the obstacle height adjustment range must be less than or equal to the patient's joint mobility threshold.

[0057] Training dose constraints are used to control the intensity of virtual reality rehabilitation training to comply with clinical treatment standards. For example, the task time limit must comply with clinical treatment guidelines, the single training time shall not exceed 30 minutes, and the total daily training time shall not exceed 1 hour.

[0058] Safety threshold constraints are used to ensure that the intensity of sensory stimulation does not exceed the patient's individual tolerance limit, such as the intensity of tactile feedback does not exceed 50% of the patient's pain threshold.

[0059] In some embodiments, embedding medical rule constraints into a task parameter and flow state output mapping model includes: For motion function constraints, hard constraints can be added by adding a parameter range restriction function to the model output layer.

[0060] For training dose constraints, soft constraints can be implemented by adding penalty terms to the reinforcement learning reward function. For example, to limit the training duration, a time penalty term can be added to the reinforcement learning reward function. When the cumulative training time exceeds the preset duration, the reward value drops sharply.

[0061] For safety threshold constraints, a safety activation function can be added to the last layer of the Actor network to constrain the network structure.

[0062] When solving the optimal parameter combination based on the mapping model between task parameters and flow state output, it is possible to balance the improvement of multiple flow dimensions within the constraints of the above-mentioned medical rules. For example, while limiting the height of obstacles, the "behavioral awareness fusion" dimension can be enhanced through aesthetic parameters to avoid single parameter adjustment exceeding the safety threshold.

[0063] like Figure 3 FIG. 1 is a flow chart of an adaptive control method for virtual reality rehabilitation training based on the aforementioned constructed flow dual-mapping model.

[0064] In step S101, the virtual reality rehabilitation device is activated, and the training task and task parameters are determined based on the patient's information to conduct rehabilitation training. Multi-dimensional data from the patient during training is collected in real time, such as through wearable devices, VR controllers, eye trackers, etc., to collect the patient's physiological signals, behavioral performance data, and subjective experience data.

[0065] In step S102, a pre-built flow dual-map model is obtained.

[0066] In step S103, the multidimensional data of the patient acquired in step S101 is input into the patient data and flow state input mapping model in the flow dual mapping model to obtain the current flow state of the patient.

[0067] In step S104, the patient's current flow state is compared with a preset target flow state to determine whether the task parameters need to be adjusted.

[0068] If the current flow state does not reach the preset target flow state, the task parameters and flow state output mapping model are used to solve the optimal parameter combination that meets the medical rule constraints; the task parameters in the virtual reality rehabilitation training task are adjusted according to the optimal parameter combination.

[0069] In some embodiments, finding the optimal parameter combination that satisfies the medical rule constraints includes the following steps: The current flow state is subtracted from the target flow state to calculate the difference in each dimension. Based on the preset allowable error threshold, the set of dimensions that need to be optimized is selected, which is the set of target dimensions.

[0070] Through the back propagation of the task parameters and the flow state output mapping model, the gradient of each task parameter with respect to the target dimension is calculated, and the gradients are arranged in descending order according to their absolute values ​​to generate an adjustment priority sequence.

[0071] The top-K parameters are selected from the adjustment priority sequence as optimization variables, and the remaining parameters are fixed to their current values. For example, K=3.

[0072] In the feasible solution space limited by the constraints of medical rules, with the goal of minimizing the difference between the current flow state and the target flow state, an iterative optimization algorithm is used to optimize the optimization variables and obtain the optimal parameter combination.

[0073] In some embodiments, the iterative optimization algorithm optimization step includes: Set the maximum number of iterations and the convergence threshold, take the current parameter combination as the optimal solution, and the current difference as the initial difference.

[0074] In each iteration, based on the current parameters and gradient information, a candidate parameter combination is generated along the negative gradient direction. The candidate parameter combination is checked item by item for medical rule constraints. The candidate parameter combination that has passed the check is input into the task parameters and the flow state output mapping model to obtain a predicted flow state. The new difference between the predicted flow state and the target flow state is calculated. If the new difference is less than the original difference, the candidate parameter combination is updated to the current optimal solution. Iterate until the preset maximum number of iterations is reached or the new difference is less than the preset convergence threshold, and output the optimal solution, that is, the optimal parameter combination. Exemplarily, an iteration interval is set, such as performing an iteration every 500ms.

[0075] Based on the optimal parameter combination, the task parameters in virtual reality rehabilitation training tasks are regulated, ultimately achieving precise maintenance of the flow state and improving the participation and efficacy of rehabilitation training.

[0076] In some embodiments, adjusting each task parameter includes: Adjust avatar parameters, including adjusting one or more of the avatar's limb proportions, range of motion, and motion-guiding markers. Examples include adjusting the scaling factor to suit the patient's motor abilities, providing visual guidance of joint range of motion, and dynamically highlighting key muscle groups.

[0077] Adjust narrative parameters, including adjusting one or more of the following: mission progress, mission difficulty, and mission environment story elements. For example, adjust mission level unlocking conditions, the number of subtasks, and the mission completion time limit, and set scene details related to the patient's life.

[0078] Adjust the task mechanism parameters, including adjusting one or more of the virtual obstacle attributes, task time limit, and interaction mechanism. For example, adjust the vertical height of virtual steps and gullies, the countdown threshold for completing the goal, the action response sensitivity, and the content and frequency of error correction prompts.

[0079] Adjust aesthetic parameters, including adjusting the color contrast of the virtual scene and the visual prominence of interface elements. For example, use high-contrast colors to highlight key paths, select weather environments such as rainy, sunny, and snowy days, and adjust the transparency of the UI interface (the visual weight of progress bars and tooltips).

[0080] Adjust music parameters, including background music type, tempo, and prompt audio frequency. For example, add voice guidance, select instrument timbres, adjust the interval between sound effects for key actions, and the synchronization coefficient between background music tempo and action frequency.

[0081] Adjust multimodal stimulation parameters, including adjusting the intensity of multisensory feedback and one or more environmental interference elements. For example, adjust the vibration amplitude of the vibrating vest, the degree of randomization of virtual object movement, release specific scents based on specific scenarios, and adjust the scent concentration.

[0082] In step S105, if the current flow state reaches the preset target flow state, the current flow state is maintained for rehabilitation training without any adjustment.

[0083] Corresponding to the above method, the present invention also provides a virtual reality rehabilitation training adaptive control system based on the heart flow dual mapping model. When the system is executed, the steps of the above method can be implemented. The system includes: Model building module, used to build and maintain the flow dual map model.

[0084] The closed-loop control module is used to control the parameters of the virtual reality rehabilitation training task in real time based on the flow dual-mapping model, so that the patient's flow state reaches the preset target flow state.

[0085] The virtual reality training module includes a task scene construction unit and a human-computer interaction interface, which is used to execute and present virtual reality rehabilitation training tasks.

[0086] The present invention will be further described below with reference to a specific embodiment.

[0087] In this example, basic information such as age and gender, as well as clinical assessment data such as the Fugl-Meyer motor scale, were collected from 100 stroke patients. Physiological signals such as heart rate and respiratory rate, as well as behavioral performance data such as stride length, range of motion, and obstacle avoidance success rate, were simultaneously acquired during different VR walking tasks. Subjective flow scores across nine dimensions, including challenge-skill balance, focus, and sense of control, were collected using the VR-embedded Flow State Scale-2 questionnaire. Based on this collected information, a mapping model between patient data and flow state input was constructed, revealing significant correlations between stride length difference and the "challenge-skill balance" dimension, and between heart rate and the "sense of control" dimension. A mapping model between task parameters and flow state output was also constructed to determine the weighted impact of obstacle height on the "challenge-skill balance" dimension and music rhythm on the "focus" dimension.

[0088] In the adaptive control method, the system sets initial task parameters as an obstacle height of 10 cm and a music tempo of 80 bpm. The currently collected stride difference of 12 cm and heart rate of 95 beats per minute are input into the patient data and the flow state input mapping model, predicting the current flow state to be 4 points in the "Challenge-Skill Balance" dimension and 5 points in the "Sense of Control" dimension (total score of 40 points), which is significantly lower than the preset target flow state (≥ 6 points in each dimension, total score of 60 points). Based on the task parameters and the flow state output mapping model, the system calculates the optimal parameter combination and adjusts the task parameters: reducing the obstacle height to 8 cm to optimize the "Challenge-Skill Balance" dimension, and adjusting the music tempo to 60 bpm to improve the "Concentration" dimension. After the task parameters were adjusted, the patient's stride difference was reduced to 8 cm, the heart rate dropped to 78 beats / minute, and the flow score increased to 66 points (7 points for "challenge-skill balance", 7 points for "sense of control", and ≥6 points for other dimensions). The "data collection-state assessment-parameter adjustment" closed loop was completed every 300ms, allowing the patient to maintain a high flow state to improve compliance with walking training.

[0089] The present invention will be further described below with reference to a specific embodiment.

[0090] In this example, baseline information and clinical data, including Berg Balance Scale scores, were collected from 80 Parkinson's disease patients. Physiological signals such as heart rate and skin conductivity, as well as behavioral performance data such as center of gravity shift amplitude and balance reaction speed, were simultaneously acquired during VR tasks. Subjective scores were also collected using a nine-dimensional flow questionnaire. Based on this collected data, a mapping model between patient data and flow state input was constructed. Significant negative correlations were found between center of gravity shift amplitude and the "sense of control" dimension, and between heart rate variability (HRV) and the "intrinsic motivation" dimension. A mapping model between task parameters and flow state output was also constructed to determine the weighted impact of virtual obstacle movement speed on the "challenge-skill balance" dimension, and the influence of scene ground texture complexity on the "focus" dimension.

[0091] In the adaptive control method, the system initialized the task parameters with a virtual obstacle speed of 1.0 m / s and a ground texture complexity of level 3. The patient's data, including a 4 cm center of gravity deviation and decreased HRV, were then incorporated into the flow state input mapping model. The predicted current flow state was 5 points for "sense of control" and 6 points for "focus" (total score of 70), which was below the target threshold (≥ 7 points for each dimension, total score of 70). Based on the task parameters and the flow state output mapping model, the system calculated the optimal parameter combination and adjusted the task parameters: the obstacle speed was reduced to 0.8 m / s to improve the "sense of control" dimension, and the ground texture was simplified to level 2 to reduce cognitive load and thus improve the "focus" dimension. After the adjustments, the patient's center of gravity deviation was reduced to 2.5 cm, HRV returned to normal, and the flow score increased to 80 points (7 points for "sense of control" and "focus"). Closed-loop control was completed every 400 ms. Through precise adaptation of task parameters, the system improved the balance training effect of Parkinson's patients.

[0092] Corresponding to the above method, the present invention also provides an electronic device, which includes a computer device, the computer device includes a processor and a memory, the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the electronic device implements the steps of the method described above.

[0093] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned method. The computer-readable storage medium may be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium known in the art.

[0094] It should be understood by those skilled in the art that the various exemplary components, systems, and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software, or a combination of both. Whether to implement the system in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention. When implemented in hardware, it may be, for example, an electronic circuit, an application-specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments used to perform the required tasks. The programs or code segments can be stored in a machine-readable medium or transmitted over a transmission medium or communication link via a data signal carried in a carrier wave.

[0095] It should be understood that the present invention is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, a detailed description of known methods is omitted. In the above embodiments, several specific steps are described and illustrated as examples. However, the method of the present invention is not limited to the specific steps described and illustrated. Those skilled in the art may make various changes, modifications, and additions, or change the order of the steps after understanding the spirit of the present invention.

[0096] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or replace features of other embodiments.

[0097] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. A method for adaptive control of virtual reality rehabilitation training based on a dual-map model of flow, characterized in that: The method comprises: Real-time collection of multi-dimensional data of patients during virtual reality rehabilitation training tasks; Obtaining a pre-built dual-flow mapping model, the dual-flow mapping model comprising a patient data-flow state input mapping model and a task parameter-flow state output mapping model; wherein the patient data-flow state input mapping model is used to characterize the nonlinear mapping relationship between the patient's multidimensional data and the flow state, the patient's multidimensional data comprising at least basic patient information, clinical assessment data, and physiological signals, behavioral performance data, and subjective experience data in each virtual reality rehabilitation training task; and the task parameter-flow state output mapping model is provided with medical rule constraints and is used to characterize the mapping relationship between the task parameter combination of virtual reality rehabilitation training and the flow state; In one control, the multidimensional data is input into the patient data and flow state input mapping model to calculate the current flow state; If the current flow state does not reach the preset target flow state, the task parameters and the flow state output mapping model are used to solve the optimal parameter combination that meets the medical rule constraint conditions; the task parameters in the virtual reality rehabilitation training task are adjusted according to the optimal parameter combination; and the adjustment is repeated until the current flow state reaches the target flow state; If the current flow state reaches the preset target flow state, the current flow state is maintained to continue the virtual reality rehabilitation training task.

2. The method according to claim 1, characterized in that The method further includes constructing a mapping model between the patient data and the flow state input, wherein the mapping model between the patient data and the flow state input is obtained based on the following steps: Collecting and preprocessing multidimensional data of patients to construct a training set; labeling the training set with a label of a true flow state; Constructing an initial model, wherein the initial model adopts a multi-output machine learning model; The initial model takes the patient data in the training set as input and outputs a predicted flow state; The initial model is trained using the training set, and the initial model is optimized with the goal of minimizing the error between the predicted flow state and the actual flow state, ultimately obtaining a mapping model between the patient data and the flow state input.

3. The method according to claim 2, characterized in that The true flow state is quantified based on the following steps: A preset flow state scale is used to perform multi-dimensional scoring of the flow experience in each virtual reality rehabilitation training task to form a flow state vector as the true flow state; wherein the flow state scale includes the following dimensions: challenging skill balance, behavioral awareness integration, clear goals, clear feedback, concentration, sense of control, weakening of self-awareness, distorted time perception, and intrinsic motivation perception.

4. The method according to claim 1, wherein The method further includes constructing a mapping model between the task parameters and the flow state output, wherein the mapping model between the task parameters and the flow state output is obtained based on the following steps: Constructing a task parameter system, and using a preset sampling method for each task parameter in the task parameter system to generate multiple task parameter combinations; Constructing a state space using the task parameter combination, the patient's real-time state data, and the flow state, constructing an action space using the adjustment amount of the task parameters, and constructing a reward function based on the flow state improvement amount and the medical rule constraint condition; An Actor-Critic network is used to construct a reinforcement learning model. The state vector of the state space is input into the Actor network to generate an action probability distribution, and the state vector is input into the Critic network to generate a state value estimate. A mean square error loss is constructed based on the state value estimate. With the goal of minimizing the mean square error loss, the parameters of the Actor network and the Critic network are updated, and finally a mapping model between the task parameters and the flow state output is obtained.

5. The method according to claim 4, characterized in that The task parameter system includes a variety of adjustable task parameters, and the task parameters at least include avatar parameters, narrative parameters, task mechanism parameters, aesthetic parameters, music parameters and multimodal stimulation parameters.

6. The method according to claim 1, characterized in that The medical rule constraints include but are not limited to motor function constraints, training dose constraints and safety threshold constraints; wherein the motor function constraints are used to limit the control range of the task parameters to not exceed the tolerable threshold of the patient's current physiological function; the training dose constraints are used to control the intensity of the virtual reality rehabilitation training to comply with clinical treatment standards; and the safety threshold constraints are used to ensure that the intensity of sensory stimulation does not exceed the patient's individualized tolerance limit.

7. The method according to claim 1, characterized in that The task parameters and the flow state output mapping model are used to solve the optimal parameter combination that meets the medical rule constraint conditions, including: Calculating a difference between the current flow state and the target flow state; Based on the back propagation gradient of the task parameters and the flow state output mapping model, the contribution of each task parameter to the improvement of the flow state is calculated, and an adjustment priority sequence is generated; Filtering the first preset number of task parameters with the greatest contribution as optimization variables according to the adjustment priority sequence; In a feasible solution space that satisfies the constraints of the medical rule, an iterative optimization algorithm is used to optimize the optimization variables with the goal of minimizing the difference to obtain the optimal parameter combination; Among them, the iterative optimization algorithm performs the following operations in one iteration: generates a candidate parameter combination according to the current gradient direction; inputs the candidate parameter combination that passes the medical rule constraint condition into the task parameter and flow state output mapping model to obtain a predicted flow state; calculates a new difference between the predicted flow state and the target flow state; if the new difference is less than the original difference, updates the optimal parameter combination; iterates until a preset number of iterations is reached or the new difference converges.

8. The method according to claim 5, characterized in that Regulating the task parameters in the virtual reality rehabilitation training task according to the optimal parameter combination includes collaboratively adjusting at least two types of task parameters, wherein the adjustment method of each task parameter includes: Adjusting the avatar parameters, including adjusting one or more of the avatar's limb proportions, range of motion, and action guide markers; Adjusting the narrative parameters, including adjusting one or more of task progress, task difficulty, and task environment story elements; Adjusting the task mechanism parameters, including adjusting one or more of virtual obstacle attributes, task time limit, and interaction mechanism; Adjusting the aesthetic parameters, including adjusting one or more of the color contrast of the virtual scene and the visual prominence of the interface elements; Adjusting the music parameters, including adjusting one or more of the background music type, rhythm, and prompt audio frequency; Adjusting the multimodal stimulation parameters includes adjusting one or more of multisensory feedback intensity and environmental interference elements.

9. A virtual reality rehabilitation training adaptive control system based on the heart flow dual mapping model, characterized in that: When the system is executed, the steps of the method according to any one of claims 1 to 8 are implemented, and the system comprises: Model building module, used to build and maintain the flow dual-map model; A closed-loop control module is used to control the parameters of the virtual reality rehabilitation training task in real time based on the heart flow dual mapping model so that the patient's heart flow state reaches a preset target heart flow state; The virtual reality training module includes a task scene construction unit and a human-computer interaction interface, which is used to execute and present virtual reality rehabilitation training tasks.

10. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images