Psychological intervention adaptive training method and system driven by virtual reality scene

Abstract

The application discloses a virtual reality scene driven psychological intervention adaptive training method and system, relates to the virtual reality intervention training field, and has the technical scheme as follows: a virtual training scene with configurable parameters is generated according to psychological evaluation data; a second training target vector with logical complementarity with the first training target vector is acquired; second behavior data of a second user is received, and the second behavior data is mapped into virtual role behavior in a virtual training scene corresponding to the first user; first behavior data of the first user in the virtual training scene is collected, and the first behavior data is transmitted to the second user, so that the first behavior data is mapped into virtual role behavior in a virtual training scene corresponding to the second user. The application can not only quickly construct a virtual role, but also enhance the authenticity and reaction efficiency of the virtual role, reduce the required computing resources of the virtual reality scene driving, and improve the skill transfer rate.

Description

Technical Field

[0001] This invention relates to the field of virtual reality intervention training, and more specifically, to a method and system for adaptive training of psychological intervention driven by virtual reality scenarios. Background Technology

[0002] Virtual reality (VR) technology has shown significant potential in the field of psychological intervention. By constructing immersive and controllable environments, it can provide intervention and guidance for psychological problems such as social phobia, anxiety disorders, and post-traumatic stress disorder. Currently, mainstream VR psychological intervention technologies mainly use preset scenarios and AI (artificial intelligence) generated characters, enabling individualized exposure therapy and cognitive training.

[0003] However, with the surge in demand for mental health services, traditional approaches face significant challenges in terms of realism, dynamic adaptability, and collaborative intervention within groups. For instance, current VR-based psychological intervention technologies rely on AI-generated virtual characters, whose behavioral patterns are simplistic and lack the unpredictability of human interaction. This can easily lead to users experiencing a false sense of security during social training, resulting in low skill transfer rates. Furthermore, most VR psychological intervention technologies only support single-person training, and the few that support multiple users generally use fixed role assignments, failing to dynamically adapt to differences in user abilities. When user goals conflict—for example, withdrawn individuals need to practice rejection, while assertive individuals need to practice listening—the system lacks a goal fusion mechanism, easily causing training interruptions due to conflict.

[0004] Therefore, how to research and design a virtual reality scene-driven psychological intervention adaptive training method and system that can overcome the above-mentioned defects is an urgent problem that we need to solve. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a method and system for adaptive training of psychological intervention driven by virtual reality scenarios. By performing logically complementary correlation matching on training target vectors, the real behavioral data (voice, actions) of one user is anonymized and mapped in real time to the NPC behavior of another user in a virtual scenario. This replaces traditional AI-generated characters, which can not only quickly construct virtual characters, but also enhance the realism and reaction efficiency of virtual characters, reduce the computing resources required for virtual reality scenario-driven training, and improve skill transfer rate.

[0006] The above-mentioned technical objective of the present invention is achieved through the following technical solution:

[0007] Firstly, a virtual reality-driven method for psychological intervention and adaptive training is provided, including the following steps:

[0008] Obtain the psychological assessment data of the first user and the first training target vector;

[0009] A virtual training scenario with configurable parameters is generated based on the psychological assessment data;

[0010] Obtain at least one second training target vector that is logically complementary to the first training target vector, and associate the first user with the second user to which the second training target vector belongs;

[0011] Receive the second behavior data of the second user in the virtual training scene, and map the second behavior data to the virtual character behavior in the virtual training scene corresponding to the first user;

[0012] Collect the first behavior data of the first user in the virtual training scene, and transmit the first behavior data to the second user, so as to map the first behavior data to the virtual character behavior in the virtual training scene corresponding to the second user.

[0013] Furthermore, the psychological assessment data includes one or more of the following: subjective self-assessment data, behavioral performance data, physiological signal data, task performance data, and real-time dynamic data.

[0014] Furthermore, the parameters include the level of environmental oppression, the intensity of the virtual character's aggression, and the task failure threshold.

[0015] Furthermore, the logical complementarity specifically refers to:

[0016] If the first user is to be trained in a first skill, and the second user is to be trained in a second skill;

[0017] When the implementation of the first skill depends on the output of the second skill, and the implementation of the second skill depends on the output of the first skill, then there is a logical complementarity between the first skill and the second skill.

[0018] Furthermore, the method also includes:

[0019] Calculate the Euclidean distance between the first training target vector and the second training target vector to obtain the target conflict index;

[0020] When the target conflict index is greater than a set threshold, a balanced target vector is calculated based on the first training target vector and the second training target vector.

[0021] The balanced target vector is used to simultaneously replace the first training target vector and the second training target vector, resulting in the final training target vector.

[0022] Furthermore, the formula for calculating the equilibrium target vector is as follows:

[0023] ;

[0024] in, This is the clipping function; The target value for balancing dimension i in the target vector; Let i be the initial target value for the first user in dimension i; Let i be the initial target value for the second user in dimension i; This is the tolerance factor for the second user, and its value is greater than 0.

[0025] Furthermore, the method also includes:

[0026] Based on the balance target vector, the mapping behavior data is corrected in real time, and the deviation compensation amount is output.

[0027] The corrected behavioral data is determined based on the deviation compensation amount.

[0028] Furthermore, the first behavioral data and / or the second behavioral data include physiological data and motion data;

[0029] The physiological data is one or more of skin conductance, heart rate variability and brain beta wave energy value, and the physiological data is output as three types of features: anxiety index, focus and emotional arousal after real-time noise reduction by edge computing nodes.

[0030] The motion data includes one or more of the following: motion delay time, voice tremor frequency, and eye movement focus drift.

[0031] Furthermore, the method also includes:

[0032] Scene separation is triggered when the training target deviation of the first user or the second user increases over N consecutive iterations:

[0033] The first user will independently train the first training objective in the corresponding virtual training scenario;

[0034] In addition, the second user will independently train the second training objective in the corresponding virtual training scenario.

[0035] Secondly, a virtual reality scene-driven psychological intervention adaptive training system is provided. This system is used to implement the virtual reality scene-driven psychological intervention adaptive training method as described in any one of the first aspects, including:

[0036] The data acquisition module is used to acquire the psychological assessment data of the first user and the first training target vector;

[0037] The scene construction module is used to generate virtual training scenes with configurable parameters based on the psychological assessment data;

[0038] The complementary association module is used to obtain at least one second training target vector that is logically complementary to the first training target vector, and to associate the first user with the second user to which the second training target vector belongs.

[0039] The behavior mapping module is used to receive the second behavior data of the second user in the virtual training scene, and map the second behavior data to the virtual character behavior in the virtual training scene corresponding to the first user;

[0040] The feedback mapping module is used to collect the first behavior data of the first user in the virtual training scene and transmit the first behavior data to the second user, so as to map the first behavior data to the virtual character behavior in the virtual training scene corresponding to the second user.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. The virtual reality scene-driven psychological intervention adaptive training method provided by the present invention, through logical complementary correlation matching of training target vectors, maps the real behavioral data (voice, action) of one user to the NPC behavior of another user in the virtual scene in real time after anonymization processing, replacing the traditional AI-generated character. It can not only quickly construct virtual characters, but also enhance the realism and reaction efficiency of virtual characters, reduce the computing resources required for virtual reality scene-driven training, and improve the skill transfer rate.

[0043] 2. In order to detect the target difference by means of a serious deviation from the training target during iterative training, this invention uses the target conflict index to detect the target difference. When the target difference is large, the pruning function is used to optimize and generate a balanced target vector. This can ensure that the balanced target is within the acceptable range for both parties and avoid causing excessive pressure on any user, effectively reducing the conflict training interruption rate.

[0044] 3. By constructing a recursive behavior calibrator, when the target deviation is detected to reach a certain condition, the present invention corrects the mapping behavior data in real time according to the balanced target vector and outputs the deviation compensation amount. The corrected behavior data is determined according to the deviation compensation amount, which can reduce the internal error in the training process and ensure the safe and stable operation of intervention training. Attached Figure Description

[0045] The accompanying drawings, which are included to provide a further understanding of embodiments of the invention and form part of this application, do not constitute a limitation thereof. In the drawings:

[0046] Figure 1 This is a flowchart from Embodiment 1 of the present invention;

[0047] Figure 2This is a system block diagram in Embodiment 2 of the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0049] Example 1: A virtual reality scene-driven method for psychological intervention and adaptive training, such as... Figure 1 As shown, it includes the following steps:

[0050] S1: Obtain the psychological assessment data of the first user and the first training target vector;

[0051] S2: Generate virtual training scenarios with configurable parameters based on psychological assessment data;

[0052] S3: Obtain at least one second training target vector that is logically complementary to the first training target vector, and associate the first user with the second user to which the second training target vector belongs;

[0053] S4: Receive the second behavior data of the second user in the virtual training scene, and map the second behavior data to the virtual character behavior in the virtual training scene corresponding to the first user;

[0054] S5: Collect the first user's first behavior data in the virtual training scene, and transmit the first behavior data to the second user, so as to map the first behavior data to the virtual character behavior in the virtual training scene corresponding to the second user.

[0055] In step S1, the psychological assessment data includes one or more of the following: subjective self-assessment data, behavioral performance data, physiological signal data, task performance data, and real-time dynamic data.

[0056] In some examples, subjective self-report data can be one or more of the following: anxiety level (GAD-7), depression level (PHQ-9), social avoidance scale (LSAS), which can be collected through standardized psychological scales and the obtained values ​​are normalized to [0,1].

[0057] In some examples, behavioral performance data can be one or more of speech fluency (words / minute), eye contact duration, and limb stiffness index, which can be collected through head-mounted display eye tracking, gesture recognition, etc., and the obtained values ​​are normalized to [0,1].

[0058] In some examples, physiological signal data may be one or more of the following: electrical skin response (EDA), heart rate variability (HRV), and frontal electroencephalography (β / θ waves), which can be acquired by biosensors on wristbands or headbands, and the obtained values ​​are normalized to [0,1].

[0059] In some examples, task performance data can be one or more of the following: number of retries for frustrating tasks, social response delay time, and error decision rate. These can be obtained through virtual scene tracking and the resulting values ​​are normalized to [0,1].

[0060] In some examples, real-time dynamic data can be one or more of the current stress peak, immediate emotional arousal, and cognitive load intensity, which can be obtained through real-time analysis of physiological signals, such as EDA derivative calculation, HRV spectrum 0.04-0.15Hz energy, and the obtained values ​​are converted to [-1,1].

[0061] In step S2, configurable parameters include, but are not limited to, environmental oppression level, virtual character aggression intensity, and task failure threshold.

[0062] In some examples, the level of environmental oppression can be determined based on data such as anxiety index, voice tremor frequency, and skin conductance response. The determined level of environmental oppression can be adjusted by adjusting scene data such as ambient light, background noise, and distance between characters. For example, the higher the anxiety index value, the darker the ambient light; the higher the voice tremor frequency value, the closer the virtual audience is; and the higher the skin conductance response value, the stronger the background noise.

[0063] In some examples, the aggressiveness of a virtual character can be determined based on a stress index, such as the higher the LF / HF power ratio in the HRV spectrum, the faster the NPC (non-player character) approaches.

[0064] In step S3, the first training target vector is composed of the first user's training targets in multiple dimensions, and the complementary training targets refer to the fact that the target needs of the two users in VR intervention training form a complementary structure in terms of psychological function.

[0065] In some examples, logical complementarity specifically means: if a first user is to be trained on a first skill and a second user is to be trained on a second skill; when the implementation of the first skill depends on the output of the second skill, and the implementation of the second skill depends on the output of the first skill, then there is logical complementarity between the first skill and the second skill.

[0066] For example, for socially withdrawn individuals, the training goal is to improve their ability to refuse, which requires the other party to provide unreasonable requests; for assertive communicators, the training goal is to reduce aggression, which requires the other party to provide gentle refusal; for anxious individuals, the training goal is to practice confident public speaking, which requires the other party to provide focused feedback from a virtual audience; and for individuals with attention deficit disorder, the training goal is to practice listening skills, which requires the other party to provide speech content input.

[0067] This invention addresses the challenge of achieving perfect complementarity between the first and second training target vectors when training objects are limited and training objectives are complex. Directly mapping behavioral data in such cases can easily lead to significant deviations from the training objective within a limited number of training iterations. Therefore, this invention calculates the Euclidean distance between the first and second training target vectors to obtain a target conflict index. When the target conflict index exceeds a set threshold, a balanced target vector is calculated based on the first and second training target vectors. This balanced target vector then simultaneously replaces both the first and second training target vectors, serving as the final training target vector.

[0068] It should be noted that the threshold can be set according to the user's required accuracy. For example, if the user requires high training accuracy, the threshold should be set to be smaller. In addition, the threshold can also be set according to the user's required training consistency. For example, if a higher number of training iterations is required, the threshold should be set to be smaller.

[0069] In some examples, the target conflict index is calculated by first calculating the Euclidean distance between the training target vectors of the two users, then determining the maximum possible distance, and finally dividing the Euclidean distance by the maximum possible distance to obtain the value in the interval [-1, 1]. It should be noted that the maximum possible distance is determined when all dimensions are completely opposite, i.e., the first user's values ​​are all 1s and the second user's values ​​are all -1s, in which case the conflict index is 1.

[0070] Specifically, the formula for calculating the target conflict index is as follows:

[0071] ;

[0072] in, Target conflict index; The number of target dimensions; Let i be the initial target value for the first user in dimension i; Let i be the initial target value for the second user in dimension i.

[0073] To ensure that the balance target is within an acceptable range for both parties and to avoid excessive pressure on any user, this invention employs a pruning function for target balancing. The formula for calculating the balance target vector is as follows:

[0074] ;

[0075] in, This is the clipping function; This is the tolerance factor for the second user, and its value is greater than 0.

[0076] In this invention, the clipping function is defined as: if Less than or equal to ,but Values ;like Greater than or equal to ,but Values ;like Greater than and less than Then the value is .

[0077] In step S4, the first behavioral data and / or the second behavioral data include physiological data and action data; the physiological data is one or more of skin conductance, heart rate variability and EEG beta wave energy value, and the physiological data is output as three types of features: anxiety index, focus and emotional arousal after real-time noise reduction by edge computing nodes; the action data is one or more of action delay time, voice tremor frequency and eye movement focus drift.

[0078] In some examples, considering the difference between the balancing target vector and the original training target vector, in order to reduce the internal error during the training process, the present invention also corrects the mapping behavior data in real time based on the balancing target vector and outputs the deviation compensation amount; and determines the corrected behavior data based on the deviation compensation amount.

[0079] Specifically, the formula for calculating the deviation compensation amount is as follows:

[0080] ;

[0081] in, This is the deviation compensation amount; The dynamic compensation coefficients for the recursive calibrator; To balance the target vector; The user's original behavior vector, such as the first training target vector and the second training target vector.

[0082] To ensure that the corrected behavior does not deviate excessively from the balance target, and to prevent psychological impact from sudden and drastic corrections, the calculation formula for determining the corrected behavioral data based on the deviation compensation amount in this invention is as follows:

[0083] ;

[0084] in, This refers to the user's corrected behavioral data in dimension i. The user's original behavior in dimension i; Let i be the amount of deviation compensation for dimension i; This is the constraint threshold.

[0085] In step S5, to avoid ineffective intervention training, the present invention also triggers scenario separation when the training target deviation of the first user or the second user increases in consecutive N iterations: the first user is independently trained for the first training target in the corresponding virtual training scenario; and the second user is independently trained for the second training target in the corresponding virtual training scenario.

[0086] It should be noted that the virtual reality scene-driven psychological intervention adaptive training method described in this invention can, when applied to two or more virtual characters, split the behavior of a single user to provide it to different other users. Similarly, the virtual reality scene of a single user can only receive a portion of the behavior of other users.

[0087] Furthermore, the behavioral data in this invention includes, but is not limited to, language, facial expressions, and actions. The mapping of behavioral data to virtual character behavior is mainly achieved by performing the above behaviors through a desensitized human body model in a virtual scene, in order to realize character simulation.

[0088] Example 2: A virtual reality scene-driven psychological intervention adaptive training system, which is used to implement the virtual reality scene-driven psychological intervention adaptive training method described in Example 1, such as... Figure 2 As shown, it includes a data acquisition module, a scene construction module, a complementary association module, a behavior mapping module, and a feedback mapping module.

[0089] The system includes: a data acquisition module for acquiring psychological assessment data and a first training target vector from a first user; a scene construction module for generating a virtual training scene with configurable parameters based on the psychological assessment data; a complementary association module for acquiring at least one second training target vector that is logically complementary to the first training target vector and associating the first user with the second user to whom the second training target vector belongs; a behavior mapping module for receiving second behavior data from the second user in the virtual training scene and mapping the second behavior data to the virtual character behavior in the virtual training scene corresponding to the first user; and a feedback mapping module for collecting first behavior data from the first user in the virtual training scene and transmitting the first behavior data to the second user to map the first behavior data to the virtual character behavior in the virtual training scene corresponding to the second user.

[0090] Working principle: This invention performs logically complementary correlation matching on the training target vectors, and after anonymizing the real behavior data (voice, action) of one user, it maps it in real time to the NPC behavior of another user in a virtual scene. This replaces the traditional AI-generated characters, which can not only quickly build virtual characters, but also enhance the realism and reaction efficiency of virtual characters, reduce the computing resources required for virtual reality scene driving, and improve the skill transfer rate.

[0091] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0092] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0093] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0094] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0095] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A virtual reality scene-driven psychological intervention adaptive training system, characterized in that, Includes the following steps: The data acquisition module is used to acquire the psychological assessment data of the first user and the first training target vector; The scene construction module is used to generate virtual training scenes with configurable parameters based on the psychological assessment data; The complementary association module is used to obtain at least one second training target vector that is logically complementary to the first training target vector, and to associate the first user with the second user to which the second training target vector belongs. The behavior mapping module is used to receive the second behavior data of the second user in the virtual training scene, and map the second behavior data to the virtual character behavior in the virtual training scene corresponding to the first user; The feedback mapping module is used to collect the first behavior data of the first user in the virtual training scene and transmit the first behavior data to the second user, so as to map the first behavior data to the virtual character behavior in the virtual training scene corresponding to the second user; The logical complementarity specifically refers to: If the first user is to be trained in a first skill, and the second user is to be trained in a second skill; When the implementation of the first skill depends on the output of the second skill, and the implementation of the second skill depends on the output of the first skill, then there is a logical complementarity between the first skill and the second skill.

2. The virtual reality scene-driven psychological intervention adaptive training system according to claim 1, characterized in that, The psychological assessment data includes one or more of the following: subjective self-assessment data, behavioral performance data, physiological signal data, task performance data, and real-time dynamic data.

3. The virtual reality scene-driven psychological intervention adaptive training system according to claim 1, characterized in that, The parameters include the level of environmental oppression, the intensity of the virtual character's aggression, and the task failure threshold.

4. The virtual reality scene-driven psychological intervention adaptive training system according to claim 1, characterized in that, The system also includes: Calculate the Euclidean distance between the first training target vector and the second training target vector to obtain the target conflict index; When the target conflict index is greater than a set threshold, a balanced target vector is calculated based on the first training target vector and the second training target vector. The balanced target vector is used to simultaneously replace the first training target vector and the second training target vector, resulting in the final training target vector.

5. The virtual reality scene-driven psychological intervention adaptive training system according to claim 4, characterized in that, The formula for calculating the equilibrium target vector is: ； in, This is the clipping function; The target value for balancing dimension i in the target vector; Let i be the initial target value for the first user in dimension i; Let i be the initial target value for the second user in dimension i; This is the tolerance factor for the second user, and its value is greater than 0.

6. The virtual reality scene-driven psychological intervention adaptive training system according to claim 4, characterized in that, The system also includes: Based on the balance target vector, the mapping behavior data is corrected in real time, and the deviation compensation amount is output. The corrected behavioral data is determined based on the deviation compensation amount.

7. The virtual reality scene-driven psychological intervention adaptive training system according to claim 1, characterized in that, The first behavioral data and / or the second behavioral data include physiological data and motion data; The physiological data is one or more of skin conductance, heart rate variability and brain beta wave energy value, and the physiological data is output as three types of features: anxiety index, focus and emotional arousal after real-time noise reduction by edge computing nodes. The motion data includes one or more of the following: motion delay time, voice tremor frequency, and eye movement focus drift.

8. The virtual reality scene-driven psychological intervention adaptive training system according to claim 1, characterized in that, The system also includes: Scene separation is triggered when the training target deviation of the first user or the second user increases over N consecutive iterations: The first user will independently train the first training objective in the corresponding virtual training scenario; In addition, the second user will independently train the second training objective in the corresponding virtual training scenario.

Images