Intelligent psychological tutoring integrated platform for psychological element universe

Abstract

The invention relates to the technical field of VR, AI and psychological therapy crossing, and discloses a psychological universe intelligent psychological tutoring integrated platform which comprises a holographic sensing module used for acquiring multi-modal data of a user in real time; the intelligent psychological engine module is used for receiving and analyzing the multi-modal data and generating a multi-dimensional psychological state vector; and the dynamic meta universe scene and personalized intervention module is used for adjusting virtual scene parameters and behaviors of the intelligent non-player character in real time according to the multi-dimensional psychological state vector. According to the method, multi-modal data such as behaviors, eye movement and physiology are acquired and deployed at edge computing nodes, and attention weight calculation is adopted to carry out weighted fusion and comprehensive analysis on the multi-modal data, so that a unified multi-dimensional psychological state vector is generated; therefore, the technical defects that in the prior art, data are dispersed, and effective integrated analysis cannot be carried out are overcome, and objective, quantitative and comprehensive evaluation on the psychological state of the user is achieved.

Description

Technical Field

[0001] This invention relates to the interdisciplinary fields of VR, AI, and psychotherapy, and in particular to the integrated intelligent psychological counseling platform of the Psychological Metaverse. Background Technology

[0002] In recent years, with the continuous development of virtual reality (VR) and metaverse technologies, their application in the field of mental health has become increasingly profound. Virtual reality technology can be used to construct highly immersive, controllable, and safe virtual environments for users to conduct various psychological interventions and counseling training. This provides an effective supplement or alternative to traditional psychotherapy methods.

[0003] For example, in exposure therapy, users with specific phobias (such as acrophobia or social phobia) can be guided to gradually confront their feared objects in virtual scenarios; in the treatment of post-traumatic stress disorder (PTSD), specific traumatic memory-related scenarios can be recreated for desensitization therapy; and in social skills training, a variety of social situations can be simulated for users to practice repeatedly.

[0004] In order to assess the user's condition, some existing technical solutions have begun to integrate various sensors into virtual reality systems to collect objective data on the user during the experience. For example, they can obtain physiological signals such as heart rate and skin conductance through biosensors, or obtain the user's behavior and eye movement data through trackers built into the headset.

[0005] However, existing technologies still have the following shortcomings in practical applications:

[0006] Limited scenario and poor adaptability: Traditional systems mainly rely on preset, static treatment scripts and scenarios, and cannot dynamically adjust and personalize according to the user's real-time physiological state (such as heart rate and eye movement data) and psychological state (such as anxiety level and attention changes) during the treatment process, resulting in limited treatment effects.

[0007] Data silos and insufficient analysis: Multimodal data generated during treatment (such as high-precision physiological signals, user behavior data in the VR environment, eye-tracking data, and therapist records) are stored in a scattered manner and lack effective integration. Existing systems lack a powerful analysis engine to perform real-time, comprehensive correlation analysis of this data, making it difficult to form a comprehensive assessment of the user's psychological state and make accurate intervention decisions.

[0008] Interaction limitations and weak group support: Existing systems mostly focus on single-person interaction experiences, and are significantly insufficient in supporting group therapy. It is difficult to effectively capture and analyze the complex social dynamics in multi-person interactions (such as distance between members, eye contact, and dialogue patterns), which limits the effective implementation of group therapy in VR environments.

[0009] System latency impacts immersion and therapeutic efficacy: The delay (often exceeding 1 second) between user emotion / state recognition (based on sensors and AI analysis) and system intervention (such as adjusting the scene or NPC behavior) severely undermines the immersive experience of VR and reduces the immediacy and effectiveness of treatment, especially in exposure therapy or social context training that requires rapid response. Summary of the Invention

[0010] The purpose of this invention is to provide an integrated intelligent psychological counseling platform for the Psychological Metaverse, which solves the problems of existing technologies such as limited scene diversity and poor adaptability, data silos and insufficient analysis, limited interaction and weak group support, as well as system latency affecting immersion and therapeutic efficacy.

[0011] To achieve the above objectives, the first aspect of this invention provides a psychological metaverse intelligent psychological counseling integrated platform, comprising:

[0012] A holographic perception module is used to acquire the user's multimodal data in real time. The multimodal data includes behavioral data and eye movement data collected by at least one virtual reality (VR) head-mounted display device equipped with a 6-DOF spatial positioning module and a high-precision eye-tracking module, as well as physiological data collected by at least one biosensor.

[0013] The intelligent psychological engine module, deployed on an edge computing node, is used to receive and analyze the multimodal data to generate a multidimensional psychological state vector representing the user's current psychological state.

[0014] It also includes a dynamic metaverse scene and personalized intervention module, which is used to adaptively adjust the virtual scene parameters and the behavior of intelligent non-player characters (NPCs) in the platform in real time based on the multidimensional psychological state vector.

[0015] In one embodiment, when the intelligent psychological engine module receives and analyzes the multimodal data, it first performs feature extraction operations on the multimodal data acquired by the holographic perception module within a preset-width sliding time window using a temporal feature extraction unit to generate standardized feature data. This feature extraction operation can be defined by the following formula:

[0016] ;

[0017] In the formula, For at any time The extracted first Standardized feature data; For the first Each sensor at the original time point The raw data stream collected; For a sliding time window, where This is the preset width of the time window; In order to target the Preset feature extraction operators for each raw data stream.

[0018] In one specific embodiment, the intelligent psychological engine module further includes:

[0019] The attention weight calculation unit is used to calculate the evaluation weight of each standardized feature data based on a preset query vector and the standardized feature data, using an attention weight calculation model. The evaluation weight can be defined by the following formula:

[0020] ;

[0021] In the formula, where, Let be a weight vector, whose elements represent the weights used in evaluating the first weight. When considering a psychological dimension, the attention weights are assigned to each standardized feature data. To represent the assessment of the first A query vector representing a psychological dimension of intent; Let be the transpose of the key matrix derived from all standardized feature data, where It is the transpose operator; The dimension of the key vector is used as a scaling factor; It is a normalized exponential function.

[0022] The feature fusion unit is used to perform weighted fusion of the standardized feature data according to the evaluation weights to calculate one or more mental state components constituting the multidimensional mental state vector. This weighted fusion process can be defined by the following formula:

[0023] ;

[0024] In the formula, For a moment Calculated The value of each psychological state component; In order to evaluate the first When assigning a psychological component, it is allocated to the first... Weights of standardized feature data; For the first Standardized feature data; In order to target the Preset activation functions for each psychological component; This represents the total number of standardized feature data.

[0025] Preferably, the intelligent psychological engine module is further used to calculate a cognitive load component as a psychological state component in the multidimensional psychological state vector. The calculation of this cognitive load component can be defined by the following formula:

[0026] ;

[0027] In the formula, Cognitive load components; For the user's task error rate; For the user's task response time; For the user's pupil diameter data; , , These are the corresponding preset weighting coefficients.

[0028] In one embodiment, when adjusting the virtual scene parameters, the dynamic metaverse scene and personalized intervention module specifically performs the following steps: determining a target scene parameter vector based on the multidimensional psychological state vector; and using a preset smoothing coefficient, smoothly approximating the current parameter vector of the virtual scene to the target scene parameter vector. This smooth approximation process can be defined by the following formula:

[0029] ;

[0030] In the formula, For at any time The current parameter vector of the scenario; For at any time The target scene parameter vector; This is a mapping function used to calculate the target scene parameter vector based on the mental state vector; This refers to the multidimensional psychological state vector; It is a smoothing coefficient with a value between 0 and 1.

[0031] In one embodiment, when adjusting the behavior of the intelligent NPC, the dynamic metaverse scene and personalized intervention module specifically involves: performing a vector operation between a preset weight vector associated with a specific intervention task and the multidimensional psychological state vector to output the current intervention task with the highest priority from multiple preset intervention tasks. This decision-making process can be defined by the following formula:

[0032] ;

[0033] In the formula, An index of the intervention tasks with the highest priority determined by the decision; In order to be with the first The weight vector associated with each intervention task; This refers to the multidimensional psychological state vector; To obtain the parameter that maximizes the value of the expression The operation.

[0034] Preferably, the dynamic metaverse scene and personalized intervention module is further configured to: activate the intervention node when the value of a psychological state component in the multidimensional psychological state vector meets a preset activation threshold, thereby driving the intelligent NPC to perform a corresponding specific intervention action. This activation condition can be defined by the following formula:

[0035] ;

[0036] In the formula, For the first The activation status of each intervention node, where 1 represents activation and 0 represents inactivation; For the first The value of a specific psychological state component; In order to target the The preset activation threshold for each intervention node.

[0037] In a specific application scenario, when the holographic perception module acquires multimodal data from multiple users, the intelligent psychological engine module is further used to: calculate a social interaction score between any two users based on the physical distance, mutual gaze index, and dialogue interaction index among the multiple users. This score can be defined using the following formula:

[0038] ;

[0039] In the formula, For users With users Social interaction score between them; For users With users The physical distance between them; This serves as an indicator of mutual observation between the two. This serves as an indicator of the dialogue and interaction between the two parties. , , Preset weighting coefficients for distance, gaze, and dialogue metrics, respectively; It is an exponentially decaying function with respect to distance, and its value decreases as the distance increases.

[0040] Preferably, the platform is also used to: generate an interaction heatmap matrix based on the social interaction score, to characterize the global social dynamics of the group composed of the multiple users. The generation of this matrix can be defined by the following formula:

[0041] ;

[0042] In the formula, An interactive heatmap matrix; For users With users Social interaction score between them; This represents the total number of users in the group.

[0043] A second aspect of this invention provides a method for operating a psychological metaverse intelligent psychological counseling integrated platform, the method comprising the following steps:

[0044] Step a, acquire the user's multimodal data in real time, including behavioral data and eye-tracking data collected through a virtual reality (VR) head-mounted display device, as well as physiological data collected through biosensors;

[0045] Step b: At the edge computing node, the multimodal data is received and analyzed to generate a multidimensional psychological state vector representing the user's current psychological state.

[0046] Step c: Based on the multidimensional psychological state vector, the virtual scene parameters and the behavior of intelligent non-player characters (NPCs) in the platform are adaptively adjusted in real time.

[0047] In one embodiment, by repeatedly executing the method, it can also be used to track a user's long-term treatment progress. This progress can be quantified by a treatment progress index. The index can be defined by the following formula:

[0048] ;

[0049] In the formula, For users to complete the first Treatment progress index after each treatment session; For the first During the next session, under specific standardized scenarios, the psychological state components The average value; To determine the psychological component during the initial baseline treatment, under the same standardized scenario. The average value.

[0050] In summary, the present invention has at least one of the following beneficial technical effects:

[0051] 1. This invention acquires multimodal data such as behavioral, eye-tracking, and physiological data, and deploys it on edge computing nodes. It uses attention weight calculation to perform weighted fusion and comprehensive analysis on the multimodal data to generate a unified multidimensional psychological state vector. This solves the technical defects of existing technologies, such as scattered data and inability to effectively integrate and analyze data, and realizes an objective, quantitative, and comprehensive assessment of the user's psychological state.

[0052] 2. This invention uses multidimensional psychological state vectors as the basis for decision-making, and continuously and adaptively adjusts the virtual scene parameters and the behavior of intelligent non-player characters. Therefore, it can overcome the limitations of traditional systems that rely on static scripts and cannot respond to real-time changes in user status. This allows intervention measures to be dynamically matched with the user's immediate psychological state, thereby improving the pertinence and effectiveness of the intervention.

[0053] 3. By explicitly deploying the intelligent psychological engine module on edge computing nodes, bringing it closer to the user terminal, this invention significantly reduces the system latency from data collection to intervention feedback. Compared with traditional cloud computing architectures, it ensures the immediacy of closed-loop feedback in perception, evaluation, and intervention, thus guaranteeing the user's immersive experience in the virtual reality environment.

[0054] 4. This invention quantifies the social interaction score between users based on the physical distance, mutual gaze index, and dialogue interaction index among multiple users, and further generates an interaction heatmap matrix. This solves the technical problem of existing technologies being unable to capture and analyze the social dynamics of multiple people, provides objective and visualized data support for group psychological counseling, and expands the application depth of virtual reality technology in group therapy scenarios. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of the system architecture of the present invention;

[0056] Figure 2 This is a schematic diagram of the method flow of the present invention.

[0057] Among them, 100 is the holographic perception module; 200 is the intelligent psychological engine module; 300 is the dynamic metaverse scene and personalized intervention module; 400 is the user terminal device; 500 is the edge computing node; and 600 is the cloud server. Detailed Implementation

[0058] See attached document Figure 1 , Figure 1 This is a schematic diagram of the system architecture of the Psychological Metaverse Intelligent Psychological Counseling Integrated Platform according to an embodiment of the present invention. The platform includes: one or more user terminal devices 400, one or more edge computing nodes 500, and a cloud server 600.

[0059] User terminal device 400 and edge computing node 500 interact with each other via a low-latency wireless communication network, such as Wi-Fi 6 or 5G network. Edge computing node 500 connects to cloud server 600 via a standard network interface.

[0060] In this embodiment, the platform's functions are implemented through three core modules deployed on different physical devices, including: a holographic perception module 100, an intelligent psychological engine module 200, and a dynamic metaverse scene and personalized intervention module 300.

[0061] The holographic sensing module 100 is physically integrated into the user terminal device 400. This module is used to acquire the user's multimodal data in real time and transmit the multimodal data to the edge computing node 500 through a low-latency wireless communication network.

[0062] The intelligent psychological engine module 200 and the dynamic metaverse scene and personalized intervention module 300 are both deployed on the edge computing node 500.

[0063] The intelligent psychological engine module 200 takes a multimodal data stream from the holographic perception module 100 as input and outputs a multidimensional psychological state vector representing the user's current psychological state. This vector can be represented in the following form:

[0064] ;

[0065] In the formula, For a moment A multidimensional mental state vector; For the first in the vector A psychological state component, such as a cognitive load component or an emotional valence component; This represents the total dimension of the vector; It is the transpose operator;

[0066] The dynamic metaverse scene and personalized intervention module 300 takes as input a multidimensional psychological state vector generated by the intelligent psychological engine module 200. This module performs decision calculations based on vectors and outputs scene adjustment instructions and intelligent non-player character (NPC) behavior instructions. These instructions are then sent back to the user terminal device 400 via a low-latency wireless communication network to drive real-time changes in the virtual reality environment.

[0067] The cloud server 600 is used to perform non-real-time tasks, including long-term storage of user anonymized data from the edge computing node 500, training or updating the algorithm model used in the intelligent psychological engine module 200, and generating interim treatment analysis reports.

[0068] In one specific embodiment, the holographic perception module 100 includes a behavioral data acquisition unit 110, an eye-tracking data acquisition unit 120, and a physiological data acquisition unit 130. All data acquired by these units is tagged with a synchronized timestamp to ensure data consistency over time.

[0069] The behavior data acquisition unit 110 is implemented through a 6-DOF spatial positioning module integrated into the virtual reality headset. This unit continuously acquires the spatial position information of the user's head and hand controllers at a preset sampling frequency (e.g., 90Hz). The position information specifically includes three-dimensional spatial coordinates (x, y, z) and three-axis rotational attitude (pitch angle, yaw angle, roll angle).

[0070] The eye-tracking data acquisition unit 120 is implemented through a high-precision eye-tracking module integrated into the virtual reality headset. This unit collects eye-tracking related data of the user, specifically including: the user's real-time gaze coordinates, the start and end points of the eye saccade path, the duration of the gaze, the real-time diameter of the pupil (in millimeters), and the blink frequency per unit time.

[0071] The physiological data acquisition unit 130 includes at least one biosensor connected to the user terminal device 400. In one specific embodiment, it specifically includes an electrocardiogram (ECG) sensor and a skin conductance sensor (GSR). The ECG sensor is used to acquire electrocardiogram signals and extract heart rate (HR) and heart rate variability (HRV) indicators from them. The GSR sensor is used to acquire skin conductance signals and extract the amplitude and frequency of skin conductance level (SCL) and skin conductance response (SCR).

[0072] Therefore, the multimodal data collected and synchronized by the behavioral data acquisition unit 110, the eye-tracking data acquisition unit 120, and the physiological data acquisition unit 130 together form the raw data stream, which is transmitted to the edge computing node 500 for further processing by the intelligent psychological engine module 200.

[0073] In one specific embodiment, there is an intelligent psychological engine module 200. This module receives multimodal data streams from the holographic perception module 100, and its core function is to generate and output a multidimensional psychological state vector that can represent the user's current psychological state in real time through a series of computational processes. .

[0074] Specifically, the intelligent psychological engine module 200 functionally includes a feature extraction unit 210, an attention weight calculation unit 220, and a feature fusion unit 230.

[0075] The feature extraction unit 210 receives the raw multimodal data stream and performs feature extraction operations on the data within a preset-width sliding time window to generate standardized feature data. This feature extraction operation can be defined by the following formula:

[0076] ;

[0077] In the formula, For at any time The extracted first Standardized feature data; For the first Each sensor at the original time point The raw data stream collected; For a sliding time window, where The preset width of this time window, for example, 2 seconds; In order to target the A preset feature extraction operator for a raw data stream; for example, when When it is an electrocardiogram signal, It can be an operator for calculating time-domain indices of heart rate variability (such as SDNN); when When it is a skin conductance signal, It can be used as an operator to calculate the amplitude of skin conductance response (SCR).

[0078] Attention weight calculation unit 220 receives standardized feature data generated by feature extraction unit 210 and calculates an evaluation weight for each feature data using an attention weight calculation model. This calculation process is defined by the following formula:

[0079] ;

[0080] In the formula, where, Let be a weight vector, whose elements represent the weights used in evaluating the first weight. When considering a psychological dimension, the attention weights are assigned to each standardized feature data. To represent the assessment of the first A query vector representing a psychological dimension of intent; Let be the transpose of the key matrix derived from all standardized feature data, where It is the transpose operator; The dimension of the key vector is used as a scaling factor; This is the normalization exponential function, used to convert the calculation results into a probability distribution that sums to 1;

[0081] Feature fusion unit 230 performs weighted fusion on standardized feature data based on the evaluation weights output by attention weight calculation unit 220 to calculate one or more mental state components constituting a multidimensional mental state vector. This weighted fusion process is defined by the following formula:

[0082] ;

[0083] In the formula, For a moment Calculated The value of each psychological state component; In order to evaluate the first When assigning a psychological component, it is allocated to the first... Weights of standardized feature data; For the first Standardized feature data; In order to target the Preset activation functions for each psychological component, such as linear unit functions or sigmoid functions; This represents the total number of standardized feature data.

[0084] In one specific embodiment, the intelligent psychological engine module 200 calculates cognitive load components. This is a psychological state component within a multidimensional psychological state vector. This component is calculated by weighting and combining the user's task error rate, task reaction time, and pupil diameter data. The calculation process is defined by the following formula:

[0085] ;

[0086] In the formula, Cognitive load components; For the user's task error rate; For the user's task response time; For the user's pupil diameter data; , , These are the corresponding preset weighting coefficients.

[0087] Finally, all mental state components calculated by the feature fusion unit 230 (e.g. (Emotional valence component, arousal component, etc.) together constitute the moment. Multidimensional mental state vector This vector is then transmitted to the Dynamic Metaverse Scene and Personalized Intervention Module 300 as the basis for its decision-making.

[0088] In another specific embodiment, the input to the dynamic metaverse scene and personalized intervention module 300 is a real-time multidimensional psychological state vector from the intelligent psychological engine module 200. This vector is used to perform decision calculations and generate scene adjustment instructions and intelligent non-player character (NPC) behavior instructions to achieve real-time adaptive adjustment of the virtual environment.

[0089] Specifically, the dynamic metaverse scene and personalized intervention module 300 includes a scene adaptation unit 310 and an NPC behavior decision-making unit 320.

[0090] The scene adaptation unit 310 is used to adjust the background parameters of the virtual environment. This unit first adjusts the background parameters based on the input multidimensional psychological state vector. A target scene parameter vector is defined, and then a smoothing algorithm is used to drive the current scene parameter vector to approximate the target value. This smoothing approximation process is defined by the following formula:

[0091] ;

[0092] In the formula, For at any time The current parameter vector of the scene, whose components may include quantifiable scene parameters such as scene brightness, color temperature, background volume, and weather effects; For at any time The target scene parameter vector; This is a mapping function used to calculate the target scene parameter vector based on the mental state vector; A multidimensional mental state vector; It is a smoothing coefficient with a value between 0 and 1.

[0093] The NPC behavior decision-making unit 320 is used to control the behavior of one or more intelligent NPCs in the virtual environment. This unit adopts a two-stage decision-making mechanism, first determining the high-level intervention task, and then activating one or more specific atomic intervention actions.

[0094] In the high-level intervention task decision-making phase, the NPC behavioral decision-making unit 320 performs vector operations on the pre-defined weight vector associated with each specific intervention task and the current multidimensional psychological state vector to select the current intervention task with the highest priority. This decision-making process is defined by the following formula:

[0095] ;

[0096] In the formula, The index is used to determine the intervention task with the highest priority, and the set of intervention tasks is predefined, including, for example, reassurance, encouragement, and challenge. In order to be with the first The weight vector associated with each intervention task, and The dimensions are the same, and the values ​​of each component define the ideal psychological state of the user corresponding to the task. A multidimensional mental state vector; To obtain the parameter that maximizes the value of the expression The operation.

[0097] During the specific atomic intervention activation phase, if the high-level task Once identified, the system will monitor specific psychological state components related to the task. When the value of a component meets a preset activation threshold, the corresponding intervention node will be activated, driving the NPC to perform a specific intervention action. This activation condition is defined by the following formula:

[0098] ;

[0099] In the formula, For the first The activation status of an intervention node is 1, which represents activation and 0 represents inactivation. An intervention node corresponds to an atomic NPC action, such as saying a specific line of dialogue or performing a specific gesture. For the first The value of a specific psychological state component; In order to target the Each intervention node has a preset activation threshold. For example, in a soothing task, the user's anxiety component value exceeds a preset threshold. At that time, the intervention node used to perform the soothing action Activated;

[0100] In another embodiment, when the holographic perception module 100 acquires multimodal data from multiple users, the intelligent psychological engine module 200 performs additional calculations to quantify the time-series relationship between any two users (e.g., user a and user b). The intensity of social interaction. This intensity is characterized by a social interaction score, which is calculated using the following formula:

[0101] ;

[0102] In the formula, For users With users Social interaction score between them; For users With users The physical distance between them is calculated from the location information acquired by the behavior data acquisition unit 110; It is an exponentially decaying function with respect to distance, and its value decreases as the distance increases;

[0103] The mutual gaze index is a quantitative value calculated based on data from the eye-tracking data acquisition unit 120, used to characterize the degree of visual interaction between two users. In one embodiment, this index is based on the amount of eye contact between the users within a preset time window. Watch users The duration of the virtual avatar's facial area, and the user's Watch users Calculated based on the duration of the virtual avatar's facial area;

[0104] This is a metric for dialogue interaction between two users, used to characterize the activity level of verbal communication between them. It can be calculated based on the microphone and voice activity detection (VAD) algorithm on the user terminal device 400. For example, it could be the user's... In the user During the speech, the user A quantitative value for the frequency or duration of verbal responses;

[0105] , , These are preset weighting coefficients for the three indicators: distance, gaze, and dialogue, used to adjust the relative importance of different interaction channels in the total score calculation;

[0106] To characterize the overall social dynamics of the group, the intelligent psychological engine module 200 further generates an interaction heatmap matrix based on the social interaction scores between all pairs of users. The generation process of this matrix is ​​defined by the following formula:

[0107] ;

[0108] In the formula, An interactive heatmap matrix; For users With users Social interaction score between them; The total number of users in the group is represented, and this interaction heatmap matrix can be visualized for therapists, providing them with a quantitative map of the group's internal interaction patterns, subgroup structure, and the presence of isolated members.

[0109] See attached document Figure 2 , Figure 2 This is a flowchart illustrating the operation method of the integrated intelligent psychological counseling platform of the Psychological Metaverse according to another embodiment of the present invention. The method includes:

[0110] Step a: The holographic perception module 100 acquires the user's multimodal data in real time and synchronously on the user terminal device 400 through the behavior data acquisition unit 110, eye movement data acquisition unit 120 and physiological data acquisition unit 130, and sends the data stream marked with timestamps to the edge computing node 500.

[0111] Step b: The intelligent psychological engine module 200 receives a multimodal data stream. Through its internal feature extraction unit 210, attention weight calculation unit 220, and feature fusion unit 230, it performs calculations to ultimately generate and output a multidimensional psychological state vector representing the user's current psychological state. .

[0112] Step c: The dynamic metaverse scene and personalized intervention module 300 receives multidimensional psychological state vectors. Based on the vector, the scene adaptation unit 310 and the NPC behavior decision-making unit 320 perform decision calculations to generate scene adjustment instructions and NPC behavior instructions.

[0113] Furthermore, scene adjustment commands and NPC behavior commands are sent back to the user terminal device 400, where they are executed by the rendering engine and logic engine to adjust the parameters of the virtual reality environment and the behavior of the NPCs in real time. The system then returns to step a, continuously looping through this closed-loop process.

[0114] In another embodiment, by repeating the above method across multiple independent treatment sessions, the platform can also be used to track a user's long-term treatment progress. This progress is quantified using a treatment progress index. The calculation of this index is defined by the following formula:

[0115] ;

[0116] In the formula, For users to complete the first The treatment progress index after each treatment session is a dimensionless normalized value.

[0117] For the first During the next session, in a pre-defined, standardized scenario with fixed content and parameters, a pre-selected key psychological state component was analyzed. (e.g., anxiety component) The average value obtained from continuous measurement within the session;

[0118] To ensure that, during the initial baseline therapy, the user, at the time of their first baseline assessment, is assessed in the exact same standardized setting for the same psychological state component. The baseline average obtained from the measurement is used in a standardized scenario to provide a unified measurement benchmark across different sessions, thereby ensuring that the calculated progress index can objectively reflect changes in the user's own psychological state, rather than being caused by changes in the scenario. Furthermore, the treatment progress index is stored on a cloud server 600 to generate long-term efficacy analysis reports.

[0119] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. The Psychological Metaverse Intelligent Psychological Counseling Integrated Platform, characterized in that: include: A holographic perception module is used to acquire the user's multimodal data in real time. The multimodal data includes behavioral data and eye movement data collected by at least one virtual reality head-mounted display device equipped with a 6-DOF spatial positioning module and a high-precision eye-tracking module, as well as physiological data collected by at least one biosensor. The intelligent psychological engine module, deployed on an edge computing node, is used to receive and analyze the multimodal data to generate a multidimensional psychological state vector representing the user's current psychological state. The dynamic metaverse scene and personalized intervention module is used to adaptively adjust the virtual scene parameters and the behavior of intelligent non-player characters in real time based on the multi-dimensional psychological state vector.

2. The integrated intelligent psychological counseling platform of the psychological metaverse as described in claim 1, characterized in that, The intelligent psychological engine module receives and analyzes the multimodal data, including: Within a preset sliding time window, feature extraction operations are performed on the multimodal data acquired by the holographic perception module to generate standardized feature data.

3. The integrated intelligent psychological counseling platform of the psychological metaverse according to claim 2, characterized in that, The intelligent psychological engine module includes: An attention weight calculation model is used to calculate the evaluation weight of each standardized feature data based on the preset query vector and the standardized feature data. And the standardized feature data are weighted and fused according to the evaluation weights to calculate one or more psychological state components that constitute the multidimensional psychological state vector.

4. The integrated intelligent psychological counseling platform of the psychological metaverse as described in claim 3, characterized in that, The intelligent psychological engine module also includes: The cognitive load component is calculated as a psychological state component in the multidimensional psychological state vector, wherein the cognitive load component is obtained by weighting and combining the user's task error rate, task reaction time, and pupil diameter data.

5. The integrated intelligent psychological counseling platform of the psychological metaverse according to claim 1, characterized in that, In the dynamic metaverse scene and personalized intervention module, adjusting the virtual scene parameters includes: The target scene parameter vector is determined based on the multidimensional psychological state vector; And through a preset smoothing coefficient, the current parameter vector of the virtual scene is made to smoothly approximate the parameter vector of the target scene.

6. The integrated intelligent psychological counseling platform of the psychological metaverse as described in claim 1, characterized in that, In the dynamic metaverse scene and personalized intervention module, when adjusting the behavior of the intelligent non-player character, the following is included: By performing vector operations on a preset weight vector associated with a specific intervention task and the multidimensional psychological state vector, the current intervention task with the highest priority is output from multiple preset intervention tasks.

7. The integrated intelligent psychological counseling platform of the psychological metaverse according to claim 6, characterized in that, The dynamic metaverse scene and personalized intervention module also includes: When the value of the psychological state component in the multidimensional psychological state vector meets the preset activation threshold of the intervention node, the intervention node is activated to drive the intelligent non-player character to perform an intervention action.

8. The integrated intelligent psychological counseling platform of the psychological metaverse according to claim 1, characterized in that, The intelligent psychological engine module further includes: when the holographic perception module acquires multimodal data from multiple users, it calculates the social interaction score between any two users based on the physical distance between the multiple users, mutual gaze index, and dialogue interaction index.

9. The integrated intelligent psychological counseling platform of the psychological metaverse according to claim 8, characterized in that, The platform also includes: Based on the social interaction score, an interaction heatmap matrix is ​​generated, wherein each element value in the interaction heatmap matrix corresponds to the social interaction score between a pair of users, in order to characterize the global social dynamics of the group composed of the multiple users.

10. A method for operating the Psychological Metaverse Intelligent Psychological Counseling Integrated Platform, based on the Psychological Metaverse Intelligent Psychological Counseling Integrated Platform as described in any one of claims 1-9, characterized in that, Includes the following steps: Step a, acquire the user's multimodal data in real time, including behavioral data and eye-tracking data collected through a virtual reality headset, as well as physiological data collected through biosensors; Step b: At the edge computing node, the multimodal data is received and analyzed to generate a multidimensional psychological state vector representing the user's current psychological state. Step c: Based on the multidimensional psychological state vector, the virtual scene parameters and the behavior of the intelligent non-player character in the platform are adaptively adjusted in real time.

Images