Fuzzy logic control system for virtual interactive scene of college students' psychological counseling

Abstract

The present application belongs to the technical field of mental health services, and discloses a virtual interaction scene fuzzy logic control system for college student psychological counseling, which comprises: obtaining user physiological signals, interaction behavior sequences and ACT microscale responses, mapping preliminary experiential avoidance tendency preliminary membership, and performing positive correction to generate a fuzzy variable set; identifying variable characteristics and analyzing macroscopic targets into a microcosmic control target sequence, matching atomic assembly personalized ACT narrative chains to generate an intervention task package; triggering visual cognitive dissociation intervention at the VR end to generate a real-time updated dynamic value of psychological flexibility; determining a therapeutic tolerance window through two-dimensional evaluation, executing acceptance guidance and scene fine-tuning, defining an artificial intervention threshold of an extreme state and issuing a warning, and integrating into an enhanced fuzzy context package; comparing intervention effects and optimizing intervention rules, linking artificial tutoring to reversely adapt subsequent intervention task packages to form a closed-loop cultivation link.

Description

Technical Field

[0001] This invention relates to the field of mental health service technology, and more specifically, to a virtual interactive scenario fuzzy logic control system for psychological counseling of college students. Background Technology

[0002] The current problem of weak psychological flexibility among college students is becoming increasingly prominent. Offline psychological counseling suffers from limited coverage and insufficient real-time capability. Virtual interactive technology has become an important direction for exploration in college psychological counseling. However, existing solutions have not integrated Acceptance and Commitment Therapy (ACT) with fuzzy logic to build a precise control system, making it difficult to achieve personalized, dynamic, and efficient psychological counseling.

[0003] Existing virtual psychological counseling solutions still have some shortcomings: they lack the ability to quantitatively assess psychological flexibility, adopt fixed counseling strategies, cannot adapt to users' core psychological weaknesses, and lack personalized intervention; they lack dynamic control and a closed-loop mechanism for the entire intervention process, cannot adjust intervention strategies according to real-time psychological state, and cannot optimize counseling solutions through data iteration. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a virtual interactive scenario fuzzy logic control system for psychological counseling of college students, comprising: Baseline variable module: Acquires user physiological signals, interaction behavior sequences and ACT microscale responses, maps preliminary membership of experiential avoidance tendency through ACT microscale, and positively corrects through interaction behavior sequences and physiological signals, generating a five-dimensional psychological flexibility fuzzy variable set, and simultaneously encrypts and establishes a core weakness baseline profile. Task package generation module: Based on fuzzy variable set and baseline file, it identifies variable features and parses macro-objectives into micro-control target sequences. Then, it matches and assembles personalized ACT narrative chains from the ACT narrative atom library, allocates edge-cloud tasks and differentiated control division of labor between VR terminal and edge side, and generates intervention task packages. Intervention data module: Based on the micro-control target sequence, personalized ACT narrative chain and intervention task package, the VR terminal triggers visual cognitive dissociation intervention, links with the sensory feedback of wearable devices, and generates updated dynamic values ​​of psychological flexibility and intervention execution data in real time. Tolerance Encapsulation Module: Based on updated dynamic values ​​and intervention execution data, it assesses the therapeutic tolerance window through a two-dimensional evaluation, performs acceptance guidance and scenario fine-tuning, defines the threshold for artificial intervention in extreme states and issues warnings, and integrates and encapsulates the records into an enhanced fuzzy context package. Closed-loop iteration module: Based on the enhanced fuzzy context package, update the user's psychological activity development trajectory file, compare the intervention effect and optimize the intervention rules, link with human counseling to reverse adapt to subsequent intervention task packages, and form a closed-loop training link.

[0005] Furthermore, the generation method of the fuzzy variable set includes: Simultaneously collect user physiological signals, interactive behavior sequences in virtual scenarios, and ACT microscale responses of users at decision points in virtual scenarios; The ACT microscale response is mapped to the initial membership degree of empirical avoidance tendency, and combined with high-stimulation scene behavior and guided triggering behavior in the interaction behavior sequence, and positively corrected according to high and low priority. Physiological signals are converted into fuzzy membership degrees, and deviation is checked against the modified empirical avoidance tendency preliminary membership degrees. After the check passes, the final EA membership degree is obtained. Based on EA membership, combined with the ACT microscale and interactive behavior sequence, a fuzzy variable set of five dimensions of psychological flexibility—cognitive integration, experiential avoidance tendency, current contact ability, value clarity, and commitment to action—is generated through a five-dimensional weighted cross-analysis mechanism. After encryption, a baseline profile of psychological flexibility is established to mark the core shortcomings.

[0006] Furthermore, the method of identifying variable features and parsing macroscopic objectives into a sequence of microscopic control objectives includes: Based on a fuzzy variable set and baseline profile, threshold rules are used to identify the variable characteristics and core weaknesses of users' psychological flexibility. To enhance psychological flexibility, anchoring macro-level goals involves defining the guiding emphasis of macro-level goals by combining variable characteristics and core weaknesses. Furthermore, following the principle of gradual progress and priority matching, the guidance focus is broken down into a sequence of micro-control targets that can be implemented in stages and each corresponds to a single ACT intervention direction.

[0007] Furthermore, the methods for assembling personalized ACT narrative chains include: Classified according to the six core processes of ACT theory, labeled with ACT intervention direction as intervention dimension label, set difficulty level according to intervention difficulty gradient, take the core weakness of psychological flexibility as the adaptation target, pre-set standardized narrative atoms, and integrate and construct ACT narrative atom library. Based on the micro-control target sequence, narrative atoms are matched with intervention dimension labels and difficulty gradients from the ACT narrative atom library and dynamically assembled into personalized ACT narrative chains.

[0008] Furthermore, the method for generating the intervention task package includes: Based on the personalized ACT narrative chain, the interaction attributes and execution requirements of each narrative atom are analyzed; the execution tasks of each type of narrative atom are differentiated and assigned to the VR end, edge side and cloud according to the execution requirements, and executable instructions with priority are generated. Furthermore, it integrates personalized ACT narrative chains with executable instructions and encapsulates them in an encrypted package as an intervention task package.

[0009] Furthermore, the VR terminal triggers visual cognitive dissociation intervention, and the sensory feedback of the wearable device is linked in the following ways: Based on the micro-control target sequence, personalized ACT narrative chain and intervention task package, preset visual cognitive dissociation intervention trigger conditions, and bind VR terminal interaction logic and user behavior to form intervention rules; When a user's behavior meets the triggering conditions, the VR device executes intervention according to the ACT narrative chain sequence, guiding the user's operation through interactive logic; Wearable devices provide real-time sensory feedback based on the coordinated and synchronized intervention rhythm at the edge, forming a collaborative closed loop with VR intervention.

[0010] Furthermore, the methods for generating updated dynamic values ​​of psychological flexibility and intervention implementation data include: By simultaneously collecting real-time status data of device operation through VR terminals, wearable devices, and edge computing, and combining it with preset weighted calculation rules, the dynamic values ​​of five-dimensional psychological flexibility are updated in real time; the dynamic values, the progress of intervention implementation, and status data are integrated to form an intervention implementation dataset.

[0011] Furthermore, the method of determining the therapeutic tolerance window through a two-dimensional assessment and performing acceptance guidance and scenario fine-tuning includes: Based on the updated dynamic value of psychological flexibility and the intervention execution dataset, the dynamic value of experiential avoidance tendency and intervention participation are extracted to construct a two-dimensional evaluation mechanism; Furthermore, a dual-dimensional assessment mechanism is used to pre-define different assessment intervals based on threshold values, thereby determining the therapeutic tolerance window in real time. Based on the window determination results, the edge side coordinates and issues differentiated acceptance guidance instructions for each window. The VR terminal fine-tunes the virtual scene stimulation parameters as needed, and the edge side synchronously matches the window features to execute ACT acceptance guidance. Wearable devices provide sensory feedback in conjunction.

[0012] Furthermore, the methods for defining the threshold for manual intervention in extreme states and issuing early warnings include: Based on the dynamic value of psychological flexibility, the results of the therapeutic tolerance window determination, and the intervention execution dataset, extreme state monitoring indicators are extracted to form a monitoring indicator set; By combining absolute thresholds and relative change thresholds, a set of monitoring indicators is defined and calibrated to obtain the three-level artificial intervention thresholds for extreme states. The monitoring index values ​​are monitored in real time and compared with the threshold values ​​for manual intervention. Based on the comparison results, hierarchical early warnings are issued at the local terminal and the back-end management terminal. The entire process of intervention and early warning records is integrated and packaged into an enhanced fuzzy context package, and a summary of safety interventions is generated simultaneously.

[0013] Furthermore, the methods of updating the user's psychological activity development trajectory file, comparing the intervention effect and optimizing the intervention rules, and linking with manual guidance to adapt to subsequent intervention task packages include: Based on the enhanced fuzzy context package, the data is reconstructed into a structured dataset according to the intervention stage, and then incrementally updated to the user's psychological flexibility development trajectory file according to the timeline and intervention nodes. By combining longitudinal quantitative comparison and horizontal collaborative analysis, the effectiveness of this intervention is quantitatively evaluated, weak dimensions are identified, core intervention rules are iteratively optimized, and an optimized personalized intervention rule library is formed. By combining the updated development trajectory file with the quantitative evaluation results of intervention effects, and linking with the human counseling end to conduct reverse adaptation analysis, subsequent human-customized intervention needs are generated. By integrating manually customized needs with an optimized personalized intervention rule base, subsequent personalized intervention task packages are generated, forming a closed-loop cultivation link for psychological flexibility.

[0014] The technical effects and advantages of the virtual interactive scenario fuzzy logic control system for psychological counseling of college students proposed in this invention are as follows: This invention integrates ACT therapy with fuzzy logic control technology to create a precise, dynamic, and closed-loop personalized virtual psychological counseling system.

[0015] First, through multi-source data collection and fuzzy logic analysis, a five-dimensional set of fuzzy variables for psychological flexibility is generated and a baseline profile is established to accurately identify psychological shortcomings. Based on this, a personalized ACT narrative chain is assembled and an exclusive intervention task package is generated to achieve personalized intervention for each individual from the root cause. Secondly, a two-dimensional assessment mechanism is constructed to determine the treatment tolerance window in real time, implement differentiated acceptance guidance and scenario fine-tuning, and combine it with a three-level extreme state warning to achieve dynamic control of intervention intensity and balance the effectiveness and safety of guidance. Furthermore, the data from the entire process is integrated and packaged into an enhanced fuzzy context package, the psychological development trajectory file is updated and the intervention rules are optimized, and the manual counseling is linked to adapt to the subsequent task package, forming a closed-loop training link of "intervention-feedback-iteration-customization" to continuously improve the guidance effect; Meanwhile, the differentiated task division between edge and cloud ensures the real-time immersive experience of virtual intervention, and data privacy is protected through encryption technology. It is suitable for large-scale psychological counseling scenarios in universities and has the advantages of practicality, security and scalability. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the fuzzy logic control system for virtual interactive scenarios for psychological counseling of college students according to the present invention. Figure 2 This is a flowchart illustrating the process of generating a fuzzy variable set in the virtual interactive scenario fuzzy logic control system for psychological counseling of college students according to the present invention. Figure 3 This is a schematic diagram of the fuzzy logic control system for virtual interactive scenarios for psychological counseling of college students, as described in this invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example 1

[0018] Please see Figure 1 and Figure 2 As shown in this embodiment, the virtual interactive scenario fuzzy logic control system for psychological counseling of college students includes: Baseline variable module: Acquires user physiological signals, interaction behavior sequences and ACT microscale responses, maps preliminary membership of experiential avoidance tendency through ACT microscale, and positively corrects through interaction behavior sequences and physiological signals, generating a five-dimensional psychological flexibility fuzzy variable set, and simultaneously encrypts and establishes a core weakness baseline profile. Task package generation module: Based on fuzzy variable set and baseline file, it identifies variable features and parses macro-objectives into micro-control target sequences. Then, it matches and assembles personalized ACT narrative chains from the ACT narrative atom library, allocates edge-cloud tasks and differentiated control division of labor between VR terminal and edge side, and generates intervention task packages. Intervention data module: Based on the micro-control target sequence, personalized ACT narrative chain and intervention task package, the VR terminal triggers visual cognitive dissociation intervention, links with the sensory feedback of wearable devices, and generates updated dynamic values ​​of psychological flexibility and intervention execution data in real time. Tolerance Encapsulation Module: Based on updated dynamic values ​​and intervention execution data, it assesses the therapeutic tolerance window through a two-dimensional evaluation, performs acceptance guidance and scenario fine-tuning, defines the threshold for artificial intervention in extreme states and issues warnings, and integrates and encapsulates the records into an enhanced fuzzy context package. Closed-loop iteration module: Based on the enhanced fuzzy context package, update the user's psychological activity development trajectory file, compare the intervention effect and optimize the intervention rules, link with human counseling to reverse adapt to subsequent intervention task packages, and form a closed-loop training link.

[0019] Methods for generating fuzzy variable sets include: Wearable devices (such as heart rate belts and skin conductance wristbands) are used to collect users' physiological signal data in real time. The core physiological indicators collected are heart rate variability (HRV, extracting time-domain indicators SDNN and RMSSD) or skin conductance response (EDA, recording peak value and rise time), as well as respiratory rate. Through VR devices and terminal backends, all user interaction behaviors in virtual scenes are recorded. Specific dimensions include: gaze duration (focusing on virtual stimuli for ≥3 seconds is considered a valid gaze, and gaze focus stability is recorded), ACT-guided trigger behaviors (such as whether the user actively clicks the VR controller button to trigger acceptance or disengagement guidance, guiding changes in the user's breathing rate, etc.), and high-stimulation scene behaviors (referring to performance in high-stimulation exposure scenes, such as whether an exit command is triggered in a virtual speech scene, and the duration of task persistence). Interaction behavior data is stored in a triplet format of time-behavior type-behavior parameter to generate an interaction behavior sequence. At key decision points in the virtual scene (2 seconds before the virtual audience asks a question, 3 seconds after the task feedback), an embedded ACT microscale is automatically triggered by a pop-up window. Only one choice question is displayed per round, for example: "Faced with the current tension, you are more inclined to: A. Try to get rid of this feeling B. Allow it to exist and continue the task." After the user selects using a VR controller or other device, the option is recorded immediately. The ACT microscale response data is associated with scene nodes and physiological signal timestamps to avoid data disconnect. It should be noted that ACT refers to Acceptance and Commitment Therapy. The ACT Microscale focuses on psychological flexibility in ACT theory (an individual's ability to flexibly adjust their thinking, emotions, and behaviors according to their own values ​​in the current situation). The ACT Microscale is not a single fixed scale, but a general term for a class of tools. The specific content needs to be determined in conjunction with its corresponding original scale (such as the microscale version of the Acceptance and Action Questionnaire AAQ, the simplified version of the Value Questionnaire VIQ, etc.). The assessment dimensions and applicable populations of different microscales may differ. Map the ACT microscale response to the initial membership value of empirical avoidance tendency (EA). For example: the initial membership value of empirical avoidance tendency (hereinafter referred to as the initial membership value of EA) for choosing option A (trying to get rid of negative feelings) is 0.7 (high); the initial membership value of EA for choosing option B (allowing to exist) is 0.3 (low). Then, based on the principle of consistency between behavior and subjective response, and taking the interactive behavior sequence as the basis, the high-stimulation scenario behavior and the guiding trigger behavior are used to make positive corrections according to their high and low priorities. The data of high-stimulation scenario behavior is used first (60% weight), supplemented by the data of guiding trigger behavior (40% weight). The specific rules for positive correction processing are as follows: High-priority correction: If the user does not trigger the exit command in a high-stimulation scenario and the task duration is ≥60 seconds, the initial membership degree of EA will be reduced by 0.1-0.2 regardless of the ACT microscale mapping value (e.g., the initial value of 0.7 will be corrected to 0.5-0.6). Low priority correction: If a user actively triggers the ACT guidance function ≥ 2 times, and insists on completing the task after each guidance, the initial membership degree of EA will be reduced by an additional 0.05, and the minimum membership degree of EA after correction will not be lower than 0.2; Based on the built-in mapping rule library of physiological signals and fuzzy membership degrees, physiological signals are converted into fuzzy membership degrees. The rules in the mapping rule base adopt a three-dimensional mapping relationship of indicator range - membership value - corresponding psychological dimension. Example of a core rule: If the physiological indicator is heart rate variability, then the heart rate variability recovery rate (the time for HRV to recover to the baseline level) is the primary factor: <30 seconds, then the fuzzy membership degree of the physiological rigidity index is 0.3 (low); 30-60 seconds, then the fuzzy membership degree is 0.5 (medium); >60 seconds, then the fuzzy membership degree is 0.8 (high). If the physiological indicator is skin conductance response, then the peak value of the skin conductance response is the main factor: ≥2μS, then the fuzzy membership degree of the physiological excitation index is 0.7 (high); 1-2μS, then the fuzzy membership degree is 0.4 (medium); <1μS, then the fuzzy membership degree is 0.2 (low). The deviation between the corrected EA initial membership degree and the fuzzy membership degree is checked. If the deviation between the corrected EA membership degree and the physiological rigidity index membership degree is ≤0.2, the check is deemed to pass and the corrected EA initial membership degree is taken as the final EA membership degree. If the deviation is >0.2, the check is deemed to fail and the fuzzy membership degree of the physiological signal is used as the standard to fine-tune the EA initial membership degree to ensure data consistency. Based on EA membership, combined with the ACT microscale and interactive behavior sequence, a fuzzy variable set of five dimensions of psychological flexibility—cognitive integration, experiential avoidance tendency, present contact ability, value clarity, and commitment to action—is generated through a five-dimensional weighted cross-analysis mechanism. Specifically, the generation logic of each dimension of the five-dimensional weighted cross-analysis mechanism is as follows: Cognitive integration: Calculated by weighted summation of negative thought identification tendency (weight 40%), behavioral delay duration (weight 30%), and physiological arousal index (weight 30%) from the ACT microscale. Empirical avoidance tendency: The modified EA membership degree is directly used as the final value; Current contact ability: calculated based on the stability of eye focus (weight 50%) and respiratory guidance compliance (weight 50%). Respiratory guidance compliance is determined by the matching degree between the monitored respiratory rate and the rhythm of respiratory guidance. Value clarity: Based on the user's past profile (initial value of 0.4 is assigned upon first use, and subsequently updated through interaction); Commitment to action: Calculated based on task completion rate (weight 60%) and number of proactive attempts (weight 40%).

[0020] Finally, a standardized set of fuzzy variables for the five dimensions of psychological flexibility is generated, with the following format example: Cognitive fusion: 0.75, experiential avoidance tendency: 0.55, present contact ability: 0.40, value clarity: 0.40, commitment to action: 0.35; variable generation descriptions are generated simultaneously, indicating the data source and weight of each dimension. Encryption algorithms (such as AES-256) are used to encrypt the fuzzy variable set, the collected raw data, and the variable generation instructions, and user identity information is replaced with a virtual ID to avoid the leakage of real identity. After encryption, a baseline profile of psychological flexibility is established to identify core weaknesses. The core content includes: virtual user ID, fuzzy variable set, core weakness label (variable value ≥ 0.7 is considered a high weakness, 0.4-0.7 is medium, and < 0.4 is a low weakness), data collection time, and device information. The baseline profile is named according to user ID-collection batch, and an association index of variable set-raw data-encryption key is established to support subsequent traceability queries.

[0021] Methods for identifying variable characteristics and analyzing macroscopic objectives into a sequence of microscopic control objectives include: Based on the fuzzy variable set and baseline archive, the fuzzy variable set is bound to user ID and core weakness annotation information through the association index of the campus private cloud database; By using predefined threshold rules, the five-dimensional variable values ​​of the fuzzy variable set are hierarchically bound and feature extracted to identify the variable characteristics and core shortcomings of users' psychological flexibility. The core threshold rules and operations are as follows: First, set a grading threshold: a variable value ≥ 0.7 is considered high dimension (corresponding to weak psychological function), 0.4-0.7 is medium dimension (moderate function), and < 0.4 is low dimension (good function). Then, by combining the original core weakness records in the baseline file, the variable classification results are cross-validated. For example, if the baseline is labeled as high cognitive fusion, and the current cognitive fusion is 0.75, then the weakness feature is reinforced and the label "continued high cognitive fusion, priority intervention required" is added. Finally, feature combination analysis was conducted to identify the correlation features between variables as variable features. For example, the combination of "high EA (0.68) + low present contact ability (0.32)" was identified as "avoidance cognitive pattern"; the combination of "low value clarity (0.38) + low commitment to action (0.35)" was identified as "value-driven insufficiency pattern". The variables are sorted according to the number of high-dimensional variables and their impact weights, and only 1 to 2 core weaknesses are identified (impact weights are set as follows: cognitive integration and EA account for 40%, and the other three dimensions each account for 20%), which will serve as the core basis for subsequent target analysis. Based on expert experience, the macro-level objectives corresponding to the characteristics of variables and the core shortcomings are used to guide the psychological well-being of users. To enhance psychological flexibility, the macro-level objectives of this intervention are defined by combining the identified variable characteristics and core weaknesses with the pre-set guidance focus matching logic. The guidance in the example focuses on matching logic: if the core weakness is "high cognitive integration + high EA", the guidance of the macro goal is defined as "prioritizing the reduction of cognitive integration and experiential avoidance tendency, and establishing basic acceptance and dissociation ability"; If the core weakness is "low value clarity + low commitment to action", the focus of the intervention is defined as "strengthening value awareness, stimulating proactive commitment to action, and enhancing value-driven capabilities". If the variable features do not have obvious shortcomings in high dimensions (most are medium dimensions), the guidance should be defined as "balancedly improving the psychological flexibility of each dimension and optimizing the coordination of psychological functions"; Based on the defined focus of guidance, and following the principles of gradual progress, priority matching, and effective intervention, the guidance focus of macro-level objectives is broken down into a sequence of phased micro-level control objectives. Specifically: First, the decomposition principle is defined as follows: each micro-control target must correspond to a single ACT intervention direction (such as cognitive dissociation, acceptance guidance, and value clarification), and can be achieved through atomic assembly of subsequent narrative chains to avoid abstraction; Then, following the gradual principle of "initial adaptation - mid-term reinforcement - final consolidation," a three-stage breakdown is carried out; taking the core weakness "high cognitive integration + high EA" as an example: Initial goal (adaptation phase): Reduce cognitive fusion to below 0.65 and decrease avoidance behavior in EA-positive trigger scenarios; Mid-term goal (reinforcement phase): Master basic cognitive dissociation skills, stabilize the EA value below 0.6, and improve current contact ability to above 0.45; Final goal (consolidation phase): Actively accept discomfort in low-stimulation scenarios, stabilize cognitive integration and EA value in the mid-dimensional range, and form a basic response pattern from awareness to acceptance. Next, the micro-control objectives at each stage are assigned priority (Level 1 is the highest, Level 2 is medium, and Level 3 is the lowest). The intervention objectives corresponding to the core weaknesses are set as Level 1, and the auxiliary improvement objectives are set as Level 2-3. Finally, the phased and prioritized micro-control objectives are integrated along a timeline to generate a standardized and actionable sequence of micro-control objectives, with an example format: "[Phase 1 (Priority 1): Reduce cognitive integration to below 0.65; Phase 2 (Priority 1): Stabilize EA value below 0.6; Phase 2 (Priority 2): Increase current contact ability to above 0.45; ...]", providing guidance for the subsequent assembly of personalized ACT narrative chains.

[0022] Ways to assemble personalized action narrative chains include: A pre-built ACT narrative atom library is constructed and stored according to the six core processes of ACT theory (cognitive dissociation, acceptance, engagement with the present, value clarification, commitment to action, and self as context). Each narrative atom carries standardized attribute tags to ensure matching accuracy. The core attributes of a narrative atom include: intervention dimension label, difficulty level, target audience, interaction format (VR visualization, voice guidance, gesture interaction, etc.) and duration range (15-90 seconds / atom, suitable for fragmented guidance). Among them, the intervention dimension label is the ACT intervention direction, and the difficulty level is set according to the intervention difficulty gradient (adapting to the priority of micro-control goals, level 1-3, level 1 is low challenge and level 3 is high challenge), and the core shortcomings of psychological flexibility are used as the adaptation goals (such as "cognitive dissociation atom" adapting to the "high cognitive integration" goal). Based on the core attributes of narrative atoms, standardized narrative atoms are pre-defined and integrated to form the ACT narrative atom library; Examples of narrative atoms in the ACT Narrative Atom Library: Cognitive Dissociation Atom (ID: DIS-001, Difficulty Level 1, Adaptation Objective: "Reduce Cognitive Fusion", Interaction Format: "Visualization of Thought Pumice + Gesture to Push Away", Duration: 30 seconds); Acceptance Atom (ID: ACC-002, Difficulty Level 2, Adaptation Objective: "Reduce Experiential Avoidance", Interaction Format: "Guided Mindfulness Breathing + Voice Acceptance Prompt", Duration: 45 seconds). Then, based on the micro-control target sequence, narrative atoms are matched from the ACT narrative atom library according to the intervention dimension label and difficulty gradient; The matching rules are as follows: Prioritize matching narrative atoms corresponding to Level 1 priority goals (e.g., for the core goal "reducing cognitive integration", prioritize retrieving cognitive dissociation atoms); refine matching according to user's weakness variable characteristics (e.g., "high avoidance + low current contact ability", match with acceptance + current contact atoms); follow difficulty gradient adaptation (initial goals correspond to Level 1 atoms, mid-term goals correspond to Level 2 atoms, and final goals correspond to Level 3 atoms, avoiding difficulty jumps); The matching priority is sorted as follows: intervention dimension adaptability > difficulty gradient > user past interaction preferences (first use defaults to sort by difficulty). If there are multiple adaptable atoms, the selection is supplemented by the diversity of interaction forms (to avoid duplication of similar interactions). Finally, based on the principles of logical coherence, progressive difficulty, and closed-loop intervention, the matched narrative atoms are dynamically combined according to the timeline to form a unique and personalized ACT narrative chain, avoiding the discontinuity caused by the accumulation of narrative atoms. The assembly order principle is as follows: the opening atom should prioritize level 1 difficulty and focus on the present moment (such as "mindfulness breathing guidance") to establish user scenario adaptability; the core intervention section should arrange the corresponding atoms in the order of "initial goal → intermediate goal → final goal", and the atoms in the same stage should be connected in the order of "low challenge → high challenge" (such as cognitive dissociation level 1 atom → cognitive dissociation level 2 atom); the closing atom should be paired with short-duration value clarification or commitment action atom (such as "value micro-reflection") to reinforce the intervention memory points; It should be noted that the core logic of Acceptance and Commitment Therapy (ACT) is to help individuals break out of the ineffective cycle of "eliminating negative emotions" and instead focus on "identifying values ​​and taking effective action." Its six core elements can be briefly analyzed into the following interrelated dimensions: Acceptance, cognitive defusion, present-moment awareness, self-as-context, values, and committed action. In short, the essence of the six core principles of ACT is to adjust one's attitude towards one's own experiences (acceptance, dissociation), anchor oneself in the present moment and one's true self (present focus, personal background), and then initiate value-oriented actions (value, commitment to action) to ultimately achieve "psychological flexibility," that is, to be able to move in the direction of the life one desires regardless of negative experiences.

[0023] Methods for generating intervention task packages include: Based on the personalized ACT narrative chain and the micro-control target sequence, the interactive attributes and execution requirements of each narrative atom are analyzed synchronously; the interactive forms (VR visualization, voice guidance), latency requirements (high real-time performance, regular synchronization), and resource dependencies (scene models, voice files) of the narrative atoms are extracted in detail to form a lookup table of atom IDs, execution requirements, and target associations, which serves as the core basis for task allocation; Among them, the definition of high real-time requirements is as follows: VR visual interaction (such as thought float, gesture interaction) and dynamic adjustment of scene parameters (gradual light change, viewpoint switching) must meet the response latency of ≤20ms, which are classified as high-priority execution requirements. Standard requirement definition: Voice guidance, data synchronization, and intervention rhythm coordination have low latency requirements (≤100ms) and are classified as medium-low priority requirements. Based on the execution requirements of narrative atoms, differentiated division of labor is carried out according to the execution tasks of various narrative atoms, and the tasks are allocated to VR terminals, edge terminals and cloud terminals according to the execution requirements. Among them, the VR end division of labor (high real-time interactive main controller): undertakes the execution tasks of all VR visualization-type narrative atoms, including loading the corresponding scene model, parsing gesture interaction rules, and executing scene parameter adjustment instructions (such as lighting and viewing angle), while also being responsible for collecting and temporarily storing user interaction behavior data in real time; prioritizing the allocation of core intervention atoms such as cognitive dissociation and commitment action (such as the "thought pumice visualization" atom) to ensure the smoothness of core interventions. Edge-side division of labor (intervention coordination and auxiliary execution nodes): responsible for voice-guided atomic execution (such as acceptance guidance voice, transition prompts), intervention chain rhythm calibration (synchronizing the execution progress of VR and wearable devices), real-time data processing (filtering invalid behavioral data), while backing up intervention rules and dealing with temporary VR terminal failures; allocating auxiliary intervention atoms such as acceptance and contact with the present moment to form an intervention closed loop in coordination with VR terminal. Cloud-based division of labor (data storage and resource backup node): It does not participate in real-time intervention execution, but only stores the complete version of the ACT narrative atom library, user baseline files, and backups of intervention task packages. At the same time, it responds to data synchronization requests from the edge side to ensure data security and traceability and avoid local data loss. The tasks assigned through differentiated division of labor are transformed into standardized executable instructions, ensuring that the content, triggering conditions, and execution sequence of the instructions are clear and that priorities are marked to avoid resource conflicts. Among them, the executable command generation example is: VR-side command "Load the cognitive dissociation atom with ID DIS-001. When a negative voice statement from the user is detected, trigger the visualization of the thought pumice and simultaneously dim the background light to 30%"; edge-side command "2 seconds after the VR-side pumice visualization is started, play the acceptance guidance voice for 45 seconds". Priority rules for executable instructions: Core intervention-type atomic instructions are set to level 1 (highest), transition instructions and data synchronization instructions are set to level 2, and backup and log recording instructions are set to level 3, to ensure that core interventions are executed first; The personalized ACT narrative chain, the set of executable instructions, the corresponding rule set, and user ID are encrypted and packaged into a unified format ACT intervention task package. The core content of the intervention task package includes: complete information on the personalized ACT narrative chain, a list of executable instructions, a set of rules (such as preset VR interaction rules and voice playback rules), and metadata (user ID, data encryption identifier, and task package generation timestamp).

[0024] The ways in which VR-based visual cognitive dissociation intervention is triggered and sensory feedback from wearable devices are as follows: Each execution terminal (VR terminal, wearable device, edge device) synchronously loads the ACT intervention task package. Then, based on the micro-control target sequence, personalized ACT narrative chain and intervention task package, the trigger conditions for visual cognitive dissociation intervention are preset, and the VR terminal interaction logic and user behavior are bound together to form intervention rules. For example, set "detecting negative voice statements from users (via VR microphone)" and "holding the controller in a specific area for ≥2 seconds" as trigger conditions, bind the complete interaction process of "thought float generation → floating → gesture push away → dissipation", and at the same time associate the linkage rules of scene parameters (background light focuses on the float area when the float is generated). When user behavior meets the preset triggering conditions, the VR terminal initiates visual cognitive dissociation intervention, and executes the intervention step by step according to the ACT narrative chain, guiding the user's operation through interactive logic; The core process is illustrated by the "thought pumice" atom: The first step is to trigger the response: if a negative statement from the user is detected (such as "I can't do it well"), a 3D pumice model of the corresponding text is generated within 100ms. The pumice slowly floats from the edge of the scene to the center of the user's field of vision. The color and size of the pumice change dynamically with the user's physiological signals (skin conductance response) (the higher the skin conductance value, the darker the color of the pumice). The second step is interactive guidance: the user is guided to operate through voice prompts (synchronously sent from the edge side): "Try to push away these thoughts with the controller", while activating the vibration feedback of the VR controller and prompting the interactive area; The third step is real-time feedback: When the user touches the floating stone with the controller, the floating stone produces a particle dissipation effect, and a weakened sound effect is played simultaneously; if the user successfully pushes the floating stone away, the VR terminal displays a "completed" prompt, enhancing the sense of accomplishment from the intervention; if the user does not respond, the guidance prompt is repeated after 5 seconds, up to 2 times. The fourth step is scene reset: After the single narrative atomic intervention ends, the VR terminal quickly resets the scene (the floating stone model is cleared and the lighting is restored to its basic state) to prepare for the next atomic intervention and ensure a smooth connection of the narrative chain; Simultaneously, wearable devices provide real-time sensory feedback based on the coordinated and synchronized intervention rhythm at the edge, forming a collaborative closed loop with VR-end intervention to enhance immersion and intervention effectiveness, including: Respiratory vibration feedback: During cognitive dissociation intervention, the wearable wristband vibrates at a preset rhythm (e.g., 4 seconds of inhalation and 6 seconds of exhalation) to guide the user to adjust their breathing and help reduce physiological excitability. The vibration intensity is dynamically adjusted according to the user's HRV value (the lower the HRV, the more obvious the vibration). Haptic feedback linkage: When the user completes a key interaction (such as pushing away a floating stone), the bracelet vibrates twice as positive feedback; if the user exhibits avoidance behavior (such as turning off the VR view), the bracelet vibrates continuously for 3 seconds as a prompt, and at the same time, the edge side is simultaneously triggered to optimize guidance. The edge side compares the VR intervention progress with the wearable device feedback timing in real time to ensure that the feedback delay is ≤20ms. If a deviation occurs, the wearable device command timing is automatically fine-tuned to ensure coordination consistency. The edge side monitors the intervention status of the VR terminal and wearable device throughout the process, and is responsible for overall rhythm coordination and emergency response. It issues intervention progress instructions synchronously according to the ACT narrative chain to avoid the VR terminal and wearable device intervention becoming disconnected (e.g., when the VR terminal generates a floating stone, the wristband vibration is triggered synchronously for guidance). If the VR device experiences lag (e.g., response latency > 500ms), the edge device immediately pauses wearable device feedback, issues a "temporary buffer" voice prompt, and attempts to restart the VR intervention process. If restarting fails, it switches to pure voice guidance mode to ensure uninterrupted intervention. Real-time data collection of intervention status from various terminals (such as interaction completion rate and feedback success rate) is stored locally to provide a foundation for subsequent data processing.

[0025] The methods for generating updated dynamic values ​​of psychological flexibility and intervention implementation data include: Through real-time communication channels between VR terminal, wearable device and edge device, real-time status data of device operation is captured and collected synchronously, including behavioral data of VR terminal, physiological data of wearable device and status data of edge device; VR-side behavioral data: Records all user interaction behaviors during the intervention process, including cognitive dissociation intervention results (such as the success rate of pushing away the pumice stone and the duration of the interaction response), scene operation behaviors (such as the gaze focus trajectory and the number of times the controller is triggered), avoidance behavior characteristics (such as whether to close the field of vision and attempt to exit the scene), and associates them with the corresponding narrative atom IDs. Wearable device physiological data: Continuously collects indicators such as heart rate variability (HRV), skin conductance (EDA), and respiratory rate, and simultaneously records the sensory feedback response status (such as whether vibration feedback is recognized), with the device number and collection time sequence attached, and is precisely aligned with the VR terminal behavioral data timestamp; Edge-side status data: Collect intervention status of each terminal (such as whether the VR terminal is lagging, whether the wearable device feedback is normal), intervention progress (currently executing which narrative atom, remaining time), and abnormal event records (such as the number of times the guidance prompt is repeated, feedback delay events); Outlier handling (using the 3σ principle to filter out abnormal values ​​in physiological data), timestamp alignment (calibrating the timestamps of VR and wearable device data based on the edge clock), and data format standardization (converting all data into JSON format, quantizing and encoding behavioral data, retaining two decimal places for physiological data, and labeling status data as normal or abnormal to form a standardized data set) are performed on the synchronously collected situational data. Based on the processed situational data and combined with the preset weighted calculation rules, the dynamic variable values ​​of the five-dimensional psychological flexibility in the fuzzy variable set are updated in real time. The core calculation logic and basis are as follows: Cognitive fusion: The real-time value is calculated by weighting and summing the interaction results of negative thoughts on the VR terminal (weight 50%, such as the higher the success rate of pushing away the pumice, the lower the value), the peak value of skin conductance response (weight 30%), and the duration of behavioral delay (weight 20%). It is updated once for each cognitive dissociation atom intervention completed. Experienced avoidance tendency (EA): The core criteria are the frequency of avoidance behaviors on VR devices (weight 60%, such as exit attempts and number of times the field of vision is closed) and the stability of breathing rate (weight 40%). If there are no avoidance behaviors and breathing is stable, the value is reduced, and vice versa. Current contact ability: Calculated based on eye focus stability (weight 50%, percentage of time spent focusing on the intervention target) and breathing guidance cooperation (weight 50%, matching degree of respiratory rate monitored by wearable device with guidance rhythm). The value is improved if the matching degree is ≥80%. Value clarity and commitment to action: temporarily updated based on task completion rate during intervention (weight 70%) and number of proactive attempts (weight 30%). If there is no corresponding atomic intervention in the current narrative chain, the values ​​at the previous moment will be maintained, and will be adjusted synchronously after the subsequent value and action-related atomic actions are executed. After each update, verify the rationality of the values ​​to ensure that the values ​​of each dimension are in the range of 0-1. If they are outside the range, correct them according to the nearest threshold (e.g., >1.0 is corrected to 1.0). At the same time, record the basis for the update and the source of the data. By integrating dynamic values, progress information of intervention execution, and situational data, an intervention execution dataset is formed. The core content includes: intervention progress data (number of executed atoms, task completion rate, and remaining time), interaction effect data (success rate of each atomic intervention, user interaction response efficiency, proportion of non-avoidance behavior, and number of times guidance prompts are repeated), device status data (operating status of each terminal, abnormal event types, and processing results), and dynamic value change trajectory (initial value, real-time value, and change range of each dimension).

[0026] The methods for determining the therapeutic tolerance window through a two-dimensional assessment, and for implementing acceptance guidance and scenario fine-tuning, include: Based on the updated dynamic values ​​of psychological flexibility and the intervention execution dataset, the dynamic value of experiential avoidance tendency is extracted as the psychological tolerance dimension (the core target indicator of ACT acceptance intervention is the value of EA membership; the higher the value, the stronger the avoidance tendency towards discomfort and the lower the tolerance). Intervention participation is quantitatively extracted from the intervention execution dataset as the behavioral tolerance dimension. Intervention participation can be calculated by logically weighting the sum of task completion rate (60%) + interaction response efficiency (20%) + non-avoidance behavior percentage (20%), and the result is mapped to the 0-1 range (the higher the value, the higher the user's behavioral participation and tolerance to the intervention). A two-dimensional assessment mechanism is constructed based on psychological tolerance and behavioral tolerance dimensions. Then, based on the principle of moderate challenge and tolerability of intervention in the ACT theory, the two-dimensional assessment mechanism is used to pre-set intervals to define three assessment intervals: optimal, safe, and beyond the limit. The core threshold of the therapeutic tolerance window is also clarified to provide a quantitative basis for subsequent judgment. All thresholds need to be verified by previous clinical intervention data to adapt to VR immersive intervention scenarios. Specifically, the evaluation interval delineation rules (both are 0-1 intervals) and the specific thresholds for the examples of the two-dimensional evaluation mechanism are as follows: Optimal (target): Psychological tolerance threshold: 0.4~0.6; Behavioral tolerance threshold: 0.6~0.8; Core characteristics: moderate avoidance tendency, moderate participation, intervention is challenging but tolerable; Whether it is a therapeutic tolerance window: Yes (core window); Safety (Low Challenge): Psychological tolerance threshold: <0.4; Behavioral tolerance threshold: >0.8; Core characteristics: weak avoidance tendency, high participation, no significant challenge from intervention, no therapeutic benefit; Whether it is a therapeutic tolerance window: No; Exceeding limits (overtolerance): Psychological tolerance threshold: >0.6; Behavioral tolerance threshold: <0.6; Core characteristics: strong avoidance tendency, low participation, intervention challenges exceed tolerance range, easily generating resistance; Whether it is a therapeutic tolerance window: No; In a two-dimensional threshold, if a single-dimensional indicator is in the optimal range and the other dimension is in an adjacent range, it is determined to be a quasi-optimal window, requiring only minor adjustments without changing the core intervention strategy. On the marginal side, following the rhythm of single-narrative atomic intervention completion and judgment, the values ​​of dual-dimensional indicators are substituted into the dual-dimensional assessment mechanism to perform real-time determination of the treatment tolerance window. Specifically: First, extract the dynamic value of the EA membership and intervention participation of the current intervention node (the stage node when intervening through a virtual scene), match the interval threshold of the dual-dimensional evaluation mechanism, and then determine whether the current node is in the optimal, near-optimal, safe, or beyond-limit interval according to the principle of interval matching priority and near-optimal fallback, and clarify whether it is within the treatment tolerance window; finally, bind the determination results (including window type, indicator value, and matching basis) with timestamp and atomic ID, temporarily store them in the intervention execution dataset, and synchronize them to the VR terminal and wearable device to provide instruction basis for subsequent strategy execution; After each acceptance guidance and scenario fine-tuning is completed, the window is immediately re-judged before collecting the next round of dual-dimensional indicator values ​​to achieve a closed loop of "judgment-execution-rejudgment". Based on the window determination results, the edge side coordinates and issues differentiated acceptance guidance instructions for each window. The VR terminal fine-tunes the virtual scene stimulation parameters as needed (only adjusting the stimulation-related parameters without changing the core narrative atomic logic). The edge side simultaneously executes ACT acceptance guidance (matching window features to match guidance content and form), and wearable devices provide sensory feedback in conjunction. Its core guidance and fine-tuning principles are: maintaining the optimal window, increasing the challenge within the safe window, and reducing the challenge within the excessive window. The specific differentiated execution is as follows: For the optimal or near-optimal window (maintenance strategy, mildly reinforced acceptance guidance): VR scene fine-tuning: Keep the stimulation parameters of the current VR scene (such as light brightness, number of background interference elements, and field of view) unchanged, and only calibrate the scene synchronization to ensure smooth intervention; Acceptance guidance: Lightly enhanced voice guidance is sent from the edge side (aligning with the core of ACT acceptance, such as "Try to feel the discomfort in the present moment, it is just a feeling, there is no need to avoid it"), lasting 5-10 seconds, and embedded in the narrative atom transition. Wearable device linkage: Output vibration feedback according to the basic breathing rhythm (such as 4 seconds of inhalation and 6 seconds of exhalation) to help users maintain the current contact state; For the safety window (low challenge, increased scene stimulation, targeted acceptance guidance): VR scene fine-tuning: moderately increase the scene stimulation (such as adding 1-2 slight background interference elements, brightening the scene lighting, and expanding the field of view), with the fine-tuning range ≤30%, to avoid excessive stimulation; Acceptance guidance: Targeted exploratory guidance is sent from the edge side (such as "Try to feel a slight discomfort in this scene, and continue to complete the operation with it, you can do it"), lasting 10-15 seconds, and is triggered synchronously with VR scene fine-tuning; Wearable device integration: Appropriately increase the intensity of vibration feedback (20% higher than the base intensity) to guide users to focus on discomfort in the scenario and establish acceptance; For exceeding the limit window (high tolerance, reducing scene stimulation, basic soothing acceptance guidance): VR scene fine-tuning: Immediately reduce scene stimulation (e.g., remove all background distracting elements, dim the light to a soft state, narrow the field of view to focus on the core interactive area), and simplify the scene 3D model if necessary to reduce visual load; Acceptance guidance: Basic soothing guidance is given from the edge side (prioritizing mindfulness breathing guidance, such as "first adjust your breathing by following the vibration of the bracelet, feel the air going in and out of your nasal cavity, relax slowly, and the discomfort will gradually decrease"), lasting 15-20 seconds. Pause the subsequent high-challenge interaction and complete the acceptance guidance first. Wearable device linkage: Output low-frequency vibration feedback according to slow breathing rhythm (such as 5 seconds of inhalation and 7 seconds of exhalation), simultaneously collect heart rate variability data, and provide real-time feedback to the edge side to determine the user's relaxation state; After completing the acceptance guidance and scenario fine-tuning, the edge side triggers short-term effect verification, collects the dynamic value of the EA membership degree and intervention participation of the next intervention node, and verifies the effect of strategy execution. If the validation indicators show a regression to the optimal or near-optimal window (e.g., after adjusting the overlimit window, the dynamic value of EA drops to 0.55 and the intervention participation rate rises to 0.65), it is considered effective, and the current adjusted strategy is maintained; if it does not regress (e.g., EA is still > 0.6 and the intervention participation rate is < 0.6), it is considered ineffective, and it is adjusted again according to the principle of gradient fine-tuning (e.g., when fine-tuning the overlimit window for the second time, the scene stimulation is further reduced and the duration of soothing guidance is increased); If the user fails to return to the tolerance window after two consecutive fine-tunings, a "temporary buffer instruction" is issued on the edge side. The VR device switches to a low-stimulation, quiet scene, and the edge side performs 50 seconds of basic mindfulness acceptance guidance. Intervention is continued only after the user's indicators recover, in order to avoid intervention resistance.

[0027] Methods for defining and issuing warnings regarding human intervention thresholds in extreme states include: Based on the dynamic value of psychological flexibility, the results of the therapeutic tolerance window determination, the intervention execution dataset, and the equipment operation status data, the data is bound to the intervention node by user ID. Then, monitoring indicators for extreme states are extracted, including the dynamic value of EA membership (referred to as EA dynamic value, a core psychological indicator, the higher the value, the stronger the avoidance tendency), intervention participation (a core behavioral indicator), the proportion of avoidance behavior frequency (number of avoidance behaviors during the intervention ÷ total number of interactions), and the number of intervention interruption attempts (standardized and quantified in the range of 0-1, with no attempt being 0 and multiple attempts approaching 1). Then, the indicator values ​​marked as abnormal periods in the equipment operation status data are removed and replaced with the average of the effective indicators of the same intervention node to form a monitoring indicator set. Combining the clinical safety standards of ACT immersive VR intervention and the stimulation characteristics of VR scenes, a set of monitoring indicators was defined and calibrated using a combination of absolute and relative change thresholds. This resulted in a three-tiered threshold for artificial intervention in extreme states (Level 1 warning: risk alert; Level 2 warning: preparation for artificial intervention; Level 3 warning: immediate artificial intervention). The specific definition rules are as follows: Absolute threshold: A critical value is set for a single monitoring indicator. When any monitoring indicator reaches this value, the corresponding warning is triggered (e.g., EA dynamic value ≥ 0.8, intervention participation ≤ 0.3, avoidance behavior frequency ≥ 0.7, intervention interruption attempt ≥ 0.9, all directly trigger a level three warning). Relative change threshold: A critical value is set for the dynamic change of the monitoring indicator. If the change of a single monitoring indicator is ≥50% within one intervention node (e.g., the dynamic value of EA suddenly rises from 0.5 to 0.8), and another monitoring indicator reaches the critical value, a corresponding warning is triggered (e.g., a sudden increase in EA + a decrease in intervention participation triggers a level 2 warning). The values ​​of monitoring indicators are monitored in real time and compared with thresholds for manual intervention. Based on the comparison results, tiered early warnings are issued at both the local terminal and the backend management end. Example of specific execution logic: Level 1 alert: A lightweight alert notification pops up on the backend management terminal, and the alert trigger indicators and intervention nodes are recorded simultaneously. The edge side maintains the current intervention strategy, only increasing the frequency of indicator monitoring, without the need for manual intervention. Level 2 warning: A prominent alarm pops up on the backend management terminal and pushes the warning log. The VR terminal displays a light prompt icon in the corner of the scene. The wearable device emits a low-frequency vibration reminder. The scene stimulation intensity is paused on the edge side to maintain the current intervention rhythm. The human operator is notified to prepare for intervention. Level 3 warning: The backend management terminal triggers an audio-visual alarm and continuously pushes warning information. The VR terminal immediately suspends high-challenge intervention content and automatically switches to a low-stimulation quiet scene. The wearable device emits high-frequency continuous vibration. The edge side suspends the execution of all intervention strategies and triggers a manual intervention command until the user's status is confirmed by the human to resume or terminate the intervention. Once an alert is triggered, the alert level, triggering reason, handling actions, and status recovery results are recorded in real time and synchronously updated to the intervention execution dataset to form a complete alert record. After completing the definition of early warning thresholds and real-time monitoring, the relevant data and early warning records of the entire intervention process are integrated to form a full-process intervention record, which includes records at five levels: the basic information layer (user ID, number of intervention task package, intervention start and end time, and number of each terminal device), the dynamic data layer (trajectory of dynamic value changes in the five dimensions of psychological flexibility, real-time value of intervention participation, and full record of core monitoring indicators), the intervention execution layer (results of tolerance window determination, content of acceptance guidance strategy, VR scene fine-tuning parameters, and completion status of each intervention node), the safety early warning layer (early warning classification record, handling actions, and execution results of manual intervention (if any)), and the equipment status layer (full-process equipment operation status data and abnormal event handling records). The entire process of intervention records, baseline files, and the initial set of fuzzy variables are integrated and packaged into an enhanced fuzzy context package.

[0028] Updating user psychological activity development trajectory files, comparing intervention effects and optimizing intervention rules, and linking with human counseling to adapt subsequent intervention task packages include: Based on the enhanced fuzzy context package, the extracted data is reconstructed into a structured dataset according to the format of intervention stage, core indicators and related intervention tasks; Then, the structured dataset is incrementally updated to the user's psychological flexibility development trajectory archive in the form of time axis + intervention nodes; By comparing the changes in core indicators between the current intervention and the baseline or previous interventions longitudinally and by conducting cross-sectional collaborative analysis (coordinated changes in the five dimensions of psychological flexibility and the degree of matching between intervention tasks and effects), the effectiveness of this intervention is quantitatively evaluated, and the weakest dimension among the five dimensions of psychological flexibility of users is identified. Based on the evaluation results, the core rules related to the intervention (thresholds of the dual-dimensional evaluation mechanism, rules for scenario fine-tuning, rules for acceptance guidance matching, thresholds for manual intervention, etc.) are iteratively optimized and synchronized to the cloud to form an optimized personalized intervention rule library. The cloud pushes the updated development trajectory file and quantitative evaluation results of intervention effects to the human counseling end. Based on the received data, professional counselors focus on personalized aspects that cannot be solved by computer algorithms, such as the user's weak psychological flexibility and subjective intervention suitability. They conduct reverse adaptation analysis to clarify the focus of subsequent intervention, the requirements for scene or content adjustment, the difficulty gradient, and the need for human counseling collaboration. This generates customized human intervention requirements and submits them to the cloud. The cloud integrates manually customized needs with an optimized personalized intervention rule library, combines user development trajectory files, and generates customized ACT intervention task packages that are suitable for users according to the generation specifications of intervention task packages. After encryption, the packages are synchronously stored in the cloud and pushed to each intervention terminal. This forms a complete closed-loop training chain for psychological flexibility, which involves online VR intervention, followed by the generation of an enhanced fuzzy context package, then updating the profile or optimizing the rules, then reverse adaptation through human guidance, then customizing subsequent task packages, and finally conducting online VR intervention again. The chain iterates continuously with each round of intervention. Example 2

[0029] Please see Figure 3 As shown, the parts not described in detail in this embodiment are described in Embodiment 1. A fuzzy logic control method for virtual interactive scenarios in psychological counseling for university students is provided, including: S1: Obtain user physiological signals, interaction behavior sequences and ACT microscale responses, map the initial membership degree of preliminary empirical avoidance tendency through ACT microscale, and make positive corrections through interaction behavior sequences and physiological signals to generate a five-dimensional psychological flexibility fuzzy variable set, and simultaneously encrypt and establish a core weakness baseline file. S2: Based on the fuzzy variable set and baseline file, identify variable features and analyze macro-objectives into a sequence of micro-control objectives. Then, match and assemble personalized ACT narrative chains from the ACT narrative atom library, allocate edge-cloud tasks and differentiated control division of labor between VR terminal and edge side, and generate intervention task packages. S3: Based on the micro-control target sequence, personalized ACT narrative chain and intervention task package, the VR terminal triggers visual cognitive dissociation intervention, links with the sensory feedback of wearable devices, and generates updated dynamic values ​​of psychological flexibility and intervention execution data in real time. S4: Based on the updated dynamic values ​​and intervention execution data, the treatment tolerance window is determined through a two-dimensional assessment, acceptance guidance and scenario fine-tuning are performed, and the threshold for artificial intervention in extreme states is defined and an early warning is issued. The records are integrated and packaged into an enhanced fuzzy context package. S5: Based on the enhanced fuzzy context package, update the user's psychological activity development trajectory file, compare the intervention effect and optimize the intervention rules, link with manual guidance to reverse adapt to subsequent intervention task packages, and form a closed-loop training link. Example 3

[0030] This embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the operation mode of the virtual interactive scene fuzzy logic control system for psychological counseling of college students described above.

[0031] Since the electronic device described in this embodiment is the one used to implement the fuzzy logic control method for virtual interactive scenes aimed at psychological counseling for college students in this application embodiment, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the fuzzy logic control method for virtual interactive scenes aimed at psychological counseling for college students described in this application embodiment. Therefore, how the electronic device implements the method in this application embodiment will not be described in detail here. Any electronic device used by those skilled in the art to implement the fuzzy logic control method for virtual interactive scenes aimed at psychological counseling for college students in this application embodiment falls within the scope of protection of this application.

[0032] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0033] The above description is merely a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for users of ordinary technical skills, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A virtual interactive scenario fuzzy logic control system for psychological counseling of college students, characterized in that, include: Baseline variable module: Acquires user physiological signals, interaction behavior sequences and ACT microscale responses, maps preliminary membership of experiential avoidance tendency through ACT microscale, and positively corrects through interaction behavior sequences and physiological signals, generating a five-dimensional psychological flexibility fuzzy variable set, and simultaneously encrypts and establishes a core weakness baseline profile. Task package generation module: Based on fuzzy variable set and baseline file, it identifies variable features and parses macro-objectives into micro-control target sequences. Then, it matches and assembles personalized ACT narrative chains from the ACT narrative atom library, allocates edge-cloud tasks and differentiated control division of labor between VR terminal and edge side, and generates intervention task packages. Intervention data module: Based on the micro-control target sequence, personalized ACT narrative chain and intervention task package, the VR terminal triggers visual cognitive dissociation intervention, links with the sensory feedback of wearable devices, and generates updated dynamic values ​​of psychological flexibility and intervention execution data in real time. Tolerance Encapsulation Module: Based on updated dynamic values ​​and intervention execution data, it assesses the therapeutic tolerance window through a two-dimensional evaluation, performs acceptance guidance and scenario fine-tuning, defines the threshold for artificial intervention in extreme states and issues warnings, and integrates and encapsulates the records into an enhanced fuzzy context package. Closed-loop iteration module: Based on the enhanced fuzzy context package, update the user's psychological activity development trajectory file, compare the intervention effect and optimize the intervention rules, link with human counseling to reverse adapt to subsequent intervention task packages, and form a closed-loop training link.

2. The virtual interactive scene fuzzy logic control system for psychological counseling of college students according to claim 1, characterized in that, The methods for generating the fuzzy variable set include: Simultaneously collect user physiological signals, interactive behavior sequences in virtual scenarios, and ACT microscale responses of users at decision points in virtual scenarios; The ACT microscale response is mapped to the initial membership degree of empirical avoidance tendency, and combined with high-stimulation scene behavior and guided triggering behavior in the interaction behavior sequence, and positively corrected according to high and low priority. Physiological signals are converted into fuzzy membership degrees, and deviation is checked against the modified empirical avoidance tendency preliminary membership degrees. After the check passes, the final EA membership degree is obtained. Based on EA membership, combined with the ACT microscale and interactive behavior sequence, a fuzzy variable set of five dimensions of psychological flexibility—cognitive integration, experiential avoidance tendency, current contact ability, value clarity, and commitment to action—is generated through a five-dimensional weighted cross-analysis mechanism. After encryption, a baseline profile of psychological flexibility is established to mark the core shortcomings.

3. The virtual interactive scene fuzzy logic control system for psychological counseling of college students according to claim 2, characterized in that, The methods for identifying variable features and parsing macroscopic objectives into a sequence of microscopic control objectives include: Based on a fuzzy variable set and baseline profile, threshold rules are used to identify the variable characteristics and core weaknesses of users' psychological flexibility. To enhance psychological flexibility, anchoring macro-level goals involves defining the guiding emphasis of macro-level goals by combining variable characteristics and core weaknesses. Furthermore, following the principle of gradual progress and priority matching, the guidance focus is broken down into a sequence of micro-control targets that can be implemented in stages and each corresponds to a single ACT intervention direction.

4. The virtual interactive scenario fuzzy logic control system for psychological counseling of college students according to claim 3, characterized in that, The methods for assembling personalized ACT narrative chains include: Classified according to the six core processes of ACT theory, labeled with ACT intervention direction as intervention dimension label, set difficulty level according to intervention difficulty gradient, take the core weakness of psychological flexibility as the adaptation target, pre-set standardized narrative atoms, and integrate and construct ACT narrative atom library. Based on the micro-control target sequence, narrative atoms are matched with intervention dimension labels and difficulty gradients from the ACT narrative atom library and dynamically assembled into personalized ACT narrative chains.

5. The virtual interactive scene fuzzy logic control system for psychological counseling of college students according to claim 4, characterized in that, The methods for generating intervention task packages include: Based on the personalized ACT narrative chain, the interaction attributes and execution requirements of each narrative atom are analyzed; the execution tasks of each type of narrative atom are differentiated and assigned to the VR end, edge side and cloud according to the execution requirements, and executable instructions with priority are generated. Furthermore, it integrates personalized ACT narrative chains with executable instructions and encapsulates them in an encrypted package as an intervention task package.

6. The virtual interactive scene fuzzy logic control system for psychological counseling of college students according to claim 5, characterized in that, The VR terminal triggers visual cognitive dissociation intervention, and the sensory feedback of wearable devices is linked in the following ways: Based on the micro-control target sequence, personalized ACT narrative chain and intervention task package, preset visual cognitive dissociation intervention trigger conditions, and bind VR terminal interaction logic and user behavior to form intervention rules; When a user's behavior meets the triggering conditions, the VR device executes intervention according to the ACT narrative chain sequence, guiding the user's operation through interactive logic; Wearable devices provide real-time sensory feedback based on the coordinated and synchronized intervention rhythm at the edge, forming a collaborative closed loop with VR intervention.

7. The virtual interactive scene fuzzy logic control system for psychological counseling of college students according to claim 6, characterized in that, The methods for generating updated dynamic values ​​of psychological flexibility and intervention implementation data include: By simultaneously collecting real-time status data of device operation through VR terminals, wearable devices, and edge computing, and combining it with preset weighted calculation rules, the dynamic values ​​of five-dimensional psychological flexibility are updated in real time; the dynamic values, the progress of intervention implementation, and status data are integrated to form an intervention implementation dataset.

8. The virtual interactive scene fuzzy logic control system for psychological counseling of college students according to claim 7, characterized in that, The method of determining the therapeutic tolerance window through a two-dimensional assessment and implementing acceptance guidance and scenario fine-tuning includes: Based on the updated dynamic value of psychological flexibility and the intervention execution dataset, the dynamic value of experiential avoidance tendency and intervention participation are extracted to construct a two-dimensional evaluation mechanism; Furthermore, a dual-dimensional assessment mechanism is used to pre-define different assessment intervals based on threshold values, thereby determining the therapeutic tolerance window in real time. Based on the window determination results, the edge side coordinates and issues differentiated acceptance guidance instructions for each window. The VR terminal fine-tunes the virtual scene stimulation parameters as needed, and the edge side synchronously matches the window features to execute ACT acceptance guidance. Wearable devices provide sensory feedback in conjunction.

9. The virtual interactive scene fuzzy logic control system for psychological counseling of college students according to claim 8, characterized in that, The methods for defining the threshold for manual intervention in extreme states and issuing early warnings include: Based on the dynamic value of psychological flexibility, the results of the therapeutic tolerance window determination, and the intervention execution dataset, extreme state monitoring indicators are extracted to form a monitoring indicator set; By combining absolute thresholds and relative change thresholds, a set of monitoring indicators is defined and calibrated to obtain the three-level artificial intervention thresholds for extreme states. The monitoring index values ​​are monitored in real time and compared with the threshold values ​​for manual intervention. Based on the comparison results, hierarchical early warnings are issued at the local terminal and the back-end management terminal. The entire process of intervention and early warning records is integrated and packaged into an enhanced fuzzy context package, and a summary of safety interventions is generated simultaneously.

10. The virtual interactive scene fuzzy logic control system for psychological counseling of college students according to claim 9, characterized in that, The methods mentioned, including updating the user's psychological activity development trajectory file, comparing intervention effects and optimizing intervention rules, and linking manual guidance to adapt subsequent intervention task packages, include: Based on the enhanced fuzzy context package, the data is reconstructed into a structured dataset according to the intervention stage, and then incrementally updated to the user's psychological flexibility development trajectory file according to the timeline and intervention nodes. By combining longitudinal quantitative comparison and horizontal collaborative analysis, the effectiveness of this intervention is quantitatively evaluated, weak dimensions are identified, core intervention rules are iteratively optimized, and an optimized personalized intervention rule library is formed. By combining the updated development trajectory file with the quantitative evaluation results of intervention effects, and linking with the human counseling end to conduct reverse adaptation analysis, subsequent human-customized intervention needs are generated. By integrating manually customized needs with an optimized personalized intervention rule base, subsequent personalized intervention task packages are generated, forming a closed-loop cultivation link for psychological flexibility.

Images