A vr natural environment state regulation cabin system based on multi-modal physiological-brain computer interface closed-loop feedback

Abstract

The application discloses a VR natural environment state regulation cabin system based on a multi-modal physiological-brain computer interface closed-loop feedback, which integrates a multi-modal physiological-brain function signal synchronous acquisition module, a multi-modal signal processing and state index calculation module, a VR natural social environment immersive construction module, a multi-sensory environment linkage regulation module, a closed-loop neural feedback regulation module, a data management and individualized scheme generation module, a cabin environment control module and a system integration and control module. The system synchronously acquires physiological and brain electrical signals, evaluates the user state through fusion analysis, constructs an immersive VR environment accordingly, and linkage regulates multi-sensory stimulation. The core lies in that the closed-loop neural feedback module dynamically adjusts the environment and task parameters according to real-time monitoring data, and supports individualized scheme generation. The cabin environment control module guarantees the stability of the physical environment, and all the modules realize an integrated and self-adapting working closed loop under the coordination of the system integration and control module.

Description

Technical Field

[0001] This invention belongs to the field of intelligent human-computer interaction technology, and in particular relates to a VR-like natural environment state control cabin system based on multimodal physiological-brain-computer interface closed-loop feedback. Background Technology

[0002] Maintaining stable operational efficiency of personnel in special working environments characterized by long-duration, high-load, or closed and monotonous conditions (such as ocean voyages, aerospace monitoring, and precision instrument operation and monitoring) is a key technical challenge. Existing human-machine interaction and environmental control systems have the following significant technical limitations: First, the system lacks the ability to accurately and continuously perceive and fuse the operator's real-time physiological and neural signals. Although progress has been made in multimodal physiological monitoring and brain-computer interface technology, which can collect signals such as electrocardiogram and electroencephalogram, existing solutions often focus on single-point or offline state recognition and fail to perform high-precision synchronization and deep fusion of multi-source heterogeneous signals, thus failing to provide a reliable and consistent input data stream for real-time closed-loop control.

[0003] Secondly, system parameters (such as VR task difficulty and ambient light and sound stimulation intensity) are mostly static presets or only support limited manual configuration. For example, although some VR training systems can construct virtual scenes, their task flow and environmental parameters cannot be dynamically and adaptively adjusted according to the user's real-time cognitive load and physiological arousal level, resulting in a lack of flexibility in human-computer interaction and low training efficiency.

[0004] Furthermore, the loose coupling between the perception, analysis, decision-making, and execution stages makes it difficult to form an efficient real-time control closed loop. In existing solutions, signal acquisition, data analysis, and execution mechanisms (such as VR rendering engines and environmental control equipment) typically operate independently, lacking a unified coordination and control mechanism. This results in excessively high latency from signal perception to environmental response, making it impossible to achieve true real-time interaction and control.

[0005] Furthermore, the system has low integration and poor environmental adaptability. Existing equipment is mostly assembled from discrete components, resulting in a large size that makes it difficult to reliably deploy and operate stably for a long time in space-constrained and harsh environments (such as high temperature, high humidity, and high salt spray).

[0006] Finally, the environmental presentation methods are monotonous, and multi-sensory stimulation lacks coordination. Many systems only provide visual or simple auditory feedback, failing to coordinate and consistently render visual, auditory, olfactory, and light environmental stimuli in a spatiotemporal manner. This affects the sense of immersion and weakens the possibility of coordinated guidance through multi-sensory channels.

[0007] Therefore, there is an urgent need for an integrated, highly reliable intelligent control system that can synchronously collect and fuse multimodal physiological and neural signals in real time, and thereby drive the VR virtual environment and multisensory physical environment parameters to make millisecond-level adaptive adjustments, in order to solve the above-mentioned technical bottlenecks. Summary of the Invention

[0008] To address the technical problems in existing technologies, such as insufficient accuracy in perceiving the real-time physiological and cognitive states of workers, static and fixed system control parameters, and poor human-computer interaction adaptability, this invention provides a VR-like natural environment state control cabin system based on multimodal physiological-brain-computer interface closed-loop feedback.

[0009] This invention provides a VR-based natural environment state control cabin system based on multimodal physiological-brain-computer interface closed-loop feedback, integrated within a closed cabin. The system includes: The multimodal physiological-brain function signal synchronous acquisition module is used to synchronously acquire the user's peripheral physiological signals and brain function signals in real time. The multimodal signal processing and state index calculation module is communicatively connected to the multimodal physiological-brain function signal synchronous acquisition module. It is used to perform fusion analysis on the acquired multimodal data and establish an evaluation model to output state indexes and regulatory trigger signals. The VR-like natural and social environment immersive construction module is communicatively connected to the multimodal signal processing and state index calculation module, and is used to construct and present a virtual reality environment based on state indexes. The virtual reality environment includes natural environment scenes, social environment scenes, and interactive cognitive and regulatory tasks. The multi-sensory environment linkage control module is set up in conjunction with the VR-type natural and social environment immersive construction module to link and control auditory stimulation, light environment and olfactory stimulation in the closed cabin. The closed-loop neural feedback control module is connected to the multimodal physiological-brain function signal synchronous acquisition module, the multimodal signal processing and state index calculation module, the VR-like natural and social environment immersive construction module, and the multisensory environment linkage control module, respectively. It is used to dynamically adjust the presentation parameters of the virtual reality environment, the execution difficulty of the state training task, and the control parameters of the multisensory environment stimulation based on the real-time monitored physiological and brain function states. The data management and personalized solution generation module is communicatively connected to the closed-loop neural feedback control module, and is used to manage user data, build individual profiles, and generate personalized training solutions based on the current state. The cabin environment control module is used to maintain the temperature, humidity, ventilation, and clean physical environment inside the enclosed cabin. The system integration and control module communicates with all the above modules and is used to coordinate the operation and data interaction of each module.

[0010] Optionally, the multimodal physiological-brain function signal synchronous acquisition module includes: The physiological signal acquisition submodule is used to acquire the user's electrocardiogram, heart rate variability, skin conductance, respiratory rate and body temperature signals based on a multi-channel physiological recorder. The brain-computer interface submodule is used to collect the user's multichannel electroencephalogram (EEG) signals or functional near-infrared spectral signals.

[0011] Optionally, the multimodal signal processing and state index calculation module includes: The data preprocessing submodule is used to filter, denoise, and normalize the acquired raw signals, and to extract multi-dimensional features from the processed raw signals. The feature fusion submodule is used to fuse extracted multi-dimensional features using a fusion method that combines feature-level fusion and decision-level fusion. The user status assessment submodule is used to run machine learning models to identify the user's cognitive status, assess the level of cognitive load, cognitive load, and predict fatigue recovery. The control parameter mapping submodule is used to generate system control parameter adjustment instructions based on the calculated state indicators and preset rules.

[0012] Optionally, the VR-like immersive natural and social environment construction module includes: The VR display submodule is used to present virtual reality scenes to users; The virtual reality scene construction submodule is used to construct virtual reality scenes based on a library of natural environment scenes, a library of social environment scenes, and a library of interactive cognition and regulation tasks.

[0013] Optionally, the multi-sensory environment linkage control module includes: The 3D spatial audio submodule is used to render 3D spatial sound effects that match the VR scene; The dynamic rhythmic lighting environment submodule is used to adjust the color temperature and brightness of the cabin lighting to simulate the natural light rhythm. The aroma control submodule is used to control the release of specific aroma essential oils according to the user's status and adjustment needs.

[0014] Optionally, the closed-loop neural feedback modulation module includes: The real-time status monitoring submodule is used to analyze physiological and EEG signals in real time to obtain key status indicators; The adaptive control algorithm submodule is used to generate adjustment instructions based on the key state indicators and according to preset mapping rules or optimization algorithms, so as to adaptively switch VR scenes, adjust scene parameters, adjust task difficulty, or adjust environmental stimulus parameters.

[0015] Optionally, the data management and personalized solution generation module includes: The data storage submodule is used to store users' training history data, physiological EEG signals, evaluation results, and subjective evaluations; The personalized profile submodule is used to build individual profiles based on historical data, including user baseline characteristics, risk characteristics, and preference characteristics; The personalized scheme generation submodule is used to generate adjustment schemes, including target scenarios, training tasks, and initial parameters, based on current status indicators and individual profiles.

[0016] Optionally, the cabin environment control module includes: Temperature control submodule is used to regulate and maintain the temperature inside the enclosed compartment; A humidity control submodule is used to adjust and maintain the humidity inside the enclosed chamber; The ventilation system submodule is used to provide a circulation of filtered fresh air; The sound insulation and noise reduction submodule is used to reduce the interference of external noise on the cabin environment; A moisture-proof and corrosion-resistant submodule is used to protect the electronic equipment and cabin structure of the system in high-temperature, high-humidity, and high-salt environments. The lighting system submodule is used to provide basic cabin lighting, including emergency lighting. The safety protection submodule is used to monitor the safety of the cabin environment and trigger alarms and protective actions in case of abnormalities.

[0017] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computing program stored in the memory and executable on the processor, wherein the processor, when executing the computing program, controls the system.

[0018] On the other hand, the present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, is used to control the system.

[0019] Compared with the prior art, the present invention has the following advantages and technical effects: This invention utilizes a multimodal signal synchronous acquisition module and an intelligent processing module, employing hardware clock synchronization and machine learning algorithms to perform real-time fusion analysis of physiological and electroencephalogram (EEG) signals. The result is precise synchronization and efficient analysis of multi-source heterogeneous signals, transforming the system input from discrete, asynchronous data streams into a continuous data source with high spatiotemporal consistency, thus providing a high-precision, low-latency data foundation for subsequent adaptive control.

[0020] This invention utilizes a closed-loop adaptive control module to process data in real time and generate control commands using mapping rules or optimization algorithms. The effect is that it automatically drives the VR construction module and the multi-sensory control module, dynamically adjusting parameters such as scene, task difficulty, sound, light, and smell. This forms a millisecond-level real-time control closed loop of "signal input - analysis and decision-making - environmental output," fundamentally overcoming the technical shortcomings of existing systems with fixed parameters and the inability to dynamically adapt, significantly improving the system's automation and intelligence levels.

[0021] This invention presents a high-fidelity dynamic scene through a VR construction module, and uses a multi-sensory control module to coordinate the control of audio, lighting, and aroma. The effect is that the central controller ensures precise synchronization and consistency of cross-sensory stimuli in time and space, significantly improving the immersion, realism, and synergy of multi-channel stimulation in the virtual environment from a technical perspective. This provides a controllable and powerful technical means for the system to achieve specific regulatory goals (such as maintaining attention or inducing relaxation).

[0022] This invention provides a stable internal physical environment through a cabin environment control module and employs protective processes and modular design for key hardware. The result is that it effectively solves the key technical challenges of stable operation and rapid deployment of precision electronic systems in extreme environments such as high temperature, high humidity, and vibration, enabling the system to be transformed from a laboratory environment into a highly integrated field device that can be reliably deployed under harsh conditions.

[0023] This invention constructs a user characteristic model through a data management module and optimizes parameter configuration using intelligent algorithms. The result is that it can generate differentiated initial parameter sets for different users and continuously iterate and optimize the system's control strategy based on historical interaction data, enabling the overall system's control accuracy and user adaptability to continuously evolve and improve as usage progresses. Attached Figure Description

[0024] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of the system structure according to an embodiment of the present invention. Detailed Implementation

[0025] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0026] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0027] Example 1 like Figure 1 As shown, this embodiment provides a VR-based natural environment state control cabin system based on multimodal physiological-brain-computer interface closed-loop feedback, integrated within a closed cabin. The system includes: The multimodal physiological-brain function signal synchronous acquisition module is used to synchronously acquire the user's peripheral physiological signals and brain function signals in real time. The multimodal signal processing and state index calculation module is communicatively connected to the multimodal physiological-brain function signal synchronous acquisition module. It is used to perform fusion analysis on the acquired multimodal data and establish an evaluation model to output state indexes and regulatory trigger signals. The VR-like natural and social environment immersive construction module is communicatively connected to the multimodal signal processing and state index calculation module, and is used to construct and present a virtual reality environment based on state indexes. The virtual reality environment includes natural environment scenes, social environment scenes, and interactive cognitive and regulatory tasks. The multi-sensory environment linkage control module is set up in conjunction with the VR-type natural and social environment immersive construction module to link and control auditory stimulation, light environment and olfactory stimulation in the closed cabin. The closed-loop neural feedback control module is connected to the multimodal physiological-brain function signal synchronous acquisition module, the multimodal signal processing and state index calculation module, the VR-like natural and social environment immersive construction module, and the multisensory environment linkage control module, respectively. It is used to dynamically adjust the presentation parameters of the virtual reality environment, the execution difficulty of the state training task, and the control parameters of the multisensory environment stimulation based on the real-time monitored physiological and brain function states. The data management and personalized solution generation module is communicatively connected to the closed-loop neural feedback control module, and is used to manage user data, build individual profiles, and generate personalized training solutions based on the current state. The cabin environment control module is used to maintain the temperature, humidity, ventilation, and clean physical environment inside the enclosed cabin. The system integration and control module communicates with all the above modules and is used to coordinate the operation and data interaction of each module.

[0028] The technical problems to be solved in this embodiment include: how to integrate multiple devices such as multimodal physiological signal acquisition, non-invasive brain-computer interface, VR display, and multisensory environment control in a closed cabin space to achieve high integration and optimized space design; how to construct a multimodal user state accurate assessment model suitable for people in extreme environments, integrating heterogeneous data such as subjective scales, physiological signals, EEG characteristics, and cognitive behavior to achieve state trend recognition and adaptive regulation based on physiological markers; how to develop a VR-based high-immersion natural and social environment scene library and interactive cognition and regulation task library to create a multisensory linkage recovery environment in a limited cabin space; how to realize closed-loop neuromodulation technology based on real-time physiological and EEG feedback, adaptively adjusting VR scene parameters, light environment, sound environment, aromatherapy stimulation, and training difficulty according to the individual's current state to achieve an intelligent closed loop of "perception-assessment-regulation-feedback"; and how to ensure the stable operation and long-term reliability of precision electronic equipment under harsh environmental conditions such as high temperature, high humidity, and high salinity.

[0029] By addressing these technical challenges, this embodiment provides a VR-based natural environment state control cabin system based on multimodal physiological-brain-computer interface closed-loop feedback. This system employs a highly integrated modular design, integrating multiple functions such as user state assessment, VR environment presentation, multisensory stimulation control, and neurofeedback training within a closed cabin space of approximately 6-8 square meters, achieving comprehensive maintenance and enhancement of the user's physiological efficacy.

[0030] The system mainly includes the following core modules: (1) Multimodal physiological-brain function signal synchronous acquisition module: This module is responsible for real-time acquisition of multi-dimensional physiological and brain function data from users. The physiological signal acquisition submodule integrates a multi-channel physiological recorder (≥5 channels) to simultaneously acquire peripheral physiological indicators such as electrocardiogram (ECG), heart rate variability (HRV), skin conductance response (GSR / EDA), respiratory rate (RESP), and peripheral body temperature (Temp), with a sampling rate ≥256Hz and signal accuracy meeting clinical standards. The brain-computer interface submodule uses a multi-channel electroencephalogram (EEG) device (≥32 channels) or functional near-infrared spectroscopy (fNIRS) device to monitor neural activity in key brain regions such as the frontal, parietal, and temporal lobes, with an EEG sampling rate ≥500Hz, supporting multi-band analysis including Alpha (8-12Hz), Beta (13-30Hz), and Gamma (31-50Hz). All signal acquisition devices are synchronized via a unified hardware clock, with time synchronization accuracy reaching the microsecond level (<10μs), ensuring accurate correspondence of multimodal data.

[0031] The acquisition module employs wireless transmission technology to reduce cable constraints. The physiological electrodes are made of medical-grade Ag / AgCl material with an electrode impedance of <5kΩ. The EEG electrodes utilize dry or semi-dry electrode technology, with a wearing time of <5 minutes, ensuring high user comfort. The module features real-time signal quality monitoring, automatically identifying and labeling artifacts (such as eye movement, electromyography, and motion artifacts), and using an adaptive filtering algorithm for real-time noise reduction.

[0032] (2) Multimodal signal processing and state index calculation module: This module, based on multimodal data fusion and machine learning technologies, enables accurate assessment of user status and risk warning. The data preprocessing submodule filters, denoises, and normalizes the raw physiological and EEG signals, extracting multi-dimensional feature vectors such as time-domain features (e.g., HRV time-domain indices SDNN, RMSSD), frequency-domain features (e.g., HRV frequency-domain indices LF, HF, LF / HF, EEG band power), time-frequency features (e.g., wavelet coefficients, Hilbert-Huang transform), and nonlinear features (e.g., sample entropy, approximate entropy, detrended fluctuation analysis).

[0033] The feature fusion submodule employs a strategy that combines feature-level fusion and decision-level fusion. Feature-level fusion uses principal component analysis (PCA) or an autoencoder to reduce and fuse high-dimensional features; decision-level fusion uses ensemble learning methods (such as random forest, gradient boosting tree, and stacking) to integrate the outputs of multiple single-modal classifiers.

[0034] The user status assessment submodule builds multiple machine learning models, including: Cognitive state recognition model: classifies cognitive states with an accuracy of ≥90%; Cognitive load level assessment model: Outputs a cognitive load index of 0-100, with a prediction accuracy of R² ≥ 0.85; Cognitive load assessment model: assesses attention level, working memory load, and cognitive fatigue level; Fatigue recovery prediction model: Predicts the risk of fatigue recovery within the next 24 hours.

[0035] The risk warning submodule establishes a multi-level warning mechanism: State Mode A (Parameter Maintenance): Key indicators are within the baseline range, and the system maintains the current control parameters.

[0036] State Mode B (Parameter Fine-tuning): When a single indicator deviates slightly, the system automatically fine-tunes the intensity of environmental stimuli or the difficulty of the task.

[0037] Status Mode C (Program Switching): When multiple indicators deviate significantly, the system automatically switches to the preset "High Load Relief" or "Attention Enhancement" control program.

[0038] Safety protection mechanism: If extreme physiological signals (such as abnormal heart rate) are detected, the system will pause the task and issue a notification.

[0039] The system establishes dynamic thresholds based on individual baseline data and group norms, enabling it to identify abnormal deviations in individual user status and provide early warnings (24-72 hours in advance).

[0040] (3) VR-based immersive construction module for natural and social environments: This module creates a highly immersive virtual natural and social environment within a closed cabin, providing users with a space to recover their state. The VR display submodule uses a high-resolution head-mounted display (HMD, resolution ≥ 2K×2K per eye, field of view ≥ 110°, refresh rate ≥ 90Hz) or a CAVE projection system to provide a 360-degree panoramic visual experience. The system uses asynchronous time warp (ATW) and asynchronous space warp (ASW) technologies to optimize rendering performance, with a latency of <20ms, effectively preventing VR motion sickness.

[0041] The natural environment scene library contains ≥20 high-fidelity natural scenes, covering: Forest series: temperate broad-leaved forests, tropical rainforests, coniferous forests, bamboo forests, etc.; Water features: beaches, lakes, streams, waterfalls, etc. Mountains and rivers series: alpine meadows, canyons, snow-capped mountains, desert oases, etc.; Sky series: sunrise and sunset, starry sky, aurora, sea of ​​clouds, etc.

[0042] Each scene features a dynamic weather system (sunny, cloudy, rainy, etc.), day-night cycle, and adjustable parameters for seasonal changes. The scenes include realistically captured natural sound effects (birdsong, insect chirping, flowing water, ocean waves, wind, etc.) rendered in 3D space to enhance immersion.

[0043] The social environment scenario library contains ≥10 social interaction scenarios, covering: Home environment: cozy family spaces such as living room, bedroom, kitchen, and courtyard; Community environment: social spaces such as parks, cafes, libraries, and gyms; Virtual social interaction: Interacting with AI-driven virtual characters to alleviate feelings of social isolation.

[0044] The interactive cognition and regulation task library contains ≥10 training tasks: Mindfulness meditation training includes guided meditation, breath awareness, body scan, etc. Progressive muscle relaxation: Systemic muscle tension-relaxation training; Breathing rhythm regulation: breathing training based on HRV biofeedback (such as 6 breaths / minute resonant breathing). Cognitive reassessment of cognitive states: Contextualized training in cognitive behavioral therapy (CBT); Attention training: selective attention, sustained attention, and attention switching training; Working memory training: N-back task, complex span task, etc.; Executive function training: inhibitory control and cognitive flexibility training; Maintaining social connections: virtual social interaction and emotional expression training; Sleep improvement training: pre-sleep relaxation, sleep hygiene education, etc.

[0045] Each training task has multiple difficulty levels (1-10), and the system adaptively adjusts the difficulty based on the user's performance.

[0046] (4) Multi-sensory environment linkage control module: This module enables coordinated control of vision, hearing, smell, and lighting environment, enhancing the adjustment effect. The 3D spatial audio submodule employs HRTF (Head-Related Transfer Function) 3D sound technology from a surround sound system or headphones to render spatial sound effects in real time based on the VR scene, with sound source localization accuracy ≤5°, enhancing immersion and presence. The system has a built-in natural sound effect library, and the volume can be adaptively adjusted according to user preferences and user status.

[0047] The dynamic rhythmic lighting environment submodule employs an adjustable color temperature LED lighting system (color temperature range 2700K-6500K) to simulate natural day-night rhythm changes. The system adjusts lighting parameters according to time and user status. Morning mode: 6000K cool white light, brightness gradually increases, simulating sunrise to promote wakefulness; Daytime mode: 5500K natural light, maintaining high brightness and alertness; Evening mode: 4000K warm white light, brightness gradually decreases, simulating sunset to promote relaxation; Night mode: 2700K warm yellow light, low brightness, promotes melatonin secretion, and improves sleep.

[0048] The light intensity is adjustable from 10 to 1000 lux, supporting both gradual and pulse modes. Studies have shown that dynamic light regulation can effectively improve circadian rhythm disorders in enclosed environments.

[0049] The aroma control submodule uses a micro-essential oil diffusion device to release different fragrances according to the user's state and adjustment needs: Relaxing and calming: Lavender, chamomile, sandalwood, etc.; Invigorating and refreshing: peppermint, lemon, rosemary, etc.; Improved cognitive state: orange blossom, rose, bergamot, etc.; Improve sleep: Cedarwood, Marjoram, Vetiver, etc.

[0050] The diffusion device uses ultrasonic atomization technology, with precise and controllable release (0.1-5ml / hour) and adjustable aroma concentration to avoid excessive stimulation.

[0051] Multi-sensory linkage control algorithms enable coordinated regulation of vision, hearing, smell, and light: For example, in the "Forest Walk" scene, VR presents a lush forest scene, while playing 3D sound effects of birdsong and wind blowing through leaves, releasing a faint pine scent, and adjusting the light environment to soft natural light, with multi-sensory stimulation working in synergy.

[0052] (5) Closed-loop neural feedback regulation module: This module is the core innovation of the system, enabling intelligent closed-loop control that adaptively adjusts the VR environment and training parameters based on real-time physiological and EEG states. The real-time state monitoring submodule continuously analyzes the user's physiological and EEG signals, updating state indicators every second. Key monitoring indicators include: HRV high-frequency power (HF): reflects parasympathetic nerve activity and indicates the degree of relaxation; Skin conductance response amplitude (SCR amplitude): reflects sympathetic nerve activity and indicates arousal level; EEG Alpha wave power: reflects the state of relaxation and wakefulness; EEG Beta wave power: reflects cognitive activity and stress level; EEG Alpha / Beta ratio: a comprehensive indicator of relaxation-tension balance; Heart rate (HR) and respiratory rate (RR): reflect physiological arousal levels.

[0053] The adaptive control algorithm submodule dynamically adjusts the control parameters based on real-time monitoring results: VR scene adaptive switching: When a high level of cognitive load is detected in the user (HRV-LF / HF>2.5, frequent SCR), the system automatically switches from the current scene to a calmer scene (such as switching from crashing waves to a tranquil lake). When excessive relaxation or decreased attention is detected in the user (excessive alpha wave power, slower reaction time), the scene dynamics and interactivity are increased.

[0054] Real-time adjustment of scene parameters: Adjust the scene's visual parameters (color saturation, brightness, motion speed, etc.) and auditory parameters (volume, sound effect type, etc.) according to the user's physiological state. Training difficulty adaptive adjustment: In cognitive training tasks, the difficulty of the task is dynamically adjusted based on the user's performance (accuracy, reaction time) and cognitive load (Beta wave power, pupil diameter, etc.). Increase task difficulty when accuracy is >85% and cognitive load is moderate; decrease task difficulty when accuracy is <60% or cognitive load is too high.

[0055] Breathing-guided rhythm regulation: During breathing training, the system adjusts the guided rhythm based on real-time breathing signals and HRV feedback; The goal is to guide the respiratory rate to the resonant frequency (typically 5.5-6 breaths / minute), maximize HRV-HF power, and enhance parasympathetic activity.

[0056] Dynamic regulation of light environment and aroma: Adjust lighting parameters and aroma type according to the user's circadian rhythm and current user status; For example, during nighttime relaxation training, gradually reduce the color temperature and brightness, and release sleep-inducing aromas.

[0057] The neurofeedback visualization submodule presents the user's real-time physiological and EEG states in an intuitive way, such as: HRV curves and frequency domain power spectrum display; EEG frequency band power bar graph; Relaxation index (0-100) real-time value; Respiratory rhythm waveform and target rhythm guide line.

[0058] Users can learn self-regulation skills by observing this feedback information, thereby enhancing the initiative and sustainability of the regulation effect.

[0059] (6) Data Management and Personalized Solution Generation Module: This module is responsible for storing, analyzing, and managing user training data, and generating personalized training plans. The data storage submodule stores multimodal data using HDF5 or SQLite database format, including: The raw physiological and EEG signals of each training session; Status assessment results and early warning records; VR scenarios utilize historical and interactive behavior data; Training task performance data (accuracy, reaction time, etc.); Subjective evaluation data (cognitive status scores before and after training, satisfaction scores, etc.).

[0060] Data storage employs a combination of local encrypted storage and cloud backup to ensure data security and privacy protection.

[0061] The personalized profile submodule builds a user's health profile based on historical training data: Baseline characteristics: individual's physiological and EEG baseline levels and norm range; Risk characteristics: Types of health problems that are susceptible to user issues; Preference characteristics: preferred VR scene types, training task preferences, optimal training time periods, etc.; Response characteristics: The response pattern and magnitude of different regulatory measures.

[0062] The personalized solution generation submodule intelligently recommends the optimal adjustment plan based on the current evaluation results and individual profile: Adjustment target setting: Determine adjustment priorities based on the current main problems; Scene combination recommendation: Recommend the VR scene sequence that best suits the current state and individual preferences; Training task customization: Select the most relevant training task and starting difficulty level; Training plan development: Recommendations for daily training duration, frequency, and optimal time of day.

[0063] The system uses reinforcement learning algorithms to continuously optimize personalized solutions and adjusts recommendation strategies based on feedback from the adjustment effects.

[0064] The performance evaluation and report generation submodule automatically generates an evaluation report after each training session, including: Training overview: training duration, usage scenarios and tasks, parameter settings, etc. Changes in physiological and EEG indicators: Comparison of HRV, skin conductance, EEG and other indicators before and after training; Changes in subjective feelings: Comparison of cognitive state scores and cognitive load scores before and after training; Training effectiveness evaluation: Effectiveness score of this training and suggestions for improvement; Historical trend analysis: Recent (7 days, 30 days) user status change trend chart.

[0065] The report uses visual charts to facilitate user understanding and self-monitoring.

[0066] (7) Cabin environment control module: This module is responsible for maintaining the physical environment of the cabin, ensuring stable system operation and user comfort. The temperature control submodule uses a high-efficiency air conditioning system with an adjustable temperature range of 18-28℃ and an accuracy of ±0.5℃, capable of rapidly cooling down in high-temperature environments (such as tropical regions). The humidity control submodule uses a dehumidifier or humidifier with an adjustable humidity range of 40%-70%RH and an accuracy of ±5%, preventing damage to electronic equipment from high humidity environments.

[0067] The ventilation system submodule provides fresh air circulation with an air exchange rate of ≥3 times / hour. It is equipped with a HEPA high-efficiency filter to filter airborne particles and allergens, while an activated carbon filter removes odors. The ventilation system is designed for quiet operation, with a noise level of <35dB, to avoid interfering with user training.

[0068] The sound insulation and noise reduction submodule lays high-grade sound-absorbing materials (such as polyurethane sound-absorbing cotton and sound insulation felt) on the walls, top, and floor of the cabin, achieving a sound insulation effect of ≥30dB. This effectively isolates external environmental interference such as ocean waves, mechanical noise, and human voices, creating a quiet restorative space for users. The cabin door adopts a double-layer sealing structure to further enhance the sound insulation effect.

[0069] The moisture-proof and corrosion-resistant submodule is designed for high-humidity and high-salt environments. The cabin is constructed with rust- and corrosion-resistant materials (such as stainless steel, aluminum alloy, and corrosion-resistant coated steel plates) to form the frame, and the internal electronic equipment is encapsulated with an IP54 or higher protection rating. Key circuit boards are coated with a three-proof coating (moisture-proof, salt spray-proof, and mildew-proof). The cabin is equipped with an industrial-grade dehumidifier and desiccant packs to continuously control humidity and prevent moisture damage to electronic components. Regular maintenance includes desiccant replacement, seal inspection, and corrosion-resistant coating repair.

[0070] In addition to the aforementioned dynamic rhythmic lighting environment system, the lighting system submodule is also equipped with emergency lighting and safety indicator lights to ensure that users can safely evacuate the cabin in the event of a failure of the main lighting system or a power outage.

[0071] The safety protection submodule includes safety devices such as smoke detectors, CO2 concentration monitors, and emergency stop buttons. When a fire risk, abnormal air quality, or a user triggers the emergency stop button, the system automatically stops all training programs, activates the ventilation system, turns on emergency lighting, unlocks the hatch, and sends an alarm to management personnel. The cabin is equipped with fire extinguishers, first-aid kits, and other emergency supplies.

[0072] (8) System Integration and Control Module: This module serves as the "brain" of the entire system, coordinating the work of all sub-modules and enabling unified management and intelligent control. The main control computer is a high-performance workstation (Intel Core i7 or higher CPU, ≥16GB RAM, ≥512GB SSD, NVIDIA RTX 2060 or higher GPU) running custom-developed system control software. The software employs a modular architecture, is based on a hybrid C++ / Python programming style, and uses the Qt framework for its graphical user interface.

[0073] The device interface submodule provides a unified hardware interface protocol and supports multiple communication methods: USB 3.0 / 3.1 interface: for connecting physiological signal acquisition devices and EEG devices; HDMI / DisplayPort interface: Connect to VR headsets or projection devices; Bluetooth 5.0 / Wi-Fi 6: Wireless connectivity for wearable sensors and wireless EEG devices; RS-485 / Modbus: Controls environmental control equipment (air conditioners, dehumidifiers, lighting, aroma diffusers, etc.); Network interface: Supports remote monitoring, data upload, and software updates.

[0074] The data synchronization mechanism ensures precise alignment of timestamps across all acquisition and control devices. The clocks of each device are calibrated using Network Time Protocol (NTP) or Precision Time Protocol (PTP), and the acquisition time of each data packet is marked using a hardware trigger signal or software timestamp. The multimodal data time synchronization error is <10ms.

[0075] The system monitoring submodule monitors the real-time operating status of each hardware device (online / offline, signal quality, power consumption, etc.) and the working status of software modules (normal / abnormal, processing delay, etc.). When a device failure or software anomaly is detected, it automatically logs and triggers an alarm, and if necessary, activates backup equipment or degrades the operating mode to ensure system stability.

[0076] The user interface submodule provides an intuitive and user-friendly interface, including: Login interface: User identification (ID card swiping or facial recognition); Assessment interface: Guides users to complete physiological scales and cognitive task tests; Training selection interface: Displays recommended training schemes and a library of available scenarios / tasks; Training interface: VR scene presentation, real-time feedback display, training progress prompts; Report interface: Post-training performance evaluation report and historical data viewing; Settings interface: Personalized parameter adjustments (volume, brightness, scene preferences, etc.).

[0077] The interface design follows ergonomic principles, and the operation process is simple and clear, allowing ordinary users to use it independently without professional training.

[0078] (9) System workflow: The complete workflow of the system in this embodiment includes the following steps: Step 1: User Login and Identity Verification Users log in to the system by swiping their ID card or using facial recognition, and the system automatically retrieves the user's historical training data and individualized profile information.

[0079] Step 2: Baseline Assessment The system guides users through a baseline assessment, including: (a) Completion of the subjective physiological scale: takes approximately 5 minutes; (b) Baseline acquisition of physiological signals: The user wears physiological sensors and EEG electrodes and remains in a resting state for 3-5 minutes. The system acquires baseline physiological and EEG data. (c) Rapid cognitive function test: Simplified version of the Attention Network Test (ANT) or N-back working memory task, which takes about 3-5 minutes.

[0080] Step 3: User Status Assessment and Risk Classification: The system's intelligent user status assessment module integrates the multimodal data collected in step 2 and calculates three key status indicators using a machine learning model: Cognitive Load Index (CLI, 0-100), Attention Scale Index (ACI, 0-100), and Physiological Relaxation Index (PRI, 0-100). These indicators serve as inputs for the subsequent adaptive control module.

[0081] Step 4: Generation of personalized training plans: Based on the evaluation results and the user's individual profile, the system intelligently generates the optimal training plan: (a) Prioritize conditioning goals: For example, if cognitive load is high and HRV-HF power is low, prioritize relaxation training; if cognitive efficacy is low and Beta wave power is low, prioritize cognitive training. (b) Recommended VR scene combinations: Based on user preferences and current status, the most suitable scene sequence is recommended, such as "Forest walk (10 minutes) → Beach sunset (5 minutes) → Mindfulness meditation (5 minutes)"; (c) Recommended training tasks: Select appropriate training tasks according to the regulation goals, such as "breathing rhythm regulation + progressive muscle relaxation" or "attention training + working memory training"; (d) Set training parameters: initial difficulty level, training duration, closed-loop feedback sensitivity, etc.

[0082] The system displays recommended options to users, who can accept the recommendations or manually select other scenarios and tasks.

[0083] Step 5: VR-like natural environment presentation and multi-sensory interaction: When the user wears a VR headset (or enters a CAVE projection pod), the system activates the selected VR scene. Simultaneously, the multi-sensory environment linkage and control module works in concert: (a) The VR display module presents high-fidelity three-dimensional natural or social environment images; (b) The three-dimensional spatial audio module plays natural sound effects that match the scene; (c) The dynamic rhythmic lighting environment module adjusts the color temperature and brightness of the cabin lighting; (d) The aromatherapy module releases fragrances that match the scene and conditioning objectives; (e) The temperature and humidity control module maintains a comfortable physical environment.

[0084] The synergistic effect of multi-sensory stimulation maximizes immersion and recovery.

[0085] Step 6: Real-time physiological-brain function monitoring and data acquisition: Throughout the training process, the multimodal physiological-brain function signal synchronous acquisition module continuously monitors the user's status: (a) Collect ≥256 physiological signal sample points per second (ECG, skin conductance, respiration, etc.); (b) Acquire ≥500 EEG signal samples per second (32-channel EEG); (c) Record changes in VR scene parameters and user interaction behaviors (head movements, controller operations, etc.); (d) Record changes in environmental parameters (light, temperature, aroma release, etc.).

[0086] All data is tagged with precise timestamps and stored in a local database.

[0087] Step 7: Closed-loop neural feedback regulation: The closed-loop neural feedback control module analyzes the collected data in real time and dynamically adjusts the control parameters. (a) Calculate key metrics (HRV-HF power, SCR frequency, Alpha / Beta ratio, etc.) every 1-2 seconds. (b) Assess the current regulatory effect: whether the degree of relaxation has increased, whether the cognitive load is appropriate, and whether the cognitive state has improved; (c) Implement an adaptive adjustment strategy based on the evaluation results: If the user's cognitive load level does not decrease, switch to a calmer scene or reduce the intensity of the scene's stimulation. If users relax excessively or their attention decreases, increase the interactivity of the scene or increase the difficulty of the training task; If the respiratory rate is detected to be close to the target resonant frequency, provide visual or auditory positive feedback encouragement. If a sustained increase in alpha wave power is detected, it indicates that the relaxation training is effective; maintain the current parameters. (d) Visualize real-time feedback information to users (such as relaxation index, breathing guidance lines, etc. on the screen) to help users learn self-regulation.

[0088] This closed-loop process runs throughout the entire training session, ensuring that the tuning is always in an optimal state.

[0089] Step 8: Training task execution and performance monitoring: If the training program includes interactive cognitive training tasks (such as attention training, working memory training, etc.), the system presents the task stimuli in a VR environment and records the user's reaction time, accuracy, and other performance data. The system dynamically adjusts the task difficulty based on performance and cognitive load indicators. (a) If the accuracy rate is >85% for 5 consecutive trials and the reaction time is stable, the difficulty level is increased by one level; (b) If the accuracy rate for 5 consecutive trials is <60% or the reaction time is significantly prolonged, the difficulty level is reduced by one level; (c) Keep the task difficulty within the user's "zone of proximal development" so that it is challenging but not overly frustrating.

[0090] During the task, the system provides continuous and real-time feedback (correct / incorrect prompts, score display, etc.) to enhance training motivation.

[0091] Step 9: Training Completion and Effectiveness Evaluation: When the preset training time is reached (usually 15-30 minutes) or the user actively ends the training, the system performs a post-training evaluation: (a) Collect physiological and EEG data after training (remain at rest for 2-3 minutes); (b) Users fill in their subjective feelings after training (cognitive state, cognitive load level, fatigue level, etc.); (c) The system calculates the changes in indicators before and after training: HRV-HF power change percentage; Changes in the amplitude and frequency of skin conductance response; Changes in the EEG Alpha / Beta ratio; Changes in subjective cognitive load scores; Changes in cognitive state scale scores; (d) Evaluate the overall effectiveness of this training: effective / partially effective / ineffective, and analyze the possible reasons.

[0092] Step 10: Training Report Generation and Personalized Solution Updates The system automatically generates a training report, including: (a) Overview of this training: training duration, usage scenario, training task, parameter settings; (b) Comparison charts before and after training: comparison of HRV, skin conductance, EEG, and subjective scores before and after training; (c) Evaluation of training effectiveness: overall effectiveness score (0-100 points) and improvement status of individual items; (d) Improvement suggestions: Recommended adjustments for the next training session (such as changing the scene, increasing the training duration, etc.); (e) Historical trends: User status change trend charts and training frequency statistics for the past 7 days and 30 days.

[0093] The report is saved to the user's profile, and the system updates the user's individual profile and adjusts the response model based on the training data to optimize subsequent personalized recommendation algorithms.

[0094] Step 11: Data Upload and Cloud Synchronization: If the user agrees, the system can upload the anonymized training data to the cloud server for: (a) Cross-device data synchronization: Users can retrieve historical data when using different devices in different locations; (b) Big Data Analytics: Aggregating multi-user data for group characteristic analysis, optimizing evaluation models and adjustment algorithms; (c) Remote monitoring: Remotely view the user's training status and provide remote guidance.

[0095] Data transmission uses SSL / TLS encryption, and cloud storage uses AES-256 encryption, complying with data privacy protection standards (such as HIPAA, GDPR, etc.).

[0096] The system proposed in this embodiment involves the following key technologies: First, regarding multimodal data fusion and user state assessment technology. The multimodal signal processing and state index calculation module of the system in this embodiment is configured to perform deep fusion and accurate assessment of multimodal physiological-brain function data. Specifically, the data fusion and modeling unit included in this module is configured to: employ feature-level fusion algorithms to reduce and fuse high-dimensional heterogeneous features extracted from physiological signals and EEG signals, wherein the feature-level fusion algorithms include principal component analysis (PCA), independent component analysis (ICA), or autoencoders; and / or, employ decision-level fusion algorithms to integrate the outputs of multiple classifiers built based on single-modal data, wherein the decision-level fusion algorithms include ensemble learning methods such as random forests, gradient boosting trees, stacking, or Voting; simultaneously, use temporal modeling algorithms to process signal data to capture dynamic patterns of user states, wherein the temporal modeling algorithms include long short-term memory networks (LSTM), gated recurrent units (GRU), or temporal convolutional networks (TCN); furthermore, employ a personalized modeling strategy based on transfer learning, namely, pre-training a general evaluation model using group data, and then fine-tuning the general model using training data from individual users to construct a personalized evaluation model for that user, thereby achieving accurate user state evaluation under small sample conditions.

[0097] Second, regarding real-time closed-loop neurofeedback modulation technology. The closed-loop neurofeedback modulation module of this embodiment is configured to achieve adaptive adjustment based on real-time EEG-physiological feedback. The adaptive control unit included in this module is configured to: extract EEG features in real time, including online calculation of EEG frequency band power, Alpha / Beta ratio, event-related potentials (ERPs), or brain network connectivity features, with a processing latency of less than 500 milliseconds; access and apply a preset state-parameter mapping rule base, which defines the mapping relationship between user state indicators and VR scene parameters, environmental parameters, and task parameters, such as the linkage control rule of "when HRV-HF power is lower than the threshold, trigger switching to a specific forest scene, reducing color saturation, playing birdsong sound effects, and releasing lavender fragrance"; and / or run a multi-objective optimization control algorithm to balance and optimize multiple objectives such as improving recovery effect, maintaining user comfort, and avoiding overstimulation, the algorithm including fuzzy control or reinforcement learning algorithm; furthermore, it can also execute predictive modulation strategies, that is, predict the future change trend of user state based on a time-series prediction model, and adjust the modulation parameters in advance accordingly.

[0098] Third, regarding VR environment construction and multi-sensory linkage technology. In this embodiment, the VR-like natural and social environment immersive construction module and the multi-sensory environment linkage control module are configured collaboratively to create a highly immersive multi-sensory linkage environment. Specifically, the VR construction module includes a scene library sub-module that stores high-fidelity 3D models of natural scenes obtained through photogrammetry or LiDAR technology, and integrates relevant dynamic environment systems to simulate weather changes, day-night cycles, seasonal changes, and dynamic effects of vegetation and water bodies; its audio rendering sub-module performs 3D spatial audio rendering based on the Head Related Transfer Function (HRTF) and environmental acoustic models. The system integration and control module is configured to coordinate the dynamic rhythmic light environment sub-module and aroma control sub-module within the VR construction module and the multi-sensory environment linkage control module, ensuring consistency of visual, auditory, olfactory, and cabin temperature and humidity stimuli in time and space, avoiding sensory conflicts, and achieving collaboration with social interaction AI (based on Natural Language Processing (NLP) and emotion computing driven) in the virtual scene to jointly alleviate the user's sense of social isolation.

[0099] Fourth, regarding the integrated technology for adaptability to extreme environments. The cabin environment control module and hardware packaging of this embodiment are configured to adapt to extreme environments such as high temperature, high humidity, and high salinity. The moisture-proof and corrosion-resistant sub-module of the cabin environment control module uses processes including conformal coating, vacuum potting, or hermetically sealed packaging to encapsulate electronic components, achieving a protection level of IP54 to IP67. Thermal management technologies including heat pipes, phase change materials (PCM), or liquid cooling systems are employed to ensure heat dissipation for high-performance computing equipment. The cabin structure uses corrosion-resistant materials such as 316 stainless steel, aluminum-magnesium alloy, or carbon fiber composite materials, and undergoes surface treatment processes such as anodizing, powder coating, or fluorocarbon coating. Furthermore, the entire system adopts a modular, rapid deployment design, with standardized mechanical and electrical interfaces between functional modules, supporting rapid disassembly, replacement, and on-site assembly.

[0100] Fifth, regarding the intelligent matching and optimization technology for personalized training programs. The data management and personalized program generation module of this embodiment is configured to achieve intelligent matching and dynamic optimization of adjustment programs. The individual profile construction submodule of this module integrates multi-dimensional information on the user's demographic, psychological, physiological, cognitive, and behavioral preferences to construct a complete individual profile. Its personalized program generation submodule is configured to run at least one of the following intelligent algorithms: a collaborative filtering recommendation algorithm, recommending programs for the current user based on the training preferences and effects of "similar users"; a reinforcement learning optimization algorithm, modeling the selection of adjustment programs as a Markov decision process (MDP) and using algorithms such as Q-learning or deep Q-networks (DQN) to continuously optimize the recommendation strategy based on feedback from historical training effects; and a multi-armed gambling machine strategy, used to balance and optimize between exploring new programs and utilizing known effective programs. This system is particularly suitable for maintaining and improving the performance of workers in extreme environments such as high temperature, high humidity, high salinity, enclosed spaces, and social isolation.

[0101] Compared with the prior art, this embodiment has the following significant beneficial effects and technical advantages: (1) Achieve objective quantitative assessment and early warning of user status: Traditional physiological assessments primarily rely on subjective scales and clinical interviews, resulting in assessments heavily influenced by subjective factors and lacking timeliness. This embodiment integrates multi-channel physiological signal acquisition and multi-channel EEG monitoring to obtain objective physiological and brain function indicators such as heart rate variability, skin conductance response, respiratory rate, and EEG spectrum. Combined with a machine learning model, it achieves quantitative assessment of the user's state with a classification accuracy of ≥90%. More importantly, this embodiment establishes an early warning mechanism based on physiological biomarkers, capable of identifying risk signals 24-72 hours before health problems become apparent, thus gaining valuable time for timely intervention.

[0102] (2) Achieve personalized and precise adjustment, and improve the effectiveness of adjustment: Existing VR physiological regulation systems often employ a fixed, one-size-fits-all approach, failing to adapt to individual differences and changes in user state. This embodiment constructs a multi-dimensional individual profile, intelligently matching the optimal regulation plan based on the user's psychological, physiological, cognitive, and preference characteristics. More importantly, this embodiment achieves closed-loop adaptive regulation based on real-time physiological and EEG feedback. The system dynamically adjusts VR scene parameters, training difficulty, and environmental parameters according to changes in the user's state during regulation, ensuring that the regulation is always at its optimal state. Experimental data shows that the effect of closed-loop adaptive regulation is 35%-50% higher than that of a fixed approach.

[0103] (3) Break through the limitations of a single regulatory approach and form a multi-dimensional coordinated regulation: Existing technologies typically offer only a single regulatory approach (such as VR scenes or biofeedback), with limited effectiveness. This embodiment deeply integrates multiple regulatory approaches, including VR-like immersive natural and social environment experiences, non-invasive brain-computer interface neurofeedback, multi-sensory environmental regulation (auditory-olfactory-optical linkage), and interactive cognitive training, to form a multi-dimensional synergistic regulatory system. The synergistic effect of multi-sensory stimulation can maximize immersion and recovery effects. Compared to a single VR scene, multi-sensory linkage can improve subjective relaxation scores by 40% and HRV improvement by 30%.

[0104] (4) Solve the deployment challenges in extreme environments and achieve highly integrated modular design: Current control systems typically consist of multiple independent devices, resulting in complex systems that occupy a large amount of space and are difficult to deploy in resource-constrained environments. This embodiment employs a highly integrated modular design, integrating all functions such as signal acquisition, brain-computer interface, VR display, and environmental control within a 6-8 square meter enclosed chamber, reducing space usage by more than 60%. Furthermore, to address harsh environments such as high temperature, high humidity, and high salinity, technologies such as moisture-proof and corrosion-resistant packaging, efficient heat dissipation, and corrosion-resistant materials are used to ensure stable system operation under extreme conditions (MTBF ≥ 1000 hours), solving the problem of control systems being "unusable, poorly functioning, or unsustainable" in extreme environments.

[0105] (5) Establish a continuously optimized intelligent learning mechanism: This embodiment is not a static adjustment device, but an intelligent system with self-learning and continuous optimization capabilities. Through long-term accumulation of user training data, the system can: (a) continuously optimize the individualized assessment model to improve assessment accuracy; (b) use reinforcement learning algorithms to optimize personalized recommendation strategies and improve adjustment effects; and (c) discover new risk markers and adjustment response patterns to expand the knowledge base. As usage time increases, the system's intelligence level continuously improves, enabling personalized services.

[0106] (6) Data-driven effect evaluation and visual feedback: Traditional training often lacks objective effectiveness evaluation, making it difficult for users and managers to understand whether the training is effective. This implementation automatically generates a detailed effectiveness evaluation report after each training session, including before-and-after comparisons of changes in physiological indicators, EEG indicators, and subjective feelings, as well as historical trend analysis. Data visualization charts intuitively present the training effect, enhancing users' self-efficacy and training adherence.

[0107] (7) Good user experience and high compliance: This embodiment emphasizes user experience design: (a) the VR scenes are exquisite and realistic, with a strong sense of immersion and high user satisfaction; (b) the user interface is simple and user-friendly, requiring no professional training; (c) the training process is easy and enjoyable; and (d) the training duration is flexible (15-30 minutes), suitable for daily use. In practical applications, user training compliance reaches over 85%, significantly higher than traditional adjustment methods (usually <60%).

[0108] Example 2 To verify the ability of the system of the present invention to regulate the imbalance of autonomic nerve function and EEG state, a test subject who had been working under high load for a long time and whose physiological indicators showed an imbalance of autonomic nerve was selected for system testing.

[0109] System operation process and technical effects: (1) The tester logs into the system and completes the standardized baseline assessment process.

[0110] (2) The system's multimodal signal processing and state index calculation module performs fusion analysis on the above multimodal data. Based on the built-in machine learning model, the system outputs quantitative evaluation results.

[0111] (3) Based on the evaluation results, the data management and personalized solution generation module initiates the solution decision-making process. For example, when the evaluation result is a decline in attention index, the system generates and recommends a personalized training program based on the optimization goals of "increasing parasympathetic nerve activity (corresponding to increasing HRV-HF power), reducing sympathetic nerve excitability (corresponding to reducing skin conductance frequency and LF / HF ratio), and optimizing EEG frequency band balance (corresponding to increasing Alpha / Beta ratio)". The scene sequence is "temperate forest scene (10 minutes) → seaside sunset scene (5 minutes) → low-stimulation meditation scene (5 minutes)"; the training program combination is "breathing rhythm guidance program" and "progressive muscle tension adjustment program"; the total duration is set to 20 minutes.

[0112] (4) After the tester confirms the plan, the system starts. The VR-type natural and social environment immersive construction module renders a high-fidelity temperate forest scene. At the same time, the multi-sensory environment linkage control module performs collaborative control according to the plan parameters: the dynamic rhythm light environment sub-module adjusts the cabin lighting to a color temperature of 4000K and medium brightness; the aroma control sub-module starts and releases the set concentration of pine wood fragrance.

[0113] (5) The system executes the "breathing rhythm guidance program". A visual animation element (such as a periodically expanding / contracting light sphere) synchronized with the breathing command is generated in the VR scene, and the three-dimensional spatial audio submodule plays a gentle, rhythmic guidance sound in sync. The initial guidance frequency of the program is set to the test subject's baseline breathing frequency (12 breaths / minute), and according to the algorithm plan, it gradually transitions to the target frequency (6 breaths / minute) during the training process.

[0114] (6) After training begins, the closed-loop neurofeedback modulation module continues to run. Its real-time state monitoring submodule analyzes the incoming physiological and EEG data every second. After about 5 minutes, the algorithm detects changes in key indicators: the synchronization rate between the breathing curve and the guided rhythm exceeds 85%, the HF power in the HRV spectrum shows an upward trend, and the skin conductance response frequency drops to an average of 4 times / minute. Based on this, the adaptive control algorithm submodule determines that "the breathing guidance program is effective and the current parameters are appropriate" and maintains the existing control commands.

[0115] (7) Approximately 12 minutes later, the real-time status monitoring submodule detected new indicator changes: the average power of the EEG Alpha band increased by 120% compared to the baseline, the real-time calculated Alpha / Beta ratio reached 0.9, and the LF / HF ratio decreased to 2.0. The adaptive control algorithm submodule made a decision based on the preset "scene switching threshold rule" and triggered the scene switching command. The VR construction module smoothly transitioned the main scene from "temperate forest" to "seaside sunset", and the multi-sensory control module simultaneously adjusted the lighting color temperature to 3000K and switched the fragrance to a fresh marine scent.

[0116] (8) In the final stage of the “low-stimulation meditation scenario”, the system mainly performs the maintenance phase of the “breathing guidance program” and plays low-information-density guidance voice. The lighting environment submodule further dims the illuminance and maintains the color temperature at 2800K to match the low wakefulness requirements of the scenario.

[0117] (9) After training, the system automatically performed the post-training evaluation process. The collected resting-state physiological data showed that HRV-SDNN improved to 52ms (49% improvement from baseline), the LF / HF ratio decreased to 1.8 (44% improvement from baseline), and the skin conductance response frequency decreased to 3 times / minute (63% decrease from baseline). EEG data showed that the Alpha / Beta ratio increased to 1.1 (83% improvement from baseline). The subjective cognitive load score (0-10) submitted by the test subject decreased from 8 points to 4 points.

[0118] (10) The system report generation unit synthesizes the above data and outputs a structured report, concluding that "all physiological and EEG indicators showed significant changes in the expected direction during this adjustment." At the same time, the individualized profile submodule updates the test subject's profile, adding tags such as "positive response to the breathing rhythm guidance program" and "significant increase in alpha waves in forest-like scenarios" to optimize future program recommendations.

[0119] Long-term data recording: After the test subject used the system cyclically for 4 weeks, at least 5 times a week, the long-term data recording showed that after the test subject used the system cyclically, his objective physiological indicators such as HRV-SDNN and Alpha / Beta ratio showed a stable optimization trend, and his cognitive task performance indicators such as reaction time in the attention network test were improved simultaneously.

[0120] Example 3 To verify the ability of the system of the present invention to monitor and adaptively train physiological and EEG indicators related to cognitive function, a test subject with a cognitive function index deviation due to long-term high-intensity cognitive work was selected for system testing.

[0121] System operation process and technical effects: (1) The test subject completed the system baseline assessment. The quantitative baseline data are as follows: the subjective fatigue perception score was 8 points (0-10 scale); in terms of cognitive task performance, the reaction time for the conflict effect of the attention network test (ANT) was 125ms, and the accuracy rate of the N-back working memory task (2-back) was 68%; in terms of physiological signals, multiple time-domain and frequency-domain indicators of heart rate variability (HRV) were lower than the conventional reference range; in terms of EEG signals, the average power of the Theta band was higher than the group baseline, and the average power of the Beta band was lower. The above data were input into the system assessment module.

[0122] (2) The system's multimodal signal processing and state index calculation module analyzes the multimodal data. Based on its built-in cognitive efficacy assessment model, the system outputs quantitative results: the overall cognitive efficacy score is below the threshold, and the cognitive fatigue index is high.

[0123] (3) Based on the evaluation results, the data management and personalized solution generation module starts the solution generation program. Based on the adjustment goal of "optimizing the performance of attention and working memory related tasks and adjusting the EEG spectrum characteristics (increasing Beta power and decreasing Theta power)," the system generates a personalized solution: the main scene is "alpine meadow"; the training program combination is "selective attention task program" and "dynamic working memory load task program", the initial difficulty parameter is set to medium level; and the total duration is set to 25 minutes.

[0124] (4) After the tester confirms, the system starts. The VR construction module renders a high-fidelity, high-view alpine meadow scene. The multi-sensory environment linkage control module executes the settings simultaneously: the light environment submodule adjusts the lighting to a color temperature of 5500K and a high illuminance mode; the aroma control submodule releases a set concentration of mint fragrance.

[0125] (5) The system first executes the "Selective Attention Task Program" (using the Stroop task paradigm). Visual stimuli with color conflicts are presented in a random sequence set by the algorithm in the VR scene, and the test subject responds through the input device. The system records the reaction time and accuracy in real time. At the same time, the real-time status monitoring submodule continuously analyzes the EEG signals and extracts features such as the amplitude of the P300 component related to attention.

[0126] (6) During training, the closed-loop neurofeedback control module continued to operate. In the early stages of the task (approximately the first 2 minutes), the average accuracy rate was monitored at 72%, the average reaction time was 550ms, and no significant improvement was observed in the Beta power of the EEG. After approximately 5 minutes, an improvement in task performance was observed (the average accuracy rate increased to 82%, and the average reaction time shortened to 480ms), while the EEG signal showed an upward trend in Beta wave power and a decrease in Theta wave power. The adaptive control algorithm submodule determined that the current task difficulty parameters were appropriate according to the "performance-load balancing rule" and maintained them unchanged.

[0127] (7) Subsequently, the system switched to the "Dynamic Working Memory Loading Task Program" (using the N-back paradigm). The initial difficulty parameter of the program was set to "2-back". The system monitored the accuracy and reaction time of each trial in real time. According to the preset adaptive rule: if the accuracy of 5 consecutive trials is >85%, the program will increase the difficulty parameter to "3-back"; if the accuracy of 5 consecutive trials is <60%, it will decrease to "1-back". The difficulty parameter was dynamically adjusted twice according to this rule throughout the training phase.

[0128] (8) In the later stages of training (around the 20th minute), the real-time status monitoring submodule detected that several indicators exceeded the load threshold: the Beta wave power continued to climb to the warning value, the frequency of skin conductance response increased significantly, and the average reaction time was 15% longer than the previous time period. Based on this, the adaptive control algorithm submodule triggered the "intermittent recovery protocol". The system immediately paused the current task program, the VR scene automatically switched to the low cognitive load "mountain stream" scene, the light environment was adjusted to a soothing mode, and a standardized "breathing rhythm guidance program" lasting 2 minutes was started. After the program ended, the system automatically resumed the previous training task, and the test subject's subsequent task performance indicators returned to within the load threshold.

[0129] (9) After training, the system performed a post-training assessment. Data comparison showed that the reaction time for the ANT conflict effect was shortened to 100ms (a 20% improvement compared to before training); the accuracy of the N-back task (2-back) improved to 79% (an 11% improvement compared to before training); and the subjective fatigue score decreased to 5. Physiological and EEG data also showed that some HRV indicators and the Beta / Theta power ratio showed an improving trend.

[0130] (10) The system report generates the conclusion that "the performance indicators of the cognitive task have been specifically improved". The individualized profile submodule updates the test subject's profile and records its response pattern and load threshold to the N-back task adaptive rules.

[0131] Long-term data recording: The test subject underwent periodic training over 6 weeks according to the training plan generated by the system. Subsequent evaluation showed that the ANT conflict effect reaction time stabilized at 85ms, the 2-back accuracy of the N-back task improved to 88%, and the 3-back accuracy improved from a baseline of 55% to 75%. Long-term data trends support the system's sustained effectiveness in modulating specific cognitive function-related indicators.

[0132] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A VR-based natural environment state control cabin system based on multimodal physiological-brain-computer interface closed-loop feedback, characterized in that, Integrated within a closed compartment, the system includes: The multimodal physiological-brain function signal synchronous acquisition module is used to synchronously acquire the user's peripheral physiological signals and brain function signals in real time. The multimodal signal processing and state index calculation module is communicatively connected to the multimodal physiological-brain function signal synchronous acquisition module. It is used to perform fusion analysis on the acquired multimodal data and establish an evaluation model to output state indexes and regulatory trigger signals. The VR-like natural and social environment immersive construction module is communicatively connected to the multimodal signal processing and state index calculation module, and is used to construct and present a virtual reality environment based on state indexes. The virtual reality environment includes natural environment scenes, social environment scenes, and interactive cognitive and regulatory tasks. The multi-sensory environment linkage control module is set up in conjunction with the VR-type natural and social environment immersive construction module to link and control auditory stimulation, light environment and olfactory stimulation in the closed cabin. The closed-loop neural feedback control module is connected to the multimodal physiological-brain function signal synchronous acquisition module, the multimodal signal processing and state index calculation module, the VR-like natural and social environment immersive construction module, and the multisensory environment linkage control module, respectively. It is used to dynamically adjust the presentation parameters of the virtual reality environment, the execution difficulty of the physiological signal training task, and the control parameters of the multisensory environment stimulation based on the real-time monitored physiological and brain function states. The data management and personalized solution generation module is communicatively connected to the closed-loop neural feedback control module, and is used to manage user data, build individual profiles, and generate personalized training solutions based on the current state. The cabin environment control module is used to maintain the temperature, humidity, ventilation, and clean physical environment inside the enclosed cabin. The system integration and control module communicates with all the above modules and is used to coordinate the operation and data interaction of each module.

2. The system according to claim 1, characterized in that, The multimodal physiological-brain function signal synchronous acquisition module includes: The physiological signal acquisition submodule is used to acquire the user's electrocardiogram, heart rate variability, skin conductance, respiratory rate and body temperature signals based on a multi-channel physiological recorder. The brain-computer interface submodule is used to collect the user's multichannel electroencephalogram (EEG) signals or functional near-infrared spectral signals.

3. The system according to claim 1, characterized in that, The multimodal signal processing and state index calculation module includes: The data preprocessing submodule is used to filter, denoise, and normalize the acquired raw signals, and to extract multi-dimensional features from the processed raw signals. The feature fusion submodule is used to fuse extracted multi-dimensional features using a fusion method that combines feature-level fusion and decision-level fusion. The user status assessment submodule is used to run machine learning models to classify and quantify user status. The control parameter mapping submodule is used to generate system control parameter adjustment instructions based on the calculated state indicators and preset rules.

4. The system according to claim 1, characterized in that, The VR-based immersive natural and social environment construction module includes: The VR display submodule is used to present virtual reality scenes to users; The virtual reality scene construction submodule is used to construct virtual reality scenes based on a library of natural environment scenes, a library of social environment scenes, and a library of interactive cognition and regulation tasks.

5. The system according to claim 1, characterized in that, The multi-sensory environment linkage control module includes: The 3D spatial audio submodule is used to render 3D spatial sound effects that match the VR scene; The dynamic rhythmic lighting environment submodule is used to adjust the color temperature and brightness of the cabin lighting to simulate the natural light rhythm. The aroma control submodule is used to control the release of specific aroma essential oils according to the user's status and adjustment needs.

6. The system according to claim 1, characterized in that, The closed-loop neural feedback modulation module includes: The real-time status monitoring submodule is used to analyze physiological and EEG signals in real time to obtain key status indicators; The adaptive control algorithm submodule is used to generate adjustment instructions based on the key state indicators and according to preset mapping rules or optimization algorithms, so as to adaptively switch VR scenes, adjust scene parameters, adjust task difficulty, or adjust environmental stimulus parameters.

7. The system according to claim 1, characterized in that, The data management and personalized solution generation module includes: The data storage submodule is used to store users' training history data, physiological EEG signals, evaluation results, and subjective evaluations; The personalized profile submodule is used to build individual profiles based on historical data, including user baseline characteristics, risk characteristics, and preference characteristics; The personalized scheme generation submodule is used to generate adjustment schemes, including target scenarios, training tasks, and initial parameters, based on current status indicators and individual profiles.

8. The system according to claim 1, characterized in that, The cabin environment control module includes: Temperature control submodule is used to regulate and maintain the temperature inside the enclosed compartment; A humidity control submodule is used to adjust and maintain the humidity inside the enclosed chamber; The ventilation system submodule is used to provide a circulation of filtered fresh air; The sound insulation and noise reduction submodule is used to reduce the interference of external noise on the cabin environment; A moisture-proof and corrosion-resistant submodule is used to protect the electronic equipment and cabin structure of the system in high-temperature, high-humidity, and high-salt environments. The lighting system submodule is used to provide basic cabin lighting, including emergency lighting. The safety protection submodule is used to monitor the safety of the cabin environment and trigger alarms and protective actions in case of abnormalities.

9. An electronic device comprising a memory, a processor, and a computing program stored in the memory and executable on the processor, characterized in that, When the processor executes the computing program, it is used to control the system as described in any one of claims 1-8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it is used to control the system as described in any one of claims 1-8.

Images