Psychological stress monitoring method and system based on virtual reality

Abstract

The invention relates to the technical field of virtual reality and health monitoring, and particularly discloses a psychological stress monitoring method and system based on virtual reality. The system comprises a physiological signal acquisition module, a virtual reality environment generation module, a stress reaction quantification module, a self-adaptive pressure regulation and control module and a pressure state evaluation module, pressure source elements in a virtual reality scene are dynamically adjusted by acquiring multi-modal physiological signals of a user in real time, the stress level of the user is quantified, and the pressure state of the user is evaluated. And psychological stress grade evaluation is output based on the personalized stress response model, so that closed-loop psychological stress monitoring with high signal-to-noise ratio and high ecological efficiency is realized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of virtual reality and health monitoring, and particularly relates to a psychological stress monitoring method and system based on virtual reality. BACKGROUND

[0002] Psychological state monitoring and intervention technology is an important part of modern health management, aiming to evaluate individual psychological load and emotional state through objective physiological indicators, and to provide scientific basis for psychological health maintenance. Virtual reality technology, as an immersive human-computer interaction method, provides a new technical approach for the induction and quantitative analysis of psychological state by constructing a highly simulated three-dimensional environment.

[0003] Psychological stress monitoring based on virtual reality is a hot research direction at present, and its core goal is to use virtual reality environment as a standardized stress source to synchronously collect physiological response data of users, so as to realize objective and quantitative evaluation of psychological stress level.

[0004] The prior art usually adopts a pre-set static virtual reality scene to simply associate with biological sensor data, and this method has limitations. The stimulation mode of the static scene is fixed and lacks pertinence, and it is difficult to effectively induce specific stress response of different users, resulting in low signal-to-noise ratio of physiological data and insufficient monitoring sensitivity.

[0005] At the same time, due to the inability to simulate the complex characteristics of dynamic evolution of stress sources in the real world, the ecological validity of the existing monitoring environment is seriously insufficient, and the physiological response mode obtained is difficult to accurately reflect the real state of the user under dynamic changing stress. Therefore, how to construct a virtual reality monitoring environment that can adaptively induce effective stress response and has high ecological validity has become a technical problem to be solved in the field. SUMMARY

[0006] The technical problem to be solved by the present application is to overcome the defects in the prior art that the virtual reality psychological stress monitoring uses a static scene stimulation mode which is fixed and lacks pertinence, is difficult to effectively induce specific stress response of users, results in low signal-to-noise ratio of physiological data, has insufficient monitoring sensitivity, and due to the inability to simulate the dynamic evolution characteristics of stress sources in the real world, the ecological validity is seriously insufficient, and to provide a psychological stress monitoring method and system based on virtual reality which can adaptively induce effective stress response and has high ecological validity.

[0007] The present application provides a psychological stress monitoring system based on virtual reality, which comprises a physiological signal acquisition module, a virtual reality environment generation module, a stress response quantification module, an adaptive stress regulation module and a stress state evaluation module.

[0008] The physiological signal acquisition module is used to collect multimodal physiological signals of users during virtual reality experiences in real time, including electrocardiogram signals, skin conductance signals, respiratory waveforms, and electroencephalogram signals.

[0009] The virtual reality environment generation module is used to build and present a virtual reality scene containing a variety of dynamic pressure source elements. The pressure source elements in the scene can be dynamically adjusted according to preset rules or real-time feedback.

[0010] The stress response quantification module is used to preprocess and extract features from the collected multimodal physiological signals, and calculate stress response indicators that reflect the degree of activation of the user's autonomic nervous system.

[0011] The adaptive stress regulation module is used to dynamically adjust the presentation intensity, complexity, and evolution rate of stress source elements in the virtual reality environment generation module based on the real-time stress response indicators output by the stress response quantification module, so as to maintain the user's immersive experience within the preset target stress level range.

[0012] The stress state assessment module is used to build a personalized stress response model for users based on time series data of stress response indicators and adaptive regulation history, and output the final psychological stress level assessment result.

[0013] Furthermore, the physiological signal acquisition module includes an electrocardiogram (ECG) acquisition unit, a skin conductance acquisition unit, a respiratory monitoring unit, and an electroencephalogram (EEG) acquisition unit.

[0014] The ECG acquisition unit acquires ECG signals through two dry electrodes placed on the user's chest and calculates the ratio of high-frequency power to low-frequency power for beat-by-beat heart rate and heart rate variability.

[0015] The skin conductance acquisition unit measures the skin conductance level and its fluctuation frequency through two silver chloride electrodes fixed to the distal phalanx of the user's index finger.

[0016] The respiratory monitoring unit acquires respiratory waveforms and extracts respiratory rate and respiratory rhythm disorder by using a respiratory plethysmography sensor wrapped around the user's chest and abdomen.

[0017] The EEG acquisition unit records EEG signals from the prefrontal and parietal lobe regions using an 8-lead EEG cap worn on the user's head, and analyzes the relative power spectral density of alpha and beta waves.

[0018] Furthermore, the virtual reality environment generation module includes a scene database, a stress source element library, and a real-time rendering engine.

[0019] The scenario database stores at least five basic virtual environment templates, covering social assessment scenarios, time pressure scenarios, multitasking scenarios, uncertainty threat scenarios, and physical hazard simulation scenarios.

[0020] The stress source element library contains more than 20 dynamic stress source elements that can be called independently or in combination. These elements specifically include the virtual character's gaze direction and density, the complexity and update frequency of task instructions, the intensity of abrupt changes in ambient lighting and sound, the probability of obstacles appearing on the path and movement speed, and the remaining time prompts for the countdown.

[0021] The real-time rendering engine selects specific elements from the pressure source element library and performs real-time fusion and rendering according to the specified intensity parameters based on the control instructions issued by the adaptive pressure control module, generating a virtual reality scene stream with continuous dynamic pressure gradient.

[0022] Furthermore, the stress response quantification module includes a signal preprocessing submodule, a feature extraction submodule, and an index fusion submodule. The signal preprocessing submodule performs second-order Butterworth bandpass filtering on the raw physiological signal to remove power frequency interference and baseline drift, and performs amplitude normalization on the signals of each channel.

[0023] The feature extraction submodule extracts the standard deviation of adjacent RR interval sequences from the preprocessed ECG signal, extracts the peak count of skin conductance response within a specific number from the skin conductance signal, extracts the coefficient of variation of inspiratory and expiratory durations from the respiratory signal, and extracts the ratio of the power of the theta band to the beta band in the prefrontal cortex from the EEG signal.

[0024] The index fusion submodule uses a linear weighting method based on principal component analysis to fuse multiple extracted physiological features into a scalar value in the range of 0 to 100, called the comprehensive stress index. The higher the index value, the stronger the activation of the user's autonomic nervous system.

[0025] Furthermore, the adaptive pressure control module includes a target pressure range setting unit, a control strategy decision-making unit, and an instruction generation unit.

[0026] The target stress range setting unit allows system operators to set a comprehensive stress index target range for users based on the monitoring purpose. The lower limit of this range is the minimum stimulus intensity that can ensure a measurable physiological response, while the upper limit is the maximum tolerable intensity that avoids excessive stress and discomfort.

[0027] The control strategy decision unit continuously receives the real-time comprehensive stress index sent by the stress response quantification module and compares it with the target stress range. When the real-time index is less than the lower limit of the target range, the decision unit generates a control strategy to enhance the stress stimulus. When the real-time index is higher than the upper limit of the target range, the decision unit generates a control strategy to reduce the stress stimulus. The control strategy is specifically manifested as an adjustment scheme for the intensity parameters and dynamic characteristics of one or more stress source elements in the virtual reality environment.

[0028] The instruction generation unit converts the control strategy into standardized control instructions that can be parsed by the virtual reality environment generation module. The instructions include the identifier of the pressure source element to be adjusted, the target intensity value, and the change transition duration parameter.

[0029] Furthermore, the stress state assessment module includes a data caching unit, a model update unit, and a level assessment unit. The data caching unit stores the comprehensive stress index time series data and the corresponding adaptive control command history for the most recent 10 minutes in a 2-second cycle.

[0030] The model update unit uses a recursive least squares algorithm to update a second-order autoregressive model representing the user's individual stress response characteristics online based on cached data. The parameters of this model reflect the user's baseline response level, response gain, and recovery speed to virtual reality stress stimuli.

[0031] The rating assessment unit uses an updated personal stress response model to conduct a multi-dimensional analysis of the user's average stress level, stress response variability, and stress recovery ability during a complete monitoring session. The final output is a psychological stress assessment conclusion with five levels, which is accompanied by a detailed analysis of the scores for each dimension.

[0032] As one embodiment of the present invention, the control strategy decision unit of the adaptive pressure control module adopts an algorithm based on the proportional-integral-derivative control principle to achieve precise pressure level maintenance.

[0033] The algorithm sets the midpoint of the target stress range as the set point and uses the real-time comprehensive stress index as the process variable to calculate the adjustment amount of the stress stimulus intensity.

[0034] The proportional term is used to quickly respond to the current deviation between the comprehensive stress index and the set point, the integral term is used to eliminate persistent steady-state errors, and the derivative term is used to predict the changing trend of the comprehensive stress index and implement advance correction.

[0035] By adjusting the three gain parameters of the proportional-integral-derivative controller, the system can adapt to the different stress sensitivities of different users, achieving smooth and precise stress experience control.

[0036] As one embodiment of the present invention, the design of the stress source element library in the virtual reality environment generation module follows the multi-dimensional principle of psychological stress theory.

[0037] Each stressor element is assigned three attribute parameters: cognitive load, emotional arousal, and time pressure.

[0038] When the real-time rendering engine receives a control command, it calculates the target value required for each dimension according to the overall pressure intensity required by the command and the preset weight allocation algorithm. Then, it selects the pressure source elements whose attribute combinations are closest to the target value from the element library for instantiation and rendering.

[0039] This multi-dimensional mapping mechanism ensures that virtual stress scenarios are not only controllable in intensity, but also scientific and diverse in the composition of stress traits.

[0040] As one embodiment of the present invention, the level assessment unit in the pressure state assessment module uses a fuzzy logic reasoning mechanism to process the multi-dimensional features output by the model.

[0041] This unit defines three input language variables, corresponding to average stress level, stress response variability, and stress recovery ability, respectively. Each variable is divided into three fuzzy sets.

[0042] At the same time, one output language variable, namely the psychological stress level, was defined and divided into 5 fuzzy sets.

[0043] The evaluation unit incorporates 25 fuzzy rules based on expert knowledge to describe the nonlinear relationship between input features and output levels.

[0044] Through three steps—fuzzification, rule evaluation, and defuzzification—the system can transform continuous model output features into discrete stress levels that conform to the judgment habits of human experts, thereby improving the interpretability and practicality of the evaluation results.

[0045] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention integrates physiological signal acquisition, virtual reality environment generation, stress response quantification, adaptive stress regulation, and stress state assessment to form a complete closed-loop monitoring system. Its core innovation lies in the introduction of an adaptive stress regulation mechanism based on real-time physiological feedback. This mechanism dynamically adjusts the stressor elements in the virtual reality scene according to the user's current stress level, ensuring that the applied stress stimulus is always within an ideal range that both induces specific physiological responses and avoids excessive stress. This dynamic adaptive stimulus presentation mode fundamentally overcomes the shortcomings of static scene stimuli—fixed stimuli and poor targeting—and improves the signal-to-noise ratio of physiological signal acquisition and the sensitivity of the monitoring system.

[0046] 2. The virtual reality environment generation module constructed in this invention possesses high dynamism and multi-dimensional stressor simulation capabilities. Its stressor element library contains various dynamic elements that can be called independently or in combination, and the element attribute design follows multi-dimensional psychological stress theories such as cognitive load, emotional arousal, and time pressure. Combined with a real-time rendering engine, the system can generate complex scenes simulating the dynamic evolution characteristics of real-world stressors. This highly ecologically valid stress environment makes the collected user physiological response patterns closer to their actual state under real-world dynamic stress, greatly improving the authenticity and extrapolation validity of the monitoring results.

[0047] 3. The stress assessment module of this invention does not simply rely on instantaneous physiological indicators, but rather on complete monitoring session data, performing multi-dimensional comprehensive analysis through online updates of the user's personalized stress response model. This model can capture the user's unique stress baseline, response gain, and recovery dynamics characteristics. Combined with fuzzy logic reasoning, the assessment results not only provide an overall stress level, but also offer analyses of multiple dimensions such as average stress level, volatility, and recovery ability. This analysis method based on personalized models and multi-dimensional characteristics makes the final psychological stress assessment more accurate, comprehensive, and possesses stronger individual relevance and clinical reference value. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the overall technical architecture of the virtual reality-based psychological stress monitoring system proposed in this invention. Figure 2 This is a schematic diagram of the core principle framework of the adaptive pressure regulation mechanism in this invention; Figure 3 This is a diagram of the multi-dimensional pressure source element mapping and rendering logic framework of the virtual reality environment generation module in this invention; Figure 4 This is a logical flowchart of the stress response quantification and index fusion in this invention; Figure 5 This is a schematic diagram of the multi-dimensional assessment principle of the pressure state assessment module in this invention, which is based on a personalized model and fuzzy logic. Detailed Implementation

[0049] The virtual reality-based psychological stress monitoring system comprises five core modules: a physiological signal acquisition module, a virtual reality environment generation module, a stress response quantification module, an adaptive stress regulation module, and a stress state assessment module. These modules constitute a complete closed-loop monitoring system. Please refer to the appendix for its overall architecture. Figure 1 .

[0050] The physiological signal acquisition module is responsible for capturing multimodal physiological signals of the user during the virtual reality experience in real time.

[0051] The virtual reality environment generation module is responsible for constructing and presenting virtual scenes that include dynamic stressor elements. The stress response quantification module is responsible for processing physiological signals and calculating indicators reflecting the degree of activation of the autonomic nervous system.

[0052] The adaptive stress regulation module is responsible for dynamically adjusting the stress stimulation in the virtual scene based on real-time stress indicators.

[0053] The stress status assessment module is responsible for building a personalized user model based on historical data and outputting the final stress level assessment.

[0054] The physiological signal acquisition module contains four dedicated acquisition units.

[0055] The ECG acquisition unit acquires raw ECG signals through two dry electrodes placed on the user's chest.

[0056] This unit has a built-in analog front-end amplifier and analog-to-digital converter to amplify and digitize microvolt-level ECG signals.

[0057] The digital signal processor performs QRS complex detection on the electrocardiogram waveform, accurately identifies the peak position of each R wave, and calculates the continuous R wave interval, i.e., the RR interval sequence.

[0058] Based on the RR interval sequence, the system further calculates the frequency domain index of heart rate variability, specifically extracting the ratio of power in the high-frequency band (0.15Hz to 0.4Hz) to power in the low-frequency band (0.04Hz to 0.15Hz).

[0059] The skin conductance acquisition unit measures skin conductivity levels using two silver chloride electrodes fixed to the distal phalanx of the user's index finger. The unit applies a constant 0.5V DC excitation voltage and measures changes in microcurrent passing through the skin tissue.

[0060] The signal sampling rate was 32 samples / s. After moving average filtering, two key features were extracted: the baseline value of skin conductance and the number of peak values ​​of skin conductance response within a specific time window. The peak value was defined as an event in which the conductance value increased by more than 0.02 μS and lasted for more than 1s.

[0061] The respiratory monitoring unit uses a respiratory inductive plethysmography sensor, which consists of a braided conductive fiber strip wrapped around the user's chest and abdomen.

[0062] The sensor detects changes in thoracic cavity volume caused by respiratory movements, which in turn causes changes in its own inductance, modulating the frequency of the output AC signal. The demodulation circuit converts the frequency change into a voltage waveform, i.e., the respiratory waveform.

[0063] The digital signal processor extracts the respiratory rate from the respiratory waveform, measured in breaths per minute, and calculates the coefficient of variation between the inspiratory and expiratory durations as an indicator of respiratory rhythm disorder.

[0064] The EEG acquisition unit records EEG signals using an 8-lead EEG cap worn on the user's head.

[0065] The electrode layout covers the FP1FP2FPZ channels in the frontal lobe and the P3P4PZ channels in the parietal lobe, with the reference electrode connected to both earlobes.

[0066] The signal amplifier gain is set to 10,000 times, and the bandwidth is 0.5Hz to 50Hz.

[0067] The analog-to-digital conversion accuracy is 16 bits, and the sampling rate is 256 samples / second. The digital signal processor performs a fast Fourier transform on the EEG signal of each channel, calculates the power spectral density, and focuses on analyzing the relative power ratio of the theta band (4Hz to 8Hz) and the beta band (13Hz to 30Hz) in the prefrontal region.

[0068] The virtual reality environment generation module is responsible for constructing dynamic stress scenarios; please refer to the appendix for its internal structure. Figure 3 This module contains three core sub-parts.

[0069] The scenario database stores at least five basic virtual environment templates.

[0070] The social assessment scenario simulates public speaking or interview situations, including virtual audience roles and evaluation feedback mechanisms.

[0071] The time pressure scenario simulates an emergency task completion situation, including a dynamic countdown and progress prompts.

[0072] The multi-task processing scenario simulates a complex working environment, including parallel task instruction queues and resource contention.

[0073] Uncertainty threat scenarios simulate potentially dangerous environments, including random event triggers and ambiguous clues.

[0074] The physical hazard simulation scenarios simulate high or edge situations, incorporating visual depth cues and balance challenges.

[0075] The pressure source element library contains more than 20 dynamic pressure source elements that can be called independently or in combination.

[0076] These elements have multidimensional attributes, including cognitive load dimension, emotional arousal dimension, and time pressure dimension.

[0077] Specific elements include the virtual character's gaze direction and density parameters. The gaze direction can be set to direct gaze, avoidance, or randomness, while the density parameter controls the number of virtual characters per unit area.

[0078] The task instruction elements include complexity parameters and update frequency parameters. Complexity is defined by the number of task steps and the number of logical conditions, while update frequency is defined by the time interval between the issuance of new instructions.

[0079] Ambient lighting and sound elements include abrupt change intensity parameters, defined as the instantaneous amplitude and rate of change in brightness or loudness.

[0080] The path obstacle element includes an occurrence probability parameter and a movement speed parameter. The occurrence probability determines the probability of a new obstacle being generated in each time period, while the movement speed controls the rate at which obstacles approach the user.

[0081] The countdown element includes parameters for how the remaining time is indicated, such as a numerical display, a progress bar, or an audible alarm.

[0082] The real-time rendering engine receives control instructions from the adaptive pressure control module. The instructions are in JSON format and include the target pressure intensity value, dimension weight coefficient, and transition duration parameter.

[0083] The engine calculates the required target values ​​for each dimension based on the target pressure intensity and weight coefficients, and selects the element whose attribute combination is closest to the target value from the pressure source element library for instantiation.

[0084] The rendering engine is developed using Unity3D and supports a 90fps frame rate to ensure visual immersion and real-time interactivity.

[0085] The stress response quantification module processes raw physiological signals and calculates comprehensive stress indices. Please refer to the appendix for the processing flow. Figure 4 This module contains three sequentially executed submodules. The signal preprocessing submodule receives multi-channel physiological data streams.

[0086] For ECG signals, a second-order Butterworth bandpass filter with a passband frequency of 0.5 Hz to 35 Hz is applied to suppress power line interference and baseline drift.

[0087] For the skin conductance signal, a 0.1Hz high-pass filter is applied to eliminate slow baseline changes.

[0088] For respiratory signals, a 0.1Hz to 1Hz bandpass filter is used to preserve the fundamental and harmonic components of the respiratory signal.

[0089] For EEG signals, a 0.5Hz high-pass filter and a 50Hz notch filter are used to eliminate DC offset and power supply interference.

[0090] All channel signals are then amplitude normalized, mapping the original voltage values ​​to a range of -1 to 1.

[0091] The feature extraction submodule extracts four core physiological features from the preprocessed signal.

[0092] The standard deviation of adjacent RR interval sequences was extracted from the electrocardiogram signal. The calculation window was a 30s sliding window with a step size of 2s.

[0093] Peak counts of skin conductance responses were extracted from the skin conductance signal, with a statistical window of 60 s and a threshold of 0.02 μS.

[0094] The coefficient of variation of inspiratory and expiratory durations was extracted from the respiratory signals. The calculation window was 60 seconds, and the coefficient of variation was defined as the ratio of the standard deviation to the mean.

[0095] The power ratio of the theta band to the beta band in the prefrontal cortex was extracted from the EEG signal. The calculation window was 10 s, and the power spectrum was estimated using the Welch method.

[0096] The indicator fusion submodule uses a linear weighting method based on principal component analysis to fuse multiple features into a single comprehensive stress index.

[0097] First, a four-dimensional feature vector is constructed, and each feature is standardized using the Z-score. Then, the feature vector is projected onto the direction of the first principal component using a pre-computed principal component analysis transformation matrix.

[0098] The projected values ​​are linearly mapped to the range of 0 to 100. The overall stress index is updated every 2 seconds, with higher values ​​indicating stronger activation of the autonomic nervous system.

[0099] The adaptive stress regulation module dynamically adjusts stress stimulation based on real-time stress indicators. Please refer to the appendix for its core principles. Figure 2 This module contains three functional units.

[0100] The target pressure range setting unit provides a user interface for system operators to configure parameters.

[0101] The operator sets a target range for the comprehensive stress index, with a lower limit typically between 25 and 30 to ensure a measurable physiological response, and an upper limit typically between 70 and 75 to avoid excessive stress and discomfort. The median of the target range is used as the control setpoint.

[0102] The control strategy decision-making unit continuously receives real-time comprehensive stress indices from the stress response quantification module. This unit employs an algorithm based on the proportional-integral-derivative control principle to achieve precise stress level maintenance.

[0103] The algorithm uses the real-time comprehensive stress index as a process variable and the target median as a set point to calculate the adjustment amount of the stress stimulus intensity.

[0104] The proportional term coefficient is set to 0.5, the integral time constant is set to 60s, and the derivative time constant is set to 10s.

[0105] The control algorithm outputs a stress stimulus intensity adjustment value, ranging from -1 to 1. A negative value indicates a weakening of the stimulus, while a positive value indicates an strengthening of the stimulus. The regulation strategy decision unit converts this adjustment value into a specific adjustment scheme for one or more stressor elements in the virtual reality environment.

[0106] For example, when increased pressure is needed, the density of virtual characters may be increased by 10%, the complexity of the task may be increased by one level, and the remaining countdown time may be shortened by 15%.

[0107] The instruction generation unit converts the adjustment scheme into standardized control instructions that can be parsed by the virtual reality environment generation module.

[0108] The instructions are serialized using ProtocolBuffers and contain an array of string identifiers for pressure source elements, an array of floating-point numbers representing target intensity values, and an array of integers representing the transition duration.

[0109] The instructions are sent to the rendering engine via a gigabit Ethernet interface with a transmission latency of less than 5ms.

[0110] The pressure status assessment module performs the final pressure rating assessment; please refer to the appendix for the assessment principle. Figure 5 This module comprises three collaborative units. The data caching unit stores the comprehensive stress index time-series data for the most recent 10 minutes, along with the corresponding adaptive control command history, in a 2-second cycle. Data storage employs a circular buffer structure with a capacity of 300 sample points.

[0111] Each data point includes a timestamp, a comprehensive stress index value, and a hash value of the control instruction. The model update unit uses a recursive least squares algorithm to update the user's personalized stress response model online. The model is a second-order autoregressive structure, mathematically expressed as: ; This represents the overall stress index at the current moment. and This represents the historical values ​​from the previous two moments. Indicates the intensity of the stress stimulus input. This represents the modeling error. Model parameters. , , The algorithm updates the data every 30 seconds using a recursive least squares algorithm, with a forgetting factor set to 0.99. Parameters The autocorrelation characteristics reflecting the user's stress response, parameters The parameters reflect the periodic fluctuations of the reaction. It reflects the increase in the user's sensitivity to stress stimuli.

[0112] The rating assessment unit utilizes an updated individual stress response model for multidimensional analysis. Three key features are extracted: average stress level (mean of the composite stress index during the monitoring session), stress response volatility (standard deviation of the composite stress index), and stress resilience (model parameters). The reciprocal of.

[0113] The evaluation unit uses a fuzzy logic reasoning mechanism to process these three features.

[0114] Three input language variables are defined: average stress level is divided into three fuzzy sets: low, medium and high; stress response fluctuation is divided into three fuzzy sets: stable, moderate and severe; and stress recovery ability is divided into three fuzzy sets: fast, medium and slow.

[0115] The output language variable, psychological stress level, is defined as being divided into five fuzzy sets: very low, lower, medium, higher, and very high.

[0116] The system has 25 built-in fuzzy rules based on expert knowledge. For example, if the average stress level is high, the stress response is highly volatile, and the stress recovery ability is slow, then the psychological stress level is very high.

[0117] The reasoning process employs the Mamdani fuzzy reasoning method, and defuzzification uses the centroid method. The final output is a psychological stress assessment conclusion divided into 5 levels, along with a detailed analysis report of the scores for each dimension.

[0118] The system operation process is as follows: The user wears a physiological signal acquisition device and a virtual reality head-mounted display. After system initialization, the virtual reality environment generation module presents an initial neutral scene.

[0119] The physiological signal acquisition module begins to continuously acquire multimodal physiological signals.

[0120] The stress response quantification module calculates the comprehensive stress index every 2 seconds. The adaptive stress regulation module generates regulation commands based on the deviation between the comprehensive stress index and the target range, and sends them to the virtual reality environment generation module.

[0121] The virtual reality environment generation module dynamically adjusts the pressure source elements in the scene according to instructions.

[0122] This closed-loop process continues to operate, creating an adaptive stress stimulation environment.

[0123] The stress status assessment module continuously caches data and updates the model throughout the monitoring session, and generates a final assessment report after the session ends.

[0124] The entire system is implemented using a mix of C++ and Python programming and runs on a high-performance workstation equipped with an Intel i7 processor and an NVIDIA RTX 3080 graphics card, ensuring real-time processing and rendering performance.

[0125] In practice, the system hardware connections adopt a modular design.

[0126] The physiological signal acquisition device communicates with the host via Bluetooth 5.0, with a transmission interval of 50ms. The virtual reality head-mounted display connects to the host via a DisplayPort 1.4 interface, supporting a 120fps refresh rate.

[0127] All software modules exchange data via shared memory and message queues to ensure low-latency communication. The system provides a graphical configuration interface, allowing operators to set monitoring parameters, view real-time data streams, and export evaluation reports.

[0128] The system implements multiple security protection mechanisms.

[0129] The physiological signal acquisition circuit has electrical isolation and current limiting protection to ensure user safety.

[0130] The brightness and contrast of the virtual reality scene are automatically limited to a safe range for the eyes. The adaptive control module has an emergency stop function; when the overall stress index is detected to be greater than 10 seconds, it automatically switches to the lowest stress scene and issues an alarm.

[0131] All user data is stored in encrypted form and complies with medical data privacy protection standards.

[0132] The system calibration and verification process includes two stages: equipment calibration and algorithm verification.

[0133] During the equipment calibration phase, a standard signal generator is used to generate test signals with known amplitude and frequency to verify the gain, linearity, and frequency response of each physiological channel.

[0134] During the algorithm validation phase, a standard population was recruited to conduct stress-induced experiments, and the results of the systematic evaluation were analyzed for consistency with the results of interviews with clinical psychologists to ensure the validity of the evaluation.

[0135] The system undergoes a full calibration every six months and a quick self-test before each use.

[0136] The system's scalability design supports multi-user parallel monitoring. Through containerized deployment, a single workstation can support up to four independent monitoring sessions simultaneously.

[0137] Each session runs in an isolated Docker container with independent resource allocation and data storage.

[0138] System administrators can monitor the status of all sessions and perform unified configuration and management through a web interface. This architecture is suitable for group stress assessments and large-scale clinical research scenarios.

[0139] This embodiment provides an optimized implementation scheme for specific user groups. The scheme is primarily designed for elderly users and users with specific anxiety disorders, with optimizations specifically implemented in the design of stressor elements, physiological signal processing algorithms, and safety protection mechanisms.

[0140] The stress source element library in the virtual reality environment generation module has been adaptively adjusted for elderly users. The default density of virtual characters in social assessment scenarios has been reduced by 30% to avoid excessive social stress.

[0141] In time-pressure scenarios, the countdown timer is extended by 50%, providing more time to react.

[0142] In multi-tasking scenarios, the number of parallel tasks is limited to two to reduce cognitive load. In uncertain threat scenarios, clear warnings are provided to reduce anxiety caused by ambiguity.

[0143] In physically hazardous simulated scenarios, heightened sensory cues are reduced to prevent dizziness.

[0144] For users with anxiety disorders, the stress factor library has added a category of soothing elements, including natural landscape scenes, guided meditation audio, and rhythmic breathing cues.

[0145] These factors can be incorporated into the monitoring process as a means of pressure regulation.

[0146] The physiological signal acquisition module adjusts the signal processing parameters to suit the physiological characteristics of elderly users.

[0147] The heart rate variability analysis of the ECG acquisition unit incorporates an age compensation factor, adjusting the interpretation range of the high-frequency power to low-frequency power ratio based on the user's actual age.

[0148] To address the potential reduction in sweat gland reactivity, the skin conductance acquisition unit adjusted the peak detection threshold from 0.02 μS to 0.01 μS.

[0149] To address potential irregularities in breathing patterns, the respiratory monitoring unit extends the calculation window for respiratory rhythm disorder from 60 seconds to 120 seconds, thereby improving the stability of the indicators.

[0150] To address potential EEG signal attenuation, the amplifier gain in the EEG acquisition unit has been increased from 10,000 times to 15,000 times to ensure signal quality.

[0151] The feature extraction algorithm of the stress response quantification module is optimized for population specificity. For elderly users, the ECG feature extraction adds a QT interval variability index to calculate the standard deviation of consecutive QT intervals.

[0152] Skin conductance feature extraction enhances the analysis of the slow drift trend of skin conductance levels, and linear regression is used to fit the baseline changes within 60 minutes.

[0153] Respiratory feature extraction increases the coefficient of variation of tidal volume, and the change in respiratory volume for each breath is calculated by integrating the respiratory waveform.

[0154] The EEG feature extraction incorporates a global field synchronization index to analyze the spatial consistency of multi-channel EEG signals.

[0155] For users with anxiety disorders, the feature extraction adds a startle reflex index, which assesses the degree of startle response by analyzing the acceleration of heart rate 100ms after a sudden sound stimulus.

[0156] The control strategy of the adaptive pressure regulation module is adjusted gently.

[0157] The proportional-integral-derivative (PID) controller parameters were retuned, with the proportional term coefficient reduced from 0.5 to 0.3, the integral time constant extended from 60 s to 120 s, and the derivative time constant extended from 10 s to 20 s. This adjustment resulted in a smoother system response, preventing abrupt changes in pressure levels.

[0158] The control strategy decision unit adds safety verification logic. When the rate of change of the comprehensive stress index is detected to be greater than 10 points / s, the control command is suspended for 5 seconds to wait for the system to stabilize.

[0159] The instruction generation unit adds a priority mechanism to ensure that instructions that reduce stress stimulation are executed before those that enhance it.

[0160] The fuzzy logic reasoning system of the stress state assessment module is expanded with expert rules.

[0161] For older users, 10 new specific rules have been added to take into account the impact of age-related physiological changes on stress response.

[0162] For example, if the average stress level is moderate, the stress response is stable, and the age is over 65, the psychological stress level should be adjusted to a lower level.

[0163] For users with anxiety disorders, 15 new specific rules have been added to take into account the impact of anxiety traits on baseline stress levels.

[0164] For example, if the average stress level is high, the stress recovery ability is rapid, and there is a history of anxiety, then the psychological stress level should be adjusted to moderate.

[0165] The grade assessment unit output report now includes a population-specific interpretation guide to help operators correctly understand the clinical significance of assessment results in specific populations.

[0166] The system hardware configuration has been modified to be age-appropriate.

[0167] The physiological signal acquisition electrodes are disposable electrodes pre-coated with conductive gel, simplifying the wearing process.

[0168] The respiratory plethysmography sensor uses a more flexible material to improve comfort.

[0169] The virtual reality headset weighs as little as 300g and comes with an adjustable headband and extra padding.

[0170] The system interface font size has been increased by 1.5 times, the icon size has been increased by 50%, and voice navigation is now available.

[0171] The system's security protection mechanisms have been strengthened.

[0172] For elderly users, an abnormal heart rate detection function has been added. When the heart rate is continuously detected to be greater than 140 beats / min or less than 40 beats / min, the monitoring will automatically stop and an alarm will be issued.

[0173] For users with anxiety disorders, an early panic attack detection function has been added. By analyzing the sudden increase in respiratory rate and the explosive rise in skin conductance, relief measures can be intervened in advance before the comprehensive stress index reaches the threshold. The system's emergency stop threshold has been lowered from 85 to 75 to increase the safety margin.

[0174] The system validation and calibration process has been expanded to include a population-specific validation step.

[0175] We recruited elderly volunteers and volunteers with anxiety disorders to conduct specialized tests and collected group-specific physiological data to establish a reference database.

[0176] The algorithm parameters were re-optimized based on population data to ensure the accuracy of the evaluation.

[0177] The system provides a group selection switch, allowing operators to select the appropriate algorithm mode based on user type for precise monitoring.

[0178] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0179] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A psychological stress monitoring system based on virtual reality, characterized in that, include: The physiological signal acquisition module is used to acquire multimodal physiological signals of the user in real time during the virtual reality experience. The multimodal physiological signals include electrocardiogram signals, skin conductance signals, respiratory waveforms, and electroencephalogram signals. The virtual reality environment generation module is used to construct and present a virtual reality scene containing multiple dynamic pressure source elements, which are dynamically adjusted according to preset rules or real-time feedback. The stress response quantification module is used to preprocess and extract features from the collected multimodal physiological signals, and calculate stress response indicators that reflect the degree of activation of the user's autonomic nervous system. The adaptive stress regulation module is used to dynamically adjust the presentation intensity, complexity, and evolution rate of stress source elements in the virtual reality environment generation module based on the real-time stress response indicators output by the stress response quantification module, so as to maintain the user's immersive experience within the preset target stress level range. The stress state assessment module is used to build a personalized stress response model for users based on time series data of stress response indicators and adaptive regulation history, and output the final psychological stress level assessment result.

2. The psychological stress monitoring system based on virtual reality according to claim 1, characterized in that, The physiological signal acquisition module includes an electrocardiogram acquisition unit, a skin conductance acquisition unit, a respiratory monitoring unit, and an electroencephalogram acquisition unit; The ECG acquisition unit acquires ECG signals through dry electrodes placed on the user's chest and calculates the ratio of high-frequency power to low-frequency power for beat-by-beat heart rate and heart rate variability. The skin conductance acquisition unit measures the skin conductance level and its fluctuation frequency through a silver-silver chloride electrode fixed to the distal phalanx of the user's index finger. The respiratory monitoring unit acquires respiratory waveforms and extracts respiratory rate and respiratory rhythm disorder by using a respiratory plethysmography sensor wrapped around the user's chest and abdomen. The EEG acquisition unit records EEG signals from the prefrontal and parietal lobe regions using an 8-lead EEG cap worn on the user's head, and analyzes the relative power spectral density of alpha and beta waves.

3. The psychological stress monitoring system based on virtual reality according to claim 1, characterized in that, The virtual reality environment generation module includes a scene database, a stress source element library, and a real-time rendering engine. The scenario database stores various basic virtual environment templates, covering social assessment scenarios, time pressure scenarios, multitasking scenarios, uncertainty threat scenarios, and physical hazard simulation scenarios; The stress source element library contains a variety of dynamic stress source elements that can be called independently or in combination. The dynamic stress source elements include the virtual character's gaze direction and density, the complexity and update frequency of task instructions, the intensity of abrupt changes in ambient lighting and sound, the probability of the appearance of obstacles on the path and the movement speed, and the remaining time prompt of the time countdown. The real-time rendering engine selects specific elements from the pressure source element library and performs real-time fusion and rendering according to the specified intensity parameters based on the control instructions issued by the adaptive pressure control module, generating a virtual reality scene stream with a continuous dynamic pressure gradient.

4. A virtual reality-based psychological stress monitoring system according to claim 1, characterized in that, The stress response quantification module includes a signal preprocessing submodule, a feature extraction submodule, and an index fusion submodule; The signal preprocessing submodule performs second-order Butterworth bandpass filtering on the raw physiological signal to remove power frequency interference and baseline drift, and performs amplitude normalization on the signals of each channel. The feature extraction submodule extracts the standard deviation of adjacent RR interval sequences from the preprocessed electrocardiogram signal, extracts the peak count of skin conductance response within a specific number from the skin conductance signal, extracts the coefficient of variation of inspiratory and expiratory durations from the respiratory signal, and extracts the ratio of the power of the theta band to the beta band in the prefrontal cortex from the electroencephalogram signal. The index fusion submodule uses a linear weighting method based on principal component analysis to fuse multiple extracted physiological features into a scalar value, called the comprehensive stress index. The higher the index value, the stronger the activation of the user's autonomic nervous system.

5. A virtual reality-based psychological stress monitoring system according to claim 1, characterized in that, The adaptive pressure control module includes a target pressure range setting unit, a control strategy decision unit, and an instruction generation unit. The target stress range setting unit allows the system operator to set a comprehensive stress index target range for the user according to the monitoring purpose. The lower limit of the range is the minimum stimulus intensity that can induce a measurable physiological response, while the upper limit is the maximum tolerable intensity that avoids excessive stress and discomfort. The control strategy decision unit continuously receives the real-time comprehensive stress index sent by the stress response quantification module and compares it with the target stress range; when the real-time index is less than the lower limit of the target range, the decision unit generates a control strategy to enhance stress stimulation. When the real-time index is higher than the upper limit of the target range, the decision-making unit generates a control strategy to reduce the stress stimulus. The control strategy is specifically manifested as an adjustment scheme for the intensity parameters and dynamic characteristics of one or more pressure source elements in the virtual reality environment. The instruction generation unit converts the control strategy into standardized control instructions that can be parsed by the virtual reality environment generation module. The instructions include the identifier of the pressure source element to be adjusted, the target intensity value, and the change transition duration parameter.

6. A virtual reality-based psychological stress monitoring system according to claim 5, characterized in that, The control strategy decision unit uses an algorithm based on the proportional-integral-derivative control principle to achieve precise pressure level maintenance; The algorithm sets the midpoint of the target pressure range as the set point and uses the real-time comprehensive stress index as the process variable to calculate the adjustment amount of the stress stimulus intensity. The proportional term is used to quickly respond to the current deviation between the comprehensive stress index and the set point, the integral term is used to eliminate the continuous steady-state error, and the derivative term is used to predict the changing trend of the comprehensive stress index and implement advance correction. By adjusting multiple gain parameters of the proportional-integral-derivative controller, the system can adapt to the varying stress sensitivities of different users, achieving smooth and precise stress experience control.

7. A virtual reality-based psychological stress monitoring system according to claim 3, characterized in that, Each stressor element in the stressor element library is assigned three-dimensional attribute parameters, including cognitive load dimension, emotional arousal dimension, and time pressure dimension. When the real-time rendering engine receives a control command, it calculates the target value required for each dimension according to the overall pressure intensity required by the command and the preset weight allocation algorithm. Then, it selects the pressure source elements whose attribute combinations are closest to the target value from the element library for instantiation and rendering.

8. A virtual reality-based psychological stress monitoring system according to claim 1, characterized in that, The pressure status assessment module includes a data caching unit, a model update unit, and a level assessment unit. The data caching unit stores the comprehensive stress index time series data and the corresponding adaptive control command history at a fixed period on a rolling basis. The model update unit uses a recursive least squares algorithm to update a second-order autoregressive model representing the user's personal stress response characteristics online based on cached data. The parameters of the model reflect the user's baseline response level, response gain, and recovery speed to virtual reality stress stimuli. The rating assessment unit uses an updated personal stress response model to conduct multi-dimensional analysis of the user's average stress level, stress response variability, and stress recovery ability during a complete monitoring session, and finally outputs a psychological stress assessment conclusion divided into multiple levels.

9. A virtual reality-based psychological stress monitoring system according to claim 1, characterized in that, The virtual reality-based psychological stress monitoring system constitutes a complete closed-loop monitoring system; The physiological signal acquisition module captures the user's multimodal physiological signals in real time during the virtual reality experience; The virtual reality environment generation module constructs and presents a virtual scene containing dynamic stress source elements; The stress response quantification module processes physiological signals and calculates indicators reflecting the degree of activation of the autonomic nervous system; The adaptive stress control module dynamically adjusts the stress stimulation in the virtual scene based on real-time stress indicators. The pressure status assessment module builds a personalized model for users based on historical data and outputs the final pressure level assessment.

10. A method for monitoring psychological stress based on virtual reality, characterized in that, Psychological stress monitoring is achieved using the virtual reality-based psychological stress monitoring system described in any one of claims 1 to 9.

Images