High-altitude falling simulation system based on VR equipment

Abstract

The invention discloses a high-altitude falling simulation system based on VR equipment, and relates to the technical field of virtual reality simulation, and the system comprises a data collection module which collects the pupil distance and initial sight line data of a user; the visual disturbance module is used for dynamically adjusting local visual disturbance of a corresponding area in the high-altitude virtual scene based on the gaze data to form a visual disturbance effect; the psychological index calculation module is used for calculating an anxiety index and a focus drift speed by adopting a time sequence convolutional neural network according to the visual disturbance effect and the eye movement behavior of the user to obtain an anxiety index and a focus drift speed index; and the disturbance adjustment module adaptively adjusts the frequency and amplitude of the visual disturbance effect by adopting a method of combining fuzzy logic control and adaptive gain adjustment to obtain disturbance rhythm parameters. According to the method, the personalized matching ability and training effectiveness in psychological behavior intervention and virtual high-altitude adaptation training are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual reality simulation, in particular to a high-altitude fall simulation system based on VR equipment. Background Art

[0002] With the continued development of virtual reality (VR) technology, its applications in psychological intervention, behavioral training, immersive experiences, and crisis response are becoming increasingly widespread. It demonstrates particularly strong scene-reconstruction capabilities in high-altitude environment simulation. Existing high-altitude simulations primarily rely on 3D modeling engines and image rendering algorithms. These systems utilize preset scenes, viewpoint control, and stereoscopic display devices to construct virtual high-altitude scenes, eliciting fear responses in users, which can then be used for behavioral assessment or psychological adaptation training. Furthermore, some systems incorporate motion capture or simple physiological signal monitoring to passively record individual responses. However, based on currently available technical literature and product proposals, the linkage mechanisms between immersion construction and interactive response in these systems remain limited. Particularly in complex scenarios, these systems lack the ability to provide real-time feedback and intervention on changes in the user's psychological state, making it difficult to support personalized adaptation.

[0003] The main shortcoming of existing high-altitude simulation technology is the lack of a linkage mechanism for dynamic adjustment based on the user's real-time psychological reaction characteristics. Specifically, most current virtual reality high-altitude simulation solutions only achieve static scene presentation or simple interactive triggering at the visual effect level, and are unable to combine the dynamic behavioral characteristics (such as eye movement trajectories, gaze patterns) and psychological reactions (such as anxiety levels and attention drift) generated by users during the simulation process for real-time recognition and response. This separation makes it impossible to personalize the simulation when facing user emotional states such as fear and alertness, and it is difficult to form an effective immersive behavioral assessment and adaptation training process. Therefore, how to build a virtual reality high-altitude simulation solution that can perceive the user's psychological and behavioral characteristics in real time and dynamically adjust visual stimulation accordingly has become one of the key bottlenecks in the current development of the technology. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a high-altitude fall simulation system based on VR equipment to solve the problem that it is impossible to achieve adaptive adjustment of visual stimulation based on the user's psychological behavior state.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: The invention provides a high-altitude fall simulation system based on a VR device, which comprises: a data collection module for collecting user pupil distance and initial sight data; a gaze capture module for driving a three-dimensional modeling engine to dynamically construct a high-altitude virtual scene according to the initial sight data, and capturing the user's gaze point position and gaze duration in real time to generate gaze data; a visual perturbation module for dynamically adjusting local visual perturbations of corresponding areas in the high-altitude virtual scene based on the gaze data to form a visual perturbation effect; a psychological index calculation module for calculating an anxiety index and a focus drift speed according to the visual perturbation effect and the user's eye movement behavior using a temporal convolutional neural network to obtain anxiety index and focus drift speed indicators; a perturbation adjustment module for adaptively adjusting the frequency and amplitude of the visual perturbation effect using a method combining fuzzy logic control with adaptive gain adjustment to obtain perturbation rhythm parameters; and a dynamic simulation module for triggering visual perspective transformation and viewing angle rotation using the perturbation rhythm parameters to simulate the visual effect of a high-altitude fall, and for labeling behavioral characteristics in real time in combination with the gaze data, the anxiety index and the focus drift speed indicator to form an immersive experience report.

[0007] As a preferred solution of the high-altitude fall simulation system based on VR equipment of the present invention, the specific steps of collecting the user's pupil distance and initial line of sight data are as follows: Collect images of the user's eyes, use binocular geometry analysis to extract the center points of the left and right pupils, and calculate the distance between the left and right pupils to obtain the user's pupil distance; The coordinates of the gaze starting point, binocular disparity information, head posture, gaze stability parameters and timestamp are collected to form the initial gaze data.

[0008] As a preferred solution of the high-altitude fall simulation system based on VR equipment of the present invention, the following specific steps are taken to generate gaze data: The user's pupil distance and the initial line of sight data are spatially transformed and superimposed, and fused using a rotation matrix to obtain the line of sight direction vector. Ray intersection is then performed in virtual space to obtain the spatial coordinate information of the area the user is facing. According to the spatial coordinate information of the user's facing area, the 3D modeling engine is driven to call high-altitude scene resources to generate a high-altitude virtual scene; The spatial coordinate information and duration of the user's facing area in the high-altitude virtual scene are recorded in real time and integrated into gaze data in chronological order.

[0009] As a preferred solution of the high-altitude fall simulation system based on VR equipment of the present invention, wherein: according to the spatial coordinate information of the user-facing area, the three-dimensional modeling engine is driven to call the high-altitude scene resources to generate the high-altitude virtual scene. The specific steps are as follows: Using the octree spatial index and multi-level scene management method, we retrieve and filter the environmental elements and detail data of the corresponding area in the high-altitude scene resource library to obtain the scene element dataset. Classify and sort the scene element data set, call the graphics rendering interface to load and build the three-dimensional environment of the high-altitude virtual scene, and generate real-time rendering frame data; The loading and rendering of high-altitude virtual scene details are dynamically adjusted based on real-time rendering frame data to generate a high-altitude virtual scene.

[0010] As a preferred solution of the high-altitude fall simulation system based on VR equipment described in the present invention, the gaze point heat map analysis method is used to calculate the spatial density of the gaze data, obtain the coordinate set of the gaze hotspot area, and dynamically adjust the visual elements of the corresponding area in the high-altitude virtual scene to form a visual disturbance effect.

[0011] As a preferred solution of the high-altitude fall simulation system based on VR equipment of the present invention, the anxiety index and focus drift speed index are obtained by the following specific steps: Collect user eye movement behavior data such as gaze trajectory, gaze point location, pupil diameter change, and gaze duration, and form a multi-dimensional time series data sequence in chronological order; Input the multidimensional time series data sequence into the time series convolutional neural network to extract the multidimensional time series feature representation reflecting the response of the visual perturbation effect; The multi-dimensional temporal feature representation is correlated with the visual perturbation effect, and time alignment and feature matching are used to perform synchronous matching on the time axis, extracting the change correlation pattern and obtaining the predicted values ​​of anxiety index and focus drift speed. The predicted values ​​were normalized using the z-score standardization method and dynamically adjusted in combination with the multidimensional time series feature representation to obtain the anxiety index and focus drift speed index.

[0012] As a preferred solution of the high-altitude fall simulation system based on VR equipment described in the present invention, the multidimensional time series data sequence is input into the time series convolutional neural network to extract the multidimensional time series feature representation reflecting the response of the visual perturbation effect. The specific steps are as follows: The multi-dimensional time series data sequence is input into a multi-layer one-dimensional convolutional neural network, local dynamic features are extracted through a sliding time window, and a preliminary feature tensor is output; Using multi-scale fusion and context modeling methods, combined with residual connections and attention mechanisms, the initial feature tensors are weighted integrated and time-series optimized to generate a time-series feature sequence of behavior change trends; Through the fully connected layer, the time series feature sequence is dimensionally compressed and feature integrated to obtain a multi-dimensional time series feature representation.

[0013] As a preferred solution of the high-altitude fall simulation system based on VR equipment of the present invention, the specific steps of obtaining the disturbance rhythm parameters are as follows: The fuzzy logic control method is used to fuzzify the anxiety index and focus drift speed index, and then input them into the preset fuzzy rules for fuzzy aggregation and defuzzification to derive the fuzzy control output value. According to the fuzzy control output value, the adaptive gain adjustment method is used to adjust the frequency and amplitude of the visual disturbance effect to obtain the disturbance rhythm parameters.

[0014] As a preferred solution of the high-altitude fall simulation system based on VR equipment of the present invention, wherein: the use of perturbation rhythm parameters to trigger visual perspective transformation and viewing angle rotation to simulate the high-altitude fall visual effect is specifically performed as follows: According to the disturbance rhythm parameters, the visual perspective transformation matrix is ​​calculated using the matrix transformation method, and the angular velocity calculation method is used to obtain the angular velocity of the viewing angle rotation; Update the projection parameters of the high-altitude virtual scene based on the visual perspective transformation matrix, dynamically adjust the position and orientation of the virtual camera based on the angular velocity of the view rotation, and generate the high-altitude virtual scene transformation frame and view rotation trajectory; By adopting the time synchronization and visual sequence fusion method, the transformed picture frames of the high-altitude virtual scene and the perspective rotation trajectory are dynamically integrated to form a visual effect of simulating a high-altitude fall.

[0015] As a preferred solution of the high-altitude fall simulation system based on VR equipment of the present invention, the steps of forming the immersive experience report are as follows: Based on gaze data, anxiety index, and focus drift speed indicators, multi-dimensional statistical analysis and time series analysis are performed to identify abnormal gaze patterns and anxiety changes, and output behavioral feature labels and behavioral change trend data; Using the support vector machine algorithm, we classify and identify behavioral change trend data, analyze psychological states and behavioral types, and form an immersive experience report.

[0016] The beneficial effects of the present invention are as follows: by integrating fuzzy logic and adaptive gain control, intelligent dynamic adjustment of visual disturbance intensity and frequency is achieved, and an immersive adaptation mechanism with closed-loop psychological state adjustment capabilities is constructed. The system can perceive changes in the user's anxiety state and gaze behavior in real time, dynamically optimize the intensity of visual intervention while processing the ambiguity of psychological reactions, and thus accurately control the stimulation level in the high-altitude fall simulation. This effectively avoids stress reactions caused by excessive stimulation or training failures caused by insufficient stimulation, significantly improving the personalized matching ability and training effectiveness in psychological behavioral intervention and virtual high-altitude adaptation training. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 Schematic diagram of the high-altitude fall simulation system based on VR equipment in the present invention.

[0019] Figure 2 This is a flow chart of collecting pupil distance and line of sight data in the present invention.

[0020] Figure 3 This is a flow chart for generating the high-altitude falling visual effect in the present invention.

[0021] Figure 4 This is a flow chart for the dynamic construction of the high-altitude virtual scene in the present invention. DETAILED DESCRIPTION

[0022] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0023] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0025] Reference Figures 1 to 4 This embodiment provides a high-altitude fall simulation system based on a VR device, comprising the following steps: The data collection module collects the user's pupil distance and initial line of sight data.

[0026] The user's binocular images are collected, the positions of the left and right pupil centers are extracted using binocular geometry analysis methods, and the distance between the left and right pupils is calculated to obtain the user's pupil distance.

[0027] Specifically, after collecting images of the user's two eyes, the left eye image and the right eye image are grayscaled respectively to reduce interference caused by changes in lighting; the eye contour is extracted in each image using the edge detection method, and then the left eye pupil boundary and the right eye pupil boundary are located based on the Hough circle transform or ellipse fitting method, and the coordinates of the left eye pupil center and the right eye pupil center are calculated respectively; through the binocular geometry analysis method, under the premise of knowing the distance between the binocular cameras, combined with the positions of the left and right eye pupil centers in the image coordinate system, the three-dimensional coordinates of the two pupil centers in space are calculated by triangulation; based on the Euclidean distance between the two spatial coordinate points, the user's pupil distance is obtained.

[0028] The coordinates of the gaze starting point, binocular disparity information, head posture, gaze stability parameters and timestamp are collected to form the initial gaze data.

[0029] Specifically, a binocular camera is used to capture a sequence of synchronized binocular images. Image processing methods are used to extract the coordinates of the left and right pupil centers. Combined with binocular geometry analysis methods, the user's gaze focus in three-dimensional space is calculated as the gaze starting point coordinates. Based on the known pupil center pixel coordinates in the binocular images, the baseline length and focal length parameters of the binocular camera are combined to calculate the horizontal pixel difference between the left and right eye gaze focuses on the image plane to obtain binocular disparity information. A facial feature point detection method is used to extract a set of three-dimensional facial feature points. A pose estimation algorithm (such as the PNP method) is used to calculate the rotation vector and displacement vector of the head in space to obtain head pose data. A sequence of gaze point coordinates is continuously acquired within a short time window, and the spatial fluctuation range of the gaze point within the time window is calculated to obtain the gaze stability parameter. Upon completion of each gaze data acquisition, a high-precision system clock time is synchronously recorded as the timestamp for the corresponding data. It should also be noted that all acquisition steps are completed within the same time frame to ensure that the gaze starting point coordinates, binocular disparity information, head pose, gaze stability parameter, and timestamp correspond to consistent gaze behavior states, thereby forming the initial gaze data.

[0030] The gaze capture module drives the 3D modeling engine to dynamically build a high-altitude virtual scene based on the initial line of sight data, and captures the user's gaze point position and gaze duration in real time to generate gaze data.

[0031] The user's pupil distance and the initial line of sight data are spatially transformed and superimposed, and fused using a rotation matrix to obtain the line of sight direction vector. Ray intersection is then extended in the virtual space to obtain the spatial coordinate information of the area the user is facing.

[0032] Specifically, the user's left and right pupil center positions are collected. The user's gaze center position is calculated based on the geometric relationship between the left and right pupils and used as the basic input for the initial gaze data. Combined with the posture angle information output by the user's head posture sensor, the initial gaze data is converted to a unified coordinate system in three-dimensional space. A rotation matrix is ​​constructed to represent the head posture change, and the rotation matrix is ​​applied to the initial gaze data to complete the unified adjustment of spatial coordinates. The initial gaze data, which has undergone rotational fusion processing, is expanded into a direction vector. Starting from the user's current head center position, a gaze extension ray is generated in virtual space in the direction of the starting point. Based on a preset 3D scene model, the ray is intersected with the scene surface, and the coordinates of the nearest intersection point are extracted. This coordinate is then identified as the spatial region facing the user. The intersection coordinates are used as the spatial coordinate information for the region currently facing the user, providing basic spatial reference data for subsequent visual feature analysis and dynamic content adjustment.

[0033] It should also be explained that the preset three-dimensional scene model collects the geometric shape, spatial position and boundary information of each static and dynamic object in the virtual environment, and uniformly converts it into a three-dimensional coordinate description form, such as obtaining the size parameters and relative positions of objects such as desktops and panels; uses the boundary volume hierarchy structure or octree method to construct a spatial index for the three-dimensional grid data to improve the efficiency of ray intersection; normalizes and aligns the three-dimensional descriptions of all objects to integrate a complete three-dimensional scene data set; finally, loads the complete three-dimensional scene data set as a reference space for use in line of sight extension and spatial interaction calculations.

[0034] According to the spatial coordinate information of the user's facing area, the 3D modeling engine is driven to call high-altitude scene resources to generate a high-altitude virtual scene.

[0035] Furthermore, the octree spatial index and multi-level scene management methods are used to retrieve and filter the environmental elements and detail data of the corresponding areas in the high-altitude scene resource library to obtain a scene element dataset.

[0036] Specifically, according to the spatial coordinate information of the user-facing area, the target index area containing the spatial coordinates is located in the high-altitude scene resource library, the spatial coordinates are used as the query key value, and the octree spatial index is called to perform position matching in the three-dimensional space; in the octree spatial index, the voxel node to which the spatial coordinates belong is determined level by level, and the root node is traversed downward to the leaf node to lock the smallest voxel unit containing the spatial coordinates; the scene element index information in the high-altitude scene resource library associated with the leaf node is read, and the multi-level scene management method processing flow is entered; in the multi-level scene management method, the scene elements are classified according to the scene resource type (such as architecture, weather, terrain, etc.) and the level identifier, and irrelevant scene elements are filtered according to the level of the current user-facing area; the resource path is called within the level, and the environmental elements and detail data that coincide with or are adjacent to the spatial coordinates in the high-altitude scene resource library are read accordingly, where the detail data includes texture information, auxiliary structure models, dynamic simulation parameters, and environmental interaction values ​​for enhancing the accuracy and realism of the scene. A set of local scene elements is formed; resource elements with high correlation with the user's gaze direction in the local scene element set are further screened based on indicators such as spatial distance and perspective correlation, and non-visible elements not involved in the current frame rendering are eliminated; finally, the environmental elements and detail data retained after the above processing are integrated to form a scene element dataset.

[0037] It should also be explained that the specific steps for constructing a high-altitude scene resource library are: extracting representative bird's-eye view scene images from existing high-resolution remote sensing images and oblique photogrammetry results, and organizing them into a high-altitude scene resource library for virtual interaction through three-dimensional reconstruction and semantic annotation.

[0038] In this multi-level scene management approach, resource metadata is extracted based on the spatial attributes and semantic types of scene resources, and classified into categories such as architectural, terrain, and meteorological. Then, based on resource category and spatial scale, a level identifier is assigned within a pre-set hierarchical coding system—for example, architectural resources are assigned L2 and meteorological resources L3. This level identifier allows for hierarchical management of all scene resources. Once the user has determined the region, resources irrelevant to the current level are filtered based on their level, thus achieving hierarchical classification and efficient access to scenes.

[0039] Classify and sort the scene element data set, call the graphics rendering interface to load and build the three-dimensional environment of the high-altitude virtual scene, and generate real-time rendering frame data.

[0040] Specifically, the scene element dataset is classified based on the element type information in the scene element dataset, and the architectural elements, meteorological elements, terrain elements, texture elements and dynamic special effect elements in the scene element dataset are respectively assigned to the corresponding resource queues; they are sorted in ascending order according to the relative distance between the spatial position and the spatial coordinate information of the user-facing area; the graphics rendering interface is called for the resource queue after ascending sorting, and the 3D model files, texture map files and animation control parameters corresponding to each type of resource are loaded in turn; in the resource loading process, the management function of the graphics rendering interface is first called to load the triangular mesh models of the architectural elements and terrain elements, and then The texture mapping function is used to match the corresponding texture elements with the loaded triangular mesh model and bind the texture map; the lighting calculation function of the graphics rendering interface is used to simulate the light and shadow of the three-dimensional mesh model according to the scene time and the direction of the virtual light source, and then the virtual camera parameters are set according to the spatial coordinate information of the user-facing area, including the observation position, orientation, near clipping plane and far clipping plane, and perspective projection transformation is performed; the processed elements are rendered into three-dimensional image frames, and dynamic special effects elements are dynamically loaded in a frame-by-frame loop and inserted into the current frame image; finally, real-time rendering frame data including buildings, terrain, weather and special effects content is output.

[0041] It should also be noted that when sorting in ascending order based on the relative distance between the spatial position and the spatial coordinate information of the area the user is facing, the relative distance refers to the three-dimensional spatial distance between the position coordinates of each environmental element in the high-altitude scene resource library and the spatial coordinates corresponding to the area the user is facing. The three-dimensional spatial distance can be calculated based on the difference in three-dimensional coordinates and is used to reflect the proximity of the scene element to the user's gaze area. By sorting the three-dimensional spatial distance in ascending order, the resource elements in the local scene are prioritized from near to far, providing a reference basis for the subsequent scene loading order, rendering processing order, or interactive response order.

[0042] The loading and rendering of high-altitude virtual scene details are dynamically adjusted based on real-time rendering frame data to generate a high-altitude virtual scene.

[0043] Specifically, the image resolution, frame rate and rendering time in the real-time rendering frame data are collected as performance indicators, which are provided by the graphics rendering interface and the rendering timestamp; the collected performance indicators are compared with the preset performance threshold to determine whether they need to be adjusted; the spatial distance of the environmental elements is calculated based on the spatial coordinates of the user-facing area and the position of the virtual camera, and the octree spatial index is used to quickly locate the spatial unit where the element is located, and evaluate the spatial priority; based on the performance indicators and spatial priority, the multi-level texture map and different triangular mesh model files are dynamically switched to adjust the texture resolution and mesh complexity; the graphics rendering interface is called to reload the texture and mesh data, and the light and shadow effects are recalculated when the scene time or lighting changes; the observation parameters of the virtual camera are updated in real time according to the user's head posture and viewing angle parameters; finally, the adjusted scene elements are rendered to the frame buffer through the graphics rendering interface, and the current high-altitude virtual scene is output.

[0044] The specific steps for setting the preset performance threshold should also be explained: collecting the target device's frame rate and rendering time data under typical rendering tasks; evaluating processor performance parameters (such as main frequency and number of cores) and graphics processing capability parameters (such as GPU computing power and video memory bandwidth); combining the subjective acceptance range of screen smoothness and response delay in the preset performance threshold to determine the user's acceptable performance lower limit; setting the frame rate threshold to no less than the example value of 30 frames per second and the rendering time threshold to no more than the example value of 33 milliseconds based on the statistical analysis results; The specific steps of comparing the collected performance indicators with the preset performance threshold are as follows: comparing the frame rate in the collected real-time rendering frame data with the preset frame rate threshold. If the frame rate is lower than the preset frame rate threshold, it is determined that the performance does not meet the requirements; at the same time, comparing the collected single-frame rendering time with the preset rendering time threshold. If the single-frame rendering time exceeds the preset rendering time threshold, it is determined that the performance does not meet the requirements; when the frame rate is not lower than the preset frame rate threshold and the single-frame rendering time does not exceed the preset rendering time threshold, it is determined that the performance meets the requirements.

[0045] The spatial coordinate information and duration of the user's facing area in the high-altitude virtual scene are recorded in real time and integrated into gaze data in chronological order.

[0046] Specifically, the spatial coordinate information of the area the user is facing is obtained in real time, and the timestamp of each spatial coordinate change is recorded; the duration between adjacent timestamps is calculated to determine the length of time the user gazes at the same spatial coordinate; the time periods of continuous gaze at the same spatial coordinate are merged to generate a gaze duration record; all gaze data are sorted in chronological order to form a complete gaze data sequence; for example, if the user gazes at a spatial coordinate continuously for more than 0.5 seconds, it is regarded as a valid gaze event, and the multiple sampled gaze data within the time period are merged, and finally the gaze data containing the spatial coordinates and the corresponding gaze duration is output.

[0047] It should also be explained that the specific steps for obtaining the spatial coordinate information of the area the user is facing in real time are as follows: continuously collect the user's head posture data, including position and orientation information, through the head tracking device; calculate the corresponding spatial coordinates according to the head orientation in combination with the virtual camera parameters; record the calculated spatial coordinates and corresponding timestamp once every fixed time interval (for example, 10 milliseconds in the example); transmit the collected data stream to the processing unit in real time to ensure that the spatial coordinate information of the area the user is currently facing is obtained continuously and without loss.

[0048] The visual perturbation module dynamically adjusts the local visual perturbation of the corresponding area in the virtual scene based on the gaze data to form a visual perturbation effect.

[0049] The gaze point heat map analysis method is used to calculate the spatial density of the gaze data to obtain the coordinate set of the gaze hotspot area, and the visual elements of the corresponding area in the virtual scene are dynamically adjusted to form a visual disturbance effect.

[0050] Specifically, the spatial coordinates of each gaze point in the gaze data are mapped to pre-divided spatial grid units, and the number of gaze points in each grid unit is counted to calculate the spatial density; a preset gaze density threshold is used to screen out grid units with higher gaze density, and the center coordinates are extracted as the gaze hotspot area coordinate set; based on the gaze hotspot area coordinate set, the visual elements in the corresponding virtual scene are determined; according to the set visual perturbation rules, the visual elements are dynamically adjusted, including changing attributes such as color, brightness or texture details, thereby forming a visual perturbation effect.

[0051] It should also be explained that the specific steps for presetting the gaze density threshold are: counting the number of all gaze points in the spatial grid unit and calculating the gaze density of each grid unit; collecting the gaze density data of all grid units and sorting them by numerical value; setting the gaze density threshold based on the statistical analysis results, for example, selecting the upper ten percent of the gaze density distribution as the gaze density threshold; screening out the grid units with a gaze density higher than the gaze density threshold, and extracting the center coordinates as the gaze hotspot area.

[0052] The specific steps for setting visual perturbation rules are as follows: for the visual elements corresponding to the gaze hotspot area, obtain the current color, brightness and texture detail parameters; define the perturbation amplitude, such as the percentage of increase or decrease in color brightness, and the adjustment level of texture detail; dynamically change the color value, brightness value or replace it with a texture map of different detail levels according to the perturbation amplitude; apply the adjusted visual attributes to the corresponding visual elements to complete the dynamic update of the visual perturbation effect.

[0053] The psychological index calculation module uses a temporal convolutional neural network to calculate the anxiety index and focus drift speed based on the visual disturbance effect and user eye movement behavior, and obtains the anxiety index and focus drift speed indicators.

[0054] Collect user eye movement behavior data such as gaze trajectory, gaze point position, pupil diameter change and gaze duration, and form a multi-dimensional time series data sequence in chronological order.

[0055] Specifically, a high-frequency sampling device is used to collect user gaze trajectory data at a frequency of 60 times per second, recording the three-dimensional spatial coordinates of each sampling time point. A gaze detection algorithm is used to identify valid gaze events, and the gaze point location and gaze duration are collected. Only valid gaze events with a gaze duration greater than or equal to 200 milliseconds are retained. An optical pupil tracking device is used to record the pupil diameter at each sampling time point with an example measurement accuracy of 0.01 mm. The change in pupil diameter between adjacent time points is calculated and the pupil diameter change value is extracted. The timestamp data of gaze trajectory, gaze point location, pupil diameter change, and gaze duration are stored in their respective time series lists. All timestamps are extracted, merged, and sorted in ascending order. For each ascending time stamp, the corresponding multidimensional time series data is searched from each time series. If missing, nearest neighbor interpolation or previous value filling is used. The multidimensional time series data items are combined in timestamp order to form a synchronized multidimensional time series data sequence. The timestamp error is controlled within 1 millisecond, as an example.

[0056] The multidimensional time series data sequence is input into the time series convolutional neural network to extract the multidimensional time series feature representation reflecting the response to the visual perturbation effect.

[0057] Furthermore, the multidimensional time series data sequence is input into a multi-layer one-dimensional convolutional neural network, local dynamic features are extracted through a sliding time window, and a preliminary feature tensor is output.

[0058] Specifically, the multidimensional time series data sequence is divided into sliding time windows according to the set length and step size. In the example, the time window length is set to 300 milliseconds and the step size is 100 milliseconds. Multiple overlapping windows can be divided within the unit time; for the multidimensional time series data in each divided sliding time window, the first convolution layer of the multi-layer one-dimensional convolutional neural network is input, and the settings of the convolution kernel size of 30, the step size of 1, the number of output channels of 32, and the padding method of the example are used to perform a one-dimensional convolution operation to extract local time series features; the convolution result of each window is passed to the next convolution layer in turn to continue to extract higher-dimensional time series features. In the example, the convolution kernel size of the second convolution layer is 15 and the number of channels is 64, and the convolution kernel size of the third convolution layer is 7 and the number of channels is 128; the convolution output results of all time periods are spliced ​​in chronological order to obtain a preliminary feature result of a three-dimensional structure, where the dimensions correspond to the number of time periods, the length of data in each time period, and the number of time series features of each data segment.

[0059] The multi-scale fusion and context modeling methods are adopted, combined with residual connection and attention mechanism, to perform weighted integration and time series optimization processing on the preliminary feature tensor to generate a time series feature sequence of behavior change trends.

[0060] Specifically, the preliminary feature tensor is input into three one-dimensional convolution paths with different convolution kernel sizes, and the convolution kernel sizes are set to 3, 5 and 7 respectively, and the number of output channels of each convolution path is set to 64; the output features of the three convolution paths are spliced ​​in the channel dimension to form a fused feature tensor; the fused feature tensor is input into the bidirectional long short-term memory network for contextual temporal modeling, the hidden state dimension is set to 128, and the contextual feature sequence is output; the channel attention mechanism is applied to the contextual feature sequence, the channel statistics are extracted by global average pooling, the two-layer fully connected network is connected to calculate the channel weight, and each channel of the contextual feature sequence is weighted by multiplying the corresponding weight coefficient; the weighted contextual feature sequence is added element-by-element to the preliminary feature tensor in the same dimension, and the residual connection operation is completed to obtain the fused temporal feature tensor; the fused temporal feature tensor is input into the one-dimensional convolution layer for channel compression, the convolution kernel size is 3, and the number of output channels is 64, and finally a temporal feature sequence of behavioral change trends is generated.

[0061] It should also be explained that the training and modeling process of inputting the fused feature tensor into the bidirectional long short-term memory network for contextual temporal modeling is as follows: the fused feature tensors are arranged in chronological order to form an input sequence, which is input into the forward and backward hidden layers of the bidirectional long short-term memory network. The forward hidden layer processes the sequence information in forward chronological order, and the backward hidden layer processes the sequence information in reverse chronological order. During the training phase, training samples with labeled behavioral states or psychological state categories are used. By minimizing the cross-entropy loss function between the predicted category and the true label, the weight parameters and bias parameters in the bidirectional long short-term memory network are iteratively updated in combination with the backpropagation algorithm and the time backpropagation algorithm. After the training is completed, the trained bidirectional long short-term memory network is used to infer the newly input fused feature tensor, calculate the forward and backward hidden states respectively, and then perform splicing or weighted fusion on the bidirectional hidden states to extract the state representation containing the complete temporal context dependency for subsequent behavioral recognition or psychological state analysis.

[0062] Through the fully connected layer, the time series feature sequence is dimensionally compressed and feature integrated to obtain a multi-dimensional time series feature representation.

[0063] Specifically, each context feature vector in the context feature sequence is taken as the input of a time step and is sent to the fully connected layer in sequence for linear transformation. The dimension of the weight matrix of the linear transformation is 256×64, and the dimension of the bias vector is 64 dimensions. Each context feature vector is multiplied by the corresponding weight matrix and the bias is added to obtain a compressed time step feature vector. The compressed feature vectors of all time steps are arranged in chronological order to form a compressed time series feature sequence. The compressed time series feature sequence is then pooled along the time dimension. The pooling method is average pooling or maximum pooling. The pooling window is, for example, the entire sequence length, and the global feature vector is extracted. The global feature vector is spliced ​​with the compressed time series feature sequence to form a multi-dimensional time series feature representation.

[0064] The multi-dimensional temporal feature representation is correlated with the visual disturbance effect and analyzed. Time alignment and feature matching are used to perform synchronous matching on the time axis, and the change correlation pattern is extracted to obtain the predicted values ​​of anxiety index and focus drift speed.

[0065] Specifically, the multidimensional temporal feature representation and the visual disturbance effect data with timestamps are synchronously aligned according to the timestamps. For the parts with mismatched timestamps, linear interpolation or previous value filling method is used to fill in the data. At each time point, the rate of change of the spatial coordinates of the gaze point in the multidimensional temporal feature representation is calculated as the focus drift speed, and the weighted average of the change amplitude of the pupil diameter and the change amplitude of the gaze duration is calculated as the original indicator of the anxiety index. The sliding time window method is used to smooth the focus drift speed and the original indicator of the anxiety index. The window size example is 5 time points, and the step size example is 1 time point. According to the aligned time points, the correlation coefficient between the multidimensional temporal feature representation and the change in the intensity of the visual disturbance effect is calculated in each time window, and the time period with significant correlation is screened out. The focus drift speed and the original indicator of the anxiety index in the time period with significant correlation are respectively averaged and output as the predicted value of the focus drift speed and the predicted value of the anxiety index for the corresponding time period.

[0066] The predicted values ​​were normalized using the z-score standardization method and dynamically adjusted in combination with the multidimensional time series feature representation to obtain the anxiety index and focus drift speed index.

[0067] Specifically, the mean and standard deviation of the anxiety index prediction value sequence and the focus drift speed prediction value sequence are calculated respectively, and the z-score normalization method is used to obtain the normalized standardized prediction value sequence; the normalized standardized prediction value sequence is combined with the multidimensional time series feature representation of the corresponding time point, and the weighted average or linear mapping method is used to dynamically adjust the standardized prediction value sequence. The adjustment results are used to correct the original anxiety index indicator sequence and the focus drift speed indicator sequence to obtain the dynamically corrected anxiety index indicator sequence and the focus drift speed indicator sequence; the dynamically corrected indicator sequence is output in chronological order as the final anxiety index and focus drift speed indicator.

[0068] The disturbance adjustment module adopts a method combining fuzzy logic control and adaptive gain adjustment to adaptively adjust the frequency and amplitude of visual disturbances and obtain disturbance rhythm parameters.

[0069] The fuzzy logic control method is used to fuzzify the anxiety index and focus drift speed indicators, and input them into the preset fuzzy rules for fuzzy aggregation and defuzzification to derive the fuzzy control output value.

[0070] Specifically, a fuzzy logic control method is adopted to perform fuzzy processing on the anxiety index and focus drift speed indicators respectively. The fuzzy processing includes mapping the anxiety index and focus drift speed to a set of fuzzy language variables according to a preset membership function. For example, the anxiety index is divided into three fuzzy language values ​​of "low", "medium" and "high", and the focus drift speed is divided into three fuzzy language values ​​of "slow", "moderate" and "fast"; the fuzzification result is substituted into a preset fuzzy rule set as an input variable, and the output trend corresponding to different input variable combinations is described in the form of "if-then"; the activated fuzzy rules are fuzzy aggregated to integrate the output trend, and then defuzzified by the weighted average method to obtain a fuzzy control output value used to characterize the state response.

[0071] The steps for constructing the preset membership functions and fuzzy rule sets should also be explained: The membership functions for the anxiety index and focus drift speed indicator are constructed. For the anxiety index, the input range is set to [0, 1], and three membership functions are defined, corresponding to the fuzzy linguistic variables "low," "medium," and "high," respectively. The membership functions can be described using triangular or trapezoidal functions. For example, "low" can correspond to a triangular function with a center value of 0.2 and a range of ±0.1; "medium" corresponds to a triangular function with a center value of 0.5 and a range of ±0.15; and "high" corresponds to a triangular function with a center value of 0.8 and a range of ±0.1. For the focus drift speed indicator, the input range is set to [0, 10] (unit: degrees / second), and three membership functions are defined, corresponding to the fuzzy linguistic variables "slow", "moderate", and "fast", respectively. For example, the center value of "slow" is 2, and the expansion range is ±1; the center value of "moderate" is 5, and the expansion range is ±1.5; the center value of "fast" is 8, and the expansion range is ±1.

[0072] A preset set of fuzzy rules is constructed. Each fuzzy rule uses an "if-then" statement to represent the mapping relationship between a combination of input variables and an output value. For example, if the anxiety index is "low" and the focus drift speed is "slow," the output is "stable"; if the anxiety index is "medium" and the focus drift speed is "moderate," the output is "vigilant"; if the anxiety index is "high" and the focus drift speed is "fast," the output is "abnormal"; if the anxiety index is "medium" and the focus drift speed is "fast," the output is "vigilant," and so on. By enumerating all possible combinations of fuzzy linguistic variables, the fuzzy rule set is established for subsequent fuzzy reasoning and control output calculation.

[0073] According to the fuzzy control output value, the adaptive gain adjustment method is used to adjust the frequency and amplitude of the visual disturbance effect to obtain the disturbance rhythm parameters.

[0074] Specifically, according to the fuzzy control output value, the frequency and amplitude threshold is preset, and the adaptive gain coefficient is calculated. The adaptive gain coefficient is adjusted according to the deviation between the fuzzy control output value and the preset frequency and amplitude threshold. A discrete update method with a time step of 1 second is used to iteratively update the frequency and amplitude parameters until the frequency and amplitude parameters are stable within the preset frequency and amplitude threshold range. The frequency and amplitude of the adjusted visual disturbance effect are output as the disturbance rhythm parameters.

[0075] The steps for presetting the frequency and amplitude thresholds should also be explained: collect a large amount of visual perturbation experimental data, including users' subjective comfort evaluations at different frequencies and amplitudes; perform statistical analysis on the experimental data to calculate the distribution of user comfort scores corresponding to the frequency and amplitude combinations; determine the frequency and amplitude ranges in which the comfort scores meet the preset standards (for example, 80% of users have no discomfort reaction); finally, set the minimum and maximum values ​​in the range as the upper and lower limits of the frequency and amplitude thresholds, respectively. For example, the frequency threshold is 0.5 Hz to 5 Hz, and the amplitude threshold is 0 to 1.

[0076] The dynamic simulation module uses perturbation rhythm parameters to trigger visual perspective transformation and view rotation to simulate the visual effect of falling from a high altitude. It combines gaze data, anxiety index, and focus drift speed indicators to annotate behavioral characteristics in real time and form an immersive experience report.

[0077] According to the disturbance rhythm parameters, the matrix transformation method is used to calculate the visual perspective transformation matrix, and the angular velocity calculation method is used to obtain the angular velocity of the viewing angle rotation.

[0078] Specifically, a two-dimensional or three-dimensional visual perspective transformation matrix is ​​constructed based on the perturbation rhythm parameters. The specific steps include: calculating the corresponding translation, rotation and scaling transformation matrix elements based on the frequency and amplitude information in the perturbation rhythm parameters; for example, using the amplitude parameter to determine the translation offset in the transformation matrix, and using the frequency parameter to determine the change amplitude of the rotation angle; and combining each basic transformation matrix in sequence through matrix multiplication to obtain a complete visual perspective transformation matrix; The angular velocity calculation method is adopted, and the specific steps include: collecting the viewing angle data at consecutive moments, calculating the angular difference of the viewing angles at adjacent moments, and the angular unit is radians or degrees; calculating the angular velocity according to the time interval to obtain the viewing angle rotation speed. For example, if the viewing angle difference is 5 degrees and the time difference is 0.1 seconds, the angular velocity is 50 degrees per second; calculating the corresponding angular velocity for the viewing angle rotation in multiple directions respectively; and using the visual perspective transformation matrix and angular velocity as output to describe the spatial transformation and rotation dynamics under visual disturbance.

[0079] The projection parameters of the high-altitude virtual scene are updated based on the visual perspective transformation matrix, and the position and orientation of the virtual camera are dynamically adjusted in combination with the angular velocity of the view rotation to generate the high-altitude virtual scene transformation frame and view rotation trajectory.

[0080] Specifically, the near plane distance and the far plane distance of the viewing cone are calculated by extracting specific elements in the visual perspective transformation matrix, and the horizontal viewing angle and aspect ratio are inferred by using the relevant values ​​in the matrix to obtain the corresponding projection matrix; the current viewing angle rotation angular velocity of the virtual camera is collected and expressed as an angular velocity vector around a fixed coordinate axis; according to the viewing angle rotation angular velocity, the rotation increment of the virtual camera at the current time step is calculated using the Euler integral method or quaternion interpolation, and the rotation increment of the current time step is applied to the orientation matrix of the virtual camera at the previous moment to obtain an updated orientation matrix; the position change of the virtual camera in the three-dimensional coordinate system is synchronously calculated, for example, the displacement vector is obtained by multiplying the rotation axis direction by the moving step length, and the new position is obtained by superimposing it with the current coordinates; the view matrix is ​​constructed based on the updated virtual camera position and orientation, and the view matrix is ​​multiplied by the projection matrix to obtain a high-altitude virtual scene transformation picture frame; the virtual camera position and orientation in each frame are recorded in chronological order to generate a viewing angle rotation trajectory.

[0081] By adopting the time synchronization and visual sequence fusion method, the transformed picture frames of the high-altitude virtual scene and the perspective rotation trajectory are dynamically integrated to form a visual effect of simulating a high-altitude fall.

[0082] Specifically, each frame of the high-altitude virtual scene transformation picture frame is time-sorted according to the generation timestamp, and the perspective rotation trajectory points composed of the virtual camera position and orientation at the corresponding time point are extracted to ensure that the high-altitude virtual scene transformation picture frame and the perspective rotation trajectory correspond one-to-one on the time axis; at each time point, the corresponding perspective rotation trajectory point is used as the reference perspective parameter of the current frame, and the image content of the high-altitude virtual scene transformation picture frame of the current frame is read; through the posture-based image synthesis method, the virtual camera posture (including three-dimensional coordinate position and three-dimensional Euler angle direction) in the current perspective rotation trajectory point is applied to the current picture frame, and the image in the frame is subjected to perspective distortion correction and rotation transformation to generate a simulated image obtained from posture observation; all the picture frames that have undergone posture fusion processing are combined into an image sequence in time series order to form a high-altitude virtual scene transformation picture frame sequence after dynamic fusion; according to the amplitude of the virtual camera posture change between consecutive picture frames, a corresponding motion blur kernel is generated, and further dynamic blur processing is applied to each frame in the picture sequence to output the synthesized simulated high-altitude falling visual effect.

[0083] Based on gaze data, anxiety index and focus drift speed indicators, multi-dimensional statistical analysis and time series analysis are performed to identify abnormal gaze patterns and anxiety changes, and output behavioral feature labels and behavioral change trend data.

[0084] Specifically, the gaze data are arranged by timestamp, and basic features such as gaze position coordinates, gaze duration, and gaze point change frequency at each moment are extracted. The anxiety index and focus drift speed index are aligned according to corresponding time periods to form a three-dimensional time series dataset consisting of gaze data, anxiety index, and focus drift speed index. Sliding window statistical processing is performed on the three-dimensional time series dataset to calculate statistical features such as average gaze duration, gaze point jump frequency, anxiety index mean, standard deviation, and focus drift speed mean and maximum within each window. Frequency amplitude thresholds are used, such as when the standard deviation of the anxiety index exceeds 0.25 or the gaze jump frequency exceeds 5 times per second, to mark the corresponding window as abnormal. The labeled results are clustered and merged in chronological order to form continuous behavioral change segments. For each behavioral change segment, a behavioral feature label is assigned based on the distribution of statistical indicators, such as "tension," "avoidance," and "focus shift." At the same time, curve fitting is performed on the anxiety index and focus drift speed for consecutive time periods to extract the first-order derivative change trend to represent the behavioral change trend data. Finally, the behavioral feature label and behavioral change trend data for the corresponding time period are output.

[0085] Using the support vector machine algorithm, we classify and identify behavioral change trend data, analyze psychological states and behavioral types, and form an immersive experience report.

[0086] Specifically, the output behavior change trend data is collected, including behavior feature labels, corresponding timestamps, and feature dimension values ​​(such as gaze concentration, anxiety index change rate, focus drift speed mean, etc.); the input feature vector of the support vector machine algorithm is constructed according to the preset classification target. The feature vector includes the numerical combination of the above three dimensions in each behavior time slice, and the known psychological state category labels (such as tension, relaxation, concentration, anxiety) and behavior type category labels (such as frequent saccades, stable gaze, and continuous wandering) in the training data set are annotated; the labeled training samples are input into the support vector machine algorithm for classification. Training, using linear kernel function or radial basis function to construct the hyperplane, and determine the optimal parameter C value and γ value through cross-validation, for example, C=1.2, γ=0.1; input the behavior change trend data to be identified into the trained support vector machine algorithm to obtain the classification results of the corresponding psychological state and behavior type; merge the classification results in chronological order to generate the corresponding psychological state evolution sequence and behavior type evolution sequence; map the evolution sequence into the content of the immersive experience report, which includes time period division, dominant psychological state labeling, typical behavior type annotation and fluctuation event identification.

[0087] It should also be explained that the preset classification targets are determined based on the needs of behavior recognition and the purpose of psychological state analysis. Specifically, a large amount of training sample data labeled with behavior types and psychological states is collected, and the behavior categories and psychological state categories are divided using statistical analysis methods. For example, the behavior categories are set as "calm", "nervous" and "anxious", and the psychological state categories are set as "normal", "mildly abnormal" and "severely abnormal". According to the classification targets, the classification label categories and corresponding relationships of the support vector machine algorithm are clarified to form a standardized target set for training and classification.

[0088] In summary, the present invention achieves intelligent dynamic adjustment of visual disturbance intensity and frequency by integrating fuzzy logic and adaptive gain control, and constructs an immersive adaptation mechanism with closed-loop adjustment capabilities for psychological states. It can perceive the user's anxiety state and gaze behavior changes in real time, dynamically optimize the intensity of visual intervention while processing the ambiguity of psychological reactions, and thus accurately control the stimulation level in the high-altitude fall simulation. It can effectively avoid stress reactions caused by excessive stimulation or training failures caused by insufficient stimulation, and significantly improve the personalized matching capabilities and training effectiveness in psychological behavioral intervention and virtual high-altitude adaptation training.

[0089] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A high-altitude fall simulation system based on VR equipment, characterized by: include, Data collection module, collecting user's pupil distance and initial sight data; The gaze capture module drives the 3D modeling engine to dynamically construct a high-altitude virtual scene based on the initial line of sight data, and captures the user's gaze point location and gaze duration in real time to generate gaze data; The visual perturbation module dynamically adjusts the local visual perturbation of the corresponding area in the high-altitude virtual scene based on the gaze data to form a visual perturbation effect; The psychological index calculation module uses a temporal convolutional neural network to calculate the anxiety index and focus drift speed based on the visual disturbance effect and user eye movement behavior, and obtains the anxiety index and focus drift speed indicators; The disturbance adjustment module uses a method combining fuzzy logic control and adaptive gain adjustment to adaptively adjust the frequency and amplitude of the visual disturbance effect to obtain the disturbance rhythm parameters; The dynamic simulation module uses perturbation rhythm parameters to trigger visual perspective transformation and view rotation to simulate the visual effect of falling from a high altitude. It combines gaze data, anxiety index, and focus drift speed indicators to annotate behavioral characteristics in real time and form an immersive experience report.

2. The high-altitude fall simulation system based on VR equipment according to claim 1, characterized in that: The specific steps of collecting user pupil distance and initial sight data are as follows: Collect images of the user's eyes, use binocular geometry analysis to extract the center points of the left and right pupils, and calculate the distance between the left and right pupils to obtain the user's pupil distance; The coordinates of the gaze starting point, binocular disparity information, head posture, gaze stability parameters and timestamp are collected to form the initial gaze data.

3. The high-altitude fall simulation system based on VR equipment according to claim 1, characterized in that: The specific steps of generating gaze data are as follows: The user's pupil distance and the initial line of sight data are spatially transformed and superimposed, and fused using a rotation matrix to obtain the line of sight direction vector. Ray intersection is then performed in virtual space to obtain the spatial coordinate information of the area the user is facing. According to the spatial coordinate information of the user's facing area, the 3D modeling engine is driven to call high-altitude scene resources to generate a high-altitude virtual scene; The spatial coordinate information and duration of the user's facing area in the high-altitude virtual scene are recorded in real time and integrated into gaze data in chronological order.

4. The high-altitude fall simulation system based on VR equipment according to claim 3, characterized in that: According to the spatial coordinate information of the user's facing area, the 3D modeling engine is driven to call the high-altitude scene resources to generate the high-altitude virtual scene. The specific steps are as follows: Using the octree spatial index and multi-level scene management method, we retrieve and filter the environmental elements and detail data of the corresponding area in the high-altitude scene resource library to obtain the scene element dataset. Classify and sort the scene element data set, call the graphics rendering interface to load and build the three-dimensional environment of the high-altitude virtual scene, and generate real-time rendering frame data; The loading and rendering of high-altitude virtual scene details are dynamically adjusted based on real-time rendering frame data to generate a high-altitude virtual scene.

5. The high-altitude fall simulation system based on VR equipment according to claim 1, characterized in that: The gaze point heat map analysis method is used to calculate the spatial density of the gaze data to obtain the coordinate set of the gaze hotspot area, and the visual elements of the corresponding area in the high-altitude virtual scene are dynamically adjusted to form a visual disturbance effect.

6. The high-altitude fall simulation system based on VR equipment according to claim 1, characterized in that: The specific steps for obtaining the anxiety index and focus drift speed index are as follows: Collect user eye movement behavior data such as gaze trajectory, gaze point location, pupil diameter change, and gaze duration, and form a multi-dimensional time series data sequence in chronological order; Input the multidimensional time series data sequence into the time series convolutional neural network to extract the multidimensional time series feature representation reflecting the response of the visual perturbation effect; The multi-dimensional temporal feature representation is correlated with the visual perturbation effect, and time alignment and feature matching are used to perform synchronous matching on the time axis, extracting the change correlation pattern and obtaining the predicted values ​​of anxiety index and focus drift speed. The predicted values ​​were normalized using the z-score standardization method and dynamically adjusted in combination with the multidimensional time series feature representation to obtain the anxiety index and focus drift speed index.

7. The high-altitude fall simulation system based on VR equipment according to claim 6, characterized in that: The multi-dimensional time series data sequence is input into the time series convolutional neural network to extract the multi-dimensional time series feature representation that reflects the response of the visual perturbation effect. The specific steps are as follows: The multi-dimensional time series data sequence is input into a multi-layer one-dimensional convolutional neural network, local dynamic features are extracted through a sliding time window, and a preliminary feature tensor is output; Using multi-scale fusion and context modeling methods, combined with residual connections and attention mechanisms, the initial feature tensors are weighted integrated and time-series optimized to generate a time-series feature sequence of behavior change trends; Through the fully connected layer, the time series feature sequence is dimensionally compressed and feature integrated to obtain a multi-dimensional time series feature representation.

8. The high-altitude fall simulation system based on VR equipment according to claim 1, characterized in that: The specific steps of obtaining the disturbance rhythm parameters are as follows: The fuzzy logic control method is used to fuzzify the anxiety index and focus drift speed index, and then input them into the preset fuzzy rules for fuzzy aggregation and defuzzification to derive the fuzzy control output value. According to the fuzzy control output value, the adaptive gain adjustment method is used to adjust the frequency and amplitude of the visual disturbance effect to obtain the disturbance rhythm parameters.

9. The high-altitude fall simulation system based on VR equipment according to claim 1, characterized in that: The method of using the disturbance rhythm parameters to trigger the visual perspective transformation and the viewing angle rotation to simulate the visual effect of falling from a high altitude is as follows: According to the disturbance rhythm parameters, the visual perspective transformation matrix is ​​calculated using the matrix transformation method, and the angular velocity calculation method is used to obtain the angular velocity of the viewing angle rotation; Update the projection parameters of the high-altitude virtual scene based on the visual perspective transformation matrix, dynamically adjust the position and orientation of the virtual camera based on the angular velocity of the view rotation, and generate the high-altitude virtual scene transformation frame and view rotation trajectory; By adopting the time synchronization and visual sequence fusion method, the transformed picture frames of the high-altitude virtual scene and the perspective rotation trajectory are dynamically integrated to form a visual effect of simulating a high-altitude fall.

10. The high-altitude fall simulation system based on VR equipment according to claim 1, characterized in that: The specific steps for forming the immersive experience report are as follows: Based on gaze data, anxiety index, and focus drift speed indicators, multi-dimensional statistical analysis and time series analysis are performed to identify abnormal gaze patterns and anxiety changes, and output behavioral feature labels and behavioral change trend data; Using the support vector machine algorithm, we classify and identify behavioral change trend data, analyze psychological states and behavioral types, and form an immersive experience report.

Images