A multi-user virtual reality experience system

Abstract

The application relates to the field of virtual reality, and discloses a multi-user virtual reality experience system, which comprises a real-time positioning system, the real-time positioning system is used for capturing the spatial coordinates and action postures of wearing devices of multiple users, and real-time synchronization response of virtual roles and real actions is realized according to action information and position tracking information of each user; the real-time positioning system is also used for driving the virtual roles corresponding to each user to move in a virtual scene; a multi-modal environment simulation component; the multi-modal environment simulation component is used for simulating various physical environment changes in the virtual scene; a space base; a space construction component dynamically adjusts a space topology structure through an electric sliding rail system to restore the movement and interaction of the user in the virtual environment in a real environment; and a scene generation system generates a virtual scene matched with the position, environment and wall body layout of the user in real time. The method has the following effects: the technical problem of large space mapping error in multi-user collaborative interaction is solved.

Description

Technical Field

[0001] This invention relates to the field of virtual reality, and more particularly to a multi-user virtual reality experience system. Background Technology

[0002] Virtual Reality (VR) technology, as an innovative application system integrating computer graphics, simulation technology, and human-computer interaction, enables users to achieve immersive spatial perception by constructing a three-dimensional digital environment. Relying on head-mounted display devices and spatial positioning systems, this technology can generate biomimetic virtual scenes across multiple sensory dimensions, including vision and hearing, providing users with a deeply immersive experience. Its core feature lies in its groundbreaking natural interaction mechanism—users can not only control the environment through biometric inputs such as limb movements and gesture recognition, but also obtain real-time tactile feedback, force simulation, and other multimodal interactive responses.

[0003] In existing technologies, multi-user collaborative interaction systems suffer from fundamental spatial mapping errors: the local coordinate systems independently constructed by each VR terminal lack a global spatial registration mechanism, causing the transformation matrix between the user's physical position and the virtual coordinate system to fail to maintain orthogonality. When multiple users interact in physical space, the coordinate origins of different devices drift excessively, resulting in asymmetric deviations between the rigid body kinematic constraints of the virtual avatar and the real physical space. This spatial misalignment significantly increases the probability of overlapping collision volumes between users, posing a significant risk of physical contact.

[0004] Furthermore, existing VR experiences are mainly limited to fixed experience areas, and there is a lack of real-time interaction and feedback between user actions and the physical environment. Environmental elements are only presented visually, lacking multimodal physical stimulation and immersive feeling.

[0005] Solving the above-mentioned technical problems is a technical challenge that needs to be overcome by those skilled in the art. Summary of the Invention

[0006] The present invention provides a multi-user virtual reality experience system to at least partially solve the above-mentioned technical problems.

[0007] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a multi-user virtual reality experience system, comprising:

[0008] A real-time positioning system is used to capture the spatial coordinates and motion postures of multiple users' worn devices, and to achieve synchronized response between virtual characters and real actions in real time based on the motion information and location tracking information of each user; the real-time positioning system is also used to drive the virtual characters of each user to move in the virtual scene.

[0009] A multimodal environment simulation component; the multimodal environment simulation component is used to simulate various physical environment changes in a virtual scene;

[0010] A spatial base; the spatial base includes a spatial construction component and a scene generation system; the spatial construction component dynamically adjusts the spatial topology through an electric sliding rail system to reproduce the user's movement and interaction in the virtual environment in the real environment; the scene generation system generates virtual scenes in real time that match the user's position, environment, and wall layout.

[0011] In one optional implementation, the spatial coordinates and motion postures of multiple users' wearing devices are captured, and the virtual character's response to real-time actions is synchronized based on each user's motion information and location tracking information, including:

[0012] The user's wearable device collects angular velocity, acceleration, and magnetic field data to form a raw inertial data stream; an optical positioning data stream is generated by capturing the coordinates of optical marker points using an optical marker array in the ceiling and a top-mounted infrared camera.

[0013] The original optical data stream and inertial data stream are clock-synchronized, and a unified timestamp is added to both types of data;

[0014] The clock-synchronized optical positioning data and inertial data are input into the extended Kalman filter engine, and the fused pose data stream is output through dynamic weight allocation.

[0015] A spatiotemporal calibration algorithm is applied to the fused pose data stream to eliminate the origin drift between the local coordinate system and the global coordinate system of each user device, maintain the orthogonality of the virtual and real space transformation matrix, and generate calibrated pose data.

[0016] The system monitors the optical marker capture status in real time. When occlusion occurs, it starts the inertial navigation pre-integration model based on the calibrated pose data to predict the current pose and maintain positioning continuity. After the occlusion is removed, the predicted pose is converged to the optical coordinate system through the Lie group manifold optimization algorithm to generate anti-occlusion pose data.

[0017] For multi-user interaction scenarios, real-time deviation analysis is performed on pose data after anti-occlusion processing, and a group collision-free pose data stream is generated by compensating for rotation and translation components between devices through coordinate mapping algorithm.

[0018] Based on the collision-free pose data stream of the group, the virtual character is driven to achieve a mapping response to the user's real actions.

[0019] In one optional implementation, the multimodal environment simulation component includes:

[0020] An elevator device; the elevator device is used to simulate vertical movement perception in a virtual scene;

[0021] Risk control device; the risk control device is used to recreate climate change in a virtual scene;

[0022] Temperature control device; the temperature control device is used to reproduce temperature changes in a virtual scene;

[0023] Pneumatic device; the pneumatic device is used to recreate pneumatic impacts in a virtual scene;

[0024] Vibration device; the vibration device is used to reproduce the physical effects of vibration in a virtual scene;

[0025] A spraying device; the spraying device is used to simulate the water flow environment of a virtual scene.

[0026] In one optional embodiment, the elevator device includes: a four-dimensional motion base, an intelligent drive system, a dual-domain control model, and an environmental linkage module. The four-dimensional motion base employs a combination of servo motors and harmonic reducers, achieving composite motion of vertical lifting, horizontal translation, and rotation through a six-degree-of-freedom motion platform. The intelligent drive system, based on a fuzzy PID control algorithm, generates six-dimensional spatial commands according to virtual scene motion parameters, driving the four-dimensional motion base to simulate vertical lifting, composite motion, and inertial simulation scenarios. The dual-domain control model acquires user center-of-gravity offset data in real-time through a motion capture unit in the virtual domain, inputting it to a motion calculation engine to generate virtual scene motion parameters. In the physical domain, the dual-domain control model monitors the pressure distribution at the user's contact point with the platform in real-time through a pressure sensor array, correcting the platform's posture based on pressure changes. The environmental linkage module works in conjunction with a temperature control device to simulate a 0.6°C temperature decrease for every 10m increase in the vertical lifting scenario. The environmental linkage module also works in conjunction with a wind control device to generate unidirectional airflow in high-speed motion scenarios to enhance immersion.

[0027] The wind control device includes an adjustable fan array and an intelligent control module; the adjustable fan array is driven by a variable frequency motor and simulates a natural wind field through the Bernoulli effect; the intelligent control module is equipped with an LSTM neural network algorithm to analyze the climate parameters of the virtual scene in real time and generate dynamic control commands for wind speed and wind direction.

[0028] The temperature control device includes: a regional gradient temperature control array, a dynamic thermal radiation adjustment system, and a multimodal collaborative interface; the regional gradient temperature control array adopts a semiconductor thermopile module matrix; the dynamic thermal radiation adjustment system is based on an inverse square decay model and adjusts the thermal radiation intensity according to the distance between the user and the virtual heat source; the multimodal collaborative interface is linked with the wind control device and the sprinkler device: each wind speed corresponds to a temperature compensation of 0.5℃; when simulating rain scenarios, it reduces the user's body surface temperature by 2-5℃;

[0029] The pneumatic device includes a high-pressure injection system and a fluid dynamics simulation module; the high-pressure injection system integrates a high-pressure air pump and a multi-directional nozzle array to simulate fluid impact force; the fluid dynamics simulation module is used to calculate the injection parameters of the high-pressure injection system.

[0030] The vibration device includes a distributed servo motor array and a multi-waveform synthesis engine; no fewer than 6 servo motors are deployed on the user standing platform to form a three-dimensional vibration matrix; the multi-waveform synthesis engine decomposes the vibration parameters of the virtual scene into earthquake simulation or explosion impact simulation.

[0031] In one alternative implementation, the space-building components include: modular wall units and an electric sliding rail system;

[0032] Methods for achieving virtual-to-real-world matching using a spatial base include:

[0033] The modular wall units are driven by an electric sliding rail system to control the translation and rotation of the wall in the X, Y, and Z axes.

[0034] Wall splicing and separation are achieved using an electromagnetic adsorption interface; multiple walls are coordinated to move synchronously through a distributed PLC control network.

[0035] The system monitors the user's location in real time. When the user enters a range of 1.2m from the wall, a safety buffer mechanism is triggered to pause or adjust the wall's movement.

[0036] Methods for scene generation systems to generate virtual scenes that match the user's location in real time include:

[0037] The physical space is modeled in three dimensions using an octree spatial index, and the coordinates of the physical space are scaled proportionally to the coordinates of the virtual scene using a perspective projection transformation algorithm.

[0038] The dynamic rendering engine is used to perform differentiated rendering of different radius areas based on user eye tracking data;

[0039] When a user moves to within 1.5m of the virtual scene boundary, the adjacent virtual scene modules are preloaded through the spatial mapping algorithm, and the spatial construction components are driven to adjust the wall layout to maintain the topological consistency between the virtual scene and the physical space.

[0040] Real-time acquisition of wall position sensor data, and synchronization of physical wall position to virtual scene through coordinate mapping algorithm to generate virtual obstacle model that matches the wall layout of the real environment;

[0041] Methods for achieving virtual-real environment synchronization through collaborative control units include:

[0042] It receives user spatial coordinates and pose data provided by a real-time positioning system, and ensures the orthogonality between the physical coordinate system and the virtual coordinate system through a Lie group transformation matrix algorithm;

[0043] When the system detects that the user's movement speed exceeds the preset speed, it triggers the space construction component to adjust the wall positions and the scene generation system to update the virtual scene, thereby achieving dynamic matching between virtual environment movement and real space topology transformation.

[0044] In one alternative implementation, the electric sliding rail system in the space building component further performs the following steps when driving the modular wall units to reconstruct the topology:

[0045] The system trains a long short-term memory network model based on historical pose data to predict the movement trajectory of each user in the first time period in real time; and introduces a priority weight factor into the path planning to allocate obstacle avoidance priority according to the user's distance from the wall, movement speed, line of sight and action type.

[0046] When a potential collision risk is detected between two or more users, a group collaborative avoidance strategy is activated, and the positions of multiple walls are adjusted synchronously through a distributed PLC control network.

[0047] When a user enters the 1.2m danger zone of the wall's movement path, the electric sliding rail system switches to a low-speed, slow-motion mode and renders a warning visual effect in the virtual scene to alert the user.

[0048] In one alternative implementation, when a potential collision risk is detected between two or more users, a group cooperative avoidance strategy is activated, synchronously adjusting the positions of multiple walls via a distributed PLC control network, including:

[0049] The system acquires the current spatial coordinates, direction of motion, and speed information of each user, and predicts the motion trajectory in the first time period based on a long short-term memory network model.

[0050] Cross-analysis is performed on the movement trajectories of all users in pairs to calculate the shortest interval distance and the time of occurrence. If the shortest interval distance is less than a preset safety threshold and the time of occurrence is within the second time period, it is determined that there is a potential collision risk between the two.

[0051] After identifying a potential collision risk, a wall adjustment command is sent to the distributed PLC control network. The command includes the wall number to be moved, the target location, and the priority identifier. The distributed PLC control network generates multiple movement paths for the walls based on the wall adjustment command to ensure that the wall movement does not interfere with the normal interaction and walking paths of other users.

[0052] During the wall movement, the system continuously monitors changes in the user's pose and dynamically updates the wall's movement path. If the user's behavior changes and the collision risk is eliminated, the wall movement operation is canceled.

[0053] After the wall is moved, the scene generation system updates the obstacle model in the virtual scene in real time and guides the user to avoid areas with potential collision risks through visual cues.

[0054] In one optional implementation, the multimodal environment simulation component further includes a multimodal sensing synchronization engine; the multimodal sensing synchronization engine is used to coordinate the physical feedback timing between the elevator device, the wind control device, the temperature control device, the pneumatic device, the vibration device, and the sprinkler device; the multimodal environment simulation component can be used to perform the following steps:

[0055] Receive event trigger signals in the virtual scene and parse the event type and intensity parameters;

[0056] A preset response template is invoked based on the event type; the response template includes wind speed, temperature, vibration frequency, air pressure change, and water mist density.

[0057] The execution commands of each device are time-aligned using a unified timeline controller;

[0058] In the explosion simulation scenario, the vibration device and pneumatic device are first triggered to generate a shock wave effect. After 0.2 seconds, the wind control device and spray device are activated to simulate dust and debris splashing. Finally, the local temperature is raised by the temperature control device to enhance the thermal impact.

[0059] In extreme weather simulation scenarios, by combining wind speed and temperature linkage strategies, for each increase in wind force level, the body surface temperature of users in the corresponding area is reduced by 0.5℃, and the impact force of raindrops is simulated by a pneumatic device, which is combined with a spray device to achieve a water mist effect.

[0060] The intensity of the environmental simulation is reduced when abnormal user physiological indicators are detected.

[0061] In one alternative implementation, the electric slide rail system is also capable of performing the following steps:

[0062] The wall planning model is trained using historical user behavior data; the wall planning model takes user density distribution, movement trends and interaction hotspots as state inputs and outputs the optimal wall layout configuration strategy.

[0063] Before each wall reconstruction task is initiated, the wall planning model predicts the walls that most likely need to be adjusted and their target locations, and generates a priority scheduling table; the movement of the walls is controlled based on the priority scheduling table.

[0064] Under the premise of satisfying obstacle avoidance and synchronous control, calculate the minimum energy consumption path of the wall from the current pose to the target pose;

[0065] After the wall is moved, the wall position sensor is used to verify whether the actual layout matches the expected configuration. If there is a deviation, the position is adjusted.

[0066] Compared with the prior art, the present invention has at least the following beneficial effects: it solves the technical problem of large spatial mapping errors in multi-user collaborative interaction. Attached Figure Description

[0067] Figure 1 This is a system block diagram of a multi-user virtual reality experience system provided by the present invention;

[0068] Figure 2 This is a schematic diagram of the structure of a multi-user virtual reality experience system provided by the present invention. Detailed Implementation

[0069] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0070] Reference Figure 1 The first embodiment of the present invention provides a multi-user virtual reality experience system, comprising:

[0071] A real-time positioning system is used to capture the spatial coordinates and motion postures of multiple users' worn devices, and to achieve synchronized response between virtual characters and real actions in real time based on the motion information and location tracking information of each user; the real-time positioning system is also used to drive the virtual characters of each user to move in the virtual scene.

[0072] A multimodal environment simulation component; the multimodal environment simulation component is used to simulate various physical environment changes in a virtual scene;

[0073] A spatial base; the spatial base includes a spatial construction component and a scene generation system; the spatial construction component dynamically adjusts the spatial topology through an electric sliding rail system to reproduce the user's movement and interaction in the virtual environment in the real environment; the scene generation system generates virtual scenes in real time that match the user's position, environment, and wall layout.

[0074] Specifically, the real-time positioning system uses multiple sensors to work together to capture the spatial coordinates and motion postures of multiple users wearing devices (such as head-mounted displays and controllers), and achieves millisecond-level synchronous response between virtual characters and real actions based on the motion information and location tracking information of each user, thus synchronously driving the virtual characters of each user to move in the virtual scene.

[0075] The hardware architecture of the real-time positioning system adopts a top-mounted optical-inertial hybrid positioning scheme. An infrared optical marker array (orthogonal grid arrangement, 36-49 marker points per square meter, diameter 2-3mm, reflectivity >95%) is deployed on the ceiling. The user equipment has a built-in nine-axis inertial sensing component (including a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer, with a sampling frequency ≥1000Hz). The optical marker array is a high-density matrix of infrared reflective dots deployed on the ceiling. In conjunction with a top-mounted infrared camera, it can calculate the absolute coordinates of the device in real time. The nine-axis inertial sensing component continuously outputs attitude data by sensing the device's acceleration, angular velocity, and geomagnetic signals.

[0076] The positioning principle of a real-time positioning system: The optical positioning unit captures marker points using distributed top-mounted cameras (such as the Mars2H type, with a honeycomb arrangement covering a 12m×12m area) and calculates the absolute coordinates of the device; the inertial navigation unit monitors the motion state through sensor data. Both achieve microsecond-level clock synchronization via the PTP precise time protocol and fuse data using the extended Kalman filter (EKF) algorithm. Optical data dominates low-frequency pose correction (10Hz update), while inertial data handles high-frequency motion prediction (1000Hz interpolation), ultimately achieving more precise accuracy and sub-millimeter-level synchronization accuracy. The extended Kalman filter algorithm is a recursive state estimation algorithm that can dynamically adjust weights to optimize positioning results and reduce error accumulation by fusing uncertain sensor data.

[0077] The multimodal environment simulation component has a built-in physics engine and interaction protocol to simulate various physical environment changes in virtual scenes, such as temperature and humidity, airflow, vibration, and tactile feedback.

[0078] The multimodal environment simulation component implements the following functions: It is equipped with an environmental parameter acquisition matrix (real-time analysis of physical parameters such as temperature, airflow, and lighting) to drive devices for elevator displacement, wind pressure control, temperature and humidity regulation, aerodynamic feedback, vibration simulation, and fluid spraying. For example, in a virtual rainforest scene, the system automatically maintains a perceived temperature of 28°C and intermittently sprays water mist; in a virtual snowfield scene, it generates a low perceived temperature and superimposes directional cold air. The physics engine calculates the dynamic matching of tactile feedback intensity with virtual scene parameters in real time, ensuring consistent perception across sensory channels. The physics engine is a software module that simulates the physical laws of the real world through mathematical models. It can calculate the motion, collision, and force feedback of objects, making interactions in the virtual scene conform to real physical logic.

[0079] The spatial base includes spatial construction components and a scene generation system, which achieves accurate mapping between physical space and virtual scene through the collaboration of hardware and algorithms.

[0080] Spatial Construction Components: Utilizing modular splicing units, the spatial topology is dynamically adjusted via an electric sliding rail system, replicating the user's movement and interaction in the virtual environment in the real world. For example, the electric sliding rails can drive modular walls to quickly reconstruct the spatial layout; a 10-meter physical corridor can be visually extended to appear as a 100-meter corridor through mirror reflections and lighting effects; and mirror reflection devices at corners can create the illusion of overlapping spaces. The electric sliding rail system is a motor-driven track device that controls the movement and combination of modular walls, enabling real-time adjustments to the physical spatial layout to match the geometric features of the virtual scene.

[0081] Scene Generation System: Based on real-time positioning data, multimodal environmental data, and modular wall layouts, the system generates virtual scenes in real time that match the user's location, environment, and wall layout. The system maximizes the use of limited scenes through spatial mapping algorithms. For example, when the user moves, millimeter-level positioning data synchronously updates the virtual view, environmental sensors trigger corresponding climate simulation devices, and physical walls synchronously change their spatial topology. Finally, a 3D scene that interacts with real-world environmental parameters is generated in the dynamic rendering engine. The spatial mapping algorithm establishes a correspondence between physical and virtual spatial coordinates. Through topology optimization technology, it can achieve infinite extension of the virtual scene within a limited physical space, such as mapping a 10×10m physical space into a large virtual open world.

[0082] In this application, the real-time positioning system establishes a coordinate mapping between physical and virtual spaces, the multimodal environment simulation component converts virtual scene parameters into physical feedback, and the dynamic topology adjustment of the spatial base ensures consistency between the virtual and physical space boundaries. Compared with existing technologies, this solution solves the problems of large spatial mapping errors and high collision risks in multi-user collaborative interaction, achieving sub-millimeter-level positioning accuracy, multimodal physical feedback (touch, temperature, wind, etc.), and dynamic spatial adaptation, significantly improving the safety and immersion of group interaction.

[0083] In one embodiment, the step of capturing the spatial coordinates and motion postures of multiple users' wearing devices, and achieving real-time synchronization between virtual characters and real actions based on each user's motion information and location tracking information, includes:

[0084] The user's wearable device collects angular velocity, acceleration, and magnetic field data to form a raw inertial data stream; an optical positioning data stream is generated by capturing the coordinates of optical marker points using an optical marker array in the ceiling and a top-mounted infrared camera.

[0085] The original optical data stream and inertial data stream are clock-synchronized, and a unified timestamp is added to both types of data;

[0086] The clock-synchronized optical positioning data and inertial data are input into the extended Kalman filter engine, and the fused pose data stream is output through dynamic weight allocation.

[0087] A spatiotemporal calibration algorithm is applied to the fused pose data stream to eliminate the origin drift between the local coordinate system and the global coordinate system of each user device, maintain the orthogonality of the virtual and real space transformation matrix, and generate calibrated pose data.

[0088] The system monitors the optical marker capture status in real time. When occlusion occurs, it starts the inertial navigation pre-integration model based on the calibrated pose data to predict the current pose and maintain positioning continuity. After the occlusion is removed, the predicted pose is converged to the optical coordinate system through the Lie group manifold optimization algorithm to generate anti-occlusion pose data.

[0089] For multi-user interaction scenarios, real-time deviation analysis is performed on pose data after anti-occlusion processing, and a group collision-free pose data stream is generated by compensating for rotation and translation components between devices through coordinate mapping algorithm.

[0090] Based on the collision-free pose data stream of the group, the virtual character is driven to achieve a mapping response to the user's real actions.

[0091] Specifically, based on the nine-axis inertial sensing component (integrating a three-axis accelerometer, a three-axis gyroscope, and a three-axis magnetometer) in the user-worn device, angular velocity, acceleration, and magnetic field data of the device are collected at a sampling frequency of ≥1000Hz to form a raw inertial data stream. The nine-axis inertial sensing component is an integrated motion sensor, in which the three-axis accelerometer is used to sense the linear acceleration of the device along the X / Y / Z axes, the three-axis gyroscope monitors the rotational angular velocity, and the three-axis magnetometer provides a geomagnetic direction reference. The three work together to output the real-time motion status of the device.

[0092] Based on an optical marker array (orthogonal grid arrangement, 36-49 infrared reflective markers per square meter, 2-3mm in diameter, reflectivity >95%) deployed in the ceiling, a top-mounted infrared camera (such as the Mars2H type, with a honeycomb distribution covering the target area) captures the coordinates of the optical markers, generating a raw optical positioning data stream. The optical marker array and the top-mounted infrared camera constitute a visual positioning unit. By capturing images of the markers with the camera, computer vision algorithms are used to calculate the absolute position of the device in the global coordinate system.

[0093] The raw optical and inertial data streams are clock-synchronized using PTP (Programmable Time Protocol) to achieve microsecond-level clock synchronization. A unified timestamp is added to both types of data to ensure consistency across different sensors in the time dimension. PTP is a high-precision clock synchronization protocol that enables devices distributed across a network to maintain clock synchronization, preventing positioning data misalignment caused by clock deviations.

[0094] The clock-synchronized optical positioning data and inertial data are input into the Extended Kalman Filter (EKF) engine, which outputs a fused pose data stream through a dynamic weight allocation mechanism. Specifically, optical data dominates low-frequency pose correction (10Hz update) to eliminate accumulated errors; inertial data is responsible for high-frequency motion prediction (interpolated using 1000Hz sampled data) to handle fast-moving scenarios. The EKF algorithm is a recursive state estimation algorithm that transforms the state estimation problem of a nonlinear system into a linear one. Through a prediction-update iterative process, it fuses uncertain sensor data to optimize the pose estimation results.

[0095] A spatiotemporal calibration algorithm is applied to the fused pose data stream to eliminate origin drift between the local and global coordinate systems of each user device, maintain the orthogonality of the virtual-real space transformation matrices, and generate calibrated pose data. Through multiple rounds of iterative optimization, the spatiotemporal calibration algorithm calculates the rotation matrix and translation vector between the local and global coordinate systems, ensuring that the rigid body kinematic constraints of the virtual avatar are consistent with the real physical space and avoiding virtual scene misalignment caused by coordinate deviations.

[0096] The system monitors the optical marker capture status in real time. When occlusion occurs (such as the user's body obscuring the marker), the inertial navigation pre-integration model is activated based on the calibrated pose data to predict the current pose and maintain positioning continuity. The inertial navigation pre-integration model calculates the device's trajectory over a short period by integrating acceleration and angular velocity, thus avoiding positioning interruptions caused by optical occlusion.

[0097] After the occlusion is removed, the predicted pose is converged to the optical coordinate system using a Lie group manifold optimization algorithm, generating occlusion-resistant pose data. The Lie group manifold optimization algorithm is a differential geometry-based optimization method that can optimize within a Lie group space composed of rotation matrices, quickly aligning the inertial predicted pose with the optical positioning data and eliminating accumulated errors.

[0098] For multi-user interaction scenarios, real-time deviation analysis is performed on the pose data after anti-occlusion processing. A coordinate mapping algorithm is used to compensate for the rotation and translation components between devices, generating a collision-free pose data stream for the group. The coordinate mapping algorithm establishes a transformation relationship between the multi-user device coordinate system and the global coordinate system, uniformly adjusting the pose data of each device to ensure that the position of the virtual avatar in the virtual scene is consistent with the real physical space, reducing the probability of collision volume overlap.

[0099] Based on the aforementioned collision-free pose data stream, the system drives the virtual character to achieve millisecond-level mapping response to the user's real movements. The system uses a dynamic rendering engine to transform the pose data into the virtual character's motion trajectory and posture changes, synchronizing the virtual character's movements with the user's real movements and enhancing the immersiveness of the VR experience. The dynamic rendering engine is a software module that generates 3D images in real time, rapidly updating the virtual scene based on the input pose data to ensure visual feedback matches the user's actions.

[0100] The solution proposed in this application improves the positioning accuracy to the 0.5mm level through multi-source data fusion, spatiotemporal calibration and anti-occlusion strategy. At the same time, it controls the overlap rate of virtual collision boxes during multi-user interaction within the theoretical range, solves the risk of physical contact caused by spatial misalignment in the prior art, and achieves sub-millimeter level synchronization accuracy and low latency transmission, providing a precise spatial perception foundation for multi-user VR collaborative interaction.

[0101] In one embodiment, the multimodal environment simulation component includes:

[0102] An elevator device; the elevator device is used to simulate vertical movement perception in a virtual scene;

[0103] Risk control device; the risk control device is used to recreate climate change in a virtual scene;

[0104] Temperature control device; the temperature control device is used to reproduce temperature changes in a virtual scene;

[0105] Pneumatic device; the pneumatic device is used to recreate pneumatic impacts in a virtual scene;

[0106] Vibration device; the vibration device is used to reproduce the physical effects of vibration in a virtual scene;

[0107] A spraying device; the spraying device is used to simulate the water flow environment of a virtual scene.

[0108] In one embodiment, the elevator device includes: a four-dimensional motion base, an intelligent drive system, a dual-domain control model, and an environmental linkage module. The four-dimensional motion base employs a combination of servo motors and harmonic reducers, achieving composite motion of vertical lifting, horizontal translation, and rotation through a six-degree-of-freedom motion platform. The intelligent drive system, based on a fuzzy PID control algorithm, generates six-dimensional spatial commands according to virtual scene motion parameters, driving the four-dimensional motion base to simulate vertical lifting, composite motion, and inertial simulation scenarios. The dual-domain control model acquires user center-of-gravity offset data in real-time through a motion capture unit in the virtual domain, inputting it to a motion calculation engine to generate virtual scene motion parameters. In the physical domain, the dual-domain control model monitors the pressure distribution at the user's contact point with the platform in real-time through a pressure sensor array, correcting the platform's posture based on pressure changes. The environmental linkage module works in conjunction with a temperature control device to simulate a 0.6°C temperature decrease for every 10m increase in the vertical lifting scenario. The environmental linkage module also works in conjunction with a wind control device to generate unidirectional airflow in high-speed motion scenarios to enhance immersion.

[0109] The wind control device includes an adjustable fan array and an intelligent control module; the adjustable fan array is driven by a variable frequency motor and simulates a natural wind field through the Bernoulli effect; the intelligent control module is equipped with an LSTM neural network algorithm to analyze the climate parameters of the virtual scene in real time and generate dynamic control commands for wind speed and wind direction.

[0110] The temperature control device includes: a regional gradient temperature control array, a dynamic thermal radiation adjustment system, and a multimodal collaborative interface; the regional gradient temperature control array adopts a semiconductor thermopile module matrix; the dynamic thermal radiation adjustment system is based on an inverse square decay model and adjusts the thermal radiation intensity according to the distance between the user and the virtual heat source; the multimodal collaborative interface is linked with the wind control device and the sprinkler device: each wind speed corresponds to a temperature compensation of 0.5℃; when simulating rain scenarios, it reduces the user's body surface temperature by 2-5℃;

[0111] The pneumatic device includes a high-pressure injection system and a fluid dynamics simulation module; the high-pressure injection system integrates a high-pressure air pump and a multi-directional nozzle array to simulate fluid impact force; the fluid dynamics simulation module is used to calculate the injection parameters of the high-pressure injection system.

[0112] The vibration device includes a distributed servo motor array and a multi-waveform synthesis engine; no fewer than 6 servo motors are deployed on the user standing platform to form a three-dimensional vibration matrix; the multi-waveform synthesis engine decomposes the vibration parameters of the virtual scene into earthquake simulation or explosion impact simulation.

[0113] Specifically, the elevator device consists of a four-dimensional motion base, an intelligent drive system, a dual-domain control model, and an environmental linkage module, and achieves physical restoration of the virtual scene through mechatronics design.

[0114] The four-dimensional motion base utilizes a combination of servo motors and harmonic reducers to achieve composite motions of vertical lifting, horizontal translation, and rotation through a six-degree-of-freedom motion platform. The servo motors provide power through precise speed and torque control, while the harmonic reducers achieve high-precision speed reduction via flexible gears, ensuring a positioning accuracy of ±0.5mm for the motion platform. The six-degree-of-freedom motion platform is a mechanical structure capable of translation along the X, Y, and Z axes and rotation around these three axes, simulating multi-dimensional motion in reality.

[0115] Intelligent drive system: Based on fuzzy PID control algorithm, it generates six-dimensional spatial commands according to the motion parameters of virtual scene (such as elevator lifting speed and acceleration) to drive the four-dimensional motion base to simulate vertical lifting scene, compound motion scene (such as the forward tilting feeling when the elevator accelerates and starts) and inertial simulation scene (such as the inertial backward tilting when the elevator stops suddenly).

[0116] The dual-domain control model includes a virtual domain and a physical domain.

[0117] Virtual Domain: Real-time data on the user's center of gravity shift is acquired via motion capture units (such as a nine-axis IMU sensor array) and input into the motion processing engine to generate motion parameters for the virtual scene (such as the acceleration curve of an elevator). The motion capture unit can accurately monitor changes in the user's body posture.

[0118] Physical Domain: A pressure sensor array (deployed on the platform surface) monitors the pressure distribution at the user's contact points with the platform in real time, and corrects the platform's posture based on pressure changes (e.g., when the user's center of gravity leans forward, the platform tilts forward synchronously to simulate real force). The pressure sensor array can sense the user's weight distribution on the platform, ensuring the comfort and safety of the motion simulation.

[0119] The environmental linkage module works in conjunction with the temperature control device: In vertical lifting scenarios, the simulated temperature decreases by 0.6℃ for every 10m of ascent, achieved through a regional gradient temperature control array. The environmental linkage module also works in conjunction with the wind control device: In high-speed motion scenarios, it generates unidirectional airflow to enhance immersion; for example, when an elevator ascends at high speed, the top fan array blows airflow in the same direction as the movement.

[0120] The wind control device includes an adjustable fan array and an intelligent control module, which realizes the restoration of natural wind field through fluid dynamics design.

[0121] Adjustable fan array: Driven by variable frequency motors, it simulates natural wind fields through the Bernoulli effect, such as generating eddies or gusts by varying the speeds of different fans. The variable frequency motors can adjust their speed in real time, controlling the wind speed to continuously change within the range of 0-15 m / s, with a wind speed accuracy of ±0.5 m / s.

[0122] Intelligent control module: Equipped with an LSTM neural network algorithm, it analyzes virtual scene climate parameters (such as wind speed, wind direction, and weather type) in real time and generates dynamic control commands for wind speed and direction. For example, when the virtual scene is a typhoon, the system controls the fan array to simulate the typhoon effect with a wind speed of 10-12 m / s and a periodically changing wind direction.

[0123] The temperature control device consists of a regional gradient temperature control array, a dynamic thermal radiation adjustment system, and a multimodal collaborative interface, enabling precise thermal stimulation feedback.

[0124] Regional gradient temperature control array: Employing a semiconductor thermopile module matrix, it achieves a temperature control accuracy of ±0.5℃ within a 1m×1m area, simulating temperature gradients in virtual scenarios (such as localized temperature rise when near a virtual fire source). The semiconductor thermopile modules can rapidly achieve heating or cooling with a response time of <1s.

[0125] The dynamic thermal radiation adjustment system adjusts the thermal radiation intensity based on the distance between the user and the virtual heat source, using an inverse square attenuation model. For example, when the user is 2m away from the virtual heat source, the thermal radiation intensity is 100W / m²; when the user is 1m away from the heat source, the intensity increases to 400W / m².

[0126] The multimodal collaborative interface is used to link with wind control devices: each wind speed level (e.g., 3 m / s is level 1) corresponds to a temperature compensation of 0.5℃. For example, when the wind speed is 6 m / s, the temperature decreases by 1℃, simulating the "wind cooling effect". The multimodal collaborative interface is also used to link with sprinkler systems: when simulating rain scenarios, it reduces the user's body surface temperature by 2-5℃, and enhances the realism by combining water mist spraying.

[0127] The pneumatic device includes a high-pressure injection system and a fluid dynamics simulation module for simulating fluid impact force.

[0128] High-pressure jet system: Integrates a high-pressure air pump (output pressure 0.5-1MPa) and a multi-directional nozzle array (covering 360° direction) to simulate the impact force of fluids such as water and air. For example, in a virtual scene, when a user touches the high-pressure water gun, the nozzle array jets air at a pressure of 0.8MPa to simulate the impact of water.

[0129] Fluid dynamics simulation module: Based on computational fluid dynamics (CFD) algorithms, it calculates the injection parameters of the high-pressure injection system (such as pressure, flow rate, and injection angle) to ensure that the simulated impact force is consistent with the fluid characteristics in the virtual scene (such as the difference in impact force under different water pressures).

[0130] The vibration device includes a distributed servo motor array and a multi-waveform synthesis engine to achieve vibration feedback in multiple scenarios.

[0131] Distributed servo motor array: Deploy no fewer than 6 servo motors (e.g., 6 motors forming a 2×3 matrix) on the user's standing platform. The precise rotation of these motors generates a three-dimensional vibration matrix with a vibration frequency range of 1-100Hz and an amplitude accuracy of ±0.1mm. The servo motors can precisely control the rotation angle and speed, ensuring the accuracy of the vibration simulation.

[0132] Multi-waveform synthesis engine: Decomposes the vibration parameters of the virtual scene (such as earthquake magnitude and explosion intensity) into earthquake simulation (low-frequency, continuous vibration) or explosion impact simulation (high-frequency, pulsed vibration). For example, in a magnitude 5 earthquake in the virtual scene, the system generates low-frequency vibrations of 2-5Hz with an amplitude of 5-10mm; in an explosion scene, it generates high-frequency pulsed vibrations of 50-100Hz with a duration of 0.5-1s.

[0133] Compared to existing VR systems that only present environmental changes visually (such as temperature that only displays numerical values), this solution achieves precise coupling of multimodal physical feedback through innovative designs such as a six-degree-of-freedom motion platform for an elevator device (most existing technologies are three degrees of freedom), LSTM dynamic wind field control for a wind control device (existing technologies have fixed wind speeds), and an inverse square thermal radiation model for a temperature control device (existing technologies have no distance attenuation). This provides a foundation for multi-sensory interaction in multi-user VR experiences.

[0134] In one embodiment, the space construction component includes: modular wall units and an electric sliding rail system;

[0135] Methods for achieving virtual-to-real-world matching using a spatial base include:

[0136] The modular wall units are driven by an electric sliding rail system to control the translation and rotation of the wall in the X, Y, and Z axes.

[0137] Wall splicing and separation are achieved using an electromagnetic adsorption interface; multiple walls are coordinated to move synchronously through a distributed PLC control network.

[0138] The system monitors the user's location in real time. When the user enters a range of 1.2m from the wall, a safety buffer mechanism is triggered to pause or adjust the wall's movement.

[0139] Methods for scene generation systems to generate virtual scenes that match the user's location in real time include:

[0140] The physical space is modeled in three dimensions using an octree spatial index, and the coordinates of the physical space are scaled proportionally to the coordinates of the virtual scene using a perspective projection transformation algorithm.

[0141] The dynamic rendering engine is used to perform differentiated rendering of different radius areas based on user eye tracking data;

[0142] When a user moves to within 1.5m of the virtual scene boundary, the adjacent virtual scene modules are preloaded through the spatial mapping algorithm, and the spatial construction components are driven to adjust the wall layout to maintain the topological consistency between the virtual scene and the physical space.

[0143] Real-time acquisition of wall position sensor data, and synchronization of physical wall position to virtual scene through coordinate mapping algorithm to generate virtual obstacle model that matches the wall layout of the real environment;

[0144] Methods for achieving virtual-real environment synchronization through collaborative control units include:

[0145] It receives user spatial coordinates and pose data provided by a real-time positioning system, and ensures the orthogonality between the physical coordinate system and the virtual coordinate system through a Lie group transformation matrix algorithm;

[0146] When the system detects that the user's movement speed exceeds the preset speed, it triggers the space construction component to adjust the wall positions and the scene generation system to update the virtual scene, thereby achieving dynamic matching between virtual environment movement and real space topology transformation.

[0147] Specifically, the spatial construction components include modular wall units and an electric sliding rail system. The modular wall units employ a standardized splicing structure, enabling rapid assembly and disassembly via electromagnetic adsorption interfaces. The electric sliding rail system, deployed on the ceiling and floor, is driven by servo motors and can control the translation and rotation of the walls in the X, Y, and Z axes. The electromagnetic adsorption interface is a device that uses electromagnetic force to connect the walls; when energized, it generates a strong magnetic force to attract the walls, and when de-energized, it can be easily separated.

[0148] The synchronous movement of multiple walls is coordinated by a distributed PLC control network. Each wall unit is equipped with an independent PLC controller, and the synchronous movement of multiple walls is achieved through industrial Ethernet.

[0149] The system monitors the user's location in real time. When the user enters a certain range from the wall, a safety buffer mechanism is triggered to pause or adjust the wall's movement speed to prevent the user from colliding with the moving wall.

[0150] The scene generation system uses an octree spatial index to perform 3D mesh modeling of the physical space. A perspective projection transformation algorithm is used to scale the physical space coordinates to the virtual scene coordinates, achieving a mapping between the finite physical space and the infinite virtual scene. The octree spatial index is a 3D spatial partitioning data structure that can efficiently manage the position and attributes of spatial objects.

[0151] By leveraging a dynamic rendering engine based on user gaze tracking data, differentiated rendering is applied to areas of varying radii to optimize computational resource allocation. The dynamic rendering engine can adjust rendering resources in real time according to user attention, improving system efficiency while maintaining image quality in critical areas.

[0152] When a user moves to a certain range within the boundaries of the virtual scene, adjacent virtual scene modules are preloaded through a spatial mapping algorithm. Simultaneously, the spatial construction components adjust the wall layout to maintain topological consistency between the virtual scene and the physical space. The spatial mapping algorithm achieves infinite virtual expansion of a finite physical space through coordinate transformation and scene stitching technology.

[0153] Real-time data acquisition from wall position sensors is used to synchronize the physical wall positions to the virtual scene via a coordinate mapping algorithm, generating a virtual obstacle model that matches the wall layout in the real environment. Wall position sensors typically employ magnetic or optical gratings to monitor the wall's 3D coordinates in real time. The collaborative control unit receives user spatial coordinates and pose data from the real-time positioning system and uses a Lie group transformation matrix algorithm to ensure the orthogonality between the physical and virtual coordinate systems, avoiding spatial mapping deviations.

[0154] When the system detects that the user's movement speed exceeds the preset speed, it triggers the space construction component to adjust the wall positions and the scene generation system to update the virtual scene, thereby achieving dynamic matching between virtual environment movement and real space topology transformation.

[0155] In this application, the proposed solution achieves dynamic reconstruction of physical space through modular walls and an electric sliding rail system. Combined with octree modeling and spatial mapping algorithms, it improves the efficiency of virtual scene expansion. A safety buffer mechanism reduces the risk of collisions between users and walls, and the Lie group transformation algorithm reduces coordinate transformation errors, significantly outperforming the origin offset in existing technologies. This achieves an immersive interactive experience of "limited space, unlimited scenes."

[0156] In one embodiment, when the electric sliding rail system in the space construction component drives the modular wall units to reconstruct the topology, it also performs the following steps:

[0157] The system trains a long short-term memory network model based on historical pose data to predict the movement trajectory of each user in the first time period in real time; and introduces a priority weight factor into the path planning to allocate obstacle avoidance priority according to the user's distance from the wall, movement speed, line of sight and action type.

[0158] When a potential collision risk is detected between two or more users, a group collaborative avoidance strategy is activated, and the positions of multiple walls are adjusted synchronously through a distributed PLC control network.

[0159] When a user enters the 1.2m danger zone of the wall's movement path, the electric sliding rail system switches to a low-speed, slow-motion mode and renders a warning visual effect in the virtual scene to alert the user.

[0160] Specifically, when the electric sliding rail system drives the modular wall units to reconstruct the topology, it achieves dynamic spatial adaptation in multi-user interaction scenarios through motion trajectory prediction, priority obstacle avoidance strategy and safety buffer mechanism.

[0161] The electric sliding rail system uses a Long Short-Term Memory (LSTM) network model trained on historical pose data to predict the motion trajectory of each user in the first time period in real time. The historical pose data comes from user spatial coordinates and motion posture data (including temporal information such as position, velocity, acceleration, and limb movements) collected by a real-time positioning system. The LSTM model is a recurrent neural network capable of handling dependencies in long-term sequence data, predicting future motion trajectories by learning users' historical motion patterns.

[0162] The system incorporates priority weighting factors in path planning, assigning obstacle avoidance priorities based on the user's distance from the wall, movement speed, line of sight, and action type. Specifically: distance weighting means the closer the user is to the wall, the higher the weight, and the wall must prioritize avoiding that user's path; speed weighting means faster-moving users have higher priority, requiring the system to adjust the wall's trajectory in advance to avoid high-speed collisions; line of sight weighting means when the user's gaze is focused on the wall's movement area, it's considered a high-attention scenario, increasing obstacle avoidance priority; action type weighting means emergency actions (such as quick dodges or falling tendencies) or interactive actions (such as reaching out to touch the wall) have the highest priority, triggering the wall's movement to stop abruptly or change direction. These priority weighting factors are integrated using a fuzzy logic algorithm to generate an obstacle avoidance priority ranking for wall movement, ensuring safe interaction in multi-user scenarios.

[0163] When a potential collision risk is detected between two or more users, the electric sliding rail system initiates a collaborative avoidance strategy, synchronously adjusting the positions of multiple walls through a distributed PLC control network. The specific process is as follows:

[0164] Collision risk detection: Based on multi-user pose data provided by the real-time positioning system, the GJK algorithm is used to accurately determine the collision and overlap between users and walls, and between users themselves.

[0165] Collaborative path planning: The DuelingDQN network algorithm in reinforcement learning is used to generate the optimal wall movement trajectory, minimizing spatial reconstruction delay while avoiding user collisions;

[0166] Distributed synchronous control: Control commands are transmitted to the PLC controllers of each wall unit via industrial Ethernet to achieve millisecond-level synchronous motion adjustment of multiple walls, ensuring the consistency of obstacle avoidance for a group of users.

[0167] When a user enters the 1.2m danger zone within the wall's movement path, the electric sliding rail system switches to a low-speed, gentle-motion mode and renders a warning visual effect in the virtual scene to alert the user. Specifically:

[0168] Low-speed slow movement mode: The wall movement speed is reduced to less than 30% of the normal speed, ensuring that users have enough reaction time to avoid the moving wall;

[0169] Visual warning rendering: The scene generation system displays a flashing border or color gradient effect at the corresponding wall position in the virtual scene. This visual feedback enhances the user's perception of the wall's movement and reduces the risk of collision.

[0170] The above solution avoids the interaction limitations caused by fixed walls or passive obstacle avoidance in existing technologies.

[0171] In one implementation, when a potential collision risk is detected between two or more users, a group cooperative avoidance strategy is activated, synchronously adjusting the positions of multiple walls via a distributed PLC control network, including:

[0172] The system acquires the current spatial coordinates, direction of motion, and speed information of each user, and predicts the motion trajectory in the first time period based on a long short-term memory network model.

[0173] Cross-analysis is performed on the movement trajectories of all users in pairs to calculate the shortest interval distance and the time of occurrence. If the shortest interval distance is less than a preset safety threshold and the time of occurrence is within the second time period, it is determined that there is a potential collision risk between the two.

[0174] After identifying a potential collision risk, a wall adjustment command is sent to the distributed PLC control network. The command includes the wall number to be moved, the target location, and the priority identifier. The distributed PLC control network generates multiple movement paths for the walls based on the wall adjustment command to ensure that the wall movement does not interfere with the normal interaction and walking paths of other users.

[0175] During the wall movement, the system continuously monitors changes in the user's pose and dynamically updates the wall's movement path. If the user's behavior changes and the collision risk is eliminated, the wall movement operation is canceled.

[0176] After the wall is moved, the scene generation system updates the obstacle model in the virtual scene in real time and guides the user to avoid areas with potential collision risks through visual cues.

[0177] Specifically, when a potential collision risk is detected between two or more users, the system initiates a group collaborative avoidance strategy, synchronously adjusting the positions of multiple walls through a distributed PLC control network to achieve safe space reconstruction in multi-user interaction scenarios.

[0178] The system acquires the current spatial coordinates, direction of motion, and speed information of each user through a real-time positioning system. The real-time positioning system adopts a top-mounted optical-inertial hybrid positioning scheme, using a ceiling-mounted optical marker array and a nine-axis inertial sensor component built into the user equipment to achieve sub-millimeter positioning accuracy and millisecond-level data updates.

[0179] Based on the Long Short-Term Memory (LSTM) network model, a user motion prediction model is trained using historical pose data to predict the motion trajectory in the first time period in the future.

[0180] Cross-analysis is performed on the movement trajectories of all users in pairs. The intersection points of the trajectories are determined by computational geometry algorithms (such as line segment intersection detection), and the shortest interval distance and the time of occurrence are calculated.

[0181] If the shortest distance between any two users is less than a preset safety threshold, and this shortest distance occurs within the second time period, a potential collision risk is determined to exist between them. The safety threshold and the second time period can be adjusted according to scenario requirements. For example, the safety threshold can be set to 1.2 meters (a safe distance to avoid physical contact), and the second time period can be set to 2 seconds (to allow time for obstacle avoidance reaction).

[0182] After identifying a potential collision risk, the system sends a wall adjustment command to the distributed PLC control network. The command includes the wall number to be moved, the target location, and a priority identifier, where the priority identifier is determined according to the user's collision risk level (e.g., emergency collision risk corresponds to high priority adjustment).

[0183] The distributed PLC control network generates multiple movement paths for the walls based on wall adjustment commands and real-time user location data provided by the positioning system. Path planning follows these principles: prioritizing avoidance of users' current walking paths to ensure that wall movement does not interfere with other users' normal interactions; and employing a shortest path and minimum movement strategy to reduce spatial reconfiguration time.

[0184] During the wall's movement, the system continuously monitors changes in the user's posture (updated in real-time via a real-time positioning system) and dynamically updates the wall's movement path based on the latest data. For example, when the user changes their walking direction to avoid a potential collision area, the system replans the wall's movement trajectory to avoid unnecessary wall adjustments.

[0185] If a change in user behavior (such as sudden acceleration or steering) eliminates the collision risk, the system cancels the wall movement operation and stops the current wall movement to save energy and reduce spatial layout interference.

[0186] After the wall is moved, the scene generation system collects wall position sensor data in real time, synchronizes the physical wall position to the virtual scene through a coordinate mapping algorithm, updates the obstacle model in the virtual scene, and ensures consistency between the virtual and real spaces.

[0187] The scene generation system guides users to avoid areas with potential collision risks in virtual scenes through visual cues (such as highlighting potential collision areas and pointing arrows to safe paths), thereby enhancing the safety of user interaction.

[0188] In one embodiment, the multimodal environment simulation component further includes a multimodal perception synchronization engine; the multimodal perception synchronization engine is used to coordinate the physical feedback timing between the elevator device, the wind control device, the temperature control device, the pneumatic device, the vibration device, and the sprinkler device; the multimodal environment simulation component can be used to perform the following steps:

[0189] Receive event trigger signals in the virtual scene and parse the event type and intensity parameters;

[0190] A preset response template is invoked based on the event type; the response template includes wind speed, temperature, vibration frequency, air pressure change, and water mist density.

[0191] The execution commands of each device are time-aligned using a unified timeline controller;

[0192] In the explosion simulation scenario, the vibration device and pneumatic device are first triggered to generate a shock wave effect. After 0.2 seconds, the wind control device and spray device are activated to simulate dust and debris splashing. Finally, the local temperature is raised by the temperature control device to enhance the thermal impact.

[0193] In extreme weather simulation scenarios, by combining wind speed and temperature linkage strategies, for each increase in wind force level, the body surface temperature of users in the corresponding area is reduced by 0.5℃, and the impact force of raindrops is simulated by a pneumatic device, which is combined with a spray device to achieve a water mist effect.

[0194] The intensity of the environmental simulation is reduced when abnormal user physiological indicators are detected.

[0195] Specifically, this multimodal environment simulation component coordinates the timing of physical feedback from various devices through a multimodal perception synchronization engine. Based on the virtual scene event trigger signal, it analyzes the event type and intensity parameters, calls response templates containing parameters such as wind speed and temperature, and aligns the execution time of instructions for each device through a unified time axis controller. This enables scenarios such as in an explosion simulation where vibration and pneumatic devices are triggered first to generate shock waves, followed by the activation of wind control and spray devices 0.2 seconds later to simulate dust and debris splashing, and the raising of local temperature through a temperature control device. It also provides feedback in extreme weather scenarios where wind force and temperature are linked, simulating the impact force of raindrops and the effect of water mist. Furthermore, it can downgrade the intensity of environmental simulation when the user's physiological indicators are abnormal, thereby significantly improving the timing consistency of physical feedback in the virtual scene and the sense of sensory immersion, reducing the risk of user physiological discomfort, and achieving coordinated matching of multimodal physical stimuli.

[0196] In one embodiment, the electric slide rail system is also capable of performing the following steps:

[0197] The wall planning model is trained using historical user behavior data; the wall planning model takes user density distribution, movement trends and interaction hotspots as state inputs and outputs the optimal wall layout configuration strategy.

[0198] Before each wall reconstruction task is initiated, the wall planning model predicts the walls that most likely need to be adjusted and their target locations, and generates a priority scheduling table; the movement of the walls is controlled based on the priority scheduling table.

[0199] Under the premise of satisfying obstacle avoidance and synchronous control, calculate the minimum energy consumption path of the wall from the current pose to the target pose;

[0200] After the wall is moved, the wall position sensor is used to verify whether the actual layout matches the expected configuration. If there is a deviation, the position is adjusted.

[0201] Specifically, the electric sliding rail system trains a wall planning model using historical user behavior data. It takes user density distribution, movement trends, and interaction hotspots as inputs and outputs to determine the optimal wall layout strategy. Before wall reconstruction, it predicts the walls that need to be adjusted and the target locations, and generates a priority scheduling table. When controlling the movement of the walls, it calculates the path with the minimum energy consumption. After the movement is completed, it verifies the layout deviation through sensors and makes adjustments accordingly. This achieves intelligent prediction and energy-saving control of wall layout, improves the efficiency and rationality of space reconstruction, and reduces system energy consumption.

[0202] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A multi-user virtual reality experience system, characterized in that, include: A real-time positioning system is used to capture the spatial coordinates and motion postures of multiple users' worn devices, and to achieve synchronized response between virtual characters and real actions in real time based on the motion information and location tracking information of each user; the real-time positioning system is also used to drive the virtual characters of each user to move in the virtual scene; A multimodal environment simulation component; the multimodal environment simulation component is used to simulate various physical environment changes in a virtual scene; Space base; The spatial base includes a spatial construction component and a scene generation system; the spatial construction component dynamically adjusts the spatial topology through an electric sliding rail system to reproduce the user's movement and interaction in the virtual environment in the real environment; the scene generation system generates virtual scenes in real time that match the user's position, environment, and wall layout. The space construction components include: modular wall units and an electric sliding rail system; Methods for achieving virtual-to-real-world matching using a spatial base include: The modular wall units are driven by an electric sliding rail system to control the translation and rotation of the wall in the X, Y, and Z axes. Wall splicing and separation are achieved using an electromagnetic adsorption interface; multiple walls are coordinated to move synchronously through a distributed PLC control network. The system monitors the user's location in real time. When the user enters a range of 1.2m from the wall, a safety buffer mechanism is triggered to pause or adjust the wall's movement. Methods for scene generation systems to generate virtual scenes that match the user's location in real time include: The physical space is modeled in three dimensions using an octree spatial index, and the coordinates of the physical space are scaled proportionally to the coordinates of the virtual scene using a perspective projection transformation algorithm. The dynamic rendering engine is used to perform differentiated rendering of different radius areas based on user eye tracking data; When a user moves to within 1.5m of the virtual scene boundary, the adjacent virtual scene modules are preloaded through the spatial mapping algorithm, and the spatial construction components are driven to adjust the wall layout to maintain the topological consistency between the virtual scene and the physical space. Real-time acquisition of wall position sensor data, and synchronization of physical wall position to virtual scene through coordinate mapping algorithm to generate virtual obstacle model that matches the wall layout of the real environment; Methods for achieving virtual-real environment synchronization through collaborative control units include: It receives user spatial coordinates and pose data provided by a real-time positioning system, and ensures the orthogonality between the physical coordinate system and the virtual coordinate system through a Lie group transformation matrix algorithm; When the user's movement speed is detected to exceed the preset speed, the space construction component is triggered to adjust the wall position and the scene generation system updates the virtual scene, so as to achieve dynamic matching between virtual environment movement and real space topology transformation; When the electric sliding rail system in the space construction component drives the modular wall units to reconstruct the topology, it also performs the following steps: The system trains a long short-term memory network model based on historical pose data to predict the movement trajectory of each user in the first time period in real time; and introduces a priority weight factor in path planning to allocate obstacle avoidance priority according to the user's distance from the wall, movement speed, line of sight and action type. When a potential collision risk is detected between two or more users, a group collaborative avoidance strategy is activated, and the positions of multiple walls are adjusted synchronously through a distributed PLC control network. When a user enters the 1.2m danger zone of the wall's movement path, the electric sliding rail system switches to a low-speed, slow-motion mode and renders a warning visual effect in the virtual scene to alert the user. When a potential collision risk is detected between two or more users, a group cooperative avoidance strategy is activated. This involves synchronously adjusting the positions of multiple walls via a distributed PLC control network, including: The system acquires the current spatial coordinates, direction of motion, and speed information of each user, and predicts the motion trajectory in the first time period based on a long short-term memory network model. Cross-analysis is performed on the movement trajectories of all users in pairs to calculate the shortest interval distance and the time of occurrence. If the shortest interval distance is less than a preset safety threshold and the time of occurrence is within the second time period, it is determined that there is a potential collision risk between the two. After identifying a potential collision risk, a wall adjustment command is sent to the distributed PLC control network. The command includes the wall number to be moved, the target location, and the priority identifier. The distributed PLC control network generates multiple movement paths for the walls based on the wall adjustment command to ensure that the wall movement does not interfere with the normal interaction and walking paths of other users. During the wall movement, the system continuously monitors changes in the user's pose and dynamically updates the wall's movement path. If the user's behavior changes and the collision risk is eliminated, the wall movement operation is canceled. After the wall is moved, the scene generation system updates the obstacle model in the virtual scene in real time and guides the user to avoid areas with potential collision risks through visual cues.

2. The multi-user virtual reality experience system according to claim 1, characterized in that, The process of capturing the spatial coordinates and motion postures of multiple users' wearing devices, and achieving real-time synchronization between virtual characters and real-world actions based on each user's motion information and location tracking information, includes: The user's wearable device collects angular velocity, acceleration, and magnetic field data to form a raw inertial data stream; an optical positioning data stream is generated by capturing the coordinates of optical marker points using an optical marker array in the ceiling and a top-mounted infrared camera. The original optical data stream and inertial data stream are clock-synchronized, and a unified timestamp is added to both types of data; The clock-synchronized optical positioning data and inertial data are input into the extended Kalman filter engine, and the fused pose data stream is output through dynamic weight allocation. A spatiotemporal calibration algorithm is applied to the fused pose data stream to eliminate the origin drift between the local coordinate system and the global coordinate system of each user device, maintain the orthogonality of the virtual and real space transformation matrix, and generate calibrated pose data. The system monitors the optical marker capture status in real time. When occlusion occurs, it starts the inertial navigation pre-integration model based on the calibrated pose data to predict the current pose and maintain positioning continuity. After the occlusion is removed, the predicted pose is converged to the optical coordinate system through the Lie group manifold optimization algorithm to generate anti-occlusion pose data. For multi-user interaction scenarios, real-time deviation analysis is performed on pose data after anti-occlusion processing, and a group collision-free pose data stream is generated by compensating for rotation and translation components between devices through coordinate mapping algorithm. Based on the collision-free pose data stream of the group, the virtual character is driven to achieve a mapping response to the user's real actions.

3. The multi-user virtual reality experience system according to claim 2, characterized in that, The multimodal environment simulation component includes: An elevator device; the elevator device is used to simulate vertical movement perception in a virtual scene; Risk control device; the risk control device is used to recreate climate change in a virtual scene; Temperature control device; the temperature control device is used to reproduce temperature changes in a virtual scene; Pneumatic device; the pneumatic device is used to recreate pneumatic impacts in a virtual scene; Vibration device; the vibration device is used to reproduce the physical effects of vibration in a virtual scene; A spraying device; the spraying device is used to simulate the water flow environment of a virtual scene.

4. A multi-user virtual reality experience system according to claim 3, characterized in that, The elevator system comprises a four-dimensional motion base, an intelligent drive system, a dual-domain control model, and an environmental linkage module. The four-dimensional motion base uses a combination of servo motors and harmonic reducers to achieve composite motions of vertical lifting, horizontal translation, and rotation via a six-degree-of-freedom motion platform. The intelligent drive system, based on a fuzzy PID control algorithm, generates six-dimensional spatial commands according to virtual scene motion parameters, driving the four-dimensional motion base to simulate vertical lifting, composite motion, and inertial simulation scenarios. The dual-domain control model acquires user center-of-gravity offset data in real-time through a motion capture unit in the virtual domain, inputting it to the motion calculation engine to generate virtual scene motion parameters. In the physical domain, the dual-domain control model monitors the pressure distribution at the user's contact point with the platform in real-time using a pressure sensor array, correcting the platform's posture based on pressure changes. The environmental linkage module works in conjunction with a temperature control device to simulate a 0.6°C temperature decrease for every 10m increase in the vertical lifting scenario. The environmental linkage module also works in conjunction with a wind control device to generate unidirectional airflow in high-speed motion scenarios to enhance immersion. The wind control device includes an adjustable fan array and an intelligent control module; the adjustable fan array is driven by a variable frequency motor and simulates a natural wind field through the Bernoulli effect; the intelligent control module is equipped with an LSTM neural network algorithm to analyze the climate parameters of the virtual scene in real time and generate dynamic control commands for wind speed and wind direction. The temperature control device includes: a regional gradient temperature control array, a dynamic thermal radiation adjustment system, and a multimodal collaborative interface; the regional gradient temperature control array adopts a semiconductor thermopile module matrix; the dynamic thermal radiation adjustment system is based on an inverse square decay model and adjusts the thermal radiation intensity according to the distance between the user and the virtual heat source; the multimodal collaborative interface is linked with the wind control device and the sprinkler device: each wind speed corresponds to a temperature compensation of 0.5℃; when simulating rain scenarios, it reduces the user's body surface temperature by 2-5℃; The pneumatic device includes a high-pressure injection system and a fluid dynamics simulation module; the high-pressure injection system integrates a high-pressure air pump and a multi-directional nozzle array to simulate fluid impact force; the fluid dynamics simulation module is used to calculate the injection parameters of the high-pressure injection system. The vibration device includes a distributed servo motor array and a multi-waveform synthesis engine; no fewer than 6 servo motors are deployed on the user standing platform to form a three-dimensional vibration matrix; the multi-waveform synthesis engine decomposes the vibration parameters of the virtual scene into earthquake simulation or explosion impact simulation.

5. A multi-user virtual reality experience system according to claim 4, characterized in that, The multimodal environment simulation component also includes a multimodal perception synchronization engine; the multimodal perception synchronization engine is used to coordinate the physical feedback timing between elevator devices, wind control devices, temperature control devices, pneumatic devices, vibration devices, and sprinkler devices; the multimodal environment simulation component can be used to perform the following steps: Receive event trigger signals in the virtual scene and parse the event type and intensity parameters; A preset response template is invoked based on the event type; the response template includes wind speed, temperature, vibration frequency, air pressure change, and water mist density. The execution commands of each device are time-aligned using a unified timeline controller; In the explosion simulation scenario, the vibration device and pneumatic device are first triggered to generate a shock wave effect. After 0.2 seconds, the wind control device and spray device are activated to simulate dust and debris splashing. Finally, the local temperature is raised by the temperature control device to enhance the thermal impact. In extreme weather simulation scenarios, by combining wind speed and temperature linkage strategies, for each increase in wind force level, the body surface temperature of users in the corresponding area is reduced by 0.5℃, and the impact force of raindrops is simulated by a pneumatic device, which is combined with a spray device to achieve a water mist effect. The intensity of the environmental simulation is reduced when abnormal user physiological indicators are detected.

6. A multi-user virtual reality experience system according to claim 5, characterized in that, The electric slide rail system can also perform the following steps: The wall planning model is trained using historical user behavior data; the wall planning model takes user density distribution, movement trends and interaction hotspots as state inputs and outputs the optimal wall layout configuration strategy. Before each wall reconstruction task is initiated, the wall planning model predicts the walls that most likely need to be adjusted and their target locations, and generates a priority scheduling table; the movement of the walls is controlled based on the priority scheduling table. Under the premise of satisfying obstacle avoidance and synchronous control, calculate the minimum energy consumption path of the wall from the current pose to the target pose; After the wall is moved, the wall position sensor is used to verify whether the actual layout matches the expected configuration. If there is a deviation, the position is adjusted.

Images