Interaction system and method for 5DVR large-space scene

Abstract

The invention relates to the technical field of VR (virtual reality) scene interaction, in particular to an interaction system and method for a 5DVR (digital video recorder) large-space scene, and the system comprises a nonlinear path mapping unit which is based on a space positioning technology and a path redirection mechanism and is combined with a time-varying space deformation tensor mapping algorithm and a virtual gravitational gradient to expand the virtual movement distance of the 5DVR space scene; the environment synchronous feedback unit performs multi-sensory synchronous feedback on user behaviors and scene changes; the user collaborative guidance unit is based on a user behavior trajectory modeling technology and a path prediction method, combines a behavior curvature estimation modeling method and a double-graph structure intention prediction algorithm, and synchronously predicts path conflicts and target contention of multi-user interaction in a 5DVR space scene; and the behavior environment response unit is used for receiving natural behavior input of the user and driving state evolution and scene switching in the 5DVR virtual environment. According to the interactive system and method for the 5DVR large-space scene, the problems of multi-user target contention and virtual path distortion abrupt change are solved.

Description

Technical Field

[0001] The present invention relates to the field of VR scene interaction technology, and in particular to an interaction system and method for 5DVR large space scenes. Background Art

[0002] The 5DVR spatial scene interaction system is designed to achieve immersive multi-user virtual behavior synchronization and high-freedom expansion control of virtual paths in physical space, controlling the user's movement in physical space to be naturally mapped to the continuous path in virtual space, and realizing unlimited expansion of virtual space in a restricted physical environment, multi-user collaborative interaction and integration of intelligent environment feedback.

[0003] Existing 5DVR spatial scene interaction systems usually have difficulty balancing the nonlinear expansion of users' virtual paths with the dynamic matching of natural behaviors. Due to the limited real-world large space and the frequent need for users to move long distances, navigate continuously, and traverse multiple people in parallel in VR, virtual path overflow, distortion, and multi-person interaction conflicts may occur during dynamic behaviors such as turning, avoiding obstacles, and sudden stops. Therefore, an interaction system and method for 5DVR large-space scenes are provided. Summary of the Invention

[0004] The purpose of the present invention is to provide an interactive system and method for 5DVR large-space scenes, so as to solve the problems raised in the above background technology, that is, due to the limited real large space and the frequent need of users in VR to move long distances, continuously navigate, and cross multiple people in parallel, which may cause virtual path overflow, distortion, mutation, and multi-person interaction conflicts during users' dynamic behaviors such as turning, avoiding obstacles, and sudden stops.

[0005] To achieve the above objectives, the present invention provides an interactive system for 5DVR large space scenes, comprising:

[0006] A nonlinear path mapping unit, which is based on spatial positioning technology and path redirection mechanism, combined with a time-varying spatial deformation tensor mapping algorithm and a virtual gravitational gradient, to expand the virtual movement distance and path freedom of the 5DVR space scene;

[0007] It also includes an environmental synchronization feedback unit, which is used to provide multi-sensory synchronous feedback on user behavior and scene changes;

[0008] The system also includes a user collaborative guidance unit, which is based on user behavior trajectory modeling technology and path prediction method, combined with behavior curvature estimation modeling method and dual-graph structure intention prediction algorithm, to simultaneously predict path conflicts and target contention of multi-user interactions in 5DVR space scenes;

[0009] It also includes a behavior environment response unit, which is used to receive the user's natural behavior input, drive the state evolution and scene switching in the 5DVR virtual environment, and feed back to the environment synchronization feedback unit.

[0010] As a further improvement of the present technical solution, the nonlinear path mapping unit includes a path mapping tensor control module and a target driving gravity guidance module;

[0011] The path mapping tensor control module generates a nonlinear space deformation tensor field based on the user's motion trajectory and space boundary conditions, which is used to deform the physical space path into a longer virtual path;

[0012] The target-driven gravity guidance module combines the obstacles and virtual target points of the 5DVR space scene to construct a virtual gravity gradient, and integrates the virtual gravity gradient into the nonlinear space deformation tensor field to obtain virtual path control.

[0013] As a further improvement to this technical solution, the nonlinear space deformation tensor field is based on spatial positioning technology and path redirection mechanism, and is used to perform nonlinear deformation mapping on the physical space path to expand the range of the physical space path. The dynamic change rate of the physical space path is combined with the dynamic change rate of the virtual space path to obtain a deformation tensor, and based on the deformation tensor, the user's position vector in the physical space and virtual space is combined to obtain a nonlinear space deformation tensor field, as follows:

[0014] ;

[0015] ;

[0016] in, For time; is the time increment; is the deformation tensor; For time The dynamic change rate of the virtual space path; For time The dynamic change rate of the physical space path; For time The inverse rate of change of the physical space path; is the nonlinear space deformation tensor field; For time The user's position vector in the virtual space; For time The user's position vector in physical space; For time The user's position vector in physical space.

[0017] As a further improvement of this technical solution, the target-driven gravity guidance module combines the obstacles and virtual target points of the 5DVR space scene to construct a virtual gravity gradient, and integrates the virtual gravity gradient into the nonlinear space deformation tensor field to obtain the original virtual path control, as follows:

[0018] ;

[0019] ;

[0020] in, is the virtual gravitational gradient; The attractiveness of the target; is the target attractiveness weight; is the magnitude of the obstacle repulsion force; is the obstacle repulsion weight; is the virtual target point; Index of obstacles; For the obstacles; For time The original virtual path control.

[0021] As a further improvement of this technical solution, the environment synchronization feedback unit uses real-time graphics rendering technology, spatial audio rendering technology and tactile feedback technology to synchronously feedback the visual and tactile errors and sound direction of the original virtual path control. The environment synchronization feedback unit includes a behavioral visual feedback module, a behavioral tactile feedback module and a behavioral sound feedback module;

[0022] The behavior visual feedback module obtains the user's visual error through the visual sensor in the VR device, and the behavior visual feedback module feeds back the user's visual error under the original virtual path control to the behavior environment response unit;

[0023] The behavioral tactile feedback module acquires the user's tactile error through the tactile sensor in the VR device, and the behavioral tactile feedback module feeds back the user's tactile error under the original virtual path control to the behavioral environment response unit;

[0024] The behavior sound feedback module obtains the user's voice direction through the audio sensor in the VR device, and the behavior sound feedback module feeds back the user's voice direction under the original virtual path control to the behavior environment response unit.

[0025] As a further improvement of this technical solution, the user collaborative guidance unit includes a behavior trend modeling module and an intention conflict scheduling module;

[0026] The behavior trend modeling module predicts the user's behavior intention based on the user's behavior trajectory and spatial position in the 5DVR space scene through a behavior curvature estimation modeling method;

[0027] The intention conflict scheduling module uses a dual-graph structure intention prediction algorithm and combines the behavioral intention of each user to determine whether the behaviors of multiple users in the 5DVR space scene will cause path conflicts and target contention.

[0028] As a further improvement of this technical solution, the behavioral curvature estimation modeling method is to construct a user intention prediction feature vector based on the user's behavioral curvature and acceleration changes to predict the user's behavioral intention. The behavioral curvature estimation modeling method is specifically as follows:

[0029] S3.1.1. Set the user's physical space location to ;

[0030] S3.1.2. Calculate the curvature of user behavior ;

[0031] S3.1.3. Time based on the user's physical location Derivatively obtain the user's acceleration change ;

[0032] S3.1.4. Combining the curvature of user behavior and the acceleration change, we can obtain the user intention prediction feature vector:

[0033] ;

[0034] in, Indexing users; For the Users at time Intent prediction feature vector; For the Users at time The user's orientation angle.

[0035] As a further improvement to this technical solution, the dual-graph structure intention prediction algorithm is based on the fusion of the user's spatial adjacency graph and behavioral intention graph, and uses a graph neural network to iteratively propagate the user's behavioral intention and spatial position. It predicts whether the behavior of multiple users in the 5DVR spatial scene will cause path conflict and target contention. The dual-graph structure intention prediction algorithm is specifically as follows:

[0036] S3.2.1. Define the user's spatial adjacency graph as , the adjacency matrix of its spatial adjacency graph is ;

[0037] S3.2.2, define the user's behavioral intention diagram as , the adjacency matrix of its behavior intention graph is ;

[0038] S3.2.3. Based on the adjacency matrix of the spatial adjacency graph and the behavioral intention graph, the graph neural network is used to iteratively propagate the user's behavioral intention and spatial position to obtain the graph neural network output , which is the Users at time prediction of future intentions;

[0039] S3.2.4. Based on the prediction of each user's future intention, determine in real time whether user interactions will conflict and trigger the user's original virtual path control scheduling through the conflict weight matrix.

[0040] As a further improvement to this technical solution, in S3.2.4, the conflict weight matrix is ​​obtained based on the interaction priority between the user and the target object. It is used to trigger the user's original virtual path control scheduling when the target object is approached and controlled by multiple users' future intention predictions. The conflict weight matrix is ​​calculated as follows:

[0041] ;

[0042] in, For the User and The conflict weight matrix of the target objects; For the Users at time prediction of future intentions; For the target object at time The feature vector representation of is the total number of users; For all users and The sum of future intention predictions of target objects;

[0043] If conflicts occur between multiple users, the original virtual path control scheduling of the user is triggered through the conflict weight matrix and the original virtual path control, as follows:

[0044]

[0045] in, For the Users at time Original virtual path control; For the After the interaction conflict occurs, the user Final virtual path control; is generated according to the conflict weight matrix A user dispatches a guiding force field.

[0046] On the other hand, the present invention provides an interactive method for a 5DVR large space scene, which is used in any of the above-mentioned interactive systems for a 5DVR large space scene, comprising the following steps:

[0047] S10.1. Based on spatial positioning technology and path redirection mechanism, combined with time-varying spatial deformation tensor mapping algorithm and virtual gravity gradient, the virtual movement distance and path freedom of 5DVR space scenes are expanded;

[0048] S10.2. Provide multi-sensory synchronous feedback on user behavior and scene changes;

[0049] S10.3. Based on user behavior trajectory modeling technology and path prediction methods, combined with behavioral curvature estimation modeling methods and dual-graph structure intention prediction algorithms, path conflicts and target contention of multi-user interactions in 5DVR spatial scenarios are simultaneously predicted;

[0050] S10.4. Receive the user's natural behavior input and drive the state evolution and scene switching in the 5DVR virtual environment.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. In an interactive system and method for 5DVR large-space scenes, a time-varying nonlinear space deformation tensor field constructed based on the dynamic change rate of the physical space path and the change rate of the virtual space path can achieve continuous nonlinear mapping of the user's motion trajectory in the real physical space, thereby expanding a wide-area virtual path with directional continuity and natural behavior within a limited physical space.

[0053] 2. An interactive system and method for 5DVR large-space scenes, by constructing a dual-graph structure that integrates the spatial adjacency graph and the behavioral intention graph, and combining it with a graph neural network for multi-user intention prediction and conflict scheduling, realizes dynamic priority management and path conflict avoidance of the interactive behaviors of multiple users in the same 5DVR virtual target area. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is the overall flow chart of the present invention;

[0055] The meaning of each number in the figure is:

[0056] 1. Nonlinear path mapping unit; 11. Path mapping tensor control module; 12. Target-driven gravity guidance module; 2. Environmental synchronization feedback unit; 21. Behavioral visual feedback module; 22. Behavioral tactile feedback module; 23. Behavioral sound feedback module; 3. User collaborative guidance unit; 31. Behavioral trend modeling module; 32. Intention conflict scheduling module; 4. Behavioral environment response unit. DETAILED DESCRIPTION

[0057] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0058] Example 1:

[0059] See also Figure 1 As shown, an interactive system for 5DVR large space scenes is provided, including:

[0060] A nonlinear path mapping unit 1, which is based on spatial positioning technology and path redirection mechanism, combined with a time-varying spatial deformation tensor mapping algorithm and a virtual gravitational gradient, to expand the virtual movement distance and path freedom of the 5DVR space scene;

[0061] The nonlinear path mapping unit 1 includes a path mapping tensor control module 11 and a target driving gravity guidance module 12;

[0062] The path mapping tensor control module 11 generates a nonlinear space deformation tensor field according to the user's motion trajectory and space boundary conditions, so as to transform the physical space path into a longer virtual path;

[0063] The target-driven gravity guidance module 12 combines the obstacles and virtual target points of the 5DVR space scene to construct a virtual gravity gradient, and integrates the virtual gravity gradient into the nonlinear space deformation tensor field to obtain virtual path control.

[0064] The nonlinear space deformation tensor field is based on spatial positioning technology and path redirection mechanism, and is used to perform nonlinear deformation mapping on the physical space path to expand the range of the physical space path. The deformation tensor is obtained by combining the dynamic change rate of the physical space path with the dynamic change rate of the virtual space path. Based on the deformation tensor, the user's position vector in the physical space and the virtual space is combined to obtain the nonlinear space deformation tensor field, as follows:

[0065] ;

[0066] ;

[0067] in, For time; is the time increment; is the deformation tensor; For time The dynamic change rate of the virtual space path; For time The dynamic change rate of the physical space path; For time The inverse rate of change of the physical space path; is the nonlinear space deformation tensor field; For time The user's position vector in the virtual space; For time The user's position vector in physical space; For time The user's position vector in physical space.

[0068] In this embodiment, time The user's position vector in physical space : ;time The user's position vector in the virtual space: ;Deformation tensor It is a dynamic spatial deformation factor matrix that nonlinearly expands, bends, or scales small changes in the real path into the virtual path. By constructing a nonlinear spatial deformation tensor field, the user's path movement in the real physical space is transformed into a mapped path in the virtual space, thereby achieving infinite expansion of the virtual movement range while maintaining the real space unchanged, and ensuring the continuity of behavioral logic, path curvature, and interaction direction.

[0069] The target-driven gravity guidance module 12 combines the obstacles and virtual target points of the 5DVR space scene to construct a virtual gravity gradient, and integrates the virtual gravity gradient into the nonlinear space deformation tensor field to obtain the original virtual path control, as follows:

[0070] ;

[0071] ;

[0072] in, is the virtual gravitational gradient; The attractiveness of the target; is the target attractiveness weight; is the magnitude of the obstacle repulsion force; is the obstacle repulsion weight; is the virtual target point; Index of obstacles; For the obstacles; For time The original virtual path control.

[0073] In this embodiment, by combining the target point position and obstacle distribution in the 5DVR virtual scene, a continuous and differentiable virtual gravitational gradient is constructed to guide the user to move towards the target and automatically avoid obstacles in the virtual environment. As a navigation guide, it is superimposed on the virtual path generated by the previous layer of nonlinear space deformation tensor field, realizing a dynamic virtual path adjustment mechanism with environmental perception capabilities. The following potential function is defined in the virtual space: ; Calculate the gradient of the potential function to obtain the virtual gravitational gradient The virtual gravitational gradient is a directional vector field that points to the direction the user should go at every moment, and adjusts its strength and direction based on the target proximity and obstacle distribution. After adding the virtual gravitational gradient to the intermediate virtual path points provided by the nonlinear space deformation tensor field generated by the previous module, the final original virtual path control vector is generated.

[0074] It also includes an environment synchronization feedback unit 2, which is used to provide multi-sensory synchronization feedback on user behavior and scene changes;

[0075] The environment synchronization feedback unit 2 uses real-time graphics rendering technology, spatial audio rendering technology and tactile feedback technology to synchronously feedback the visual and tactile errors and sound direction of the original virtual path control;

[0076] The environment synchronization feedback unit 2 includes a behavior visual feedback module 21, a behavior tactile feedback module 22 and a behavior sound feedback module 23;

[0077] The behavior visual feedback module 21 obtains the user's visual error through the visual sensor in the VR device, and the behavior visual feedback module 21 feeds back the user's visual error under the original virtual path control to the behavior environment response unit 4;

[0078] The behavioral tactile feedback module 22 acquires the user's tactile error through the tactile sensor in the VR device, and the behavioral tactile feedback module 22 feeds back the user's tactile error under the original virtual path control to the behavioral environment response unit 4;

[0079] The behavior sound feedback module 23 obtains the user's voice direction through the audio sensor in the VR device, and the behavior sound feedback module 23 feeds back the user's voice direction under the original virtual path control to the behavior environment response unit 4;

[0080] In this embodiment, the user's visual error, tactile error, and sound direction under the original virtual path control are inaccurate and have errors with human perception;

[0081] The user's visual error is that under the control of the original virtual path, there is an error in the distance between the virtual objects in the virtual space scene and the user under the control of the original virtual path;

[0082] The user's tactile error is an error in which the virtual object in the virtual space scene does not come into contact with the user when the user touches the object in the actual virtual space scene under the control of the original virtual path;

[0083] The sound direction of the user is different from the direction of the virtual audio source in the virtual space scene and the sound perception position between the user under the control of the original virtual path.

[0084] The system further includes a user collaborative guidance unit 3, which is based on user behavior trajectory modeling technology and path prediction method, combined with behavior curvature estimation modeling method and dual-graph structure intention prediction algorithm, to simultaneously predict path conflicts and target contention of multi-user interactions in 5DVR space scenes;

[0085] The user collaborative guidance unit 3 includes a behavior trend modeling module 31 and an intention conflict scheduling module 32;

[0086] The behavior trend modeling module 31 predicts the user's behavior intention based on the user's behavior trajectory and spatial position in the 5DVR space scene through a behavior curvature estimation modeling method;

[0087] The intention conflict scheduling module 32 uses a dual-graph structure intention prediction algorithm and combines the behavioral intention of each user to determine whether the behaviors of multiple users in the 5DVR space scene will cause path conflicts and target contention.

[0088] The behavior curvature estimation modeling method is to construct a user intention prediction feature vector based on the user's behavior curvature and acceleration changes to predict the user's behavior intention. The behavior curvature estimation modeling method is specifically as follows:

[0089] S3.1.1. Set the user's physical space location to ;

[0090] ;

[0091] in, Indexing users; For the Users at time spatial location; For the Users at time The horizontal x-axis position; For the Users at time The horizontal y-axis position; For the Users at time The position perpendicular to the horizontal direction;

[0092] S3.1.2. Calculate the curvature of user behavior ;

[0093] ;

[0094] in, For the Users at time The curvature of behavior; For the Users at time The first derivative of the behavioral curvature; For the Users at time the second derivative of the behavioral curvature;

[0095] S3.1.3. Time based on the user's physical location Derivatively obtain the user's acceleration change ;

[0096]

[0097] in, For the Users at time The acceleration change;

[0098] S3.1.4. Combining the curvature of user behavior and the acceleration change, we can obtain the user intention prediction feature vector:

[0099] ;

[0100] in, Indexing users; For the Users at time Intent prediction feature vector; For the Users at time The user's orientation angle.

[0101] The dual-graph structure intention prediction algorithm is based on the fusion of the user's spatial adjacency graph and the behavioral intention graph, and uses a graph neural network to iteratively propagate the user's behavioral intention and spatial position. It predicts whether the behavior of multiple users in the 5DVR spatial scene will cause path conflicts and target contention. The dual-graph structure intention prediction algorithm is specifically as follows:

[0102] S3.2.1. Define the user's spatial adjacency graph as , the adjacency matrix of its spatial adjacency graph is ;

[0103] in, The elements in , Indicates the User and Users are adjacent; Indicates the User and users are not adjacent; Indexing users;

[0104] S3.2.2, define the user's behavioral intention diagram as , the adjacency matrix of its behavior intention graph is ;

[0105] in, The elements in , ;in and All are user indexes; For the Users at time Intent prediction feature vector; For the Users at time Intent prediction feature vector; is similarity calculation;

[0106] S3.2.3. Based on the adjacency matrix of the spatial adjacency graph and the behavioral intention graph, the graph neural network is used to iteratively propagate the user's behavioral intention and spatial position to obtain the graph neural network output , which is the Users at time prediction of future intentions;

[0107] The iterative propagation formula of the graph neural network is as follows:

[0108] ;

[0109] in, For the User in Feature representation of the layer; is the activation function; For the User in Feature representation of the layer; is the network weight matrix of the spatial position; is the network weight matrix of behavioral intention;

[0110] S3.2.4. Based on the prediction of each user's future intention, determine in real time whether user interactions will conflict and trigger the user's original virtual path control scheduling through the conflict weight matrix.

[0111] In S3.2.4, the conflict weight matrix is ​​obtained based on the interaction priority between the user and the target object. It is used to trigger the user's original virtual path control scheduling when the target object is approached and controlled by multiple users' future intention predictions. The conflict weight matrix is ​​calculated as follows:

[0112] ;

[0113] in, For the User and The conflict weight matrix of the target objects; For the Users at time prediction of future intentions; For the target object at time The feature vector representation of is the total number of users; For all users and The sum of future intention predictions of target objects;

[0114] If conflicts occur between multiple users, the original virtual path control scheduling of the user is triggered through the conflict weight matrix and the original virtual path control, as follows:

[0115]

[0116] in, For the Users at time Original virtual path control; For the After the interaction conflict occurs, the user Final virtual path control; is generated according to the conflict weight matrix A user dispatches a guiding force field.

[0117] In this embodiment, the specific method for the conflict between multiple users is as follows: for a certain time, if the conflict weight matrix between the user and the target object is higher than the threshold, the system determines that the user has a conflict interaction tendency with others and needs to perform original virtual path control scheduling on the user;

[0118] In this embodiment, the first User scheduling guidance field , the specific method is as follows: According to User and The conflict weight matrix of the target object , the guiding force field should be positively or negatively correlated with its value, depending on whether the user is the priority controller:

[0119] if If it is the largest compared to other users, the user is guided closer to the target;

[0120] if If it is the smallest compared to other users, the user is guided to avoid the target;

[0121] Construct the following two scheduling methods:

[0122] Scheduling method 1. users The maximum has control priority and generates the target attraction field:

[0123] ;

[0124] in, is the target attraction field; is the target attraction field weight;

[0125] Scheduling method 2. users Non-maximum does not have control priority, and a virtual repulsive guiding field is introduced:

[0126] ;

[0127] in, It is the virtual repulsive guiding field; is the weight of the virtual repulsive force guide field;

[0128] According to Whether a user is a priority controller or not, different guidance mechanisms are selected, and the final path guidance force field is:

[0129] ;

[0130] in, is generated according to the conflict weight matrix A user dispatches a guiding force field.

[0131] It also includes a behavior environment response unit 4, which is used to receive the user's natural behavior input, drive the state evolution and scene switching in the 5DVR virtual environment, and feed back to the environment synchronization feedback unit 2.

[0132] Example 2:

[0133] An interactive method for a 5DVR large space scene, used in any of the above interactive systems for a 5DVR large space scene, comprises the following steps:

[0134] S10.1. Based on spatial positioning technology and path redirection mechanism, combined with time-varying spatial deformation tensor mapping algorithm and virtual gravity gradient, the virtual movement distance and path freedom of 5DVR space scenes are expanded;

[0135] S10.2. Provide multi-sensory synchronous feedback on user behavior and scene changes;

[0136] S10.3. Based on user behavior trajectory modeling technology and path prediction methods, combined with behavioral curvature estimation modeling methods and dual-graph structure intention prediction algorithms, path conflicts and target contention of multi-user interactions in 5DVR spatial scenarios are simultaneously predicted;

[0137] S10.4. Receive the user's natural behavior input and drive the state evolution and scene switching in the 5DVR virtual environment.

[0138] The basic principles, main features, and advantages of the present invention are shown and described above. It should be understood by those skilled in the art that the present invention is not limited to the above-described embodiments. The above-described embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention, and such changes and modifications fall within the scope of the invention claimed.

Claims

1. An interactive system for 5DVR large space scenes, characterized by: include: A nonlinear path mapping unit (1), wherein the nonlinear path mapping unit (1) is based on a spatial positioning technology and a path redirection mechanism, combined with a time-varying spatial deformation tensor mapping algorithm and a virtual gravity gradient, to expand the virtual moving distance and path freedom of a 5DVR space scene; An environmental synchronization feedback unit (2), the environmental synchronization feedback unit (2) is used to provide multi-sensory synchronization feedback on user behavior and scene changes; A user collaborative guidance unit (3), wherein the user collaborative guidance unit (3) is based on user behavior trajectory modeling technology and path prediction method, combined with behavior curvature estimation modeling method and dual-graph structure intention prediction algorithm, to synchronously predict path conflicts and target contention of multi-user interactions in a 5DVR space scene; A behavior environment response unit (4) is used to receive the user's natural behavior input, drive the state evolution and scene switching in the 5DVR virtual environment, and feed back to the environment synchronization feedback unit (2).

2. The interactive system for 5DVR large space scenes according to claim 1, characterized in that: The nonlinear path mapping unit (1) comprises a path mapping tensor control module (11) and a target driving gravity guidance module (12); The path mapping tensor control module (11) generates a nonlinear space deformation tensor field according to the user's motion trajectory and space boundary conditions, and is used to map the physical space path deformation into a longer virtual path; The target-driven gravity guidance module (12) combines the obstacles and virtual target points of the 5DVR space scene to construct a virtual gravity gradient, and integrates the virtual gravity gradient into the nonlinear space deformation tensor field to obtain virtual path control.

3. The interactive system for 5DVR large space scenes according to claim 2, characterized in that: The nonlinear space deformation tensor field is based on spatial positioning technology and path redirection mechanism, and is used to perform nonlinear deformation mapping on the physical space path to expand the range of the physical space path. The deformation tensor is obtained by combining the dynamic change rate of the physical space path with the dynamic change rate of the virtual space path. Based on the deformation tensor, the user's position vector in the physical space and the virtual space is combined to obtain the nonlinear space deformation tensor field, as follows: ； ； in, For time; is the time increment; is the deformation tensor; For time The dynamic change rate of the virtual space path; For time The dynamic change rate of the physical space path; For time The inverse rate of change of the physical space path; is the nonlinear space deformation tensor field; For time The user's position vector in the virtual space; For time The user's position vector in physical space; For time The user's position vector in physical space.

4. The interactive system for 5DVR large space scenes according to claim 3 is characterized by: The target-driven gravity guidance module (12) combines the obstacles and virtual target points of the 5DVR space scene to construct a virtual gravity gradient, and integrates the virtual gravity gradient into the nonlinear space deformation tensor field to obtain the original virtual path control, as follows: ； ； in, is the virtual gravitational gradient; The attractiveness of the target; is the target attractiveness weight; is the magnitude of the obstacle repulsion force; is the obstacle repulsion weight; is the virtual target point; Index of obstacles; For the obstacles; For time The original virtual path control.

5. The interactive system for 5DVR large space scenes according to claim 4 is characterized in that: The environment synchronization feedback unit (2) uses real-time graphics rendering technology, spatial audio rendering technology and tactile feedback technology to synchronously feedback the visual and tactile errors and sound direction of the original virtual path control. The environment synchronization feedback unit (2) includes a behavioral visual feedback module (21), a behavioral tactile feedback module (22) and a behavioral sound feedback module (23); The behavior visual feedback module (21) acquires the user's visual error through the visual sensor in the VR device, and the behavior visual feedback module (21) feeds back the user's visual error under the original virtual path control to the behavior environment response unit (4); The behavioral tactile feedback module (22) acquires the user's tactile error through the tactile sensor in the VR device, and the behavioral tactile feedback module (22) feeds back the user's tactile error under the original virtual path control to the behavioral environment response unit (4); The behavior sound feedback module (23) obtains the user's voice direction through the audio sensor in the VR device, and the behavior sound feedback module (23) feeds back the user's voice direction under the original virtual path control to the behavior environment response unit (4).

6. The interactive system for 5DVR large space scenes according to claim 5, characterized in that: The user collaborative guidance unit (3) includes a behavior trend modeling module (31) and an intention conflict scheduling module (32); Wherein, the behavior trend modeling module (31) predicts the user's behavior intention through a behavior curvature estimation modeling method based on the user's behavior trajectory and spatial position in the 5DVR space scene; The intention conflict scheduling module (32) uses a dual-graph structure intention prediction algorithm and combines the behavioral intention of each user to determine whether the behaviors of multiple users in the 5DVR space scene will cause path conflicts and target contention.

7. The interactive system for 5DVR large space scenes according to claim 6, characterized in that: The behavior curvature estimation modeling method is to construct a user intention prediction feature vector based on the user's behavior curvature and acceleration changes to predict the user's behavior intention. The behavior curvature estimation modeling method is specifically as follows: S3.1.

1. Set the user's physical space location to ; S3.1.

2. Calculate the curvature of user behavior ; S3.1.

3. Time based on the user's physical location Derivatively obtain the user's acceleration change ; S3.1.

4. Combining the curvature of user behavior and the acceleration change, we can obtain the user intention prediction feature vector: ； in, Indexing users; For the Users at time Intent prediction feature vector; For the Users at time The user's orientation angle.

8. The interactive system for 5DVR large space scenes according to claim 7, characterized in that: The dual-graph structure intention prediction algorithm is based on the fusion of the user's spatial adjacency graph and the behavioral intention graph, and uses a graph neural network to iteratively propagate the user's behavioral intention and spatial position. It predicts whether the behavior of multiple users in the 5DVR spatial scene will cause path conflicts and target contention. The dual-graph structure intention prediction algorithm is specifically as follows: S3.2.

1. Define the user's spatial adjacency graph as , the adjacency matrix of its spatial adjacency graph is ; S3.2.2, define the user's behavioral intention diagram as , the adjacency matrix of its behavior intention graph is ; S3.2.

3. Based on the adjacency matrix of the spatial adjacency graph and the behavioral intention graph, the graph neural network is used to iteratively propagate the user's behavioral intention and spatial position to obtain the graph neural network output , which is the Users at time prediction of future intentions; S3.2.

4. Based on the prediction of each user's future intention, determine in real time whether user interactions will conflict and trigger the user's original virtual path control scheduling through the conflict weight matrix.

9. The interactive system for 5DVR large space scenes according to claim 8, characterized in that: In S3.2.4, the conflict weight matrix is ​​obtained based on the interaction priority between the user and the target object. It is used to trigger the user's original virtual path control scheduling when the target object is approached and controlled by multiple users' future intention predictions. The conflict weight matrix is ​​calculated as follows: ； in, For the User and The conflict weight matrix of the target objects; For the Users at time prediction of future intentions; For the target object at time The feature vector representation of is the total number of users; For all users and The sum of future intention predictions of target objects; If conflicts occur between multiple users, the original virtual path control scheduling of the user is triggered through the conflict weight matrix and the original virtual path control, as follows: ； in, For the Users at time Original virtual path control; For the After the interaction conflict occurs, the user Final virtual path control; is generated according to the conflict weight matrix A user dispatches a guiding force field.

10. An interactive method for 5DVR large space scenes, used in an interactive system for 5DVR large space scenes as claimed in any one of claims 1 to 9, characterized in that: The steps include: S10.

1. Based on spatial positioning technology and path redirection mechanism, combined with time-varying spatial deformation tensor mapping algorithm and virtual gravity gradient, the virtual movement distance and path freedom of 5DVR space scenes are expanded; S10.

2. Provide multi-sensory synchronous feedback on user behavior and scene changes; S10.

3. Based on user behavior trajectory modeling technology and path prediction methods, combined with behavioral curvature estimation modeling methods and dual-graph structure intention prediction algorithms, path conflicts and target contention of multi-user interactions in 5DVR spatial scenarios are simultaneously predicted; S10.

4. Receive the user's natural behavior input and drive the state evolution and scene switching in the 5DVR virtual environment.

Images