Virtual shopping mall shopping navigation system based on VR technology

Abstract

The invention discloses a virtual shopping mall shopping navigation system based on a VR (virtual reality) technology, and particularly relates to the technical field of virtual shopping mall shopping navigation.A collision detection failure coefficient and a state synchronization delay coefficient in the virtual shopping mall shopping navigation system are sensed in real time, and a path segment level mold-penetrating hidden danger heat map is constructed based on the real-time sensing; the method comprises the following steps: identifying and quantifying potential mode-penetrating risks of each section on a path, and carrying out iterative correction on an initial path by taking risk minimization as a target, thereby not only effectively avoiding a cross mode-penetrating area, but also giving consideration to balance of path length and navigation efficiency, and realizing full-process closed-loop control from path generation and risk perception to iterative optimization and visual presentation. Therefore, the reliability and immersion of navigation can be remarkably improved in a multi-user and high-concurrency complex VR environment.

Description

Technical Field

[0001] The present invention relates to the technical field of virtual mall shopping navigation, and more specifically, to a virtual mall shopping navigation system based on VR technology. Background Art

[0002] With the rapid development of virtual reality (VR) technology, virtual malls, as a key vehicle for immersive interactive experiences, have been widely adopted in various commercial scenarios, including retail, trade shows, and shopping guides. In these systems, users typically use VR devices to enter a highly simulated three-dimensional virtual mall space, interacting with the environment and others in real time as avatars. To enhance users' sense of spatial realism and social immersion, virtual mall systems generally employ multi-person synchronization mechanisms, enabling multiple users' avatars to collaborate and interact within the same virtual space.

[0003] However, in high-concurrency or high-load scenarios, VR systems typically employ a local-first optimization strategy to balance computing performance and bandwidth resources. The local user's avatar is treated as the primary view reference object, and its position and motion information is rendered with high precision and updated frequently. Meanwhile, the avatars of other remote users are typically "downgraded" by the system, for example by reducing the motion refresh rate, simplifying skeletal pose calculations, or using interpolation to compensate for latency. While this optimization mechanism can alleviate network latency and rendering pressure to a certain extent, due to inconsistencies in state synchronization latency and physical detection in a distributed environment, different users' local systems experience temporal lags and spatial errors in their perception of the states of other users' avatars. This leads to failed collision detection and delayed state synchronization between avatars, resulting in "cross-over" and overlapping virtual avatars. This phenomenon severely disrupts the user's perception of spatial continuity and physical realism, reducing the overall immersive quality of the system, especially in key application scenarios such as path navigation, close-range interaction, and product recommendations. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a virtual shopping mall navigation system based on VR technology to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: The virtual shopping mall navigation system based on VR technology includes a user initialization module, a navigation path planning module, a penetration hazard perception module, a path dynamic optimization module, and a navigation visualization module; The user initialization module is used to build a virtual shopping mall space based on VR 3D modeling technology, load the virtual avatar of the current online user in real time, and initialize the virtual avatar's spatial position, posture data, and motion data; The navigation path planning module is used to respond to the user's product purchase instructions and call the path search algorithm to generate an initial navigation path from the current user location to the target product area; A penetration hazard perception module is used to obtain collision detection failure information of the user avatar and state synchronization delay information between the remote avatar and the local view during the execution of the navigation path. The collision detection failure information includes a collision detection failure coefficient, and the state synchronization delay information includes a state synchronization delay coefficient. The path dynamic optimization module is used to construct a mold penetration risk heat map based on the collision detection failure coefficient and state synchronization delay coefficient, and iteratively correct the initial navigation path; The navigation visualization module is used to present the final corrected navigation path to the user in a multimodal manner.

[0006] In a preferred embodiment, the logic for obtaining the collision detection failure coefficient is as follows: For each user avatar i, at each timestamp t, record: Local observation state vector ,in is the position vector of user avatar i at time t, is the velocity vector of user avatar i at time t, is the posture vector of user avatar i at time t; System predicted state vector ,in The time advance predicted for the system; Calculate the local prediction bias vector ; Constructing a global perceptual difference map ,in is a set of user avatar nodes, each node is a user avatar, is the edge set; For any user avatar node pair , calculate the difference between the state of i from the perspective of j: ,in is the state difference, The predicted state vector of the system for user i from user j’s local perspective; Calculate the similarity weight of the user's avatar node edge: ,in is the similarity weight of the user's avatar node edge; Calculate node inconsistency entropy: ,in is the node inconsistency entropy, is the neighbor node set of user avatar node i; Calculate the collision detection failure coefficient: ,in is the collision detection failure coefficient, is the set of users who have spatial interactions with the current path, is the average node inconsistency entropy of the user's avatar node g within the current path range, ,in is the node inconsistency entropy of the user avatar node g, , is the time that the user's avatar node g stays within the current path range, is the minimum spatial distance from the user's virtual avatar node i to the midpoint of the current path, is a very small constant that prevents division by zero.

[0007] In a preferred embodiment, the logic for obtaining the state synchronization delay coefficient is as follows: For any user avatar u, at system timestamp t, the cross-view residual tensor is calculated based on the received synchronization state packets of M remote user perspectives: ,in is the cross-view residual tensor of the user avatar u, is the state synchronization error vector of the remote user virtual avatar v to the local user virtual avatar u, is the local observation state vector of the user avatar u, For the user's virtual avatar v perspective, delay Then the system predicts the state vector of the user's virtual avatar u; Get the time decay weight of the user synchronization state in the preset time window T: ,in is the time decay weight, ; Time weighting is performed on the states within the window: ,in is the time-weighted cross-view residual tensor coefficient; Environmental factors are introduced to model the interference of synchronization state. The environmental factors include the number of other users within the unit space range in the current time window. The environmental coupling factor is defined as: ,in is the environmental coupling factor of the user avatar u, is the number of other users within the unit space range in the current time window; Calculate the state synchronization delay coefficient: ,in is the state synchronization delay coefficient.

[0008] In a preferred embodiment, a mold penetration risk heat map is constructed based on the collision detection failure coefficient and the state synchronization delay coefficient, as follows: Set the initial navigation path Divide into K equal path segments: ,in is the kth path segment, k={1,2,...,K}, K is the total number of path segments; For each path segment , get the path segment A collection of all online users' avatars within a meter range : ,in Create a virtual avatar for the user With path segments The shortest distance; For each path segment , construct path segments based on the collision detection failure coefficient and state synchronization delay coefficient of the user virtual avatar in the interaction area The risk of penetration: ,in For path segments The risk value of penetration, Create a virtual avatar for the user The collision detection failure coefficient, Create a virtual avatar for the user The state synchronization delay coefficient, represent the preset proportional coefficients of the collision detection failure coefficient and the state synchronization delay coefficient, respectively, and are greater than 0, is a very small constant to prevent division by zero; Construct a mold penetration risk heat map based on the mold penetration risk value of each path : .

[0009] In a preferred embodiment, the mold penetration risk value in the mold penetration risk heat map is compared with a preset mold penetration risk value threshold, and the initial navigation path is iteratively corrected as follows: If the penetration risk value is greater than the penetration risk value threshold, an iterative correction signal is generated, and the path search algorithm is called to regenerate the navigation path from the current user location to the target product area. A minimization cost function is defined and used as an iterative condition to iteratively correct the initial navigation path to minimize the cost function: ,in is the path length of the navigation path, 、 Respectively represent the preset proportional coefficients of path length and penetration risk value, and 、 All greater than 0; If the mold penetration risk value is less than or equal to the mold penetration risk value threshold, there is no need to generate an iterative correction signal.

[0010] Technical effects and advantages of the present invention: 1. The present invention uses real-time perception of the collision detection failure coefficient and state synchronization delay coefficient in the virtual shopping mall navigation system to refine the potential cross-cutting problems caused by inconsistent states between avatars. Based on this, a heat map of cross-cutting risks is constructed at the path segment level. The potential cross-cutting risks of each segment on the path are identified and quantified. The initial path is iteratively corrected with the goal of minimizing risk. This not only effectively avoids cross-cutting areas but also strikes a balance between path length and navigation efficiency, achieving closed-loop control of the entire process from path generation and risk perception to iterative optimization and visualization. Therefore, it can significantly improve the reliability and immersion of navigation in complex multi-user, high-concurrency VR environments. Specifically, with the help of the collision detection failure coefficient and the state synchronization delay coefficient, the system accurately identifies potential high-risk areas for penetration on each path of the user's movement, and dynamically avoids the initial path through multiple rounds of iterative corrections, thereby effectively avoiding abnormal situations such as virtual avatar crossing and overlapping that destroy the sense of continuous space. While improving the accessibility and robustness of VR navigation paths, it ensures the physical consistency of multi-user collaborative interaction, greatly enhancing the user trust and usage stickiness of virtual shopping malls, and has a significant promoting effect on the quality of immersive experience in scenarios such as retail, e-commerce, and exhibitions. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 Flowchart of the system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0012] The following will clearly and completely describe 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. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0013] Embodiment: The present invention provides Figure 1 The virtual shopping mall navigation system based on VR technology shown in the figure includes a user initialization module, a navigation path planning module, a mold penetration hazard perception module, a path dynamic optimization module, and a navigation visualization module; The user initialization module is used to build a virtual shopping mall space based on VR 3D modeling technology, load the virtual avatar of the current online user in real time, and initialize the virtual avatar's spatial position, posture data, and motion data; The navigation path planning module is used to respond to the user's product purchase instructions and call the path search algorithm to generate an initial navigation path from the current user location to the target product area; A penetration hazard perception module is used to obtain collision detection failure information of the user avatar and state synchronization delay information between the remote avatar and the local view during the execution of the navigation path. The collision detection failure information includes a collision detection failure coefficient, and the state synchronization delay information includes a state synchronization delay coefficient. The path dynamic optimization module is used to construct a mold penetration risk heat map based on the collision detection failure coefficient and state synchronization delay coefficient, and iteratively correct the initial navigation path; A navigation visualization module is used to present the final corrected navigation path to the user in a multimodal manner; The user initialization module is used to build a virtual shopping mall space based on VR 3D modeling technology, load the virtual avatar of the current online user in real time, and initialize the virtual avatar's spatial position, posture data, and motion data. The details are as follows: Load pre-designed 3D scene models (such as shopping mall structures, shelves, products, etc.), and use a unified coordinate system (such as a right-handed coordinate system) to define space; use engines such as Unity3D and Unreal Engine to register object nodes in space and establish a hierarchical structure tree of scene objects; perform navigation grid processing on passable areas, impassable areas, and display areas; mark the user's birth point (entrance), guide point (path reference anchor point) and target product point (product interaction anchor point); assign a standard virtual avatar model to each newly connected user, and the virtual avatar model includes bone nodes, animation components (Animator) and action state machine; position the new user's virtual avatar at the mall entrance area or the user's most recent login location, initialize the default orientation and initial perspective to ensure that the user faces the navigation area; from VR devices (such as headsets) The system uses a wearable display, controller, and full-body motion capture equipment to obtain real-time motion data, including: HMD position and posture, left / right controller three-axis acceleration and direction, and the relative position of all joints in the body. The acquired motion data is bound to the corresponding nodes of the virtual avatar to achieve real-time posture mapping and action driving; user interaction behaviors (such as gestures, buttons, and voice) are bound to action events, such as "raising a hand to greet" triggering social animations and "opening the mouth" triggering voice animations. Each user is assigned a unique virtual identity ID and communication channel, and the virtual avatar is bound to the shopping mall scene, defining its accessible areas and interactive objects (such as clicking on products and approaching promotional areas to trigger prompts). The navigation path planning module is used to respond to the user's product purchase instructions and call the path search algorithm to generate the initial navigation path from the current user location to the target product area. The details are as follows: The user selects a product through clicks, voice, gestures, etc. The system parses the target product's unique ID and spatial anchor coordinates from the product database; identifies the target product's shelf number, floor location, area label, and other auxiliary navigation parameters; uses the current avatar's spatial center point (usually the foot or torso projection point) as the starting point of the path; and invokes a path search algorithm based on the current user's location to the target product's location, including but not limited to: breadth-first search algorithm, depth-first search algorithm, Dijkstra algorithm, and Floyd algorithm. Convert search results into a sequence of waypoints ,in is the midpoint vector of the path, including coordinates, angles, and movement modes; that is, the initial navigation path from the current user location to the target product area is obtained; A penetration hazard perception module is used to obtain collision detection failure information of the user avatar and state synchronization delay information between the remote avatar and the local view during the execution of the navigation path. The collision detection failure information includes a collision detection failure coefficient, and the state synchronization delay information includes a state synchronization delay coefficient. The collision detection failure coefficient (CFC) in this paper is a metric used to measure the failure rate of collision detection between user avatars in a virtual mall. It reflects the frequency and severity of physical contact between avatars during navigation or interaction that the system fails to accurately identify and process. This coefficient not only reflects the adaptability of the current collision detection mechanism to multi-user, high-concurrency scenarios, but also indirectly reveals the stability and accuracy of the system's physical simulation and synchronization mechanisms. A large CFC indicates that the system's collision perception mechanism in that area or path segment has significant lag or accuracy issues, failing to effectively capture physical contact events. This leads to significant "interplay" between avatars, disrupting spatial coherence and realism, and thus affecting users' immersive experience and willingness to interact. Especially in complex environments like virtual malls with multiple points of interaction and multi-person synchronization, a high CFC often indicates frequent overlap, skeletal misalignment, and interleaving of actions between users. This can lead to user distrust of the virtual environment's physical rules and interaction boundaries, reducing the system's commercial value. Conversely, a smaller collision detection failure coefficient indicates that the system's spatial collision recognition and feedback mechanisms between users are functioning well, capturing the proximity, contact, and intersection behaviors between user avatars in relatively real time. Based on the physics engine or rule model, it triggers reasonable avoidance responses, effectively preventing the occurrence of "embroidery," maintaining the stability of spatial topological relationships, and enhancing social authenticity and environmental trust between users. This is crucial for maintaining the continuity of diverse behaviors in shopping malls, such as shopping guides, product consultations, and social interactions. By quantitatively assessing the hidden dangers of avatar embroidery based on the collision detection failure coefficient, the system can predict embroidery risk hotspots during path planning and dynamically adjust the path away from areas with high collision detection failure coefficients to proactively avoid risks. Secondly, this coefficient can serve as a performance feedback indicator for the physics engine, driving adaptive upgrades to the underlying algorithm or flexible scheduling mechanisms (such as dynamic switching of frame rate priorities), thereby improving system robustness.

[0014] The logic for obtaining the collision detection failure coefficient is as follows: For each user avatar i, at each timestamp t, record: Local observation state vector ,in is the position vector of user avatar i at time t, is the velocity vector of user avatar i at time t, is the posture vector of user avatar i at time t; System predicted state vector ,in The time advance predicted for the system; Calculate the local prediction bias vector ; Constructing a global perceptual difference map ,in is a set of user avatar nodes, each node is a user avatar, is the edge set; For any user avatar node pair , calculate the difference between the state of i from the perspective of j: ,in is the state difference, The predicted state vector of the system for user i from user j’s local perspective; Calculate the similarity weight of the user's avatar node edge: ,in is the similarity weight of the user's avatar node edge; Calculate node inconsistency entropy: ,in is the node inconsistency entropy, is the neighbor node set of user avatar node i; The larger the node inconsistency entropy, the greater the difference in the state of user i "seen" by different users, that is, the more inconsistent the state perception is; Calculate the collision detection failure coefficient: ,in is the collision detection failure coefficient, is the set of users who have spatial interactions with the current path, is the average node inconsistency entropy of the user's avatar node g within the current path range, ,in is the node inconsistency entropy of the user avatar node g, , is the time that the user's avatar node g stays within the current path range, is the minimum spatial distance from the user's virtual avatar node i to the midpoint of the current path, is a minimum constant to prevent division by zero (usually ); It should be noted that the above formulas are all dimensionless and numerical calculations. Common dimensionless methods include Min-Max normalization and Z-Score normalization, which will not be described here. The state synchronization delay coefficient in this invention is a key indicator used to quantify the degree of state synchronization lag between different user avatars in a distributed virtual reality system under a network environment. This coefficient reflects the time difference and degree of consistency deviation between the actual state of other remote user avatars observed by the local user and their synchronized state. A large state synchronization delay coefficient indicates that the state data received by the remote user avatar at the current node has a significant timing lag or interpolation error, resulting in a significant deviation between the displayed virtual behavior and its actual operation. A small state synchronization delay coefficient indicates that the remote state synchronization process is relatively timely and accurate, the virtual avatar performance rendered by the system is highly consistent with its user-side behavior, and the perceptual continuity is good. In a virtual shopping mall navigation environment built based on VR technology, multiple user avatars need to collaboratively navigate and interact in the same virtual space, including parallel movement, path intersection, and target guidance. Due to the spatial proximity between virtual avatars and the high degree of intersection of movement paths, the system places higher requirements on the synchronization frequency and accuracy of key state information such as position, velocity, and posture. However, in high-concurrency scenarios, systems often synchronize remote user states only at a low frequency to optimize performance load, or employ interpolation prediction to compensate for network latency. This can cause avatars to experience "position lag," "velocity discrepancies," and "directional drift" from the user's local perspective. The core cause of these phenomena is the distributed perspective differences caused by state synchronization delays. These discrepancies are particularly pronounced in near-field interactions, and can easily lead to "interplay" between avatars. This occurs when two users fail to detect contact from their local perspectives, but their virtual bodies inevitably overlap or intersect due to state synchronization delays. By introducing a state synchronization delay coefficient, the present invention provides a measurable, feedback-based, and predictable method for assessing interplay risks. The state synchronization delay coefficient not only reflects the real-time performance of the system's distributed synchronization mechanism but also serves as a risk prediction signal during virtual interactions between users. Its introduction transforms interplay risks from a "passive" occurrence into a "perceivable and controllable" early warning issue, providing a crucial technical foundation for enhancing the stability, immersion, and security of virtual shopping mall navigation systems. The coefficient has significant engineering application value and expansion potential.

[0015] The logic for obtaining the state synchronization delay coefficient is as follows: For any user avatar u, at system timestamp t, the cross-view residual tensor is calculated based on the received synchronization state packets of M remote user perspectives: ,in is the cross-view residual tensor of the user avatar u, is the state synchronization error vector of the remote user virtual avatar v to the local user virtual avatar u, is the local observation state vector of the user avatar u, For the user's virtual avatar v perspective, delay Then the system predicts the state vector of the user's virtual avatar u; Get the time decay weight of the user synchronization state in the preset time window T: ,in is the time decay weight, ; Time weighting is performed on the states within the window: ,in is the time-weighted cross-view residual tensor coefficient; Environmental factors are introduced to model the interference of synchronization state. The environmental factors include the number of other users within the unit space range in the current time window. The environmental coupling factor is defined as: ,in is the environmental coupling factor of the user avatar u, is the number of other users within the unit space range in the current time window; Calculate the state synchronization delay coefficient: ,in is the state synchronization delay coefficient; It should be noted that the above formulas are all dimensionless and numerical calculations. Common dimensionless methods include Min-Max normalization and Z-Score normalization, which will not be described here. The path dynamic optimization module is used to construct a mold penetration risk heat map based on the collision detection failure coefficient and state synchronization delay coefficient, as follows: Set the initial navigation path Divide into K equal path segments: ,in is the kth path segment, k={1,2,...,K}, K is the total number of path segments; For each path segment , get the path segment A collection of all online users' avatars within a meter range : ,in Create a virtual avatar for the user With path segments The shortest distance; For each path segment , construct path segments based on the collision detection failure coefficient and state synchronization delay coefficient of the user virtual avatar in the interaction area The risk of penetration: ,in For path segments The risk value of penetration, Create a virtual avatar for the user The collision detection failure coefficient, Create a virtual avatar for the user The state synchronization delay coefficient, represent the preset proportional coefficients of the collision detection failure coefficient and the state synchronization delay coefficient, respectively, and are greater than 0, is a minimum constant to prevent division by zero (usually ); It should be noted that Set it according to the actual situation. For example, adopt the expert empowerment method, that is, invite experts in related fields to determine the preset proportion coefficients of various indicators through professional opinion surveys and comprehensive evaluations, for example, It can be 0.5, 0.5; Construct a mold penetration risk heat map based on the mold penetration risk value of each path : The mold penetration hazard heat map shows the risk intensity of mold penetration hazards in each section of the entire path; Compare the mold penetration risk value in the mold penetration risk heat map with the preset mold penetration risk value threshold, and iteratively correct the initial navigation path as follows: If the penetration risk value is greater than the penetration risk threshold, it means that the penetration risk of the current path segment is too high. An iterative correction signal is generated, and the path search algorithm is called to regenerate the navigation path from the current user location to the target product area. A minimization cost function is defined and used as an iterative condition to iteratively correct the initial navigation path. The minimization cost function is minimized: ,in is the path length of the navigation path, 、 Respectively represent the preset proportional coefficients of path length and penetration risk value, and 、 All greater than 0; It should be noted that after the iterative correction signal is generated and the path search algorithm is called to regenerate the navigation path from the current user location to the target product area, the process of running the module for detecting hidden dangers -> dynamic path optimization needs to be re-executed. The above formulas are all dimensionless and calculated numerically. Common dimensionless methods include Min-Max normalization and Z-Score standardization, which are not detailed here. 、 Set it according to the actual situation. For example, adopt the expert empowerment method, that is, invite experts in related fields to determine the preset proportion coefficients of various indicators through professional opinion surveys and comprehensive evaluations, for example, 、 It can be 0.5, 0.5; If the mold penetration risk value is less than or equal to the mold penetration risk value threshold, it means that the mold penetration risk of the current path segment is low and within the controllable range, and there is no need to generate an iterative correction signal; A navigation visualization module is used to present the final corrected navigation path to the user in a multimodal manner; The multimodal approach includes three-dimensional light-track guidance, ground virtual arrows, and field-of-view voice prompts; 3D light trail guidance refers to a light trail band extending along the navigation path, that is, dynamic particles or light texture voxels are evenly laid on the path curve to achieve forward flow animation. The light trail color is visually encoded according to the intensity of the penetration risk (such as red, yellow and green gradient); Ground virtual arrows dynamically generate directional arrows on the ground projection layer to help users adjust their direction in real time. For each path segment, a set of virtual arrows pointing towards the end point is generated at its starting point. When the distance is too close to the user, the arrow size is dynamically scaled to avoid visual obstruction. Field of view voice prompts refer to providing timed voice navigation prompts based on the user's current viewing angle and path segment information; The 3D light trail is the main navigation channel, the ground virtual arrow is used for instant fine-tuning, and the field of view voice prompts are used to prompt key path events; The present invention uses real-time perception of the collision detection failure coefficient and state synchronization delay coefficient in the virtual shopping mall navigation system to refine the potential cross-mold penetration problem caused by inconsistent states between avatars. Based on this, a heat map of mold penetration hazards at the path segment level is constructed to identify and quantify the potential mold penetration risks of each segment on the path. The initial path is iteratively corrected with the goal of minimizing the risk. This not only effectively avoids cross-mold penetration areas, but also takes into account the balance between path length and navigation efficiency, achieving closed-loop control of the entire process from path generation and risk perception to iterative optimization and visual presentation. Therefore, it can significantly improve the reliability and immersion of navigation in a complex VR environment with multiple users and high concurrency. Specifically, with the help of the collision detection failure coefficient and the state synchronization delay coefficient, the system accurately identifies potential high-risk areas for penetration on each path of the user's movement, and dynamically avoids the initial path through multiple rounds of iterative corrections, thereby effectively avoiding abnormal situations such as virtual avatar crossing and overlapping that destroy the sense of continuous space. While improving the accessibility and robustness of VR navigation paths, it ensures the physical consistency of multi-user collaborative interaction, greatly enhancing the user trust and usage stickiness of virtual shopping malls, and has a significant promoting effect on the quality of immersive experience in scenarios such as retail, e-commerce, and exhibitions.

[0016] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0017] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0018] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0019] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0020] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A virtual shopping mall navigation system based on VR technology, characterized by: It includes user initialization module, navigation path planning module, mold penetration hazard perception module, path dynamic optimization module, and navigation visualization module; The user initialization module is used to build a virtual shopping mall space based on VR 3D modeling technology, load the virtual avatar of the current online user in real time, and initialize the virtual avatar's spatial position, posture data, and motion data; The navigation path planning module is used to respond to the user's product purchase instructions and call the path search algorithm to generate an initial navigation path from the current user location to the target product area; A penetration hazard perception module is used to obtain collision detection failure information of the user avatar and state synchronization delay information between the remote avatar and the local view during the execution of the navigation path. The collision detection failure information includes a collision detection failure coefficient, and the state synchronization delay information includes a state synchronization delay coefficient. The path dynamic optimization module is used to construct a mold penetration risk heat map based on the collision detection failure coefficient and state synchronization delay coefficient, and iteratively correct the initial navigation path; The navigation visualization module is used to present the final corrected navigation path to the user in a multimodal manner.

2. The virtual shopping mall navigation system based on VR technology according to claim 1, characterized in that: The logic for obtaining the collision detection failure coefficient is as follows: For each user avatar i, at each timestamp t, record: Local observation state vector ,in is the position vector of user avatar i at time t, is the velocity vector of user avatar i at time t, is the posture vector of user avatar i at time t; System predicted state vector ,in The time advance predicted for the system; Calculate the local prediction bias vector ; Constructing a global perceptual difference map ,in is a set of user avatar nodes, each node is a user avatar, is the edge set; For any user avatar node pair , calculate the difference between the state of i from the perspective of j: ,in is the state difference, The predicted state vector of the system for user i from user j’s local perspective; Calculate the similarity weight of the user's avatar node edge: ,in is the similarity weight of the user's avatar node edge; Calculate node inconsistency entropy: ,in is the node inconsistency entropy, is the neighbor node set of user avatar node i; Calculate the collision detection failure coefficient: ,in is the collision detection failure coefficient, is the set of users who have spatial interactions with the current path, is the average node inconsistency entropy of the user's avatar node g within the current path range, ,in is the node inconsistency entropy of the user avatar node g, , is the time that the user's avatar node g stays within the current path range, is the minimum spatial distance from the user's virtual avatar node i to the midpoint of the current path, is a very small constant that prevents division by zero.

3. The virtual shopping mall navigation system based on VR technology according to claim 1, characterized in that: The logic for obtaining the state synchronization delay coefficient is as follows: For any user avatar u, at system timestamp t, the cross-view residual tensor is calculated based on the received synchronization state packets of M remote user perspectives: ,in is the cross-view residual tensor of the user avatar u, is the state synchronization error vector of the remote user virtual avatar v to the local user virtual avatar u, is the local observation state vector of the user avatar u, For the user's virtual avatar v perspective, delay Then the system predicts the state vector of the user's virtual avatar u; Get the time decay weight of the user synchronization state in the preset time window T: ,in is the time decay weight, ; Time weighting is performed on the states within the window: ,in is the time-weighted cross-view residual tensor coefficient; Environmental factors are introduced to model the interference of synchronization state. The environmental factors include the number of other users within the unit space range in the current time window. The environmental coupling factor is defined as: ,in is the environmental coupling factor of the user avatar u, is the number of other users within the unit space range in the current time window; Calculate the state synchronization delay coefficient: ,in is the state synchronization delay coefficient.

4. The virtual shopping mall navigation system based on VR technology according to claim 1, characterized in that: A heat map of mold penetration hazards is constructed based on the collision detection failure coefficient and the state synchronization delay coefficient, as follows: Set the initial navigation path Divide into K equal path segments: ,in is the kth path segment, k={1,2,...,K}, K is the total number of path segments; For each path segment , get the path segment A collection of all online users' avatars within a meter range : ,in Create a virtual avatar for the user With path segments The shortest distance; For each path segment , construct path segments based on the collision detection failure coefficient and state synchronization delay coefficient of the user virtual avatar in the interaction area The risk of penetration: ,in For path segments The risk value of penetration, Create a virtual avatar for the user The collision detection failure coefficient, Create a virtual avatar for the user The state synchronization delay coefficient, represent the preset proportional coefficients of the collision detection failure coefficient and the state synchronization delay coefficient, respectively, and are greater than 0, is a very small constant to prevent division by zero; Construct a mold penetration risk heat map based on the mold penetration risk value of each path : .

5. The virtual shopping mall navigation system based on VR technology according to claim 4 is characterized by: Compare the mold penetration risk value in the mold penetration risk heat map with the preset mold penetration risk value threshold, and iteratively correct the initial navigation path as follows: If the penetration risk value is greater than the penetration risk value threshold, an iterative correction signal is generated, and the path search algorithm is called to regenerate the navigation path from the current user location to the target product area. A minimization cost function is defined and used as an iterative condition to iteratively correct the initial navigation path to minimize the cost function: ,in is the path length of the navigation path, 、 Respectively represent the preset proportional coefficients of path length and penetration risk value, and 、 All greater than 0; If the mold penetration risk value is less than or equal to the mold penetration risk value threshold, there is no need to generate an iterative correction signal.

Images