An emergency evacuation simulation training method and system based on digital twinning and augmented reality, and a storage medium

Abstract

The application discloses an emergency evacuation simulation training method and system based on digital twinning and augmented reality, and a storage medium, the method generates a target virtual building corresponding to a target building in a three-dimensional virtual scene, and establishes a communication connection between a head-mounted display, a control device and the three-dimensional virtual scene; a search algorithm is used to search for a shortest path in the target virtual building to obtain global optimal path information; a continuous visual guidance track is generated according to the global optimal path information and is rendered in the display field of view of a trainee; real-time position information of each trainee in the target virtual building is acquired, and direction and distance information between each trainee and a target point is calculated. Since the digital twinning and augmented reality are combined in the method, accurate spatial anchoring is achieved, thereby providing an evacuation training scheme with VR immersion and AR information enhancement functions at a cost much lower than that of independent deployment of AR, which can be popularized to a large-scale crowd.

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of information technology and safety science, and in particular to an emergency evacuation training method, system, and storage medium based on digital twins and augmented reality. Background Technology

[0002] Emergency evacuation drills are training activities that simulate sudden emergencies and organize relevant personnel to safely, quickly, and orderly evacuate to a safe area according to a predetermined plan and route. Their purpose is to transform instinctive reactions in sudden disasters into orderly responses, thereby maximizing escape time and minimizing casualties.

[0003] Large-scale emergency evacuation training for urban public safety currently faces bottlenecks. Traditional field drills are costly, risky, and have limited coverage, while paper-based video teaching lacks immersion and interactivity. Emerging AR technology can provide virtual-real fusion guidance, but its deployment heavily relies on high-precision indoor positioning, 3D modeling, and on-site calibration, resulting in high costs, limited pathways, and difficulty in large-scale application. While VR technology can build immersive virtual environments at a lower cost, existing systems are mostly geared towards macro-behavioral analysis, lacking real-time navigation and situational information projection from the individual subject's perspective. This causes training to remain at the roaming level and cannot support effective decision-making drills.

[0004] Therefore, the existing technology still needs further improvement. Summary of the Invention

[0005] The purpose of this invention is to provide an emergency evacuation training method, system, and storage medium based on digital twins and augmented reality, aiming to overcome the problem of real-time guidance in the prior art, thereby achieving immersive evacuation training for large groups of people with real-time dynamic information support without relying on expensive AR hardware and complex on-site calibration.

[0006] The technical solution adopted by this invention to solve the technical problem is as follows: In a first aspect, the present invention provides an emergency evacuation drill training method based on digital twins and augmented reality, comprising: Based on the three-dimensional structured data of the target building, a target virtual building corresponding to the target building is generated in a three-dimensional virtual scene, and a communication connection is established between the head-mounted display and control device and the three-dimensional virtual scene. The shortest path search algorithm is used to perform a shortest path search under the target virtual building to obtain the global optimal path information; Based on the key control points contained in the global optimal path information, a continuous visual guidance trajectory is generated in the three-dimensional virtual scene, and the visual guidance trajectory is rendered in the display field of view of the trainee in the form of an overlay layer; Obtain the real-time location information of each trainee in the target virtual building, and calculate and display the direction and distance information between each trainee and the target point based on the real-time location information; The key control points are a set of path points; The steps of generating a continuous visual guidance trajectory in the 3D virtual scene based on the key control points contained in the global optimal path information, and rendering the visual guidance trajectory as an overlay layer in the trainee's display field of view include: Obtain the set of path points from the global optimal path information, and pass the path points in the set of path points to the vertex buffer of the offline renderer of the 3D virtual scene one by one; Based on the list of path inflection point coordinates corresponding to the path point set, a visual guidance trajectory is rendered in the three-dimensional virtual scene; wherein, the visual guidance trajectory is the first-person perspective guidance trajectory of each trainee in the virtual scene.

[0007] Optionally, the step of generating a target virtual building corresponding to the target building in a 3D virtual scene based on the 3D structured data of the target building, and establishing a communication connection between the head-mounted display and control device and the 3D virtual scene includes: Obtain the source file of the BIM model of the target building and import the source file of the BIM model into the Unity3D engine; The Unity3D engine parses the received source file and generates a visual graphic component for rendering a 3D scene and a physical interaction boundary component that matches the spatial occupancy of the visual graphic component, thus obtaining the target virtual building in the 3D virtual scene.

[0008] Optionally, the step of using a shortest path search algorithm to perform a shortest path search under the target virtual building to obtain globally optimal path information includes: In response to the initialization command of the target virtual building, obtain all object nodes marked as safety exits in the scene, and complete the binding of the safety exits; Obtain the current training level ID and determine the target exit corresponding to the training level ID; Starting from the trainee's current location and ending at the target exit location, via A The algorithm calculates the globally optimal path information on the navigation grid of the target virtual building.

[0009] Optionally, the starting point is the current location of the trainee, and the ending point is the location of the target exit, via A. The algorithm's step of calculating the globally optimal path information on the navigation grid of the target virtual building also includes: Judge A Does the planned path output by the algorithm meet the preset valid conditions? When A When the planned path output by the algorithm does not meet the preset valid conditions, a light carpet clearing signal is generated, and the light carpet clearing operation is controlled to terminate the rendering of the current guide path; wherein, the valid conditions include at least: the number of path points is greater than or equal to 2 inflection points.

[0010] Optionally, the line renderer is a Line Renderer component. The step of generating a continuous visual guidance trajectory in the 3D virtual scene based on the path inflection point coordinates contained in the globally optimal path information, and rendering the visual guidance trajectory as an overlay layer in the trainee's display field of view, further includes: The vertices of the Line Renderer component are updated at preset time intervals.

[0011] Optionally, the step of obtaining the real-time location information of each trainee in the target virtual building, and calculating and displaying the direction and distance information between each trainee and the target point based on the real-time location information includes: Establish a mapping relationship between world coordinates and map coordinates in a 3D virtual scene; The system determines the current floor of the trainee based on the height information of the trainee's current location, and checks whether the detected floor index is the same as the floor currently displayed on the map. If the detected floor index is different from the floor currently displayed on the map, the current map floor is updated to the floor corresponding to the floor index, and the corresponding map data is loaded. Calculate the path length between the current location of the trainee and each candidate exit, and the exit corresponding to the shortest path; Calculate the horizontal angle between the trainee's orientation and the exit direction, and map the horizontal angle to one of multiple standard direction categories according to a preset discretization strategy; wherein, the discretization strategy is to uniformly divide the 360° horizontal circumference into N direction intervals, where N is an integer greater than or equal to 4; The path length between the exit corresponding to the shortest path and the trainee, and the direction label corresponding to the angle between the horizontal plane, are combined to form a prompt display information for direction and distance.

[0012] Optionally, before the step of obtaining the real-time location information of each of the trainees in the target virtual building, and calculating and displaying the direction and distance information between each of the trainees and the target point based on the real-time location information, the method further includes: Determine whether the height information of the trainee exceeds a preset height threshold. If the height information is greater than or equal to the preset height threshold, set the display state of the prompt information to display; otherwise, set the display state of the prompt information to hide.

[0013] Optionally, the prompt display information may be prompt text, a 3D floating directional sign, or a graphic path display identifier.

[0014] Secondly, the present invention also provides an emergency evacuation drill system based on digital twins and augmented reality, comprising: The virtual scene conversion module is used to generate a target virtual building corresponding to the target building in a three-dimensional virtual scene based on the three-dimensional structured data of the target building, and to establish a communication connection between the head-mounted display and control device and the three-dimensional virtual scene. The optimal path search module is used to perform a shortest path search algorithm under the target virtual building to obtain the global optimal path information; The trajectory guidance module is used to generate a continuous visual guidance trajectory in the three-dimensional virtual scene based on the key control points contained in the global optimal path information, and to render the visual guidance trajectory in the display field of view of the trainee in the form of an overlay layer; the key control points are a set of path points; The steps of generating a continuous visual guidance trajectory in the 3D virtual scene based on the key control points contained in the global optimal path information, and rendering the visual guidance trajectory as an overlay layer in the trainee's display field of view include: Obtain the set of path points from the global optimal path information, and pass the path points in the set of path points to the vertex buffer of the offline renderer of the 3D virtual scene one by one; Based on the list of path inflection point coordinates corresponding to the path point set, a visual guidance trajectory is rendered in the three-dimensional virtual scene; wherein, the visual guidance trajectory is the first-person perspective guidance trajectory of each trainee in the virtual scene; The dynamic positioning module is used to acquire the real-time location information of each trainee in the target virtual building, and to calculate and display the direction and distance information between each trainee and the target point based on the real-time location information.

[0015] Thirdly, the present invention provides a computer storage medium, which is a computer-readable storage medium, wherein a computer program is stored in the computer storage medium, and when the computer program is executed by a computer, the computer is used to execute the emergency evacuation training method based on digital twin and augmented reality.

[0016] Beneficial effects: This invention discloses an emergency evacuation training method, system, and storage medium based on digital twins and augmented reality. The method generates a target virtual building corresponding to the target building in a 3D virtual scene based on the target building's 3D structured data, and establishes a communication connection between the head-mounted display and control devices and the 3D virtual scene. A shortest path search algorithm is used to search for the shortest path within the target virtual building to obtain globally optimal path information. Based on the key control points contained in the globally optimal path information, a continuous visual guidance trajectory is generated in the 3D virtual scene and rendered as an overlay layer in the trainees' field of vision. Real-time position information of each trainee in the target virtual building is acquired, and the direction and distance information between each trainee and the target point are calculated and displayed based on the real-time position information. Because this invention integrates digital twins and augmented reality, achieving precise spatial anchoring, it provides an evacuation training solution that combines VR immersion and AR information enhancement, and can be promoted to a large number of people, at a cost far lower than that of independent AR deployment. Attached Figure Description

[0017] Figure 1 A flowchart illustrating the steps of an emergency evacuation drill method based on digital twins and augmented reality provided by this invention; Figure 2 A schematic diagram of the overall implementation principle structure of the emergency evacuation drill method provided by the present invention; Figure 3 This is a small map simulation diagram of the method provided by the present invention in a specific application embodiment; Figure 4 Arrow simulation illustration in a specific embodiment of the method provided by the present invention Figure 1 ; Figure 5 Arrow simulation diagram in a specific embodiment of the method provided by the present invention Figure 2 ; Figure 6 Simulation of prompt text in a specific application embodiment of the method provided by the present invention Figure 1 ; Figure 7 Simulation of prompt text in a specific application embodiment of the method provided by the present invention Figure 2 ; Figure 8 The principle structure block diagram of the emergency evacuation drill system provided by the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0019] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the following claims.

[0020] With the acceleration of global urbanization, the number and scale of densely populated buildings such as subway stations, large commercial complexes, transportation hubs, and high-rise residential buildings continue to grow. In the event of emergencies such as fires, earthquakes, or terrorist attacks, organizing the safe, orderly, and rapid evacuation of large numbers of people to safe exits is a core challenge for urban public safety management. However, ordinary residents and building users generally lack evacuation knowledge and practical experience. Surveys show that most building residents lack basic understanding of the location of non-standard exits, optimal evacuation routes, and the meaning of emergency signs. This lack of crucial escape information can directly translate into evacuation delays, incorrect route selection, and even stampedes during disasters.

[0021] Consumer-grade VR headsets are now widely available, with six-degrees-of-freedom spatial tracking accuracy down to the millimeter level. The cost per unit is also significantly lower than that of AR glasses. VR devices allow users to freely walk and observe virtual buildings from a first-person perspective, creating a strong sense of spatial immersion and presence—something that traditional 2D training materials and pure AR solutions struggle to achieve.

[0022] Currently, there are three mainstream technical solutions in the industry. The first type is traditional on-site evacuation drills, which are generally organized periodically by building management. Security personnel guide participants at key points, and participants evacuate along designated routes to the assembly point. After the drill, there is usually a verbal summary and a simple headcount. The main shortcomings of this solution are: first, the number of people covered in a single drill is limited, and it cannot effectively cover the resident population of large building complexes or transportation hubs; second, to ensure safety, traditional on-site evacuation drills usually require advance notification and the deployment of protective measures, and they cannot simulate extreme disaster scenarios such as dense smoke, power outages, and structural damage, which greatly reduces the realism and suddenness of the drills; finally, traditional on-site evacuation drills are difficult to organize frequently, making it impossible to maintain high-frequency training and drills.

[0023] The second category involves disseminating emergency knowledge through two-dimensional text and video. This typically uses brochures, posters, WeChat official accounts, and short videos to spread evacuation information to the public. Content includes evacuation route maps, emergency exit locations, and fire extinguisher usage methods. The main drawback of this approach is that the information is presented in a two-dimensional, static format, disconnected from the real three-dimensional building space. Even after viewing the floor plan and entering the actual building, trainees still cannot efficiently map the routes from the plan to the three-dimensional space. Furthermore, two-dimensional text and video-based emergency knowledge dissemination lacks interactive feedback and immersive context, resulting in training effectiveness that remains at the level of knowing rather than being able to do.

[0024] The third type is a standalone AR evacuation guidance system, which mainly utilizes AR glasses or mobile AR applications to overlay virtual guidance information, such as arrows, cursors, and path indicators, in real time onto real buildings to guide users to evacuate along the optimal route. The main drawbacks of this solution are: extremely high deployment costs, requiring separate 3D modeling of each target building and deployment of indoor positioning base stations, such as UWB and WiFi RTT; complex calibration, requiring reference coordinate system calibration and manual arrangement of guidance paths; difficulty in guaranteeing indoor positioning accuracy in complex building environments due to multipath effects and signal obstruction; and the rendering of virtual guidance information requires precise registration with real physical geometry for spatial anchoring, as any positioning error will cause the guidance information to drift.

[0025] This invention aims to address the core technical bottlenecks in information presentation and real-time guidance faced by immersive simulation systems for large-scale crowd evacuation training in the field of urban public safety. Specifically, it addresses the following three major technical issues: (I) Inherent Limitations of Traditional Evacuation Training Methods. Currently, evacuation knowledge and emergency skills training for urban residents and building users mainly relies on three traditional methods: on-site fire drills, dissemination of knowledge through print and video materials, and classroom lectures. While on-site drills are the most intuitive, the number of participants is severely limited, making it impossible to effectively cover the resident population of large building complexes or transportation hubs. Furthermore, on-site drills cannot simulate real disaster scenarios such as fires, earthquakes, and stampedes, presenting significant challenges in safety risk control and incurring high organizational costs. Although print and video materials and classroom lectures have a wide reach, their information presentation is two-dimensional, static, or passive, lacking spatial immersion and real-time interactive feedback.

[0026] (II) High Cost and High Barrier to Entry for Using Augmented Reality (AR) Technology Alone for Evacuation Training. In recent years, augmented reality and mixed reality (MR) technologies have developed rapidly, capable of overlaying computer-generated visual information, such as directional arrows and path guidance, onto the real physical environment, providing users with a dynamic interactive experience that blends the virtual and real worlds. However, applying AR technology alone to evacuation training faces extremely high deployment barriers. AR devices need to obtain the user's precise location within the building space, typically relying on UWB technology to provide absolute position coordinates; simultaneously, AR experimental scenarios require high-precision 3D building models as projection anchor surfaces, and separate calibration of reference coordinate systems for different building structures; finally, the guidance information rendered by AR, such as directional indicators and exit markings, needs to be precisely registered with the actual physical geometry of the building. Taking a medium-sized subway station as an example, deploying a station-wide AR evacuation guidance system requires not only weeks of preliminary modeling and calibration work, but also faces the dual constraints of insufficient indoor positioning accuracy caused by GPS unavailability and WiFi fingerprint drift, as well as high equipment costs. Excessive cost and inconvenient access have become the core obstacles to the independent implementation of AR evacuation training.

[0027] (III) Lack of Real-Time Dynamic Information Presentation and Spatial Navigation Support When Using VR Solutions Alone. VR solutions can create highly realistic and interactive virtual environments at a lower cost and quickly construct digital twins of large buildings based on BIM models. Furthermore, the penetration rate of VR hardware, such as HTC VIVE and Meta Quest, is far higher than that of AR glasses, giving them a significant deployment advantage in evacuation training. However, existing VR evacuation simulation systems typically focus on macro-level behavioral simulation and statistical analysis of large groups of people. At the immersive training level for individual subjects, they generally lack dynamic projection of first-person perspective navigation and situational awareness information. That is, the system does not intuitively overlay the real-time decision-making support information needed by the subjects, such as path guidance, spatial orientation, and exit dynamic parameters, within their field of vision. This lack of information projection limits the training experience to merely environmental roaming, making it difficult to support effective evacuation decision-making drills and falling far short of the actual needs of individual emergency response training.

[0028] Therefore, there is an urgent need for an evacuation training method that deeply integrates BIM model-driven VR immersive environment with multimodal AR-enhanced real-time information. This method leverages the low cost and high immersion of VR to solve the problem of scene construction, and uses the concept of AR information overlay to solve the problem of real-time guidance. This would enable immersive evacuation training for large groups of people with real-time dynamic information support without relying on expensive AR hardware and complex on-site calibration.

[0029] To overcome the aforementioned problems, this invention provides an emergency evacuation training method, system, and storage medium based on digital twins and augmented reality. It constructs an interactive virtual building roaming environment by importing the 3D structural data of the target building into the Unity3D engine, while simultaneously overlaying augmented reality information. In a VR first-person perspective, it renders in real-time three types of augmented information: path guidance information indicating the route to the target exit, status prompts indicating the current spatial location and distance to the target, and directional information indicating the target exit relative to the trainee's current orientation. This is based on a light carpet-style evacuation path guidance system, a multi-floor adaptive minimap, and dynamic direction and distance prompts for reachable paths. All three types of augmented information are associated with the 3D structural data to achieve precise spatial anchoring. Therefore, at a cost far lower than that of standalone AR deployment, it provides an evacuation training solution that combines VR immersion with AR information enhancement, and can be scaled up to a large number of people.

[0030] The following description, in conjunction with the accompanying drawings, further illustrates an emergency evacuation training method, system, and storage medium based on digital twins and augmented reality disclosed in this invention.

[0031] Firstly, this invention provides an emergency evacuation drill and training method based on digital twins and augmented reality, such as... Figure 1 As shown, it includes: Step S1: Based on the three-dimensional structured data of the target building, generate a target virtual building corresponding to the target building in a three-dimensional virtual scene, and establish a communication connection between the head-mounted display and control device and the three-dimensional virtual scene.

[0032] This step first constructs a VR immersive architectural scene based on the 3D structured data of the target building. The target building is a 3D structure used for emergency evacuation drills; it could be a high-rise residential building, a subway station, or a commercial complex. The 3D structured data of the target building can be a building information model (BIM) or 3D point cloud data. Both the BIM and point cloud data contain the 3D geometric structure data of the target building; therefore, a corresponding virtual building in a 3D virtual scene can be obtained from either the BIM or point cloud data. After connecting the virtual building to the head-mounted display and control equipment, a VR immersive architectural scene is obtained.

[0033] One method for obtaining a target virtual building in a 3D virtual scene from 3D point cloud data is as follows: First, perform a LiDAR 3D scan on the existing building to obtain a high-precision point cloud. Then, reconstruct the point cloud using a mesh and import it into Unity to obtain the converted target virtual building. The advantage of using 3D point cloud data conversion is that it is suitable for older buildings lacking BIM models. Another approach is procedural building generation based on GIS and CAD data. This automatically generates 3D building models from urban GIS databases or CAD floor plans, suitable for batch generation of large-scale city-level scenes.

[0034] In detail, the steps of generating a target virtual building corresponding to the target building in a 3D virtual scene based on the target building's 3D structured data, and establishing a communication connection between the head-mounted display and control device and the 3D virtual scene include: Step S11: Obtain the source file of the BIM model of the target building and import the source file of the BIM model into the Unity3D engine.

[0035] The maturity of Building Information Modeling (BIM) technology has led to the increasing prevalence of accurate 3D digital models of large buildings. BIM models not only contain precise geometric information about the building, including the spatial location and dimensions of walls, floors, columns, stairs, doors, and windows, but also rich semantic information, such as spatial functional zoning, fire compartments, and exit attributes. This provides an ideal data foundation for constructing high-precision virtual evacuation training scenarios. Importing BIM models into real-time rendering engines such as Unity 3D allows for the rapid creation of interactive virtual building walkthrough environments. In practice, BIM models are imported into the Unity 3D engine via standard intermediate formats such as FBX or IFC.

[0036] Step S12: The Unity 3D engine parses the received source file and generates a visual graphic component for rendering a 3D scene and a physical interaction boundary component that matches the spatial occupancy of the visual graphic component, thereby obtaining the target virtual building in the 3D virtual scene.

[0037] Specifically, the visual graphical component is a Mesh renderer, and the physical interaction boundary component that matches the spatial placement of the visual graphical component is a Collider collider. After importing, Unity will automatically generate the corresponding Mesh renderer and Collider collider, providing a basis for subsequent VR roaming, AR, and other applications. The system provides a geometric basis for functions such as pathfinding and ray detection.

[0038] Step S2: Use the shortest path search algorithm to perform a shortest path search under the target virtual building to obtain the global optimal path information.

[0039] This step utilizes a search algorithm to find evacuation routes. It is the core implementation of the AR information overlay concept in VR space. By rendering a glowing strip path on the virtual building floor, it intuitively guides users to evacuate to the exit along the optimal route.

[0040] In this step, the light carpet path guidance scheme can adopt A. The algorithm performs pathfinding. A The core principle of the algorithm is to introduce a heuristic evaluation function on the basis of the traditional Dijkstra algorithm. By comprehensively evaluating the cost of each search node, the most promising node is expanded first, thereby significantly reducing the search space while ensuring the optimality of the path.

[0041] In detail, the step of using the shortest path search algorithm to perform a shortest path search under the target virtual building to obtain the globally optimal path information includes: Step S21: In response to the initialization command of the target virtual building, obtain all object nodes marked as safety exits in the scene and complete the binding of the safety exits.

[0042] During the initialization phase, the system automatically binds all safety exits in the scene through functions, so that after the virtual building scene is loaded, the system can automatically identify and remember the location of all safety exits.

[0043] The system listens for scene loading completion events. When it detects that all components of the target virtual building have been instantiated in the 3D virtual scene, it triggers an initialization command. In response to this command, the system calls the scene traversal function to recursively search for all exits in the current scene.

[0044] During the search process, the system identifies safe exits based on preset filtering criteria. These criteria include whether a preset exit identifier component is mounted, or whether its tag attribute is marked as "Exit". Exits that meet these criteria are identified as safe exits, and the system adds their reference pointers sequentially to the pre-declared list of safe exits.

[0045] After obtaining all safe exits, the system performs a binding operation. The binding process includes: traversing the aforementioned list of safe exits, extracting the instance identifier of each safe exit and its spatial coordinates in the scene's world coordinate system; associating the extracted identifier with the coordinate data, and registering it as a key-value pair in the safe exit information registry. Simultaneously, a unified interactive response interface is attached to each safe exit, which is used to receive path requests.

[0046] Once the above binding is completed, a mapping relationship is established between the identifiers of each safety exit and the spatial data in the registry. When a user triggers a navigation request, the navigation module directly queries the exit data from the registry to calculate the path, without needing to traverse and search the scene again. This mechanism reduces runtime computational overhead and improves navigation response efficiency.

[0047] Step S22: Obtain the current training level ID and determine the target exit corresponding to the training level ID.

[0048] Once a training level is acquired, its ID is obtained, and the corresponding target exit is determined based on this ID. During runtime, the target exit for the current level is determined based on its ID. Multiple exit support allows for the configuration of multiple evacuation training levels within the same VR scene, with each training session pointing to a different exit, enhancing the variability and coverage of the training.

[0049] Step S23: Taking the current position of the trainee as the starting point and the position of the target exit as the ending point, calculate the global optimal path information on the navigation grid of the target virtual building using the shortest path search algorithm.

[0050] The system will mount A The Seeker component of the pathfinding project implements path calculation. The system calls a function every 0.2 seconds, starting from the user's current location and ending at the target exit location, using A... The algorithm calculates the shortest reachable path on the navigation grid. The 0.2-second recalculation interval strikes a balance between ensuring real-time path delivery and computational overhead.

[0051] A The key formula for shortest path search is as follows: ; ; ; in, For evaluation functions; For actual cumulative costs; For heuristic functions; For any node; For the node Pointing to node A directed edge; From the starting point to the current node The edge sequence of the expanded path; For export purposes.

[0052] The Euclidean heuristic satisfies admissibility, and its key formula is as follows: ; in, For nodes To the target export The true shortest path length of A. This property guarantees A The search returned the globally optimal path.

[0053] After the path search is completed, an ordered sequence consisting of path inflection points is obtained, and its key formula is as follows: ; ; ; in, For the current frame An ordered sequence consisting of path inflection points; This refers to the user's global spatial location at the starting point.

[0054] Finally, the total length of the shortest path is obtained, which is numerically the sum of the Euclidean distances between all adjacent inflection points. The key formula is as follows: ; in, For the current frame The total length of the shortest path.

[0055] Furthermore, in specific implementation, the starting point is the current position of the trainee, and the ending point is the position of the target exit, via A. The algorithm's step of calculating the globally optimal path information on the navigation grid of the target virtual building also includes: Judge A Does the planned path output by the algorithm meet the preset valid conditions? When A When the planned path output by the algorithm does not meet the preset valid conditions, a light carpet clearing signal is generated, and the light carpet clearing operation is controlled to terminate the rendering of the current guide path; wherein, the valid conditions include at least: the number of path points is greater than or equal to 2 inflection points.

[0056] In practical implementation, fault tolerance handling should be configured when the path is unreachable. When A If the path returned by the algorithm is incorrect, or if the number of path points is less than 2 inflection points, the system will call a function to clear the light carpet display to avoid rendering invalid or incorrect guide paths.

[0057] Step S3: Based on the key control points contained in the global optimal path information, generate a continuous visual guidance trajectory in the three-dimensional virtual scene, and render the visual guidance trajectory in the form of an overlay layer in the display field of view of the trainee.

[0058] Specifically, the key control points are a set of path points; the steps of generating a continuous visual guidance trajectory in the 3D virtual scene based on the key control points contained in the global optimal path information, and rendering the visual guidance trajectory as an overlay layer in the trainee's display field of view include: Obtain the set of path points from the global optimal path information, and pass the path points in the set of path points to the vertex buffer of the offline renderer of the 3D virtual scene one by one; Based on the list of path inflection point coordinates corresponding to the path point set, a visual guidance trajectory is rendered in the three-dimensional virtual scene; wherein, the visual guidance trajectory is the first-person perspective guidance trajectory of each trainee in the virtual scene.

[0059] Furthermore, the step of generating a continuous visual guidance trajectory in the 3D virtual scene based on the path inflection point coordinates contained in the global optimal path information, and rendering the visual guidance trajectory as an overlay layer in the trainee's display field of view, also includes: Obtain the path point set from the global optimal path information, and pass the path points in the path point set to the vertex buffer of the Line Renderer component one by one. At the same time, update the vertices of the Line Renderer component at preset time intervals. According to the list of path inflection point coordinates corresponding to the path point set, render it as a visual guidance trajectory in the 3D virtual scene.

[0060] When A After a successful search returns a path, the system does not perform any intermediate data structure conversion or caching. Instead, it directly passes the path point set point by point to the vertex buffer of the Line Renderer component. During this process, the 0.2-second recalculation interval, combined with the Line Renderer's real-time vertex updates, ensures that when the user moves at a normal walking speed in VR, the light carpet path on the ground dynamically adjusts with low latency according to the user's position, fully guaranteeing visual continuity. Secondly, A The algorithm outputs a sequence of coordinates that requires no intermediate processing and is directly used as the rendering vertices of the Line Renderer. This significantly simplifies the data path from the computational result to the visual presentation, avoiding latency and accuracy loss introduced by intermediate processing. The linear segment rendering of the Line Renderer naturally corresponds to A... The inflection points and straight line segments generated on the Nav Mesh map are precisely the most reasonable guidance method in evacuation scenarios.

[0061] After the path calculation is complete, the callback function will receive A. The algorithm returns a list of coordinates of the path's inflection points. The system then assigns the Vector3 coordinate array of this path to the component, rendering it as a continuous light strip in the scene using an luminous material. Its Y-axis coordinates will be 0.1 meters above the ground, conforming to the ground surface to provide intuitive visual guidance to the user, similar to the guide lights on airport runways in the real world.

[0062] Step S4: Obtain the real-time location information of each trainee in the target virtual building, and calculate and display the direction and distance information between each trainee and the target point based on the real-time location information.

[0063] The corresponding source code files in this step will solve the user's spatial positioning problem in VR virtual architecture. For example... Figure 3 As shown, a small map is rendered using 2DCanvas to display the user's current location as an icon, such as... Figure 4 and Figure 5 As shown, the user's orientation is indicated by arrows, and automatic switching between multiple floors is supported. The minimap display interface is implemented using the Unity engine's Canvas component, which is independent of the 3D scene rendering pipeline and uses screen space coordinates for positioning. The minimap background is a preset floor plan, displayed using the Raw Image component; the user's location and direction indicators are UI graphical controls, whose position and rotation angle are driven by the virtual character's horizontal coordinates and orientation angle in the 3D scene world coordinate system, converted to the Canvas's screen coordinates after coordinate mapping.

[0064] In addition, the minimap display interface can have the following alternatives: The first approach uses a Unity Mini map camera, placing an orthographic projection camera directly above the character to render a top-down view in real time, displayed in the minimap UI. Its advantage is that it naturally includes the positional changes of all dynamic objects in the scene; its disadvantage is the need for additional rendering overhead. The second approach uses Nav Mesh-based floor detection. The current approach detects floors using a Y-axis height threshold, but this could be replaced by querying the name of the Nav Mesh area where the character is located using a function, achieving floor recognition based on semantic tags. The third approach is a 3D stereoscopic minimap, directly displaying the overall building and user position in VR space as a transparent 3D model, instead of using a 2D Canvas.

[0065] In detail, the step of obtaining the real-time location information of each of the trainees in the target virtual building, and calculating and displaying the direction and distance information between each of the trainees and the target point based on the real-time location information includes: Step S41: Establish the mapping relationship between world coordinates and map coordinates in the 3D virtual scene.

[0066] In this step, the map data for each floor will be managed through the Floor Data data structure, including floor name, floor map image, world coordinates, and floor height threshold.

[0067] The key formula for the linear mapping from world coordinates to minimap coordinates is as follows: ; ; in, The world coordinate boundary of a certain floor is a rectangular region, and the XZ plane is taken as the horizontal plane; and For the current frame User roles in the boundary area Normalized coordinates within; and For the current frame The projected coordinates of the user's real-time world position on the XZ plane.

[0068] The key formula for the anchor pixel coordinates of the minimap icon is as follows: ; in, and For the current frame The anchor pixel coordinates of the minimap icon; The width of the minimap; This represents the height of the minimap.

[0069] By matching the world coordinate boundaries with the actual building dimensions in the BIM model, a precise spatial anchoring mapping is achieved between the location of the minimap icon and the actual VR spatial location.

[0070] Step S42: Determine the current floor of the trainee based on the height information of the current trainee's location, and determine whether the detected floor index is the same as the currently displayed map floor. If the detected floor index is different from the currently displayed map floor, update the currently displayed map floor to the floor corresponding to the floor index, and load the corresponding map data.

[0071] The system will automatically determine the user's current floor by comparing the user's real-time Y-coordinate in VR with preset values ​​for each floor. When the detected floor index differs from the currently displayed map floor, a function will be automatically invoked to switch the map image, achieving seamless automatic switching of the minimap when the user moves up or down floors in VR.

[0072] Step S43: Calculate the path length between the current location of the trainee and each candidate exit, and the exit corresponding to the shortest path.

[0073] This module will be used to display dynamic information in real time, in text form, indicating the direction and distance of the nearest exit within the VR user's field of vision, serving as a third type of augmented information in addition to the light carpet and minimap.

[0074] (a) Selection of the nearest accessible route with multiple exits In the selection of the shortest reachable path across multiple exits, the function supports two exit selection strategies. When a reachable path is available in the navigation grid, the system calculates the length of the reachable path from the user's current location for each candidate exit and selects the exit with the shortest path, rather than the exit with the shortest straight-line distance, because the shortest straight-line distance may be blocked by walls and unreachable. When a reachable path is unavailable in the navigation grid, the system selects the exit with the shortest straight-line distance. The navigation grid reachability check ensures that the recommended exit is one that the user can actually reach, rather than an invalid exit that is visible but inaccessible.

[0075] The key formulas for calculating the reachable path length and selecting the nearest exit from the navigation grid are as follows: ; ; ; in, To reach the first The total path length that can be reached by the navigation grid of each exit; Let be the total number of segments at the inflection point of the i-th path; For the first A turning point; The user's initial position; For the first One export.

[0076] Subsequently, the nearest reachable exit is the one with the shortest reachable path length among all candidate exits, and its key formula is as follows: ; in, To reach the first The navigation grid for each exit can reach the minimum total path length; This represents the total number of exports.

[0077] Step S44: Calculate the horizontal angle between the trainee's orientation and the exit direction, and map the horizontal angle to one of multiple standard direction categories according to a preset discretization strategy; wherein, the discretization strategy is to uniformly divide the 360° horizontal circumference into N direction intervals, where N is an integer greater than or equal to 4.

[0078] This step uses a function to calculate the horizontal angle between the trainee's orientation and the exit direction, dividing it into N directional sectors, where N is greater than or equal to 4. In practice, N can be set to 8, resulting in 8 directional sectors. The orientation determination result is returned as a Chinese string and displayed directly in the VR user's field of view, outside the prompt text.

[0079] Step S45: Combine the path length between the exit corresponding to the shortest path and the trainee and the direction label corresponding to the horizontal plane angle to form a prompt display information of direction and distance.

[0080] Furthermore, before the step of obtaining the real-time location information of each of the trainees in the target virtual building, and calculating and displaying the direction and distance information between each of the trainees and the target point based on the real-time location information, the method further includes: Determine whether the height information of the trainee exceeds a preset height threshold. If the height information is greater than or equal to the preset height threshold, set the display state of the prompt information to display; otherwise, set the display state of the prompt information to hide.

[0081] To avoid displaying meaningless distance values ​​at the beginning of VR training, before the user has entered the main building, the system uses a gating mechanism to control the visibility of distance text. The distance text is only activated when the user's VR avatar's Y-coordinate exceeds a certain value; it is automatically hidden when it falls below that threshold. This ensures that distance guidance prompts only appear after the user has entered the effective area of ​​the building.

[0082] There are three alternative solutions for dynamic direction and distance prompts. The first solution replaces the TMP text with a 3D floating sign in VR space, generating a semi-transparent 3D arrow sign in the direction of the nearest exit, always facing the user and floating in front of their field of vision, providing intuitive directional guidance. The second solution replaces 8-direction determination with continuous angle values; the current solution outputs discrete direction labels, which can be replaced by outputting continuous angle values, providing more refined orientation information. The third solution calculates A based on path distance. The graph path approach uses Seeker.Start Path to obtain the precise graph path length, offering higher accuracy than navigation grid paths in complex buildings.

[0083] In practice, the displayed information may be prompt text, a 3D floating directional sign, or a graphic path display. For example... Figure 6 and Figure 7 The image shows a schematic diagram of how the prompt text is displayed in a virtual scene.

[0084] To achieve real-time directional arrow following, the system acquires the VR camera's real-time Y-axis Euler rotation angle and maps it to the Z-axis rotation of the minimap icon. When reverse rotation is determined to be true, the system implements a common VR minimap navigation mode where the map remains stationary while the arrow rotates, with the icon arrow always pointing in the user's current VR field of view. This process uses the VR camera instead of the character's root, because the user's viewing direction in VR is determined by their head orientation, not the direction of the character's body movement.

[0085] The key formula for icon rotation driven by the VR camera's yaw angle is as follows: ; in, For the current frame The Z-axis rotation angle of the minimap icon on the 2D Canvas plane; This is a reversal flag; For VR headset cameras in the current frame The Y-axis Euler yaw angle; This is the preset angle offset.

[0086] The key formula for the rotation quaternion corresponding to the minimap icon is as follows: ; in, For the current frame The rotation quaternion corresponding to the minimap icon.

[0087] Furthermore, in practical implementation, the system provides two public interfaces that allow the default directional distance prompt text to be overridden with a specified priority when specific events are triggered, such as when a user approaches a dangerous area, a route change occurs, or a time-sensitive reminder. The duration of the high-priority prompt can be specified by setting parameters, and the default prompt content will automatically revert to its original state after a countdown.

[0088] Combination Figure 2As shown, the method provided in this embodiment can be implemented through modules corresponding to the functions of the above four steps, forming a collaborative organic whole through Unity's scene hierarchy and public data interfaces. The BIM model scene provides navigation grid baking data for light carpet pathfinding, provides world coordinate boundaries that correspond one-to-one with the actual dimensions of buildings for the minimap, and provides obstacle geometry information for distance calculation. Light carpet navigation provides intuitive visual guidance for the user's walking, the minimap provides global spatial positioning reference for the user, and dynamic distance text provides the user with precise quantitative information. The three complement each other at the visual information level: the light carpet answers which path to take; the minimap answers the location information of the virtual character in the building; and the dynamic prompts answer the direction and distance of the nearest exit. At the same time, all augmented information is directly integrated with the BIM model data. The light carpet path is generated along the navigation grid, the minimap coordinate boundaries correspond precisely to the actual dimensions of the BIM model, and the exit position is determined by the BIM model exit marker in the scene. This method achieves precise anchoring of augmented information in VR space, eliminating the need for the complex indoor positioning base station deployment and on-site calibration process of AR solutions.

[0089] Secondly, this invention also provides an emergency evacuation drill and training system based on digital twins and augmented reality, such as... Figure 8 As shown, it includes: The virtual scene conversion module 310 is used to generate a target virtual building corresponding to the target building in a three-dimensional virtual scene based on the three-dimensional structured data of the target building, and to establish a communication connection between the head-mounted display and control device and the three-dimensional virtual scene; its function is as described in step S1.

[0090] The optimal path search module 320 is used to perform a shortest path search under the target virtual building using a shortest path search algorithm to obtain global optimal path information; its function is as described in step S2.

[0091] The trajectory guidance module 330 is used to generate a continuous visual guidance trajectory in the three-dimensional virtual scene based on the coordinates of the path inflection points contained in the global optimal path information, and to render the visual guidance trajectory in the display field of the trainee in the form of an overlay layer; its function is as described in step S3.

[0092] The dynamic positioning module 340 is used to acquire the real-time location information of each trainee in the target virtual building, calculate and display the direction and distance information between each trainee and the target point based on the real-time location information, and its function is as described in step S4.

[0093] Thirdly, the present invention provides a computer storage medium, which is a computer-readable storage medium, wherein a computer program is stored in the computer storage medium, and when the computer program is executed by a computer, the computer is used to execute the emergency evacuation training method based on digital twin and augmented reality.

[0094] Compared to existing traditional evacuation training, independent AR guidance systems, and independent VR simulation systems, the method and system provided by this invention not only overcome their respective core limitations in terms of technical approach, but also demonstrate significant technical superiority in terms of actual system deployment and user experience. Specific advantages are as follows: (i) Achieving the effect of AR augmented information with the low cost of VR breaks through the cost and accuracy bottleneck of independent AR deployment.

[0095] Independent AR evacuation guidance systems require the deployment of indoor positioning base stations, such as UWB anchors, WiFi, and RTTAP, in each target building, along with complex reference coordinate system calibration and manual arrangement of guidance content. Deployment in a single medium-sized building can take several weeks. This invention avoids the difficulties of AR solutions by directly overlaying augmented information into VR virtual space. Since the entire scene is virtual and known, the spatial anchoring of all information in VR is naturally precise. The derived coordinates of the BIM model are the world coordinates, the light carpet is generated along the navigation grid, the minimap coordinates are precisely mapped from the building boundaries, and the distance text is calculated from the precise path. The positional relationship between all augmented information and building geometry can reach centimeter-level accuracy in Unity. At the same time, a set of VR equipment is low-cost and does not require separate on-site deployment for each building; only the BIM model needs to be imported into Unity.

[0096] (ii) The three types of enhanced real-time information form a complementary evacuation guidance information matrix, covering all dimensions of spatial navigation needs.

[0097] Existing VR evacuation systems are severely inadequate in terms of information presentation, making users prone to disorientation. This invention provides three types of enhanced information that complement each other in terms of information dimensions: a light carpet navigation system includes spatial path information, allowing users to know which path to take; a mini-map contains spatial positioning information, telling users their location within the building; and dynamic direction and distance prompts provide quantitative navigation information, informing users of the nearest exit's direction and distance. These three elements work synergistically to cover all the spatial navigation information dimensions required by trainees in evacuation scenarios. The light carpet reduces cognitive load with intuitive visual guidance; users simply need to follow it. The mini-map provides global spatial awareness, preventing feelings of disorientation. Dynamic text provides precise quantitative decision support, helping trainees determine whether to switch exits.

[0098] (III) Based on A Intelligent path planning using algorithms and navigation grid reachability checks ensures the accuracy and real-time nature of guidance information.

[0099] The light carpet navigation module adopts mature A The shortest path algorithm automatically considers the collision geometry of obstacles such as walls, partitions, and stairs in a building through Unity's navigation mesh system. The calculated path naturally satisfies the physical constraint of reaching the target exit along the passable area. Compared to guidance methods based solely on straight-line distance, the navigation mesh reachability check avoids recommending invalid exits that are visible but unreachable. For example, in a U-shaped corridor, the shortest straight-line exit might be completely blocked by a wall, while... The algorithm correctly calculates the complete path around the wall. Combined with a high-frequency recalculation strategy, the path is updated in real time as the user moves, ensuring that the guidance information always accurately reflects the optimal path from the user's current location to the exit.

[0100] (iv) Multi-story adaptive small map and height gating display mechanism to ensure training effectiveness in complex multi-story buildings.

[0101] The multi-floor automatic detection and switching mechanism enables the VR training system to seamlessly cover multi-story buildings, such as the ground floor, concourse, and platform of a subway station. When users ride escalators or walk up stairs, the mini-map automatically switches to the corresponding floor plan without manual operation. Simultaneously, the height gating mechanism ensures that guidance information is only displayed after the user enters the effective area of ​​the building, avoiding misleading data during scene loading and transition phases.

[0102] (v) Modular architecture design, supporting rapid adaptation to different buildings and training scenarios. The four core modules—scene construction, light carpet navigation, minimap, and dynamic prompts—are loosely coupled using Unity's component system. Based on a unified prefab platform, each component is configured using common parameters such as the exits list, navigation mesh data, and floor configuration. Adapting the system to a new building only requires importing the building's BIM model, baking the navigation mesh, placing exit markers in the scene, configuring the minimap boundaries, and finally setting the relevant settings for each floor. The entire process does not involve modifying core code, making its adaptation efficiency far superior to the process of re-calibrating the site and deploying base stations required in standalone AR solutions.

[0103] This invention discloses an emergency evacuation training method, system, and storage medium based on digital twins and augmented reality. The method generates a target virtual building corresponding to the target building in a 3D virtual scene based on the target building's 3D structured data, and establishes a communication connection between the head-mounted display and control devices and the 3D virtual scene. A shortest path search algorithm is used to search for the shortest path within the target virtual building to obtain globally optimal path information. Based on the coordinates of path inflection points contained in the globally optimal path information, a continuous visual guidance trajectory is generated in the 3D virtual scene and rendered as an overlay layer in the trainees' field of vision. Real-time position information of each trainee in the target virtual building is acquired, and the direction and distance information between each trainee and the target point are calculated and displayed based on the real-time position information. Because this invention integrates digital twins and augmented reality, achieving precise spatial anchoring, it provides an evacuation training solution that combines VR immersion and AR information enhancement, and can be promoted to a large number of people, at a cost far lower than that of independent AR deployment.

[0104] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0105] It is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. An emergency evacuation drill method based on digital twins and augmented reality, characterized in that, include: Based on the three-dimensional structured data of the target building, a target virtual building corresponding to the target building is generated in a three-dimensional virtual scene, and a communication connection is established between the head-mounted display and control device and the three-dimensional virtual scene. The shortest path search algorithm is used to perform a shortest path search under the target virtual building to obtain the global optimal path information; Based on the key control points contained in the global optimal path information, a continuous visual guidance trajectory is generated in the three-dimensional virtual scene, and the visual guidance trajectory is rendered in the display field of view of the trainee in the form of an overlay layer; Obtain the real-time location information of each trainee in the target virtual building, and calculate and display the direction and distance information between each trainee and the target point based on the real-time location information; The key control points are a set of path points; The steps of generating a continuous visual guidance trajectory in the 3D virtual scene based on the key control points contained in the global optimal path information, and rendering the visual guidance trajectory as an overlay layer in the trainee's display field of view include: Obtain the set of path points from the global optimal path information, and pass the path points in the set of path points to the vertex buffer of the offline renderer of the 3D virtual scene one by one; Based on the list of path inflection point coordinates corresponding to the path point set, a visual guidance trajectory is rendered in the three-dimensional virtual scene; wherein, the visual guidance trajectory is the first-person perspective guidance trajectory of each trainee in the virtual scene.

2. The emergency evacuation drill method based on digital twin and augmented reality according to claim 1, characterized in that, The step of generating a target virtual building corresponding to the target building in a three-dimensional virtual scene based on the target building's three-dimensional structured data includes: Obtain the source file of the BIM model of the target building and import the source file of the BIM model into the Unity3D engine, or obtain the 3D point cloud data of the target building and import the 3D point cloud data into the Unity3D engine. The Unity3D engine parses the received source file and generates a visual graphic component for rendering a 3D scene and a physical interaction boundary component that matches the spatial occupancy of the visual graphic component, thus obtaining the target virtual building in the 3D virtual scene.

3. The emergency evacuation drill method based on digital twin and augmented reality according to claim 2, characterized in that, The step of using a shortest path search algorithm to perform a shortest path search under the target virtual building to obtain the globally optimal path information includes: In response to the initialization command of the target virtual building, obtain all object nodes marked as safety exits in the scene, and complete the binding of the safety exits; Obtain the current training level ID and determine the target exit corresponding to the training level ID; Starting from the trainee's current location and ending at the target exit location, via A The algorithm calculates the globally optimal path information on the navigation grid of the target virtual building.

4. The emergency evacuation drill method based on digital twin and augmented reality according to claim 3, characterized in that, The starting point is the current location of the trainee, and the ending point is the location of the target exit, via A. The algorithm's step of calculating the globally optimal path information on the navigation grid of the target virtual building also includes: Judge A Does the planned path output by the algorithm meet the preset valid conditions? When A When the planned path output by the algorithm does not meet the preset valid conditions, a light carpet clearing signal is generated, and the light carpet clearing operation is controlled to terminate the rendering of the current guide path; wherein, the valid conditions include at least: the number of path points is greater than or equal to 2 inflection points.

5. The emergency evacuation drill method based on digital twin and augmented reality according to claim 1, characterized in that, The line renderer is a Line Renderer component; the step of generating a continuous visual guide trajectory in the 3D virtual scene based on the key control points contained in the global optimal path information, and rendering the visual guide trajectory as an overlay layer in the trainee's display field of view, further includes: The vertices of the Line Renderer component are updated at preset time intervals.

6. The emergency evacuation drill method based on digital twin and augmented reality according to claim 1, characterized in that, The steps of obtaining the real-time location information of each trainee in the target virtual building, and calculating and displaying the direction and distance information between each trainee and the target point based on the real-time location information include: Establish a mapping relationship between world coordinates and map coordinates in a 3D virtual scene; The system determines the current floor of the trainee based on the height information of the trainee's current location, and checks whether the detected floor index is the same as the floor currently displayed on the map. If the detected floor index is different from the floor currently displayed on the map, the current map floor is updated to the floor corresponding to the floor index, and the corresponding map data is loaded. Calculate the path length between the current location of the trainee and each candidate exit, and the exit corresponding to the shortest path; Calculate the horizontal angle between the trainee's orientation and the exit direction, and map the horizontal angle to one of multiple standard direction categories according to a preset discretization strategy; wherein, the discretization strategy is to uniformly divide the 360° horizontal circumference into N direction intervals, where N is an integer greater than or equal to 4; The path length between the exit corresponding to the shortest path and the trainee, and the direction label corresponding to the angle between the horizontal plane, are combined to form a prompt display information for direction and distance.

7. The emergency evacuation drill method based on digital twin and augmented reality according to claim 6, characterized in that, Before the step of obtaining the real-time location information of each trainee in the target virtual building, and calculating and displaying the direction and distance information between each trainee and the target point based on the real-time location information, the method further includes: Determine whether the height information of the trainee exceeds a preset height threshold. If the height information is greater than or equal to the preset height threshold, set the display state of the prompt information to display; otherwise, set the display state of the prompt information to hide.

8. The emergency evacuation drill method based on digital twin and augmented reality according to claim 6, characterized in that, The displayed information may be text prompts, 3D floating directional signs, or graphic path display icons.

9. An emergency evacuation drill system based on digital twin and augmented reality, characterized in that, include: The virtual scene conversion module is used to generate a target virtual building corresponding to the target building in a three-dimensional virtual scene based on the three-dimensional structured data of the target building, and to establish a communication connection between the head-mounted display and control device and the three-dimensional virtual scene. The optimal path search module is used to perform a shortest path search algorithm under the target virtual building to obtain the global optimal path information; The trajectory guidance module is used to generate a continuous visual guidance trajectory in the three-dimensional virtual scene based on the key control points contained in the global optimal path information, and to render the visual guidance trajectory in the display field of the trainee in the form of an overlay layer; The key control points are a set of path points; The steps of generating a continuous visual guidance trajectory in the 3D virtual scene based on the key control points contained in the global optimal path information, and rendering the visual guidance trajectory as an overlay layer in the trainee's display field of view include: Obtain the set of path points from the global optimal path information, and pass the path points in the set of path points to the vertex buffer of the offline renderer of the 3D virtual scene one by one; Based on the list of path inflection point coordinates corresponding to the path point set, a visual guidance trajectory is rendered in the three-dimensional virtual scene; wherein, the visual guidance trajectory is the first-person perspective guidance trajectory of each trainee in the virtual scene; The dynamic positioning module is used to acquire the real-time location information of each trainee in the target virtual building, and to calculate and display the direction and distance information between each trainee and the target point based on the real-time location information.

10. A computer storage medium, wherein the storage medium is a computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a computer, enables the computer to perform the emergency evacuation drill method based on digital twins and augmented reality as described in any one of claims 1-8.

Images