A disaster perception and risk control-based evacuation navigation method, device, equipment and medium

Abstract

The application discloses a disaster perception and risk control-based evacuation navigation method and device, equipment and medium. The method comprises determining a pedestrian-level density according to a pedestrian position, and determining a grid-level crowd density of a grid unit based on the pedestrian-level density; constructing a passing time cost for the grid unit based on the grid-level crowd density; constructing a dynamic disaster intensity and a congestion indication for the grid unit based on a disaster arrival time; constructing a navigation cost according to the passing time cost, the dynamic disaster intensity and the congestion indication to determine a global navigation direction; determining a local interaction direction according to the pedestrian-level density; and fusing the global navigation direction and the local interaction direction to update the pedestrian speed and the pedestrian position. The application combines the disaster propagation state and the crowd density change to adjust the navigation cost and the evacuation direction in real time, so that the evacuation navigation path always matches the current disaster evolution state and the crowd distribution state, the overall evacuation efficiency is further improved while the evacuation safety is ensured, and the crowd evacuation scene of various sudden and rapidly evolving disasters can be better adapted.

Description

Technical Field

[0001] This application relates to the field of intelligent navigation technology, and in particular to an evacuation navigation method, device, equipment and medium based on disaster perception and risk control. Background Technology

[0002] The safe evacuation of crowded areas is a critical issue that has always faced in the field of public safety. Efficient and reasonable crowd evacuation strategies can significantly reduce the probability of casualties in emergencies, especially in scenarios involving rapidly evolving disasters such as fires, smoke inhalation, explosions, or structural collapses. However, existing evacuation navigation methods generally guide people to the nearest exit based on geometric distance or fixed travel costs, ignoring the state of disaster propagation. This can lead to situations where navigation paths are blocked by the spread of the disaster, or people mistakenly enter high-risk areas, making it difficult to achieve safe and efficient evacuation navigation.

[0003] Therefore, the existing technology still needs to be improved and enhanced. Summary of the Invention

[0004] The technical problem to be solved by this application is to provide an evacuation navigation method, device, equipment and medium based on disaster perception and risk control, which addresses the shortcomings of the existing technology.

[0005] To address the aforementioned technical problems, the first aspect of this application provides an evacuation navigation method based on disaster perception and risk control, wherein the evacuation navigation method based on disaster perception and risk control specifically includes: Based on geographic information, the target evacuation area is discretized into grid domains; The pedestrian density is determined based on the pedestrian location, and the grid-level crowd density of the grid cell is determined based on the pedestrian density. The passage time cost is constructed for each grid cell based on the grid-level population density. Starting from the disaster source grid cell, the disaster arrival time corresponding to each grid cell is determined, and a dynamic disaster intensity is constructed for the grid cell based on the disaster arrival time and disaster intensity parameters. A blocking indication is determined based on the dynamic disaster intensity, and the blocking indication is used to identify whether the grid cell has lost its passage conditions due to the disaster. Based on the passage time cost, dynamic disaster intensity, and congestion indication, a navigation cost is constructed. A global navigation potential field is constructed with the goal of minimizing the cumulative navigation cost from the grid cell to the exit area, and the global navigation direction is determined based on the global navigation potential field. The pedestrian density is used to determine the direction of local interaction; The pedestrian navigation direction is determined based on the global navigation direction and the local interaction direction, and the pedestrian speed and pedestrian position are updated according to the navigation direction. The process of determining pedestrian density based on pedestrian location will be repeated according to the preset conditions until the evacuation termination conditions are met.

[0006] The evacuation navigation method based on disaster perception and risk control, wherein determining pedestrian density based on pedestrian location specifically includes: Obtain the distance between a pedestrian and its neighbors; The density factor corresponding to each neighboring pedestrian is determined based on the pedestrian distance and the pedestrian projection radius, and the pedestrian-level density is determined based on the density factors of all neighboring pedestrians.

[0007] The evacuation navigation method based on disaster perception and risk control, wherein the time cost of constructing a grid cell based on the grid-level population density specifically includes: Convert the grid-level crowd density into passage speed; A travel time cost is constructed based on the travel speed, and the travel time cost is negatively correlated with the travel speed.

[0008] The evacuation navigation method based on disaster perception and risk control, wherein the step of constructing dynamic disaster intensity for grid cells based on the disaster arrival time and disaster intensity parameters specifically includes: The relative arrival time is determined based on the disaster arrival time of the grid cells and the current time; If the relative arrival time is less than zero, the dynamic disaster intensity is determined based on the relative arrival time and the maximum disaster intensity, intensity factor and time factor in the disaster intensity parameters; If the relative arrival time is greater than or equal to zero, the dynamic disaster intensity is determined based on the relative arrival time and the maximum disaster intensity, intensity factor, and intensity rise rate factor in the disaster intensity parameters.

[0009] The evacuation navigation method based on disaster perception and risk control, wherein the expression for the dynamic disaster intensity is: , in, Indicates dynamic disaster intensity. Indicates the intensity of the disaster at the time of its arrival. Indicates relative arrival time. Indicates the maximum disaster intensity. Indicates the time factor. This represents the rate of increase in intensity.

[0010] In the aforementioned evacuation navigation method based on disaster perception and risk control, the expression for the blockage indication is: , in, Indicates a blockage. Indicates relative arrival time. Indicates dynamic disaster intensity. Indicates the threshold of traffic obstruction intensity. Indicates an indicator function.

[0011] The evacuation navigation method based on disaster perception and risk control, wherein constructing a global navigation potential field with the objective of minimizing the cumulative navigation cost from the grid cell to the exit area, and determining the global navigation direction based on the global navigation potential field specifically includes: The objective function is to minimize the dynamic navigation passage cost of the continuous path from the grid cell to the exit region. The objective function is to solve the objective function to obtain the global navigation potential field of the grid cell. Obtain the potential energy gradient of the global navigation potential field, and normalize the potential energy gradient to obtain the global navigation direction.

[0012] The evacuation navigation method based on disaster perception and risk control, wherein determining the local interaction direction using the pedestrian density specifically includes: Obtain the pedestrian distance between a pedestrian and its neighboring pedestrians, and determine the spatial gradient of the pedestrian relative to its neighboring pedestrians based on the pedestrian distance; Pedestrian pressure is determined based on pedestrian density, and the local interaction force of the pedestrian is determined based on the pedestrian pressure of the pedestrian, the pedestrian pressure of its neighboring pedestrians, and the spatial gradient of the pedestrian relative to its neighboring pedestrians. The local interaction force of the pedestrian is normalized to obtain the local interaction direction of the pedestrian.

[0013] The evacuation navigation method based on disaster perception and risk control, wherein determining the pedestrian navigation direction based on the global navigation direction and the local interaction direction specifically includes: The mixing coefficient is determined based on pedestrian density. The global navigation direction and the local interaction direction are fused based on the hybrid coefficient to obtain the pedestrian navigation direction.

[0014] The evacuation navigation method based on disaster perception and risk control, wherein the update process of pedestrian speed and pedestrian position is as follows: , , in, This indicates the updated pedestrian speed. Indicates navigation direction, Indicates pedestrian density. This indicates the relationship between crowd density and pedestrian speed. Represents the smoothing coefficient. This indicates the updated pedestrian location. This indicates the pedestrian's location before the update. It represents the change over time.

[0015] The second aspect of this application provides an evacuation navigation device based on disaster perception and risk control, wherein the evacuation navigation device based on disaster perception and risk control specifically includes: The discrete module is used to discretize the target evacuation area into a grid domain based on geographic information. The density determination module is used to determine the pedestrian-level density based on the pedestrian location, and to determine the grid-level crowd density of the grid cell based on the pedestrian-level density. The navigation passage cost construction module is used to construct the passage time cost for grid cells based on the grid-level population density. The disaster propagation status construction module is used to determine the disaster arrival time corresponding to each grid cell starting from the disaster source grid cell, and to construct a dynamic disaster intensity for the grid cell based on the disaster arrival time and disaster intensity parameters, and to determine a blocking indication based on the dynamic disaster intensity. The blocking indication is used to identify whether the grid cell has lost its passage conditions due to the disaster. The global navigation direction determination module is used to construct navigation costs based on the passage time cost, dynamic disaster intensity, and congestion indication, construct a global navigation potential field with the goal of minimizing the cumulative navigation cost from the grid cell to the exit area, and determine the global navigation direction based on the global navigation potential field. A local interaction direction determination module is used to determine the local interaction direction using the pedestrian density. The update module is used to determine the pedestrian navigation direction based on the global navigation direction and the local interaction direction, and update the pedestrian speed and pedestrian position according to the navigation direction; The control module is used to re-execute the step of determining pedestrian density based on pedestrian location according to preset conditions until the evacuation termination condition is met.

[0016] A third aspect of this application provides a computer-readable storage medium storing one or more programs that can be executed by one or more processors to implement the steps in the evacuation navigation method based on disaster perception and risk control as described above.

[0017] A fourth aspect of this application provides a terminal device, which includes: a processor and a memory; The memory stores a computer-readable program that can be executed by the processor; When the processor executes the computer-readable program, it implements the steps in the evacuation navigation method based on disaster perception and risk control as described above.

[0018] Beneficial effects: 1. This application determines the disaster arrival time corresponding to each grid cell by taking the disaster source grid cell as the starting point, and constructs dynamic disaster intensity and blockage indication based on the disaster arrival time. Then, the disaster propagation status is incorporated into the dynamic navigation cost through the dynamic disaster intensity and blockage indication. This fully considers the disaster propagation status during the evacuation process, avoids the situation where the navigation path is blocked by the spread of the disaster and people accidentally enter high-risk areas, and achieves safe and efficient evacuation navigation.

[0019] 2. When constructing the dynamic navigation cost, this application also incorporates the decrease in traffic capacity caused by crowd density into the dynamic navigation cost by using the passage time cost constructed by grid-level crowd density. This fully considers the changes in crowd density and the decrease in traffic capacity, avoids continuously introducing a large number of people into the same path, improves the tail evacuation time, and thus effectively shortens the overall evacuation time.

[0020] 3. This application uses disaster arrival time, dynamic disaster intensity, and blockage indication to characterize the entire life cycle of a disaster from "arrival to intensity growth to blockage activation". It can accurately reflect the dynamic impact of disaster propagation status on the passability of evacuation routes. It also adjusts navigation costs and evacuation directions in real time in combination with disaster propagation status and population density changes, so that the evacuation navigation path always matches the current disaster evolution status and population distribution status. While ensuring evacuation safety, it further improves the overall evacuation efficiency and can better adapt to various sudden and rapidly evolving disaster evacuation scenarios.

[0021] 4. This application fuses the global navigation direction determined by the global navigation potential field and the local interaction direction determined by the pedestrian density, and uses the fused navigation direction to update the pedestrian position and pedestrian speed. This allows the navigation direction to simultaneously take into account global path cost and local congestion avoidance, and can adaptively adjust the navigation path to avoid excessive congestion in local areas, further improving the tail evacuation time and thus further shortening the overall evacuation time.

[0022] 5. This application adaptively adjusts the mixing coefficient between the global navigation direction and the local interaction direction based on pedestrian density, so that the local avoidance weight is greater in high-density scenarios and the global path navigation weight is greater in low-density scenarios, which can adapt to evacuation scenarios with different levels of congestion and take into account both evacuation efficiency and evacuation safety.

[0023] 6. This application dynamically updates the crowd density field and the global navigation potential field by dynamically updating pedestrian positions and speeds. It can respond to changes in crowd distribution in real time, adjust the global navigation direction, avoid the problem of local congestion accumulation caused by static path planning, and ensure the passage efficiency during the evacuation process. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart of an evacuation navigation method based on disaster perception and risk control provided in an embodiment of this application.

[0026] Figure 2 This is a schematic diagram of region A.

[0027] Figure 3 A schematic diagram of an evacuation navigation device based on disaster perception and risk control provided in an embodiment of this application.

[0028] Figure 4 A schematic block diagram of the terminal device provided in the embodiments of this application. Detailed Implementation

[0029] This application provides an evacuation navigation method, apparatus, equipment, and medium based on disaster perception and risk control. To make the objectives, technical solutions, and effects of this application clearer and more explicit, the following detailed description is provided with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only for explaining this application and are not intended to limit this application.

[0030] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0031] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0032] It should be understood that the sequence number and size of each step in this embodiment do not imply the order of execution. The execution order of each process is determined by its function and internal logic, and should not constitute any limitation on the implementation process of this application embodiment.

[0033] Research has shown that the safe evacuation of crowded groups is a persistent and critical issue in public safety. Efficient and rational crowd evacuation strategies can significantly reduce the probability of casualties in emergencies. Existing evacuation navigation methods are mostly based on static shortest paths or static potential fields. These methods guide people to the nearest exit based on geometric distance or fixed travel costs. While the calculations are simple, they neglect changes in crowd density and reduced traffic capacity during evacuation. In high-density scenarios, these methods can easily lead to a continuous flow of people along the same path, causing localized overcrowding. This not only significantly prolongs the evacuation time for those at the back of the queue but also extends the overall evacuation duration.

[0034] To address the aforementioned issues, in this embodiment, the target evacuation area is discretized into grid domains based on geographic information; pedestrian density is determined based on pedestrian location, and grid-level crowd density of grid cells is determined based on the pedestrian density; a passage time cost is constructed for grid cells based on the grid-level crowd density; the disaster arrival time corresponding to each grid cell is determined starting from the disaster source grid cell, and a dynamic disaster intensity is constructed for the grid cells based on the disaster arrival time and disaster intensity parameters, and a blockage indication is determined based on the dynamic disaster intensity, the blockage indication being used to identify whether the grid cell has lost its passage conditions due to the disaster; a navigation cost is constructed based on the passage time cost, dynamic disaster intensity, and blockage indication, and a global navigation potential field is constructed with the goal of minimizing the cumulative navigation cost from the grid cell to the exit area, and a global navigation direction is determined based on the global navigation potential field; a local interaction direction is determined using the pedestrian density; a pedestrian navigation direction is determined based on the global navigation direction and the local interaction direction, and pedestrian speed and pedestrian location are updated according to the navigation direction; the step of determining pedestrian density based on pedestrian location is re-executed according to preset conditions until the evacuation termination conditions are met. This application uses disaster arrival time, dynamic disaster intensity, and blockage indicators to characterize the entire life cycle of a disaster, from arrival to intensity growth to blockage activation. It can accurately reflect the dynamic impact of disaster propagation status on the accessibility of evacuation routes. Furthermore, it adjusts navigation costs and evacuation directions in real time by combining disaster propagation status and population density changes, ensuring that evacuation navigation routes always match the current disaster evolution status and population distribution status. This not only guarantees evacuation safety but also further improves overall evacuation efficiency, making it better suited for various sudden and rapidly evolving disaster evacuation scenarios.

[0035] It should be noted that the executing entity of each embodiment of this application can be a computing service system with data processing, network communication, and program execution functions, such as an electronic system capable of realizing the above functions, an evacuation navigation system based on disaster perception and risk control, etc. The following uses an evacuation navigation system based on disaster perception and risk control (hereinafter referred to as "the system") as an example to describe the following embodiments.

[0036] The application content will be further explained below with reference to the accompanying drawings and the description of the embodiments.

[0037] This embodiment provides an evacuation navigation method based on disaster perception and risk control. This method can run on terminal devices, servers, simulation platforms, or emergency command systems for crowd evacuation in densely populated residential areas, large commercial complexes, open high-density gathering spaces (such as squares, stations, airport terminals, etc.), underground spaces, subway platforms and transfer passages, civil defense projects, large parks, campuses, cultural and sports venues, and large exhibition venues. Figure 1As shown, the evacuation navigation method based on disaster perception and risk control specifically includes: S10. Discretize the target evacuation area into a grid domain based on geographic information.

[0038] Specifically, geographic information is used to reflect the spatial distribution, boundary locations, and locations of obstacles and exits of the target evacuation area. For example, the geographic information of the target evacuation area can be determined by collecting spatial data such as building outlines, roads and alleys, open spaces, exit boundaries, and obstacles. Then, based on the geographic information of the target evacuation area, the entire area can be divided into grids according to a preset grid size, converting the continuous real space into a discrete set of grid cells to obtain a grid domain. After the division is completed, each grid cell is labeled with an area identifier according to the actual spatial attributes corresponding to each grid cell. If the location of the grid cell is an obstacle, it is marked as an obstacle area; if the location is an evacuation exit, it is marked as an exit area; and the remaining pedestrian-accessible locations are marked as passable areas.

[0039] Therefore, the grid domain comprises multiple grid cells, each of which is labeled with a region identifier based on its corresponding actual spatial attributes. This region identifier can be one of three: a passable region, an obstacle region, or an exit region. For example, the target region can be discretized into spatial resolutions. (e.g., 0.1m) grid area , raster domain The center point of each grid cell in the array is denoted as . The area corresponding to the grid cell is identified as a passable area. Obstacle areas or export area .

[0040] S20. Determine the pedestrian-level density based on the pedestrian location, and determine the grid-level crowd density of the grid unit based on the pedestrian-level density.

[0041] Specifically, pedestrian-level density reflects the degree of local congestion within the range of an individual pedestrian's movement, while grid-level crowd density reflects the degree of crowd aggregation within the area covered by a grid cell. In other words, for each pedestrian, the pedestrian-level density is first determined based on their current location (i.e., pedestrian position). Then, for each grid cell, all pedestrians falling within that grid cell are counted, and the grid-level crowd density of that grid cell is determined based on the pedestrian-level density of all pedestrians within that grid cell.

[0042] In one embodiment, determining pedestrian density based on pedestrian location specifically includes: Obtain the distance between a pedestrian and its neighbors; The density factor corresponding to each neighboring pedestrian is determined based on the pedestrian distance and the pedestrian projection radius, and the pedestrian-level density is determined based on the density factors of all neighboring pedestrians.

[0043] Specifically, neighboring pedestrians are other pedestrians within the current pedestrian's preset influence range. For each current pedestrian, all neighboring pedestrians within its preset influence range are traversed first, and then the pedestrian distance between the current pedestrian and each neighboring pedestrian is calculated. This pedestrian distance can be the Euclidean distance between the current pedestrian's position and the positions of its neighboring pedestrians, etc.

[0044] After obtaining the pedestrian distance, pedestrians can be treated as SPH particles. The density factor for each neighboring pedestrian is determined using the Poly6 kernel function. Then, the density factors of all neighboring pedestrians are weighted and summed to obtain the pedestrian-level density. The expression for the pedestrian-level density can be: , , in, pedestrian pedestrian density, pedestrian pedestrian positions pedestrian Nearby pedestrians pedestrian positions Indicates the radius of the pedestrian's projection (e.g., 1.5m). This refers to the Poly6 kernel function. pedestrian and nearby pedestrians The distance between pedestrians Indicates nearby pedestrians Particle weights, such as, =1 etc.

[0045] It should be noted that in practical applications, after obtaining the pedestrian distance, other methods can be used to determine the pedestrian-level density. For example, the ratio of the pedestrian's projected radius to the pedestrian distance can be used as the density factor for the corresponding neighboring pedestrians, and then the density factors of all neighboring pedestrians can be summed to determine the pedestrian-level density of the current pedestrian. Alternatively, the pedestrian distance can be omitted, and other methods can be used directly to determine the pedestrian-level density. For instance, the number of pedestrians within a preset radius around the pedestrian can be directly obtained, and then the pedestrian-level density corresponding to that pedestrian can be calculated by combining this number with the area of ​​that radius.

[0046] This application treats pedestrians as SPH particles and uses the Poly6 kernel function to estimate the pedestrian density, which can more accurately reflect the actual level of congestion around pedestrians. Compared with the method of directly counting the number of pedestrians within a preset range, it can better match the actual feeling of individual pedestrians about the surrounding congestion, and provide a more accurate basis for subsequent global navigation direction and local interaction direction.

[0047] Furthermore, in step S20 above, after obtaining the pedestrian-level density of all pedestrians, for each grid cell, the pedestrian-level density of all pedestrians falling into that grid cell can be projected onto the center point of the grid cell to obtain the grid-level crowd density of that grid cell. The grid-level crowd density can be expressed as: , in, Represents grid-level population density. This represents the set of pedestrians that fall into a grid cell.

[0048] Of course, in practical applications, other methods can be used to determine the grid-level crowd density, such as taking the average pedestrian density of all pedestrians falling into the grid cell as the grid-level crowd density of that grid cell. Furthermore, if there are no pedestrians within a grid cell, the grid-level crowd density of that grid cell is recorded as zero.

[0049] S30. Construct the passage time cost for the grid cell based on the grid-level population density.

[0050] Specifically, the passage time cost corresponding to a grid cell is positively correlated with the grid-level crowd density. That is, the higher the grid-level crowd density of a grid cell, the higher the congestion level of the grid cell area, the higher the passage cost for pedestrians to pass through the grid cell, and the larger the corresponding passage time cost value.

[0051] In one embodiment, the step of constructing the passage time cost for the grid cell based on the grid-level population density specifically includes: Convert the grid-level crowd density into passage speed; A travel time cost is constructed based on the travel speed, and the travel time cost is negatively correlated with the travel speed.

[0052] Specifically, the passage speed reflects the rate at which pedestrians move within a grid cell. The passage speed is negatively correlated with the grid-level crowd density to reflect the decrease in throughput under high density. For example, a pre-defined speed-density negative correlation model (such as the Green Shields model, or a specific speed-density curve calibrated based on measured data from the target evacuation area) can be used to convert the grid-level crowd density into the passage speed of the corresponding grid cell. In this application, the passage speed can be expressed as:

[0053] in, Indicates the passage speed of the grid cells. Indicates free walking speed, Indicates the baseline density. Indicates the shape index, Represents grid-level population density. , and The empirical fundamental graph can be determined by fitting the constrained nonlinear least squares method.

[0054] Furthermore, after obtaining the traffic speed, a traffic time cost can be constructed based on the traffic speed, which is negatively correlated with the traffic speed. For example, the reciprocal of the traffic speed can be used as the traffic time cost, or the travel time required for a pedestrian to pass through a grid cell can be calculated based on the traffic speed and the geometric dimensions of the grid cell, and this travel time can be used as the traffic time cost. In this embodiment of the application, the traffic time cost is expressed as: , in, This represents the travel time cost of a raster cell. This represents a numerically stable term to avoid calculation anomalies where the denominator is zero when the passage speed is zero. It can usually be taken as a very small positive value.

[0055] This application embodiment utilizes grid-level crowd density to construct navigation passage cost for grid cells, using grid-level crowd density (i.e., crowd density in grid cells) as a time-varying physical field to drive the global navigation potential field. This allows the navigation cost to dynamically change with the degree of crowd congestion, avoiding the navigation system from continuously allocating pedestrians to paths where crowds have already gathered. It balances pedestrian traffic on different paths at the global level, providing a reasonable dynamic cost basis for subsequent solutions to the optimal global navigation potential field.

[0056] Meanwhile, in the process of constructing the passage time cost using grid-level crowd density as grid cell, the embodiments of this application first map the continuous grid-level crowd density field into a pedestrian velocity field, and then use the pedestrian velocity field to construct the passage time cost, thus realizing a physically consistent mapping pipeline from continuous grid-level crowd density → velocity field → passage time cost (i.e., slowness cost field), and using the passage time cost as the standard input interface for solving the navigation potential energy.

[0057] It is understood that the core point to be protected in this application embodiment is the physically consistent mapping pipeline from continuous grid-level crowd density → velocity field → travel time cost (i.e., slowness cost field). The calculation method of continuous grid-level crowd density, the mapping relationship from continuous grid-level crowd density to velocity field, and the mapping relationship from velocity field to travel time cost can be adjusted according to actual needs. The calculation method of continuous grid-level crowd density, the mapping relationship from continuous grid-level crowd density to velocity field, and the mapping relationship from velocity field to travel time cost given above are typical implementation methods provided by this application embodiment.

[0058] S40. Starting from the disaster source grid cell, determine the disaster arrival time corresponding to each grid cell, and construct a dynamic disaster intensity for the grid cell based on the disaster arrival time and disaster intensity parameters, and determine the blockage indication based on the dynamic disaster intensity.

[0059] Specifically, the disaster arrival time is the time it takes for a disaster to reach the area covered by the grid cell. This arrival time can be calculated based on the disaster type, its propagation characteristics, and the spatial environment of the grid cell. For example, for a spreading disaster, the arrival time for each grid cell can be calculated based on the disaster propagation speed and path length. For a spreading fire, a fire spread simulation algorithm can be used, combined with environmental parameters such as the distribution of combustibles and ventilation conditions within the area, to simulate the time it takes for each grid cell to be covered by the fire; this is the disaster arrival time for that grid cell.

[0060] In this embodiment, the Eikonal equation is used to determine the disaster arrival time corresponding to each grid cell. Specifically, the Eikonal equation can be expressed as: , , in, Indicates the time of arrival of the disaster. The gradient representing the arrival time of a disaster indicates the rate of change of the arrival time of a disaster in space. Represents grid cells The speed of disaster spread at the location Represents a raster domain, This represents the set of grid cells containing disaster sources (such as the grid cells containing breaches, ignition points, and earthquake sources).

[0061] Then, the Eikonal equation is solved using the fast travel method (eight-neighbor recursion) to obtain the shortest propagation time from the disaster front to each grid cell. Based on the shortest propagation time corresponding to the grid cell and the current time, the disaster arrival time corresponding to the grid cell can be determined, providing a disaster local clock for the grid cell.

[0062] Furthermore, in step S40 above, after obtaining the disaster arrival time, the relative arrival time of the disaster to the grid cell can be determined based on the disaster arrival time and the current time. The relative arrival time is the time difference between the current time and the disaster arrival time. The relative arrival time can reflect the degree of impact of the disaster on the grid cell. The dynamic disaster intensity of the grid cell can be determined by using the relative arrival time and the preset disaster intensity parameters. The dynamic disaster intensity can be used to construct a blocking indication to identify whether the grid cell has lost its passage conditions due to the disaster.

[0063] For example, the step of constructing dynamic disaster intensity for grid cells based on the disaster arrival time and disaster intensity parameters specifically includes: The relative arrival time is determined based on the disaster arrival time of the grid cells and the current time; If the relative arrival time is less than zero, the dynamic disaster intensity is determined based on the relative arrival time and the maximum disaster intensity, intensity factor and time factor in the disaster intensity parameters; If the relative arrival time is greater than or equal to zero, the dynamic disaster intensity is determined based on the relative arrival time and the maximum disaster intensity, intensity factor, and intensity rise rate factor in the disaster intensity parameters.

[0064] Specifically, the relative arrival time is the difference between the current time and the disaster arrival time, and it is used as an independent variable to construct the dynamic disaster intensity. When the relative arrival time is less than zero, it means that the disaster has not yet reached the grid cell, but there may be pre-arrival exposure caused by diffusion, leakage, or preheating. When the relative arrival time is greater than or equal to zero, it means that the disaster has reached the grid cell, the grid cell has been affected by the disaster, and the disaster intensity gradually increases over time and approaches the maximum disaster intensity.

[0065] Therefore, zero can be used as a dividing point. When the relative arrival time is less than zero, the dynamic disaster intensity can be determined directly using the relative arrival time, maximum disaster intensity, intensity factor, and time factor. The time factor can be used to control the intensity attenuation or leakage impact during the pre-arrival stage. When the relative arrival time is greater than or equal to zero, the dynamic disaster intensity can be determined using the relative arrival time, maximum disaster intensity, intensity factor, and intensity rise rate factor. The intensity rise rate factor can be used to control the intensity state of the disaster after arrival. Among them, the intensity factor is used to control the intensity level at the moment of arrival of the disaster front (i.e., the intensity at the moment of disaster arrival) so that the intensity at the moment of disaster arrival is between zero and the maximum disaster intensity.

[0066] It is understandable that disaster intensity parameters include at least the maximum disaster intensity, intensity factor, time factor, and intensity escalation rate factor. The intensity factor controls the intensity level at the moment the disaster front arrives (i.e., the intensity at the time of disaster arrival), ensuring that the intensity at the time of arrival is between zero and the maximum disaster intensity. The time factor controls the time scale of intensity decay or leakage effects during the pre-arrival phase; a larger time factor indicates more significant early exposure before the disaster front arrives. The intensity escalation rate factor controls the rate of intensity escalation after arrival; a smaller intensity escalation rate factor indicates that the disaster enters a high-intensity state more quickly after arrival.

[0067] Furthermore, the disaster intensity parameter can be determined based on the disaster type. That is, when constructing dynamic disaster intensity, the disaster type can be obtained first, then the disaster intensity parameter can be obtained based on the disaster type, and finally the dynamic disaster intensity of the grid cell can be determined based on the disaster intensity parameter and the relative arrival time. The disaster intensity parameters corresponding to different disaster types can be pre-calibrated. For example, combustible gas diffusion, flue gas, thermal radiation, and toxic gases can each correspond to different maximum disaster intensities, time factors, etc., thereby reflecting the differences in propagation speed, pre-arrival impact, intensity growth, and blocking effects of different disaster media.

[0068] In one embodiment, the expression for the dynamic disaster intensity is: , = · , in, Indicates dynamic disaster intensity. Indicates the intensity of the disaster at the time of its arrival. Indicates relative arrival time. Indicates the maximum disaster intensity. Indicates the time factor. Indicates the rate of increase in intensity. This represents the intensity factor, used to ensure that the intensity at the time of disaster arrival is between 0 and the maximum disaster intensity.

[0069] Furthermore, after a disaster reaches a grid cell, the cell may become impassable due to the disaster. Therefore, a blocking indicator can be set based on the dynamic disaster intensity. If the dynamic disaster intensity exceeds a preset passage blocking threshold, the grid cell is marked as impassable; otherwise, it is marked as passable. This allows for dynamic updates of passage conditions to adapt to changes in passage space during disaster development. The expression for the blocking indicator is: , , in, Indicates a blockage. Indicates relative arrival time. Indicates dynamic disaster intensity. Indicates the threshold of traffic obstruction intensity. Indicates an indicator function, Indicates the intensity of the disaster at the time of its arrival. Indicates the maximum disaster intensity.

[0070] Simultaneously, when the intensity crosses the blocking threshold, the blocking indication of the grid cell is activated, and the blocking activation time of the grid cell can be expressed as: , in, Indicates the moment of blocking activation.

[0071] S50. Based on the passage time cost, dynamic disaster intensity and congestion indication, construct navigation cost, construct a global navigation potential field with the goal of minimizing the cumulative navigation cost from the grid cell to the exit area, and determine the global navigation direction based on the global navigation potential field.

[0072] Specifically, the navigation cost is a weighted sum of travel time cost, dynamic disaster intensity, and congestion indication. The disaster term is directly embedded into the global navigation potential field through dynamic disaster intensity, eliminating the need for post-event penalties or barrier masks, and avoiding signal conflicts between travel time cost, dynamic disaster intensity, and congestion indication. Simultaneously, for grid cells already marked as impassable, the weight coefficient of the congestion indication is directly set to infinity, prohibiting paths from passing through that grid cell.

[0073] For example, navigation cost can be expressed as: , in, Indicates navigation cost, This represents the navigation passage cost of a grid cell. This represents the efficiency weight, used to adjust the degree of impact of crowd density on navigation costs. Indicates dynamic disaster intensity. Indicates safety weight, Indicates a blockage. This indicates the blocking weight.

[0074] Furthermore, in step S40 above, after obtaining the navigation cost of each grid cell, the global navigation potential field of all grid cells corresponding to the exit area can be set to 0. With the minimum cumulative navigation cost from other grid cells to the exit area as the optimization objective, the global navigation potential field of the entire grid domain is obtained by the fast travel method. The global navigation potential of each grid cell in the global navigation potential field is equal to the minimum cumulative navigation cost from that grid cell to the exit area.

[0075] Based on this, the construction of a global navigation potential field with the objective of minimizing the cumulative navigation cost from the grid cell to the exit region, and the determination of the global navigation direction based on the global navigation potential field, specifically includes: The objective function is to minimize the dynamic navigation passage cost of the continuous path from the grid cell to the exit region. The objective function is to solve the objective function to obtain the global navigation potential field of the grid cell. Obtain the potential energy gradient of the global navigation potential field, and normalize the potential energy gradient to obtain the global navigation direction.

[0076] Specifically, the objective function can be expressed as: , in, Represents the global navigation potential field of the grid cell. Indicates a continuous path from a grid cell to the exit area. Indicates the export region. This represents the navigation passage cost of a grid cell. Indicates efficiency weight. Indicates dynamic disaster intensity. Indicates safety weight, Indicates a blockage. This indicates the blocking weight.

[0077] Then, the minimum cumulative navigation cost for each grid cell can be obtained using a pre-defined algorithm (such as Bellman optimality recursion, fast travel, fast scan, Dijkstra's algorithm, etc.), forming a discretized global navigation potential field. The discretized global navigation potential field can be expressed as: , , in, Represents the global navigation potential field of the grid cell. Indicates the export region. Indicates a passable area. Represents the set of eight neighboring raster cells of a raster cell. Represents the neighboring raster cells of a raster cell. This represents the global navigation potential field of the neighboring grid cells. Represents grid cells Its neighboring grid cells The right to the border, This indicates the distance between a grid cell and its neighboring grid cells.

[0078] It is understandable that grid cells Its neighboring grid cells The edge weights are non-negative, and this recursion is equivalent to finding the shortest path in the grid domain, which can be solved by label correction, fast traversal, fast scan, or Dijkstra's algorithm. Using an eight-neighborhood (Moore's neighborhood) approximates isotropic propagation better than a four-neighborhood and is more stable in narrow and irregular alleyways.

[0079] Finally, the potential energy gradient of the grid cell is calculated using the discretized global navigation potential field (e.g., approximated by central difference on the grid cell). Since the navigation potential energy increases with the distance cost from the grid to the exit, the opposite direction of the potential energy gradient is the minimum cost direction pointing to the exit. This minimum cost direction is then normalized to obtain the global navigation direction of the grid cell, and this global navigation direction is used as the global navigation direction for all pedestrians included in that grid cell. The global navigation direction can be expressed as: , in, Indicates the global navigation direction of the grid cell. This represents the potential energy gradient of a grid cell. This represents a numerically stable term to avoid calculation anomalies where the denominator is zero when the passage speed is zero. It can usually be taken as a very small positive value.

[0080] This application's embodiments determine the global navigation direction of pedestrians by solving the global navigation potential field with the objective of minimizing the cumulative navigation cost from the grid cell to the exit area, thus achieving the solution of the optimal density sensing potential field on the global grid domain. Furthermore, it can directly obtain the global navigation direction of complex physical spaces (such as building complexes, indoor spaces, plazas, underground spaces, and gridded areas) that can be expressed by passable areas, obstacle areas, exit areas, grid-based crowd density, and communication time costs, without requiring redesign of the solution logic. This results in stronger adaptability and higher solution efficiency, providing stable global navigation direction support for the rapid evacuation of large-scale crowds.

[0081] S60. Determine the local interaction direction using the pedestrian density.

[0082] Specifically, the local interaction direction is used to reflect the mutual avoidance relationship between pedestrians and other pedestrians around them. When determining the local interaction direction for the current pedestrian based on the pedestrian density, the self-organizing law of crowd movement can be used to calculate the direction that needs to be avoided based on the distribution of nearby pedestrians, so as to balance the local crowd congestion and avoid collisions and congestion between pedestrians.

[0083] In one embodiment, determining the local interaction direction using the pedestrian-level density specifically includes: Obtain the pedestrian distance between a pedestrian and its neighboring pedestrians, and determine the spatial gradient of the pedestrian relative to its neighboring pedestrians based on the pedestrian distance; Pedestrian pressure is determined based on pedestrian density, and the local interaction force of the pedestrian is determined based on the pedestrian pressure of the pedestrian, the pedestrian pressure of its neighboring pedestrians, and the spatial gradient of the pedestrian relative to its neighboring pedestrians. The local interaction force of the pedestrian is normalized to obtain the local interaction direction of the pedestrian.

[0084] Specifically, the spatial gradient is used to reflect the changing trend of a pedestrian's spatial position. For the current pedestrian, all neighboring pedestrians within its preset perception range are traversed, and the pedestrian distance from the current pedestrian to each neighboring pedestrian is calculated. The spatial gradient is then determined based on the pedestrian distance.

[0085] In one implementation, the Spiky kernel function can be used to determine the spatial gradient, ensuring a stronger repulsive gradient in high-density regions. This spatial gradient can be expressed as: , , in, pedestrian Compared to nearby pedestrians Spatial gradient, This refers to the Spiky kernel function. Indicates the radius of the pedestrian's projection (e.g., 1.5m). pedestrian and nearby pedestrians The distance between pedestrians pedestrian pedestrian positions pedestrian Nearby pedestrians The location of pedestrians.

[0086] Of course, in practical applications, other methods can also be used to determine the spatial gradient, such as directly calculating the distance gradient based on the spatial position difference. This application does not limit the specific calculation method.

[0087] Furthermore, after obtaining the spatial gradient, the pedestrian pressure corresponding to each pedestrian can be calculated based on the pedestrian-level density. The higher the pedestrian-level density, the greater the local congestion pressure experienced by the pedestrian, and the greater the calculated pedestrian pressure value. The correspondence between pedestrian-level density and pedestrian pressure can be defined as follows: , , in, pedestrian pedestrian pressure, pedestrian pedestrian density, Indicates pressure stiffness (e.g., 500, etc.). This represents the pressure activation density threshold (e.g., 1.0 ped / m²), below which the repulsive force is zero. This indicates a positive truncation operator.

[0088] It is understandable that, based on the correspondence between pedestrian density and pedestrian pressure, the pedestrian pressure of each pedestrian in the grid can be obtained, thus enabling the identification of pedestrians. Pedestrian pressure on nearby pedestrians.

[0089] After obtaining pedestrian pressure and spatial gradient, the pedestrian pressure and spatial gradient are combined to determine the local interaction force. Specifically, the repulsive force exerted on the current pedestrian by neighboring pedestrians increases with the product of pedestrian pressure and spatial gradient. Therefore, the initial local interaction force of neighboring pedestrians on the current pedestrian can be determined first based on the pedestrian pressure of the current pedestrian and the pedestrian pressure of neighboring pedestrians. Then, this initial local interaction force is multiplied by the spatial gradient to obtain the intermediate local interaction force of neighboring pedestrians on the current pedestrian. Finally, the intermediate local interaction forces of all neighboring pedestrians corresponding to the current pedestrian are fused to determine the local interaction force of the current pedestrian. For example, consider pedestrians... As a pedestrian at this moment, then pedestrian The local interaction force can be expressed as: , in, pedestrian Local interaction forces, pedestrian pedestrian pressure, Indicates nearby pedestrians pedestrian pressure, pedestrian pedestrian density, Indicates nearby pedestrians pedestrian density, pedestrian Compared to nearby pedestrians Spatial gradient, Indicates nearby pedestrians Particle weights, such as, =1 etc.

[0090] Finally, the local interaction force is normalized to obtain the current pedestrian's local interaction direction. This local interaction direction accurately reflects the direction the pedestrian needs to avoid based on the distribution of nearby pedestrians, thus meeting the needs of local collision avoidance and alleviating local congestion. This local interaction direction can be represented as: , in, pedestrian The direction of local interaction, This represents a numerically stable term to avoid calculation anomalies where the denominator is zero when the passage speed is zero. It can usually be taken as a very small positive value.

[0091] It should be noted that in practical applications, other methods can also be used to obtain the local interaction direction. For example, the position vectors of all other pedestrians within the current pedestrian's field of vision relative to the current pedestrian can be obtained, and the local repulsion direction can be calculated by combining the mutual influence weights of the current pedestrian and other pedestrians. This local repulsion direction can then be used as the local interaction direction.

[0092] This application combines pedestrian density, pedestrian pressure, and spatial gradient to determine the local interaction direction. It can accurately reflect the avoidance needs of pedestrians and surrounding pedestrians based on the density environment in which each pedestrian is located, better fit the self-organizing characteristics of crowd movement, effectively avoid collisions between pedestrians, alleviate the degree of congestion in local areas, and provide accurate local interaction basis for subsequent generation of navigation directions that take into account both global evacuation efficiency and local traffic safety.

[0093] S70. Determine the pedestrian navigation direction based on the global navigation direction and the local interaction direction, and update the pedestrian speed and pedestrian position according to the navigation direction.

[0094] Specifically, the navigation direction is obtained by fusing the global navigation direction and the local interaction direction. The global navigation direction is used to guide pedestrians toward the exit to ensure the overall efficiency of evacuation, while the local interaction direction is used to guide pedestrians to avoid surrounding congestion and collisions to ensure the safety of local passage. By fusing the global navigation direction and the local interaction direction, the final pedestrian navigation direction that takes into account both global efficiency and local safety can be obtained.

[0095] In one embodiment, determining the pedestrian navigation direction based on the global navigation direction and the local interaction direction specifically includes: The mixing coefficient is determined based on pedestrian density. The global navigation direction and the local interaction direction are fused based on the hybrid coefficient to obtain the pedestrian navigation direction.

[0096] Specifically, the hybrid coefficient is used to determine the weights of the global navigation direction and the local interaction direction, so as to balance global navigation and local avoidance in different density environments. This allows pedestrians to follow the global optimal path when the pedestrian density is low, and be dominated by local repulsion forces when the pedestrian density is high, thus naturally reproducing the behavior of being unable to move against the flow of people under real congestion.

[0097] It's understandable that when pedestrian density is high, crowd congestion is high, and the need for local avoidance increases, requiring a higher weighting of local interaction directions to make pedestrians primarily driven by local repulsion forces. Conversely, when pedestrian density is low, crowd congestion is low, and the need for local avoidance is low, allowing a higher weighting of the global navigation direction to ensure pedestrians follow the globally optimal path. Therefore, when determining the mixing coefficient based on pedestrian density, it's possible to control the mixing coefficient to be positively correlated with pedestrian density. Furthermore, when fusing the global navigation direction and local interaction directions based on the mixing coefficient, the weighting coefficient of the local interaction direction can be controlled to be positively correlated with the mixing coefficient.

[0098] In one embodiment, the mixing coefficient can be expressed as: , in, Represents the mixing coefficient. This represents the truncation function. Indicates pedestrian density. This represents a low-density threshold (used as a global dominant boundary for low density, such as 0.5 ped / m²). This indicates a high-density threshold (used for high-density local dominant boundaries, such as 3.0 ped / m²). and The transition zone between free flow and congested flow can be calibrated according to the basic diagram.

[0099] Furthermore, after obtaining the blending coefficients, they can be directly used as weighting coefficients for local interaction directions, and the difference between 1 and the blending coefficients can be used as weighting coefficients for global navigation directions. Accordingly, pedestrian navigation directions can be represented as: , in, Indicates pedestrian navigation direction, Represents the mixing coefficient. Indicates the global navigation direction. Indicates the direction of local interaction.

[0100] This application embodiment adaptively adjusts the mixing coefficient by pedestrian density, thereby adaptively balancing the weights of global navigation direction and local interaction direction. This can adapt to the actual needs of crowd evacuation under different levels of congestion, avoid pedestrians forcing their way through crowds in high-density congested environments and causing congestion, and further improve the safety and smoothness of the evacuation process.

[0101] Furthermore, in step S60 above, after obtaining the pedestrian navigation direction, the pedestrian speed and the pedestrian position at the next moment can be updated based on this navigation direction to complete the navigation path update for the current moment. The update process for the pedestrian speed and pedestrian position is as follows: , , in, This indicates the updated pedestrian speed. Indicates navigation direction, Indicates pedestrian density. This indicates the relationship between crowd density and pedestrian speed. Represents the smoothing coefficient. This indicates the updated pedestrian location. This indicates the pedestrian's location before the update. It represents the change over time.

[0102] This application embodiment obtains the final navigation direction by integrating the global navigation direction and the local interaction direction. This not only ensures that pedestrians move efficiently toward the exit as a whole, improving the overall efficiency of large-scale crowd evacuation, but also meets the needs of pedestrians to avoid local congestion and collisions. It balances the global efficiency of evacuation with the safety of local passage, solving the problem that traditional evacuation navigation methods cannot balance global path optimization and local collision avoidance.

[0103] S80. Re-execute the step of determining pedestrian density based on pedestrian location according to the preset conditions until the evacuation termination conditions are met.

[0104] Specifically, the preset conditions are pre-set to control the cycle of pedestrian navigation updates. For example, they can be set to trigger pedestrian position and speed updates every preset time interval, or to recalculate pedestrian density after each position update, and then trigger updates when the pedestrian density exceeds a preset threshold. The evacuation termination condition can be set to all pedestrians having reached the exit, or the evacuation time reaching a preset maximum duration. During each round of pedestrian position and speed updates, the grid-level pedestrian density and pedestrian-level density are recalculated based on the latest pedestrian position, and the global navigation direction and local interaction direction are redefined. This updates the navigation direction, pedestrian speed, and pedestrian position, enabling dynamic updates during the evacuation process. This allows for real-time matching of density distribution changes caused by changes in crowd position, ensuring the navigation direction always adapts to the current crowd distribution and continuously guaranteeing the efficiency and safety of the evacuation process. The preset condition, triggering pedestrian position and speed updates every preset time interval, can be determined using an inflection point detection method. Specifically, this is achieved by monitoring the tail end of the evacuation time. Inflection point detection is performed on the logarithmic scaling curve of the potential field recalculation cost to determine the preset time interval. The preset time interval can cover a range of [10s, 240s], for example, the preset time interval is 60s.

[0105] In one embodiment, during the evacuation navigation method based on disaster perception and risk control, the evacuation effect can also be evaluated. This allows staff to understand the strategic effectiveness of the evacuation navigation method based on disaster perception and risk control based on the evaluation results, and to determine whether to change the evacuation navigation method based on the strategic effectiveness. Specifically, this application embodiment decomposes disaster risk into two dimensions: a cumulative soft risk exposure dimension and a hard injury event count dimension. Then, the strategic effectiveness of the evacuation navigation method is evaluated by combining the cumulative soft risk exposure dimension and the hard injury event count dimension.

[0106] Among them, the cumulative dimension of soft risk exposure is based on pedestrians. A population soft risk burden is constructed based on the cumulative hazard exposure along the evacuation trajectory. This soft risk burden quantifies the cumulative continuous exposure of all pedestrians to hazard intensity, reflecting the overall risk burden from sustained low-intensity exposure that does not result in serious injury. This population soft risk burden is expressed as: ,

[0107] in, pedestrian Cumulative disaster exposure along their evacuation trajectory pedestrian In time The dynamic disaster intensity at the pedestrian's location. pedestrian The cumulative walking time along their evacuation route. Indicates the group's soft risk burden. This represents a group of pedestrians.

[0108] The hard injury event count is the number of pedestrians whose peak hazard intensity exceeds the severe injury threshold. It is used to quantify the number of hard injury events that cause irreversible serious consequences. This hard injury event count is expressed as: , in, Indicates the count of hard damage events. Indicates the threshold for severe injury. pedestrian In time The dynamic disaster intensity at the pedestrian's location. pedestrian The cumulative walking time along their evacuation route.

[0109] Furthermore, after obtaining the group's soft risk burden and hard injury event counts, these can be evaluated separately. If either the group's soft risk burden or the hard injury event count is unsatisfactory, the navigation strategy can be adjusted to re-optimize the parameters of the overall situation field calculation, further improving the disaster risk control effect during evacuation. In practical applications, the group's soft risk burden and hard injury event counts can also be used to determine the strategy score. For example, the group's soft risk burden score and hard injury event count score can be determined separately based on the group's soft risk burden score and hard injury event count score, and finally, the group's soft risk burden score and hard injury event count score are weighted to obtain the strategy score. Then, when the score exceeds a preset risk threshold, the navigation strategy can be adjusted to re-optimize the parameters of the overall situation field calculation, further improving the disaster risk control effect during evacuation. By jointly evaluating the two dimensions of soft risk accumulation and hard injury count, the risks brought by disasters of different intensities can be comprehensively covered, accurately reflecting the actual risk control effect of the evacuation navigation strategy and providing a reliable quantitative basis for strategy optimization.

[0110] In summary, this embodiment provides an evacuation navigation method based on disaster perception and risk control. The method includes: discretizing the target evacuation area into a grid domain based on geographic information; determining the crowd density field of each grid cell based on pedestrian location, and constructing a travel time cost for each grid cell based on the crowd density field of each grid cell; constructing a dynamic navigation travel cost based on the travel time cost, and solving the global navigation potential field to determine the global navigation direction of pedestrians with the objective of minimizing the cumulative navigation cost from the grid cell to the exit area; obtaining pedestrian-level density, and using the pedestrian-level density to determine the local interaction direction; determining the pedestrian navigation direction based on the global navigation direction and the local interaction direction, and updating the pedestrian speed and pedestrian position according to the navigation direction; and re-executing the step of determining the pedestrian-level density based on pedestrian location according to preset conditions until the evacuation termination condition is met. This application utilizes the crowd density field of grid cells to construct a passage time cost field, and based on this passage time cost field, constructs a global navigation potential field with the exit having a zero potential energy boundary. This incorporates the decrease in passage capacity caused by crowd density into the dynamic navigation cost, fully considering the problems of crowd density changes and passage capacity decreases during evacuation, avoiding the continuous introduction of large numbers of people into the same path, improving tail evacuation time, and thus effectively shortening the overall evacuation time. Simultaneously, this application fuses the global navigation direction determined based on the global navigation potential field and the local interaction direction determined based on pedestrian density, and uses the fused navigation direction to update pedestrian positions and speeds. This allows the navigation direction to simultaneously consider global path cost and local congestion avoidance, enabling adaptive adjustment of the navigation path, avoiding excessive congestion in local areas, further improving tail evacuation time, and thus further shortening the overall evacuation time.

[0111] To further illustrate the effectiveness of the evacuation navigation method based on disaster perception and risk control provided in the embodiments of this application, a specific application scenario example is given below.

[0112] This specific application scenario uses region A as an example to systematically evaluate the full-scale evacuation performance of the disaster perception and risk control-based evacuation navigation method (denoted as DH-60) provided in this application embodiment, and compares it with static navigation. Region A is adjacent to commercial, educational, medical, and rail transit facilities, and has mixed residential and employment functions. Figure 2 As shown, its morphological characteristics include extremely dense building clusters, a narrow, maze-like network of alleyways, frequent dead ends, irregular intersections, and severe visual obstruction. The simulated coverage area is approximately... (about ), approximately 10 times that of the simplified network used for calibration, and at a resolution Rasterize region A at 0.1m and determine the passable area. Open pedestrian spaces such as alleyways and squares; barrier areas Areas that are inaccessible, such as building outlines and walls, including exit areas. For border evacuation exits.

[0113] The strategies for DH-60 and Static are shown in Table 1.

[0114] Table 1 Strategies for DH-60 and Static

[0115] fixed , ,right Perform a sensitivity scan ( (Pairing scenario). Evaluation metrics include tail evacuation time. , 95th percentile cumulative disaster exposure and the number of serious injuries (Severe injury threshold) ). The results of the sensitivity analysis are shown in Table 2.

[0116] Table 2 Sensitivity analysis results

[0117] As shown in Table 3, with Increased from 0.1 to 10.0, Monotonic increase of 38.0% ( s), Monotonic decrease of 29.1% ( This demonstrates a significant efficiency-security trade-off. This trade-off is strongly nonlinear. Increasing from 0.1 to 1.0 costs 8.1% of time to achieve a 20.3% reduction in exposure; while increasing from 1.0 to 10.0 costs 27.7% more time to achieve only an additional 11.1% reduction in exposure. The "inflection point" of the efficiency-safety curve is determined as the recommended operating point. .

[0118] Furthermore, DH-60 and Static were evaluated using the aforementioned soft risk exposure accumulation dimension and hard injury event count dimension, and the evaluation results are shown in Table 3.

[0119] Table 3 Evaluation Results

[0120] As shown in Table 3: Evacuation efficiency: DH-60 will From 1114s to 961s ( ).

[0121] Safety features: Group soft risk burden: DH-60 will from Down to ( ); Hard damage event count: DH-60 The severity rating dropped from 2649.5 to 2.86, achieving 99.9% elimination of serious injuries.

[0122] Furthermore, in the specific application scenarios mentioned above, spatial accumulation analysis was performed on DH-60 and Static in 50 pairing scenarios. The analysis results are as follows: Cumulative density load field: DH-60 is 10.5% lower than Static; 95th percentile positive unit: DH-60 is 24.2% lower than Static; Number of units exceeding the static threshold in the same field: DH-60 is 65.5% lower than Static; Congestion duration field: DH-60 reduced by 26.0% compared to Static; DH-60 compression by approximately 71%.

[0123] Based on the aforementioned evacuation navigation method based on disaster perception and risk control, this embodiment provides an evacuation navigation device based on disaster perception and risk control, such as... Figure 3 As shown, the evacuation navigation device based on disaster perception and risk control specifically includes: Discretization module 100 is used to discretize the target evacuation area into a grid domain based on geographic information; The density determination module 200 is used to determine the pedestrian-level density based on the pedestrian location, and to determine the grid-level crowd density of the grid cell based on the pedestrian-level density. The navigation passage cost construction module 300 is used to construct the passage time cost for the grid unit based on the grid-level crowd density. The disaster propagation status construction module 400 is used to determine the disaster arrival time corresponding to each grid cell starting from the disaster source grid cell, and to construct a dynamic disaster intensity for the grid cell based on the disaster arrival time and disaster intensity parameters, and to determine a blocking indication based on the dynamic disaster intensity. The blocking indication is used to identify whether the grid cell has lost its passage conditions due to the disaster. The global navigation direction determination module 500 is used to construct a navigation cost based on the passage time cost, dynamic disaster intensity and blockage indication, construct a global navigation potential field with the goal of minimizing the cumulative navigation cost from the grid cell to the exit area, and determine the global navigation direction based on the global navigation potential field. Local interaction direction determination module 600 is used to determine the local interaction direction using the pedestrian density; The update module 700 is used to determine the pedestrian navigation direction based on the global navigation direction and the local interaction direction, and update the pedestrian speed and pedestrian position according to the navigation direction; The control module 800 is used to re-execute the step of determining pedestrian density based on pedestrian location according to preset conditions until the evacuation termination condition is met.

[0124] Based on the above-described evacuation navigation method based on disaster perception and risk control, this embodiment provides a computer-readable storage medium storing one or more programs that can be executed by one or more processors to implement the steps in the evacuation navigation method based on disaster perception and risk control as described in the above embodiment.

[0125] Based on the aforementioned evacuation navigation method based on disaster perception and risk control, this application also provides a terminal device, such as... Figure 4 As shown, it includes at least one processor 20; a display screen 21; and a memory 22, and may also include a communications interface 23 and a bus 24. The processor 20, display screen 21, memory 22, and communications interface 23 can communicate with each other via the bus 24. The display screen 21 is configured to display a preset user guide interface in the initial setup mode. The communications interface 23 can transmit information. The processor 20 can invoke logical instructions in the memory 22 to execute the methods described in the above embodiments.

[0126] Furthermore, the logical instructions in the aforementioned memory 22 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium.

[0127] The memory 22, as a computer-readable storage medium, can be configured to store software programs, computer-executable programs, such as program instructions or modules corresponding to the methods in the embodiments of this disclosure. The processor 20 executes functional applications and data processing by running the software programs, instructions, or modules stored in the memory 22, thereby implementing the methods in the above embodiments.

[0128] The memory 22 may include a program storage area and a data storage area. The program storage area may store the operating system and application programs required for at least one function; the data storage area may store data created based on the use of the terminal device. Furthermore, the memory 22 may include high-speed random access memory (RAM) and non-volatile memory. Examples include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks, as well as transient storage media.

[0129] Furthermore, the specific process of loading and executing multiple instruction processors in the aforementioned storage medium and terminal device has been described in detail in the above method, and will not be repeated here.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An evacuation navigation method based on disaster perception and risk control, characterized in that, The evacuation navigation method based on disaster perception and risk control specifically includes: Based on geographic information, the target evacuation area is discretized into grid domains; The pedestrian density is determined based on the pedestrian location, and the grid-level crowd density of the grid cell is determined based on the pedestrian density. The passage time cost is constructed for each grid cell based on the grid-level population density. Starting from the disaster source grid cell, the disaster arrival time corresponding to each grid cell is determined, and a dynamic disaster intensity is constructed for the grid cell based on the disaster arrival time and disaster intensity parameters. A blocking indication is determined based on the dynamic disaster intensity, and the blocking indication is used to identify whether the grid cell has lost its passage conditions due to the disaster. Based on the passage time cost, dynamic disaster intensity, and congestion indication, a navigation cost is constructed. A global navigation potential field is constructed with the goal of minimizing the cumulative navigation cost from the grid cell to the exit area, and the global navigation direction is determined based on the global navigation potential field. The pedestrian density is used to determine the direction of local interaction; The pedestrian navigation direction is determined based on the global navigation direction and the local interaction direction, and the pedestrian speed and pedestrian position are updated according to the navigation direction. The process of determining pedestrian density based on pedestrian location will be repeated according to the preset conditions until the evacuation termination conditions are met.

2. The evacuation navigation method based on disaster perception and risk control according to claim 1, characterized in that, The determination of pedestrian density based on pedestrian location specifically includes: Obtain the distance between a pedestrian and its neighbors; The density factor corresponding to each neighboring pedestrian is determined based on the pedestrian distance and the pedestrian projection radius, and the pedestrian-level density is determined based on the density factors of all neighboring pedestrians.

3. The evacuation navigation method based on disaster perception and risk control according to claim 1, characterized in that, The specific cost of constructing the passage time for grid cells based on the grid-level population density includes: Convert the grid-level crowd density into passage speed; A travel time cost is constructed based on the travel speed, and the travel time cost is negatively correlated with the travel speed.

4. The evacuation navigation method based on disaster perception and risk control according to claim 1, characterized in that, The specific steps of constructing dynamic disaster intensity for grid cells based on the disaster arrival time and disaster intensity parameters include: The relative arrival time is determined based on the disaster arrival time of the grid cells and the current time; If the relative arrival time is less than zero, the dynamic disaster intensity is determined based on the relative arrival time and the maximum disaster intensity, intensity factor and time factor in the disaster intensity parameters; If the relative arrival time is greater than or equal to zero, the dynamic disaster intensity is determined based on the relative arrival time and the maximum disaster intensity, intensity factor, and intensity rise rate factor in the disaster intensity parameters.

5. The evacuation navigation method based on disaster perception and risk control according to claim 1 or 4, characterized in that, The expression for the dynamic disaster intensity is: ， in, Indicates dynamic disaster intensity. Indicates the intensity of the disaster at the time of its arrival. Indicates relative arrival time. Indicates the maximum disaster intensity. Indicates the time factor. This represents the rate of increase in intensity.

6. The evacuation navigation method based on disaster perception and risk control according to claim 1, characterized in that, The expression for the blocking indication is: ， in, Indicates a blockage. Indicates relative arrival time. Indicates dynamic disaster intensity. Indicates the threshold of traffic obstruction intensity. Indicates an indicator function.

7. The evacuation navigation method based on disaster perception and risk control according to claim 1, characterized in that, The process of constructing a global navigation potential field with the objective of minimizing the cumulative navigation cost from the grid cell to the exit region, and determining the global navigation direction based on the global navigation potential field, specifically includes: The objective function is to minimize the dynamic navigation passage cost of the continuous path from the grid cell to the exit region. The objective function is to solve the objective function to obtain the global navigation potential field of the grid cell. Obtain the potential energy gradient of the global navigation potential field, and normalize the potential energy gradient to obtain the global navigation direction.

8. The evacuation navigation method based on disaster perception and risk control according to claim 1, characterized in that, The method of determining the local interaction direction using the pedestrian-level density specifically includes: Obtain the pedestrian distance between a pedestrian and its neighboring pedestrians, and determine the spatial gradient of the pedestrian relative to its neighboring pedestrians based on the pedestrian distance; Pedestrian pressure is determined based on pedestrian density, and the local interaction force of the pedestrian is determined based on the pedestrian pressure of the pedestrian, the pedestrian pressure of its neighboring pedestrians, and the spatial gradient of the pedestrian relative to its neighboring pedestrians. The local interaction force of the pedestrian is normalized to obtain the local interaction direction of the pedestrian.

9. The evacuation navigation method based on disaster perception and risk control according to claim 1, characterized in that, The process of determining the pedestrian navigation direction based on the global navigation direction and the local interaction direction specifically includes: The mixing coefficient is determined based on pedestrian density. The global navigation direction and the local interaction direction are fused based on the hybrid coefficient to obtain the pedestrian navigation direction.

10. The evacuation navigation method based on disaster perception and risk control according to claim 1, characterized in that, The process for updating pedestrian speed and pedestrian position is as follows: ， ， in, This indicates the updated pedestrian speed. Indicates navigation direction, Indicates pedestrian density. This indicates the relationship between crowd density and pedestrian speed. Represents the smoothing coefficient. This indicates the updated pedestrian location. This indicates the pedestrian's location before the update. It represents the change over time.

11. An evacuation navigation device based on disaster perception and risk control, characterized in that, The aforementioned evacuation navigation device based on disaster perception and risk control specifically includes: The discrete module is used to discretize the target evacuation area into a grid domain based on geographic information. The density determination module is used to determine the pedestrian-level density based on the pedestrian location, and to determine the grid-level crowd density of the grid cell based on the pedestrian-level density. The navigation passage cost construction module is used to construct the passage time cost for grid cells based on the grid-level population density. The disaster propagation status construction module is used to determine the disaster arrival time corresponding to each grid cell starting from the disaster source grid cell, and to construct a dynamic disaster intensity for the grid cell based on the disaster arrival time and disaster intensity parameters, and to determine a blocking indication based on the dynamic disaster intensity. The blocking indication is used to identify whether the grid cell has lost its passage conditions due to the disaster. The global navigation direction determination module is used to construct navigation costs based on the passage time cost, dynamic disaster intensity, and congestion indication, construct a global navigation potential field with the goal of minimizing the cumulative navigation cost from the grid cell to the exit area, and determine the global navigation direction based on the global navigation potential field. A local interaction direction determination module is used to determine the local interaction direction using the pedestrian density. The update module is used to determine the pedestrian navigation direction based on the global navigation direction and the local interaction direction, and update the pedestrian speed and pedestrian position according to the navigation direction; The control module is used to re-execute the step of determining pedestrian density based on pedestrian location according to preset conditions until the evacuation termination condition is met.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, which can be executed by one or more processors to implement the steps in the evacuation navigation method based on disaster perception and risk control as described in any one of claims 1-10.

13. A terminal device, characterized in that, include: Processor and memory; The memory stores a computer-readable program that can be executed by the processor; When the processor executes the computer-readable program, it implements the steps in the evacuation navigation method based on disaster perception and risk control as described in any one of claims 1-10.

Images