A crowd evacuation method that takes into account crowd disorder and congestion.

Abstract

This invention discloses a crowd evacuation method that considers crowd disorder and congestion. It describes the state of crowd disorder and congestion using velocity entropy and spatiotemporal congestion, and incorporates these factors into the heuristic function of the A* algorithm to obtain a disorder impact factor and a congestion cost, which describe the influence of crowd disorder and congestion on the planned path. Finally, a social force model is used to drive pedestrians to evacuate along the planned path. This invention utilizes velocity entropy and spatiotemporal congestion as influence factors for the A* algorithm, ensuring that the path planned by the A* algorithm simultaneously considers the effects of crowd disorder and congestion. Then, a replanning mechanism is proposed based on velocity entropy and spatiotemporal congestion to meet real-time requirements. Finally, the social force model drives pedestrians to move along the evacuation path, thus retaining the advantages of the social force model in pedestrian-environment interaction while improving upon the lack of a pathfinding mechanism through the A* algorithm.

Description

Technical Field

[0001] This invention relates to the field of crowd evacuation and route planning, and more particularly to a crowd evacuation method that combines route planning algorithms in complex scenarios. Background Technology

[0002] Crowd evacuation is a fundamental research area in crowd simulation, which uses computer simulation technology to simulate the evacuation behavior of crowds in emergency situations. It describes crowds / pedestrians through various attributes such as weight, speed, comfort distance, field of vision, emotional state, and unit time step. Combined with technologies such as emotion contagion models, path planning algorithms, and group clustering algorithms, the evacuation results are made closer to reality and can provide reference for actual evacuation.

[0003] Based on different crowd sizes and crowd / pedestrian movement characteristics, crowd evacuation models can be categorized into: macroscopic evacuation models, mesoscopic evacuation models, and microscopic evacuation models. Macroscopic evacuation models treat the crowd as a unified continuous entity, studying its average speed, density, and other movement characteristics. While macroscopic models can reflect the movement characteristics of large-scale crowds, they neglect the interaction between individual pedestrians and the environment. Representative models include artificial potential fields and dynamic continuum models. Microscopic models study crowd evacuation behavior at the individual pedestrian level, fully considering the interaction between individuals and the environment, but their description of the overall crowd behavior is insufficient. Representative models include social force models and cellular automata models. Mesoscopic crowd evacuation aims to study the movement characteristics at both the group and individual scales simultaneously, but suffers from insufficient description of both scales and poor synchronicity between macroscopic and microscopic models. Among microscopic simulation models, social force models are widely used in crowd simulation because they can effectively simulate crowd self-organization phenomena such as "fast is slow" and the "arch effect." However, traditional social force models lack pathfinding mechanisms, and pedestrians may get trapped in local minima in complex environments. The social force model that incorporates path planning algorithms can effectively solve the problem of the lack of path-finding mechanisms in traditional social force models. This allows the social force model to be applied in complex scenarios, and the social force model that incorporates path planning algorithms is more realistic in pedestrian motion simulation.

[0004] Path planning algorithms in crowd evacuation typically use the pedestrian's location as the starting point and the destination as the ending point, combined with scene information, to plan a path from the starting point to the destination. The selection criteria for this path can include shortest distance, highest safety, and lowest congestion. The A* algorithm selects the path with the lowest cost. Based on known global environmental information, the A* algorithm rasterizes the scene, then, starting from the starting point, visits its neighboring raster according to certain rules, storing rasters that meet the conditions in the OPEN table. It calculates the cost of these rasters using a heuristic function, and finally selects the raster with the lowest cost from the OPEN table as the starting point for the next step, adding it to the CLOSE table, until the destination is found or the OPEN table is empty. Traditional A* algorithms efficiently search for the shortest path and have advantages such as low computational cost and simple principles. However, as a static, single-source algorithm, its application in crowd evacuation suffers from a lack of consideration for dynamic scenes and crowds. This is because the heuristic function of traditional A* algorithms only considers distance factors, ignoring situations where pedestrians may choose longer paths to avoid chaotic or crowded areas.

[0005] Therefore, designing an A* algorithm that simultaneously considers dynamic changes in the scene and the influencing factors of the crowd, and combining it with the social force model, has become a challenging problem in the field of crowd evacuation. Summary of the Invention

[0006] This invention provides a micro-level crowd evacuation method for complex scenarios. The method describes crowd disorder and congestion states using velocity entropy and spatiotemporal congestion, respectively, and incorporates these two concepts into the heuristic function of the A* algorithm. It proposes a disorder impact factor and a congestion cost to describe the influence of crowd disorder and congestion levels on the planned path. Then, the invention proposes a replanning mechanism that allows the A* algorithm to consider dynamic factors of the scenario and meet real-time requirements. Finally, a social force model is used to drive pedestrians to evacuate along the planned path. The method involves an A* algorithm that considers crowd disorder and congestion levels, enabling it to plan paths that take into account the impact of these factors. It also involves a replanning mechanism that allows the A* algorithm to consider dynamic changes in the scenario.

[0007] The crowd evacuation method includes the following steps:

[0008] Step 1: Initialize the scene by rasterizing it to 1m. 2 Grid size;

[0009] Step 2: Initialize pedestrian attributes and differentiate different heterogeneous pedestrians by differentiating pedestrian attributes;

[0010] Step 3: Count the number of people and their speed attributes within a unit grid area, calculate the speed entropy and spatiotemporal congestion of the grid, and then use these two concepts as the influence factors of the heuristic function of the A* algorithm to plan an evacuation path.

[0011] Step 4: Sequentially use the path nodes as the desired directions of the social force model, and drive pedestrians to evacuate along the planned paths through the social force model.

[0012] Step 5: During the evacuation process, the chaotic state of the crowd within the pedestrian's field of vision is described by velocity entropy, and the crowding state within the pedestrian's field of vision is described by spatiotemporal congestion. The magnitude of these two values ​​is determined, and it is decided whether to replan the route. If yes, step 4 is executed; otherwise, no change is made.

[0013] Furthermore, the velocity entropy is derived from the concept of information entropy in information theory. Information entropy is an effective means of quantitatively studying the diversity of complex systems; the greater the entropy, the more chaotic the system. Velocity entropy describes the distribution of the magnitude and direction of the speeds of pedestrians in a crowd. Obviously, the more uniform the distribution of the magnitude and direction of the speeds of pedestrians in a crowd, the more disordered the movement of pedestrians in that crowd; that is, the greater the velocity entropy, the more chaotic the crowd. The velocity entropy is obtained by discretizing the magnitude and direction of the speeds of pedestrians in the crowd into four intervals, then detecting the magnitude and direction of the speed of each pedestrian in the crowd, and calculating the probability of each discretized interval to obtain the velocity entropy value of the crowd.

[0014] Furthermore, the spatiotemporal congestion is a theorem description of the degree of crowding in a fixed space. The spatiotemporal congestion is calculated by: counting the number of people in the fixed space and calculating the saturation, which is the ratio of the real-time number of people to the maximum capacity of the space; and then using an exponential function to describe the spatiotemporal congestion, where the saturation is the exponential influence factor of the exponential function. The greater the spatiotemporal congestion, the more crowded the fixed space, and the slower the movement speed of pedestrians in the space.

[0015] Furthermore, the disorder influence factor is the value of velocity entropy, and the disorder influence factor is prepared by using it as the multiplier influence factor of the heuristic function of the A* algorithm.

[0016] Furthermore: The congestion cost is the cost lost under the influence of spatiotemporal congestion. The congestion cost is calculated as follows: Under the influence of spatiotemporal congestion, pedestrians will reach the destination at a slower speed and take a longer time. If pedestrians reach the destination at the desired speed, it will take a shorter time. The maximum distance that can be traveled due to this time difference is the "congestion cost", and it is used as the additive influence factor of the heuristic function of the A* algorithm.

[0017] Furthermore, the replanning mechanism is actually based on two thresholds proposed by velocity entropy and spatiotemporal congestion. The replanning mechanism works as follows: when the real-time velocity entropy or spatiotemporal congestion of the crowd within the pedestrian's field of vision is higher than the threshold, the pedestrian replans the path and uses the replanned path as the expected speed of the social force model.

[0018] Compared with the prior art, the present invention has the following advantages:

[0019] Advantage 1: This invention utilizes the advantage of the A* algorithm in efficiently searching for the shortest path, combining the social force model with the A* algorithm to improve the social force model's lack of a pathfinding mechanism;

[0020] Advantage 2: This invention introduces velocity entropy and spatiotemporal congestion to describe the degree of crowd disorder and congestion, and uses these two concepts as the influence factors of the heuristic function of the A* algorithm, so that the path planned by the A* algorithm of this invention takes into account the impact of crowd disorder and congestion on evacuation.

[0021] Advantage 3: The replanning mechanism proposed in this invention based on velocity entropy and spatiotemporal congestion improves upon the shortcomings of the traditional A* algorithm as a single-source static algorithm, which lacks consideration of dynamic changes in the scene, and ensures that pedestrians always evacuate along the optimal path during the evacuation process.

[0022] Advantage 4: This invention can provide valuable reference for evacuation in complex scenarios, and has a wide range of applications in situations where it is inconvenient to organize live drills with volunteers, such as evacuation exercises and fire protection design.

[0023] In summary, the technical solution of this invention utilizes velocity entropy and spatiotemporal congestion to describe the degree of crowd disorder and congestion, respectively, and uses these two factors as influencing factors for the A* algorithm, ensuring that the path planned by the A* algorithm simultaneously considers the effects of crowd disorder and congestion. Then, a replanning mechanism is proposed based on velocity entropy and spatiotemporal congestion to meet real-time requirements. Finally, a social force model is used to drive pedestrians to move along the evacuation path, so that the crowd evacuation method of this invention retains the advantages of the social force model in pedestrian-environment interaction, while also improving upon the lack of a pathfinding mechanism in the social force model through the A* algorithm. Attached Figure Description

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

[0025] Figure 1This is a schematic diagram of the overall structure of the evacuation method of the present invention, which takes into account the heterogeneity and crowding of the population.

[0026] Figure 2 This is a schematic diagram of the A* algorithm of the present invention, which takes into account the heterogeneity and crowding of the population.

[0027] Figure 3a and Figure 3b These are schematic diagrams illustrating the replanning mechanisms for chaotic and crowded situations in front of the crowd, respectively. Detailed Implementation

[0028] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort should fall within the scope of protection of the present invention.

[0029] like Figure 1 As shown, the present invention provides an evacuation method that considers crowd heterogeneity and congestion levels, comprising the following steps:

[0030] Step 1: Initialize scene information and pedestrian information. Scene information includes topology, exit distribution, and obstacle distribution. Pedestrian information includes homogeneous and heterogeneous attributes of pedestrians. Homogeneous attributes refer to common information of pedestrians, such as the sliding friction coefficient, human elasticity coefficient, weight, and relaxation time. Heterogeneous attributes refer to attributes that can distinguish pedestrian types, such as pedestrian comfort distance and expected speed.

[0031] Step 2, use 1m 2 Discretize the scene into grids of varying sizes and calculate the crowd velocity entropy and spatiotemporal congestion within each grid area.

[0032] Entropy theory in information theory is an effective means of quantitatively studying the diversity of complex systems. The greater the entropy, the more chaotic the system. This invention describes the degree of chaos by using the velocity entropy of people within a grid moiré neighborhood, as shown in the following formula:

[0033] E = α1 * E V + α2* E D (1)

[0034]

[0035]

[0036] α1+α2=1 (4)

[0037] Among them: E V and E D Let P be the entropy of the pedestrian's speed magnitude and direction, respectively, and n = m = 4, representing the number of discretizations of the pedestrian's speed magnitude and direction, respectively. i and P j Let α1 and α2 represent the proportions of the i-th and j-th elements in the system, respectively. Let α1 and α2 represent the weights of the two entropies, respectively. In this invention, α1 = α2 = 0.5.

[0038] Spatiotemporal congestion is a theorem describing the degree of crowding in a fixed space. First, the number of people in the fixed space is counted and the saturation is calculated. Then, an exponential function is used to describe the spatiotemporal congestion, where the saturation is the exponential factor of this function. The greater the spatiotemporal congestion, the more crowded the fixed space, and the slower the pedestrians move within that space. The formula is shown below:

[0039]

[0040]

[0041]

[0042] Where, N max This represents the maximum number of pedestrians per unit grid. According to relevant literature, the average vertical projection area per capita in my country is 0.25m². 2 The unit grid area of ​​this invention is 1m². 2 Therefore, we take N. max =4; N is the actual number of pedestrians on the grid at a certain moment in the scene; C ij It represents the spatiotemporal congestion of the grid; V o It is the pedestrian's expected speed; V crowd is the congestion speed of pedestrians on this grid, and h is the Manhattan distance from this grid to the end point.

[0043] Step 3: All pedestrians combine the scene information output in Step 1 with the velocity entropy and spatiotemporal congestion output in Step 2 to execute the A* algorithm and output an evacuation path that takes into account crowd chaos and congestion.

[0044] The A* algorithm, based on known global environment information, rasterizes the scene and then, starting from the starting point, visits its neighboring rasters according to certain rules. Rasteres that meet the conditions are stored in the OPEN table. The cost of these rasters is calculated using a heuristic function. Finally, the raster with the lowest cost is selected from the OPEN table as the starting point for the next step and added to the CLOSE table. This process continues until the destination is found or the OPEN table is empty. Figure 2 As shown. This invention incorporates velocity entropy and congestion cost into the heuristic function of the A* algorithm, as shown in the following formula:

[0045] f(n)=g(n)+h(n) (8)

[0046]

[0047]

[0048] Where n is the grid cell to be judged, g(n) is the actual cost from the starting point to point n (here, the distance between unit grid cells is taken), h(n) is the estimated cost from point n to the destination (i.e., the Manhattan distance from point n to the destination), E is the velocity entropy of the grid cell, and C is the congestion cost of the grid cell. Clearly, the greater the velocity entropy and congestion cost of a grid cell, the greater the cost for a pedestrian to pass through that grid cell. According to the rule of the A* algorithm for finding the minimum cost path, the final path will avoid cluttered and congested grid cells as much as possible.

[0049] Step 4: Use the social force model to drive pedestrians to evacuate along the path output in Step 3.

[0050] The social force model is a microscopic simulation model based on Newton's second law. It consists of three forces: a driving force pointing towards the target location, a repulsive force between pedestrians, and a repulsive force from obstacles on pedestrians. The formulas are shown below:

[0051]

[0052] The right side of the equation consists of, in order: the driving force pointing towards the target, and the repulsive force from other pedestrians and obstacles, where:

[0053]

[0054]

[0055]

[0056] m i represents weight, Represents velocity vector, vector Indicates the pedestrian's desired direction. It is a scalar representing the pedestrian's expected speed, τ. i Relaxation time represents the time it takes for a pedestrian to adapt to a change in speed; A i B represents the intensity of interaction between pedestrians. i r represents the minimum distance at which repulsive force is generated. ij It is the sum of the comfortable distances for pedestrians, d ij It is the distance between the centroids of pedestrians. Let j represent the unit direction vector pointing from j to i. yes The unit tangential direction, K is the elastic coefficient of the human body, and k is the coefficient of sliding friction. It is the relative speed of the two pedestrians. G(r ij -d ij ) is a function, and its formula is shown below:

[0057]

[0058] The repulsive force parameters of the obstacle are the same as those of the pedestrian.

[0059] Step 5: During the evacuation of pedestrians along the path, the scene information will change. When the speed entropy of the crowd or the spatiotemporal congestion within the pedestrian's field of vision exceeds the threshold, the path will be replanned, i.e., steps 3 and 4 will be executed.

[0060] Traditional A* algorithms ignore dynamic changes in the scene. When there are chaotic and crowded people on the shortest path, pedestrians behind will make decisions. If the crowd ahead is less chaotic and less crowded, pedestrians are more likely to maintain their original path. If the crowd ahead is highly chaotic and crowded, pedestrians are more likely to take a longer route to avoid the chaotic and congested crowd. This invention quantifies this decision-making process using velocity entropy and spatiotemporal congestion metrics, such as... Figure 3a and Figure 3b As shown.

[0061] The above embodiments of the present invention are merely for illustrative purposes and do not represent the superiority or inferiority of the embodiments.

[0062] In the above embodiments of the present invention, the descriptions of each embodiment have their own focus. For parts that are not described in detail in a certain example, please refer to the relevant descriptions of other embodiments.

[0063] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention 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 or all of the technical features. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method of crowd evacuation considering the degree of crowd chaos and congestion, characterized by: The disorder and congestion of the crowd are described by velocity entropy and spatiotemporal congestion. The velocity entropy and spatiotemporal congestion are combined into the heuristic function of the A* algorithm to obtain the disorder impact factor and congestion cost, which are used to describe the impact of the degree of disorder and congestion on the planned path. Finally, a social force model is used to drive pedestrians to evacuate according to the planned path. The method includes the following steps: Step 1, initialize the scene, rasterize the scene into 1 size grid; Step 2: Initialize pedestrian attributes and differentiate different heterogeneous pedestrians by differentiating pedestrian attributes; Step 3: Count the number of people and speed attributes within the unit grid area, calculate the speed entropy and spatiotemporal congestion of the grid, and then use the speed entropy and spatiotemporal congestion as the influence factors of the heuristic function of the A* algorithm, and plan the evacuation path. The formula for calculating the velocity entropy E of the grid is as follows: (1) (2) (3) (4) where, and are the size and direction entropy of pedestrian velocity, respectively, and , respectively, represent the number of discretization of the size and direction of pedestrian velocity, and respectively represent the first , proportion of the element in the system, and respectively represent the weight of the two kinds of entropy; The congestion cost C of this grid is calculated using the following formula: (5) (6) (7) wherein, is the maximum number of pedestrians per unit grid, taken ; is the actual number of pedestrians on the grid at a certain time; is the space-time congestion degree of the grid; is the expected speed of the pedestrian; is the congestion speed of the pedestrian on the grid, is the Manhattan distance from the grid to the destination; The velocity entropy and congestion cost are combined into the heuristic function of the A* algorithm, as shown in the following formula: (8) (9) (10) wherein, is the grid to be judged, is the real cost from the start point to point, is the estimated cost from point to the end point, i.e. the Manhattan distance from point to the end point, is the speed entropy of the grid; Step 4: Sequentially use the path nodes as the desired directions of the social force model, and drive pedestrians to evacuate along the planned paths through the social force model. Step 5: During the evacuation process, the chaotic state of the crowd within the pedestrian's field of vision is described by velocity entropy, and the crowding state within the pedestrian's field of vision is described by spatiotemporal congestion. The magnitude of these two values ​​is determined, and it is decided whether to replan the route. If yes, proceed to step 4; otherwise, no change is made. The confusion influence factor serves as the multiplier influence factor for the heuristic function of the A* algorithm; The congestion cost is used as an additive influence factor in the heuristic function of the A* algorithm.

2. The crowd evacuation method of claim 1, wherein the degree of chaos and congestion of the crowd is considered. In step 3, the speed magnitude and direction of pedestrians in the crowd are discretized into four intervals. Then, the speed magnitude and direction of each pedestrian in the crowd are detected, and the probability of each discretized interval is calculated to obtain the speed entropy value of the crowd.

3. The crowd evacuation method of claim 1, wherein: the crowd evacuation method further comprises: determining a degree of congestion and disorder of the crowd; and determining the evacuation route based on the degree of congestion and disorder of the crowd. In step 3, the number of people in the fixed space is counted and the saturation is calculated. The saturation is the ratio of the real-time number of people to the maximum capacity of the space. Then, the spatiotemporal congestion is described by an exponential function. The saturation is the exponential influence factor of the exponential function. The greater the spatiotemporal congestion, the more crowded the fixed space is, and the slower the pedestrians move in the space.

4. The crowd evacuation method of claim 1, wherein the degree of chaos and congestion of the crowd is considered. When the real-time velocity entropy or spatiotemporal congestion of the crowd within a pedestrian's field of vision exceeds a threshold, the pedestrian replans their path, and the replanned path is used as the expected velocity of the social force model.

Images