An indoor dense crowd emergency evacuation method considering the flying effect of fragments

Abstract

The application discloses an indoor dense crowd emergency evacuation method considering a fragment scattering effect, belongs to the technical field of emergency evacuation, and comprises the following steps: a group intelligence algorithm and a fragment scattering effect social force model are used to calculate different optimal evacuation exit schemes corresponding to a sudden event and partition crowds under different conditions; site, partition crowds and sudden event information are taken as input, and the optimal evacuation exit scheme is taken as output; all the input and output are used to train a decision tree model to obtain a real-time emergency evacuation decision model; indoor partition crowd information under a sudden event is collected, and the information is input into the real-time emergency evacuation decision model together with the site and sudden event information to obtain the optimal evacuation exit scheme; and the optimal target exit of each partition in the optimal evacuation exit scheme is transmitted to a drone of each partition, and the drone uses time-sharing broadcasting to guide the partition crowds to evacuate to the optimal exit.

Description

Technical Field

[0001] This invention relates to the field of emergency evacuation technology, and specifically to an emergency evacuation method for densely populated indoor crowds that takes into account the effect of fragmentation. Background Technology

[0002] Sudden events occurring indoors are characterized by their transient nature, strong impact, and explosive, step-like energy release. When such events occur, pedestrians around the epicenter (the hotspot) will scatter and flee in panic, disrupting the existing dynamic balance of the indoor crowd. If effective emergency evacuation of the dense crowd is not carried out promptly, it can easily lead to uncontrolled crowd behavior, resulting in stampedes and casualties. These events typically have four stages: 1. The sudden event erupts; 2. Panic-driven escape of pedestrians near the hotspot; 3. The stampede spreads from the point to the entire area; 4. A disaster resulting in casualties.

[0003] Currently, evacuation models for densely populated indoor crowds primarily employ discrete models, typically represented by cellular automata, and continuous models, typically represented by social force models. Cellular automata, limited by their inherent simplification rules for efficient computation, have limitations in describing the highly explosive and forceful characteristics of dense crowds during emergencies. They cannot accurately quantify the complex psychological and physical forces between pedestrians in a crowd. Social force models, based on Newtonian mechanics, precisely compensate for these shortcomings of cellular automata. They can continuously characterize pedestrian behavior in time and space, and quantify the psychological and physical forces between individual pedestrians. Compared to cellular automata, social force models have advantages in characterizing pedestrian speed, force, and range of action. However, existing social force models struggle to accurately describe the scattered evacuation state of crowds during emergencies, and the setting of relevant parameters is difficult to quantify. This significantly reduces the ability of social force models to characterize and objectively describe pedestrian movement behavior in emergencies, resulting in low accuracy of planned emergency evacuation routes.

[0004] Furthermore, the most crucial aspect of a complete emergency evacuation method is its application in real-world scenarios, addressing the issues of real-time crowd guidance and safe, rapid evacuation. This requires the evacuation method to not only quickly calculate the optimal evacuation exits and routes based on the situation but also to provide real-time feedback to the people inside the building, guiding densely packed groups to evacuate quickly and safely. Currently, indoor evacuation measures primarily rely on installing fixed-direction evacuation indicator lights, audible and visual alarms, and large static displays. These traditional evacuation devices largely follow local, static rules, primarily using the "nearest exit" or pre-set fixed evacuation routes. The fixed evacuation guidance information cannot be updated and dynamically guided in real-time based on the movement of people, easily leading to problems such as instantaneous overload at exits. In practical applications, they suffer from poor real-time performance, low accuracy, and a lack of dynamic adaptability. They struggle to update and feedback evacuation guidance information to the people on site in a timely manner, resulting in low evacuation efficiency. Summary of the Invention

[0005] In view of the above-mentioned shortcomings in the prior art, the emergency evacuation method for densely populated indoor crowds that takes into account the fragmentation effect provided by the present invention solves the problems of low accuracy of evacuation exits and evacuation routes and low on-site evacuation efficiency in the case of pedestrians scattering and fleeing during emergencies.

[0006] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0007] An emergency evacuation method for densely populated indoor areas, taking into account the fragmentation effect, is provided, comprising the following steps:

[0008] S1. Read the zoning information of the indoor venue and calculate the social force model of the fragmentation effect of the resultant force on pedestrians in a sudden event;

[0009] S2. Using swarm intelligence algorithm and fragmentation effect social force model, calculate the different optimal evacuation exit schemes corresponding to the location of the emergency and the location of the population in each zone under different distribution conditions. Each optimal evacuation exit scheme includes the optimal target exit for each population in each zone.

[0010] S3. Using the site environment, zoning population information, and emergency event information as inputs, and the optimal evacuation exit plan as output, the decision tree model is trained using all inputs and outputs as a training set to obtain a real-time emergency evacuation decision model.

[0011] S4. Collect population information for each zone of the indoor venue during an emergency, and input it along with the venue environment and emergency information into the real-time emergency evacuation decision model to obtain the optimal evacuation exit plan.

[0012] S5. Transmit the optimal target exit for each zone of the population in the optimal evacuation exit plan to an independent drone. All drones in each zone use time-sharing broadcasting to guide the population of their respective zones to the designated optimal exit.

[0013] Furthermore, the emergency evacuation method for densely populated indoor areas also includes constructing an evacuation route library: selecting the pedestrian path in each zone within all optimal evacuation exit schemes in step S2 whose evacuation time is closest to the average evacuation time of the entire zone's population as the optimal evacuation route; and using all optimal evacuation routes to construct an evacuation route library.

[0014] Between step S4 and step S5, the following is also included:

[0015] A1. Based on the model input and output of each partition in step S4, calculate the similarity between each partition and the corresponding input and output of each optimal evacuation path in the evacuation path library in step S3.

[0016] A2. Determine if there is an optimal evacuation path with a similarity greater than a preset threshold. If it exists, use it as the optimal evacuation path for the corresponding evacuation zone population in the optimal evacuation exit plan. If it does not exist, proceed to step A3.

[0017] A3. Input the information of the pedestrians in the population of the partition that are farthest from their optimal target exit into the social force model of the fragmentation effect, calculate the evacuation path of the pedestrian, and use it as the optimal evacuation path for the population of the partition.

[0018] Step S5 also includes transmitting the optimal evacuation route of the partition to the corresponding drone. The drone flies according to the optimal evacuation route and guides the people in its partition to the corresponding optimal exit in combination with time-sharing broadcast.

[0019] Furthermore, the expression for the social force model of the fragmentation effect is:

[0020]

[0021]

[0022] in, Let be the net force acting on pedestrian i at time t; Indicates mass as pedestrian i at the expected speed Towards the desired export direction Its own driving force; The quality of pedestrian i; Let be the expected speed of pedestrian i; Quantify the influence coefficient for the expected velocity of the fragments; For the desired speed increment; Let be the unit direction vector pointing from the current position of pedestrian i at time t to the desired exit. Let be the velocity of pedestrian i at time t; The reaction time constant; Let t be the force exerted by pedestrian j on pedestrian i. Let w be the force between pedestrian i and obstacle w at time t; This refers to the fragmentation coefficient; Let t be the psychological repulsion between pedestrians i and j. The fragmentation repulsion force between pedestrians i and j at time t; K is the positive compressive force coefficient; For contact functions; The sum of the physical radii of pedestrian i and pedestrian j's bodies; Let be the distance between the centers of mass of pedestrians i and j at time t; Let be the unit normal vector pointing from pedestrian j to pedestrian i at time t; The coefficient of sliding friction; Let be the difference in tangential velocity between pedestrians i and j at time t; For time t and The perpendicular unit tangent vector.

[0023] Furthermore, step S2 further includes:

[0024] S21. Based on the zoning crowd information and emergency information in the indoor venue, a swarm intelligence algorithm is used to generate multiple different exit schemes for the indoor venue, wherein the exit schemes include the target exits for each zone.

[0025] S22. The social force model of fragmentation effect is used to calculate the resultant force on each pedestrian in each zone in each exit scheme. Then, the time for the pedestrian to reach the target exit of its zone is calculated based on the resultant force.

[0026] S23. Take the evacuation time of the last pedestrian in each partition of each exit scheme as the fitness value of the exit scheme, and determine whether the number of iterations is less than the preset number of iterations. If so, proceed to step S24; otherwise, proceed to step S25.

[0027] S24. Return the fitness value of each exit scheme to the swarm intelligence algorithm for iteration to generate different new exit schemes, and then return to step S22.

[0028] S25. Save the exit scheme corresponding to the lowest fitness value as the optimal evacuation exit scheme. Determine whether the number of saved optimal evacuation exit schemes has reached the preset number. If so, output all optimal evacuation exit schemes; otherwise, proceed to step S26.

[0029] S26. Adjust the population information and / or emergency information for each zone, then return to step S21.

[0030] Furthermore, step S22 further includes:

[0031] S221. Calculate the resultant force on each pedestrian in each zone of each exit scheme at the current moment using the social force model of fragment scattering effect;

[0032] S222. Determine the distance of each pedestrian to the target exit of their respective zone, and calculate their acceleration at the current moment based on their mass and the net force acting on them.

[0033] S223. Calculate the displacement of each pedestrian at the current moment based on their acceleration and velocity. Then determine whether the pedestrian has reached the target exit. If yes, proceed to step S224; otherwise, update their velocity at the next moment and return to step S221.

[0034] S224. Record the evacuation time of pedestrians, and proceed to step S23 when each pedestrian in each zone of each exit plan has received their evacuation time.

[0035] Furthermore, the quantification influence coefficient of the expected velocity of the fragments was calculated. The expression is:

[0036]

[0037] in, Let be the distance between pedestrian i and the center of the emergency at time t. Let be the radius of influence of the sudden event at time t; max and min are the maximum and minimum values, respectively.

[0038] The expression is:

[0039]

[0040] in, Let be the desired exit position vector of pedestrian i at time t. Let be the position vector of pedestrian i at time t; The modulus of the orientation quantity.

[0041] Furthermore, the interaction force between pedestrian i and obstacle w is calculated. The expression is:

[0042]

[0043]

[0044] in, The force coefficient between pedestrian i and obstacle w; It is an exponential function; The range of the repulsive force between pedestrian i and obstacle w is defined; Let i be the physical radius of the pedestrian's body. Let be the distance from pedestrian i to obstacle w at time t. Let be the unit normal vector pointing from obstacle w to pedestrian i at time t; For time t and The perpendicular unit tangent vector.

[0045] Furthermore, psychological repulsion and fragmentation effect repulsion force The expressions are as follows:

[0046] ,

[0047] in, It is the psychological repulsion constant; and All are constants; >0 indicates that pedestrian i and pedestrian j have already been physically squeezed at time t, and the larger the difference, the more severe the squeezing; = 0 indicates that pedestrian i and pedestrian j just make contact at time t; <0 indicates that pedestrian i and pedestrian j did not have physical contact at time t; The constant of the fragment repulsion force; It is an exponential function;

[0048] Calculate the fragmentation coefficient The expression is:

[0049]

[0050] ,

[0051] ,

[0052] Wherein, CV is the step velocity weighting factor; CD is the fragment distance range factor; and CA is the pedestrian scattering angle factor. Let j be the speed of pedestrian j at time t and the standard walking speed of pedestrian j. The speed difference between them; Let be the velocity of pedestrian j at time t; The standard walking speed for pedestrians; The critical distance at which fragmentation occurs between pedestrians; W is the control parameter for the intensity of opposite movement; T is the attenuation coefficient for the control of same-direction movement. Let be the angle between the motion direction vector of pedestrian j at time t and the direction vector of pedestrian j pointing towards pedestrian i; Let be the angle between the motion direction vector of pedestrian i at time t and the direction vector of pedestrian i pointing towards pedestrian j.

[0053] Furthermore, the site environment includes the site area, the path distance from each zone center to each exit, and the width of each exit; the emergency information includes the emergency type, the location of the emergency, the impact index of the emergency hotspot, and the distance from each zone center to the emergency hotspot; the zoned crowd information includes the overall crowd density of the indoor site, the number of people in each zone, the crowd density in each zone, and the initial position of each pedestrian.

[0054] Step S26 adjusts the zonal population information and / or emergency information by adjusting at least one of the parameters of the zonal population information and emergency information.

[0055] Compared with the prior art, the emergency evacuation method provided by the present invention has the following beneficial effects:

[0056] (1) After dividing the indoor space into zones, this scheme uses a swarm intelligence algorithm combined with a fragmentation effect social force model to quickly calculate the optimal evacuation exit scheme, including the optimal evacuation exit for each zone, based on the zone and different scenario conditions (such as pedestrian location distribution, sudden hotspot locations, etc.). Then, a large number of optimal evacuation exit schemes are used to train the decision tree model, which can ensure that the decision tree model can accurately plan the optimal evacuation exit scheme for each zone within milliseconds based on the input data of the on-site situation when an emergency occurs. Due to the fast planning speed, the implementation of the indoor space evacuation scheme becomes possible. In addition, this scheme can also provide the optimal evacuation path for the optimal evacuation exit for each zone, with clear objectives, which can avoid panic and escape of the crowd, thereby improving evacuation efficiency.

[0057] Furthermore, this solution transmits the optimal evacuation path for each zone to an independent drone. The drone can then guide the crowd in that zone towards the optimal exit, providing accurate guidance to panicked individuals and thus improving evacuation efficiency. This invention is particularly suitable for indoor venues with multiple exits, asymmetrical exits (one main exit + multiple smaller exits), significant exit bottlenecks, and high crowd density, such as stadiums, hospitals, and subway stations.

[0058] (2) Analysis revealed that when a densely populated indoor crowd encounters a sudden event, the event's epicenter generates tremendous energy instantaneously. This high energy causes a transient step change in the behavior of surrounding pedestrians, who quickly flee the epicenter and scatter in panic. The emergency evacuation behavior of pedestrians is highly impactful and destructive, significantly affecting the dense crowd from a single point to the surrounding area. Therefore, this invention incorporates the fragmentation damage effect during an ammunition explosion into characterizing this behavior, introducing parameters such as fragment dispersion velocity, impact distance range, and dispersion angle. This scheme combines a social force model to construct a social force model (FSSFM) considering the fragmentation effect. This model accurately quantifies the behavioral impact of pedestrians near the epicenter and their influence on other pedestrians. FSSFM can enrich the description of the impact of pedestrian fragmentation on the overall crowd's movement state, based on the behavioral characteristics of pedestrians encountering a sudden event. This ensures that the combination of swarm intelligence algorithms and the fragmentation effect social force model can plan more accurate optimal evacuation exit schemes and identify optimal evacuation paths.

[0059] (3) This invention utilizes drone technology to accurately deliver different evacuation information to people in designated areas. The control center continuously directs the drones to fly to designated locations along the optimal evacuation path based on the actual situation, forming an integrated air-ground intelligent evacuation system. At the same time, the control center can also monitor the on-site evacuation situation through the cameras mounted on the drones, adjust evacuation strategies in a timely manner, and provide data support for post-event analysis. Attached Figure Description

[0060] Figure 1 A flowchart for emergency evacuation methods for densely populated indoor areas, taking into account the effect of fragmentation. Detailed Implementation

[0061] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0062] refer to Figure 1 , Figure 1 A flowchart is shown for an emergency evacuation method for densely populated indoor areas that takes into account the effect of fragmentation; such as Figure 1 As shown, the method includes steps S1 to S5.

[0063] In step S1, the zoning information of the indoor site is read and the social force model of the fragmentation effect of the resultant force on pedestrians during a sudden event is calculated. In this scheme, each zone in the indoor site must have a clear label, such as zone A, zone B, zone C, and zone D, so that pedestrians in that area can accurately know which zone they belong to in the initial state. The basic information of the indoor site includes at least the following:

[0064] 1. Scene ID; 2. Total Scene Area; 3. Origin Position Coordinates; 4. X-axis Direction; 5. Y-axis Direction; 6. Number of Regions; 7. Region ID; 8. Region Name; 9. Region Area; 10. Region Type; 11. Region Boundary Vertex Position Coordinates; 12. Number of Exits; 13. Exit ID; 14. Exit Name; 15. Exit Type; 16. Exit Vertex Position Coordinates; 17. Exit Position Coordinates; 18. Exit Width; 19. Obstacle ID; 20. Obstacle Name; 21. Obstacle Type; 22. Obstacle Vertex Position Coordinates.

[0065] In one embodiment of the present invention, the expression for the social force model of the fragmentation effect is:

[0066]

[0067]

[0068] in, Let be the net force acting on pedestrian i at time t; Indicates mass as pedestrian i at the expected speed Towards the desired export direction Driven by its own forces, the expected target exit here is the target exit allocated to each partition of the population in the exit scheme generated by the swarm intelligence algorithm. The quality of pedestrian i; Let be the expected speed of pedestrian i; Quantify the influence coefficient for the expected velocity of the fragments; For the desired speed increment; Let be the unit direction vector pointing from the current position of pedestrian i at time t to the desired exit. Let be the velocity of pedestrian i at time t; The reaction time constant; Let t be the force exerted by pedestrian j on pedestrian i. Let w be the force between pedestrian i and obstacle w at time t; This refers to the fragmentation coefficient; Let t be the psychological repulsion between pedestrians i and j. Let t be the fragment repulsion force between pedestrians i and j at time t, that is, the fragment repulsion force generated by pedestrian j on pedestrian i due to the sudden scattering and escape behavior; K is the positive compressive force coefficient; For contact functions; The sum of the physical radii of pedestrian i and pedestrian j's bodies; Let be the distance between the centers of mass of pedestrians i and j at time t; Let be the unit normal vector pointing from pedestrian j to pedestrian i at time t; The coefficient of sliding friction; Let be the difference in tangential velocity between pedestrians i and j at time t; For time t and The perpendicular unit tangent vector.

[0069] The social force model provided in this scheme, which considers the fragmentation effect, describes the transient evacuation behavior of pedestrians near a sudden hotspot as they disperse and flee in all directions. It fully considers the impact of step velocity, effective distance, and scattering angle on pedestrians. The resultant force on pedestrians calculated by this model can accurately calculate the time it takes for pedestrians to evacuate to the exit location, which provides a guarantee for the generation of the optimal evacuation exit scheme.

[0070] In implementation, this scheme preferably calculates the interaction force between pedestrian i and obstacle w. The expression is:

[0071]

[0072]

[0073] in, The force coefficient between pedestrian i and obstacle w; It is an exponential function; The range of the repulsive force between pedestrian i and obstacle w is defined; Let i be the physical radius of the pedestrian's body. Let be the distance from pedestrian i to obstacle w at time t. Let be the unit normal vector pointing from obstacle w to pedestrian i at time t. For time t and The perpendicular unit tangent vector.

[0074] In step S2, the swarm intelligence algorithm and the fragmentation effect social force model are used to calculate the different optimal evacuation exit schemes corresponding to the location of the emergency and the location of the population in different zones under different distribution conditions. Each optimal evacuation exit scheme includes the optimal target exit for each population zone.

[0075] In one embodiment of the present invention, step S2 further includes:

[0076] S21. Based on the zoning crowd information and emergency information in the indoor venue, a swarm intelligence algorithm is used to generate multiple different exit schemes for the indoor venue, wherein the exit schemes include the target exits for each zone.

[0077] The preferred swarm intelligence algorithm in this scheme is the hybrid frog leaping algorithm (SFLA). In SFLA, each frog represents a specific exit scheme. For example, if there are 4 exits in an indoor venue and N=5 indoor zones, the exit scheme [2,1,2,3,4] represents the exit numbers assigned to the crowds in the 5 zones. The crowds in zone 1 evacuate to exit 2, the crowds in zone 2 evacuate to exit 1, the crowds in zone 3 evacuate to exit 2, the crowds in zone 4 evacuate to exit 3, and the crowds in zone 5 evacuate to exit 4.

[0078] S22. Using the fragmentation effect social force model, calculate the resultant force on each pedestrian in each zone of each exit scheme, and then calculate the time for each pedestrian to reach the target exit of their zone based on the resultant force:

[0079] S221. Calculate the resultant force on each pedestrian in each zone of each exit scheme at the current moment using the social force model of fragment scattering effect;

[0080] S222. Determine the distance of each pedestrian to the target exit of their respective zone. Based on the mass and net force acting on each pedestrian at the current moment, calculate their acceleration at the current moment. ;

[0081] S223. Calculate the displacement of each pedestrian at the current moment based on their current acceleration and velocity. Then, determine whether the pedestrian has reached the target exit. If so, proceed to step S224; otherwise, update their speed for the next moment. and return to step S221;

[0082] S224. Record the evacuation time of pedestrians, and proceed to step S23 when each pedestrian in each zone of each exit plan has received their evacuation time.

[0083] Before time calculation, the coordinates of each exit inside the building are known. SFLA will pass the position of each exit from several generated exit schemes as parameters to the Fragmentation Effect Social Force Model (FSSFM). At this point, the initial position of pedestrian i is known, and it can be calculated based on... The formula is used to calculate the unit direction vector of pedestrian i at time t, pointing from its current position to the desired exit. Then, the formula of FSSFM can be used to calculate the resultant force on each pedestrian in each group of people, and then the evacuation time can be calculated.

[0084] S23. Take the evacuation time of the last pedestrian in each partition of each exit scheme as the fitness value of the exit scheme, and determine whether the number of iterations is less than the preset number of iterations. If so, proceed to step S24; otherwise, proceed to step S25.

[0085] S24. Return the fitness value of each exit scheme to the swarm intelligence algorithm for iteration to generate different new exit schemes, and then return to step S22.

[0086] S25. Save the exit scheme corresponding to the lowest fitness value as the optimal evacuation exit scheme. Determine whether the number of saved optimal evacuation exit schemes has reached the preset number. If so, output all optimal evacuation exit schemes; otherwise, proceed to step S26.

[0087] S26. Adjust the population information and / or emergency information for each zone, then return to step S21.

[0088] In step S3, the site environment, zoning population information, and emergency event information are taken as inputs, and the optimal evacuation exit plan is taken as output. The decision tree model is trained using all inputs and outputs to obtain the real-time emergency evacuation decision model.

[0089] In implementation, this plan preferably includes the site environment, the path distance from each zone center to each exit, and the width of each exit; the emergency information includes the emergency type, the location of the emergency, the impact index of the emergency hotspot, and the distance from each zone center to the emergency hotspot; the zoned crowd information includes the overall crowd density of the indoor site, the number of people in each zone, the crowd density in each zone, and the initial position of each pedestrian.

[0090] Step S26 adjusts the zonal population information and / or emergency information by adjusting at least one of the parameters of the zonal population information and emergency information.

[0091] In step S4, the population information of each zone of the indoor venue under the emergency is collected, and the information, along with the venue environment and emergency information, is input into the real-time emergency evacuation decision model to obtain the optimal evacuation exit plan.

[0092] In step S5, the optimal target exit for each zone of the population in the optimal evacuation exit plan is transmitted to an independent drone. All drones in each zone use time-sharing broadcasting to guide the population of their respective zones to evacuate to the designated optimal exit.

[0093] During implementation, the preferred method for emergency evacuation of densely populated indoor areas in this plan also includes constructing an evacuation route database: selecting the pedestrian path in each zone within all optimal evacuation exit plans in step S2 whose evacuation time is closest to the average evacuation time of the entire zone as the optimal evacuation route; and using all optimal evacuation routes to construct an evacuation route database.

[0094] Between step S4 and step S5, the following is also included:

[0095] A1. Based on the model input and output of each partition in step S4, calculate the similarity between each partition and the corresponding input and output of each optimal evacuation path in the evacuation path library in step S3.

[0096] A2. Determine if there is an optimal evacuation path with a similarity greater than a preset threshold. If it exists, use it as the optimal evacuation path for the corresponding evacuation zone population in the optimal evacuation exit plan. If it does not exist, proceed to step A3.

[0097] A3. Input the information of the pedestrians in the population of the partition that are farthest from their optimal target exit into the social force model of the fragmentation effect, calculate the evacuation path of the pedestrian, and use it as the optimal evacuation path for the population of the partition.

[0098] Step S5 also includes transmitting the optimal evacuation route of the partition to the corresponding drone. The drone flies according to the optimal evacuation route and guides the people in its partition to the corresponding optimal exit in combination with time-sharing broadcast.

[0099] In one embodiment of the present invention, the quantification influence coefficient of the expected velocity of the fragments is calculated. The expression is:

[0100]

[0101] in, Let be the distance between pedestrian i and the center of the emergency at time t. Let be the radius of influence of the sudden event at time t; max and min are the maximum and minimum values, respectively.

[0102] The expression is:

[0103]

[0104] in, Let be the desired exit position vector of pedestrian i at time t. Let be the position vector of pedestrian i at time t. If there is an obstacle (such as a wall) between pedestrian i and the desired exit, then the endpoint of the obstacle is set as the temporary desired exit, making The system points to the temporary desired exit; after pedestrian i goes around, the system returns to the desired exit position. The modulus of the orientation quantity.

[0105] Psychological Repulsion and fragmentation effect repulsion force The expressions are as follows:

[0106] ,

[0107] in, It is the psychological repulsion constant; and All are constants. This indicates the range of psychological repulsion; a larger value indicates that pedestrian i is more inclined to start avoiding others sooner. This indicates the range of the fragment repulsion force; the larger the value, the greater the impact of the fragmentation effect on pedestrian i around the sudden hotspot. >0 indicates that pedestrian i and pedestrian j have already been physically squeezed at time t, and the larger the difference, the more severe the squeezing; = 0 indicates that pedestrian i and pedestrian j just make contact at time t; <0 indicates that pedestrian i and pedestrian j did not have physical contact at time t; The constant of the fragment repulsion force; It is an exponential function;

[0108] This represents the degree of impact of pedestrian j on pedestrian i due to fragmentation, and its value range is [0,1]. =0 indicates normal movement of people indoors; =1 indicates that pedestrian i is most affected by the fragment impact from pedestrian j. Calculate the fragmentation coefficient. The expression is:

[0109]

[0110] Where (CV + CD + CA) ≤ 3, they are adjustments The three weighting factors are: CV is the step velocity weighting factor, and when CV ≥ 0.2, it indicates that pedestrian j has a fragmentation motion effect, and the fragmentation factor... This will have an impact; when CV < 0.2, it indicates that the pedestrian's velocity j has not changed significantly and does not yet show a tendency for fragmentation. At this time, the fragmentation coefficient... =0; CD is the fragment distance range factor; CA is the pedestrian scattering angle factor.

[0111] ,

[0112] in, Let be the speed difference between pedestrian j at time t and the pedestrian's standard walking speed (1.34 m / s); Let be the velocity of pedestrian j at time t; This is the standard walking speed for pedestrians.

[0113]

[0114] in, Let be the distance between the centers of mass of pedestrians i and j at time t; when <0.8m means the distance between pedestrian j and pedestrian i is less than 0.2m (assuming the sum of the physical radii of the pedestrians). =0.6m), which is very close, and the fragment distance factor reaches a maximum of 1; =1.2m represents the critical distance at which fragmentation occurs between pedestrians. This indicates that pedestrian i is within the distance range of the fragment of pedestrian j, when >1.2m indicates that pedestrian i is outside the distance range of pedestrian j's fragment.

[0115]

[0116] Where W is the control parameter for the intensity of the opposite motion; T is the control attenuation coefficient for the same-direction motion; Let be the angle between the motion direction vector of pedestrian j at time t and the direction vector of pedestrian j pointing towards pedestrian i; Let be the angle between the motion direction vector of pedestrian i at time t and the direction vector of pedestrian i pointing towards pedestrian j.

[0117] This scheme defines the cosine value of the meeting angle between pedestrian j and pedestrian i as:

[0118]

[0119] in, It is the cosine of the angle between the direction vector of pedestrian j's movement and the direction vector of j pointing towards pedestrian i, through... We can determine the position and direction of pedestrian i relative to pedestrian j. For example... When =1, or This means that pedestrian i is directly in front of pedestrian j; for example... When =-1, This means that pedestrian i is directly behind pedestrian j; It is the dot product of vectors.

[0120] when(( )or( ))and(( )or( When )), it means that pedestrian i and pedestrian j are walking towards each other on a straight line, and the influence value of the fragment scattering angle reaches its maximum.

[0121] when This indicates that pedestrian j is moving in the direction of pedestrian i, pedestrian i is in front of pedestrian j, and pedestrian j and pedestrian i are moving towards each other. In this case, pedestrian j has a very large influence on the fragmentation effect of pedestrian i. and When pedestrian i and pedestrian j are walking towards each other face to face, the fragmentation effect has the greatest impact. This mainly adjusts the sensitivity of pedestrian i and pedestrian j to the front and left / right sides when they move towards each other. When W=0, it means that pedestrian i only pays attention to the fragmentation effect caused by pedestrian j within its frontal field of vision, and completely ignores pedestrian j if it is located to the left or right of pedestrian i's field of vision. When W=1.0, it means that pedestrian i treats the fragmentation effect caused by pedestrian j within its frontal field of vision and to the left and right sides equally. In this case, CA=1 (CA=1 when CA≥1). In this scheme, the default value of W is 0.3.

[0122] when This indicates that pedestrian i is behind pedestrian j in the direction of movement. At this time, pedestrian j has a very small impact on the fragmentation effect of pedestrian i. For the sake of calculation simplicity, we set it to 0.

[0123] when This indicates that pedestrian j is moving in the direction of pedestrian i, pedestrian i is in front of pedestrian j, and pedestrian j and pedestrian i are moving in the same direction. In this case, pedestrian j has little impact on the fragmentation effect of pedestrian i. When T=0, it means that when pedestrian i is in front of pedestrian j and pedestrian j and pedestrian i are moving in the same direction, there is no fragment dispersion effect, and CA=0; when T=1, it means that the fragment dispersion effect is the same when moving in the same direction and moving in opposite directions; in this scheme, the default value of T is 0.15, that is, when pedestrian j and pedestrian i are moving in the same direction, the fragment effect formed on pedestrian i is 0.15 of that when moving in opposite directions.

[0124] This solution prioritizes indoor locations, where UWB and video data acquisition equipment are installed on-site to collect pedestrian information. Each area is assigned a drone, each equipped with a UWB terminal. The drone's real-time location can be tracked via a UWB base station. The drone's onboard camera records the crowd evacuation process, providing data support for post-event analysis. The drones are equipped with electronic screens and loudspeakers. The screens prominently display the drone's assigned area group information (e.g., Area A), its exit, exit location, and evacuation guidance route. Simultaneously, the loudspeakers guide pedestrians within the area to the designated exit, such as, "Pedestrians in Area A, please proceed to Exit 1 and follow the drone's evacuation guidance."

[0125] To avoid confusion in crowd guidance information, drones can employ time-sharing broadcasting and spatial zoning. Simultaneously, large screens and audio alerts are placed in areas A and B, also using time-sharing and zone-specific broadcasting to prevent information confusion. The information displayed on the drones is synchronized and consistent with the information on the screens. This combination of static and dynamic methods achieves efficient and adaptive evacuation flow allocation, improving crowd evacuation efficiency.

[0126] In summary, this solution, through the construction of a social force model that considers the fragmentation effect, can accurately describe the transient evacuation behavior characteristics of pedestrians near sudden hotspots as they disperse and flee in all directions, thus ensuring the accuracy of evacuation route planning; furthermore, by combining drones to dynamically guide the crowd in each zone, evacuation efficiency is improved.

Claims

1. An emergency evacuation method for densely populated indoor areas considering the fragmentation effect, characterized in that, Including the following steps: S1. Read the zoning information of the indoor venue and calculate the social force model of the fragmentation effect of the resultant force on pedestrians in a sudden event; S2. Using swarm intelligence algorithm and fragmentation effect social force model, calculate the different optimal evacuation exit schemes corresponding to the location of the emergency and the location of the population in each zone under different distribution conditions. Each optimal evacuation exit scheme includes the optimal target exit for each population in each zone. S3. Using the site environment, zoning population information, and emergency event information as inputs, and the optimal evacuation exit plan as output, the decision tree model is trained using all inputs and outputs as a training set to obtain a real-time emergency evacuation decision model. S4. Collect population information for each zone of the indoor venue during an emergency, and input it along with the venue environment and emergency information into the real-time emergency evacuation decision model to obtain the optimal evacuation exit plan. S5. Transmit the optimal target exit of each zone's population in the optimal evacuation exit plan to an independent drone. All drones in each zone use time-sharing broadcasting to guide the population of their respective zones to evacuate to the designated optimal exit. The expression for the social force model of the fragmentation effect is: in, Let be the net force acting on pedestrian i at time t; Indicates mass as pedestrian i at the expected speed Towards the desired export direction Its own driving force; The quality of pedestrian i; Let be the expected speed of pedestrian i; Quantify the influence coefficient for the expected velocity of the fragments; For the desired speed increment; Let be the unit direction vector pointing from the current position of pedestrian i at time t to the desired exit. Let be the velocity of pedestrian i at time t; The reaction time constant; Let t be the force exerted by pedestrian j on pedestrian i. Let w be the force between pedestrian i and obstacle w at time t; This refers to the fragmentation coefficient; Let t be the psychological repulsion between pedestrians i and j. The fragmentation repulsion force between pedestrians i and j at time t; K is the positive compressive force coefficient; For contact functions; The sum of the physical radii of pedestrian i and pedestrian j's bodies; Let be the distance between the centers of mass of pedestrians i and j at time t; Let be the unit normal vector pointing from pedestrian j to pedestrian i at time t; The coefficient of sliding friction; Let be the difference in tangential velocity between pedestrians i and j at time t; For time t and The perpendicular unit tangent vector.

2. The indoor dense crowd emergency evacuation method considering the fragmentation effect according to claim 1, characterized in that, It also includes constructing an evacuation route library: selecting the pedestrian path in each partition within all optimal evacuation exit schemes in step S2 whose evacuation time is closest to the average evacuation time of the entire partition as the optimal evacuation route; and using all optimal evacuation routes to construct an evacuation route library. Between step S4 and step S5, the following is also included: A1. Based on the model input and output of each partition in step S4, calculate the similarity between each partition and the corresponding input and output of each optimal evacuation path in the evacuation path library in step S3. A2. Determine if there is an optimal evacuation path with a similarity greater than a preset threshold. If it exists, use it as the optimal evacuation path for the corresponding evacuation zone population in the optimal evacuation exit plan. If it does not exist, proceed to step A3. A3. Input the information of the pedestrians in the population of the partition that are farthest from their optimal target exit into the social force model of the fragmentation effect, calculate the evacuation path of the pedestrian, and use it as the optimal evacuation path for the population of the partition. Step S5 also includes transmitting the optimal evacuation route of the partition to the corresponding drone. The drone flies according to the optimal evacuation route and guides the people in its partition to the corresponding optimal exit in combination with time-sharing broadcast.

3. The indoor dense crowd emergency evacuation method considering the fragmentation effect according to claim 1, characterized in that, Step S2 further includes: S21. Based on the zoning crowd information and emergency information in the indoor venue, a swarm intelligence algorithm is used to generate multiple different exit schemes for the indoor venue, wherein the exit schemes include the target exits for each zone. S22. The social force model of fragmentation effect is used to calculate the resultant force on each pedestrian in each zone in each exit scheme. Then, the time for the pedestrian to reach the target exit of its zone is calculated based on the resultant force. S23. Take the evacuation time of the last pedestrian in each partition of each exit scheme as the fitness value of the exit scheme, and determine whether the number of iterations is less than the preset number of iterations. If so, proceed to step S24; otherwise, proceed to step S25. S24. Return the fitness value of each exit scheme to the swarm intelligence algorithm for iteration to generate different new exit schemes, and then return to step S22. S25. Save the exit scheme corresponding to the lowest fitness value as the optimal evacuation exit scheme. Determine whether the number of saved optimal evacuation exit schemes has reached the preset number. If so, output all optimal evacuation exit schemes; otherwise, proceed to step S26. S26. Adjust the population information and / or emergency information for each zone, then return to step S21.

4. The indoor dense crowd emergency evacuation method considering the fragmentation effect according to claim 3, characterized in that, Step S22 further includes: S221. Calculate the resultant force on each pedestrian in each zone of each exit scheme at the current moment using the social force model of fragment scattering effect; S222. Determine the distance of each pedestrian to the target exit of their respective zone, and calculate their acceleration at the current moment based on their mass and the net force acting on them. S223. Calculate the displacement of each pedestrian at the current moment based on their acceleration and velocity. Then determine whether the pedestrian has reached the target exit. If yes, proceed to step S224; otherwise, update their velocity at the next moment and return to step S221. S224. Record the evacuation time of pedestrians, and proceed to step S23 when each pedestrian in each zone of each exit plan has received their evacuation time.

5. The indoor dense crowd emergency evacuation method considering the fragmentation effect according to claim 1, characterized in that, Calculate the quantification influence coefficient of the expected velocity of the fragments. The expression is: in, Let be the distance between pedestrian i and the center of the emergency at time t. Let be the radius of influence of the sudden event at time t; max and min are the maximum and minimum values, respectively. The expression is: in, Let be the desired exit position vector of pedestrian i at time t. Let be the position vector of pedestrian i at time t; The modulus of the orientation quantity.

6. The indoor dense crowd emergency evacuation method considering the fragmentation effect according to claim 1, characterized in that, Calculate the force between pedestrian i and obstacle w The expression is: in, The force coefficient between pedestrian i and obstacle w; It is an exponential function; The range of the repulsive force between pedestrian i and obstacle w is defined; Let i be the physical radius of the pedestrian's body. Let be the distance from pedestrian i to obstacle w at time t. Let be the unit normal vector pointing from obstacle w to pedestrian i at time t; For time t and The perpendicular unit tangent vector.

7. The indoor dense crowd emergency evacuation method considering the fragmentation effect according to claim 1, characterized in that, Psychological Repulsion and fragmentation effect repulsion force The expressions are as follows: ， in, It is the psychological repulsion constant; and All are constants; The constant of the fragment repulsion force; It is an exponential function; Calculate the fragmentation coefficient The expression is: ， ， Wherein, CV is the step velocity weighting factor; CD is the fragment distance range factor; and CA is the pedestrian scattering angle factor. Let the speed of pedestrian j at time t be the same as the standard walking speed of pedestrians. The speed difference between them; Let be the velocity of pedestrian j at time t; The standard walking speed for pedestrians; The critical distance at which fragmentation occurs between pedestrians; W is the control parameter for the intensity of opposite movement; T is the attenuation coefficient for the control of same-direction movement. Let be the angle between the motion direction vector of pedestrian j at time t and the direction vector of pedestrian j pointing towards pedestrian i; Let be the angle between the motion direction vector of pedestrian i at time t and the direction vector of pedestrian i pointing towards pedestrian j.

8. The indoor dense crowd emergency evacuation method considering the fragmentation effect according to claim 3, characterized in that, The site environment includes the site area, the path distance from each zone center to each exit, and the width of each exit; the emergency information includes the emergency type, the location of the emergency, the impact index of the emergency hotspot, and the distance from each zone center to the emergency hotspot; the zoned crowd information includes the overall crowd density of the indoor site, the number of people in each zone, the crowd density in each zone, and the initial position of each pedestrian. Step S26 adjusts the zonal population information and / or emergency information by adjusting at least one of the parameters of the zonal population information and emergency information.

Images