CN104239636B - Fire emergency evacuation simulation method - Google Patents
Fire emergency evacuation simulation method Download PDFInfo
- Publication number
- CN104239636B CN104239636B CN201410471493.7A CN201410471493A CN104239636B CN 104239636 B CN104239636 B CN 104239636B CN 201410471493 A CN201410471493 A CN 201410471493A CN 104239636 B CN104239636 B CN 104239636B
- Authority
- CN
- China
- Prior art keywords
- individual
- escape
- speed
- fire
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000004088 simulation Methods 0.000 title claims abstract description 36
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 55
- 238000004364 calculation method Methods 0.000 claims abstract description 27
- 230000000694 effects Effects 0.000 claims abstract description 10
- 239000000779 smoke Substances 0.000 claims description 29
- 238000009792 diffusion process Methods 0.000 claims description 18
- 238000005381 potential energy Methods 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 15
- 239000002245 particle Substances 0.000 claims description 11
- 230000003993 interaction Effects 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 8
- 238000005070 sampling Methods 0.000 claims description 4
- 238000013507 mapping Methods 0.000 claims description 3
- 230000035484 reaction time Effects 0.000 claims description 3
- 230000000007 visual effect Effects 0.000 claims 2
- 239000012530 fluid Substances 0.000 claims 1
- 230000002093 peripheral effect Effects 0.000 claims 1
- 238000011160 research Methods 0.000 abstract description 4
- 238000012795 verification Methods 0.000 abstract description 2
- 230000006399 behavior Effects 0.000 description 23
- 230000003044 adaptive effect Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 230000008033 biological extinction Effects 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 230000007246 mechanism Effects 0.000 description 3
- 230000010355 oscillation Effects 0.000 description 3
- 230000006978 adaptation Effects 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 241000282414 Homo sapiens Species 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000002485 combustion reaction Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 238000001125 extrusion Methods 0.000 description 1
- 231100000225 lethality Toxicity 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a fire emergency evacuation simulation method. On the basis of establishing a real three-dimensional fire scene, researches are made around a vector-field-based global motion planning algorithm to navigate the escape of a plurality of individuals in the fire scene in real time. The method comprises the following steps of calculating a vector-field-based safety field in real time by virtue of the algorithm, constructing a fire-scene-based crowd behavior model in combination of a social force model-based local collision avoiding algorithm, and finally providing fire escape guidance for each escape individual in the scene according to the fire-scene-based crowd behavior model. According to the fire emergency evacuation simulation method, a theoretical basis is provided, the calculation accuracy can be ensured, and verification application is developed under various experimental conditions; the method has the characteristics of low calculation complexity, high simulation effect trueness and the like, and the real-time emergency evacuation of hundreds of escape individuals can be simulated during practical application.
Description
Technical Field
The invention relates to a fire emergency evacuation simulation method.
Background
It is well known that the panic escape of people caused by fire is one of the most catastrophic collective behaviors of human beings. In a fire, people generate panic and trample and collide with each other, and finally a great deal of casualties are caused. Displaying data of the public security department: in 2011, the fire 125402 is reported all over the country, 1106 dead people and 572 injured people suffer direct property loss of 18.8 million yuan, and compared with 2010, the number of dead people is increased by 3.3%. Although many modern buildings are provided with fire early warning devices, it is difficult for these devices to truly prevent the occurrence of fire and to provide sufficient assistance for crowd evacuation and escape during fire. Therefore, the importance of emergency evacuation simulation for fire scenes is also increasing by using computer means and corresponding modeling methods.
In a motion planning method for fire emergency evacuation simulation, a flame source causing a fire is abstracted into a dangerous source with specific parameters in the current stage, and the characteristics of the flame source are utilized to influence the calculation of an escape route of an escape individual or the change of potential energy in a certain grid area in the operation process. The idea has wide applicability, but has insufficient pertinence to fire emergency evacuation simulation. In the research of the fire crowd behavior model, due to the lack of effective data of escapers in a real fire, the number of crowd behavior models aiming at a fire scene is small, and most of the models are established on the basis of experimental data. However, in a wider scope, since 2000, the panic model proposed by helling et al describes the uncoordinated movement of the pedestrian individuals in the crowd due to panic and crowding, and the model predicts the evacuation condition of the pedestrian individuals through the calculation of a mechanical model so as to simulate the occurrence of extrusion in the crowd. However, if the panic model is used in the simulation of emergency evacuation of a fire, it must be modified and adapted to the fire behavior.
The challenges faced in fire emergency evacuation simulation are mainly from two sides: the method aims at a real-time and accurate motion planning algorithm of an escaper in a fire scene and a crowd behavior model in the fire scene with high reality.
For multi-individual motion planning algorithms, high accuracy of planning often means huge computational overhead, so that real-time performance cannot be achieved. And emergency scenes such as fire and the like have the characteristic of high dynamic property, and each escape individual needs to react to the current environment in a very short time to select the motion state of the next moment, so that a motion planning algorithm aiming at the fire scene has high real-time property, and thus, the multiple escape individuals are navigated in real time. In response to this fact, the present invention proposes the concept of a safety field that reflects the hazard level of each terrain grid within the fire scene, which can provide guidance to the final safe exit for the escaper on a global scale. In addition, to improve the accuracy of the motion planning, it is necessary to calculate the local dynamics of each of the evacuees. The present invention utilizes a validated social force model to provide accurate local dynamics calculations for each individual. And navigating a plurality of bodies in the fire scene by combining a local motion planning algorithm and a global motion planning algorithm. The algorithm achieves high accuracy while real-time performance is obtained.
Due to scarcity and complexity of a crowd behavior model in a fire, fire evacuation simulation with high reality degree is difficult to realize. According to the invention, through research on documents in the fire control safety field, a smoke concentration-escape speed formula which is verified in a large quantity is used for mapping with smoke concentrations of all positions in a fire scene, and a fire escape behavior model used by the invention is provided. The model can truly show the influence of smoke in the fire on the escape individual by using relatively few computing resources.
Disclosure of Invention
The technical problem solved by the invention is as follows: aiming at the motion characteristics of escape individuals in a fire disaster, a real-time and accurate motion planning algorithm for evacuees in a fire scene and a crowd behavior model with high reality in the fire scene are constructed from the global and local aspects. A fire emergency evacuation simulation method is provided. And by using the program framework based on OpenGL, a vivid simulation effect is drawn.
The technical scheme adopted by the invention is as follows: a fire emergency evacuation simulation method comprises the following three steps:
step (1), calculating the global speed of the escape individual by using a safety field: the algorithm gives a concept of a safety field based on a vector field and is used for describing dynamic conditions in a fire scene. Then, calculating the global speed of the escape individuals in the escape system by using a global motion planning algorithm based on a safety field;
step (2), calculating the local speed of the escape individual by using the social force model: taking an escape individual as a unit, taking the global speed calculated in the step (1) as an input parameter of the step, and calculating the real-time stress condition of each individual by using a local collision avoidance algorithm based on a social force model through introducing the social force model based on the individual so as to calculate the corresponding local speed;
and (3) calculating the final escape speed of the escape individual by using the fire escape behavior model: in the step, a flame source particle system in a fire scene is constructed, sampling is carried out according to the smoke concentration of flame particles, and the final escape speed of each escape individual is calculated through a fire escape behavior model according to the global speed of the escape individual calculated in the step (1) and the local speed calculated in the step (2).
The method is based on the OpenGL framework, the calculation complexity is reduced as much as possible on the premise of ensuring the simulation reality degree, and the real-time emergency evacuation simulation of hundreds of individuals can be supported in practical application.
Further, the safety field in step (1) can use an adaptive non-uniform safety grid to describe dynamic information of a fire scene in real time, and calculate an optimal escape direction in each grid by using the danger level, so as to provide navigation for escape individuals in the fire scene. According to the method, through the design of diffusion potential energy and grid self-adaptive grade, a smoother individual escape path is generated, and the oscillation phenomenon among individuals is effectively avoided.
Furthermore, the algorithm in the step (2) is based on a social force model which can truly describe the movement criterion of the crowd in the panic state, the social force model principle is applied to each escape individual, and the collision between the escape individuals in the fire scene is controlled within a certain reasonable range, so that the escape condition of the individuals in the fire is simulated at higher precision.
Further, the model in the step (3) utilizes a verified smoke concentration-escape speed formula to constrain the movement speed of the escape individual in smoke. The model enables human-fire interaction to be possible, and provides a more real behavior model for the simulation of individual escape in a fire scene.
Furthermore, all the calculation methods in the steps (1), (2) and (3) have theoretical basis and reduce the calculation complexity on the premise of ensuring the simulation reality.
The principle of the invention is as follows:
the invention provides a method for simulating emergency evacuation of people in a fire. The method comprises the steps of on the basis of establishing a real fire three-dimensional scene, aiming at carrying out real-time escape navigation on a plurality of bodies in the fire scene, developing research around a vector field-based global motion planning algorithm, firstly calculating a vector field-based safety field in real time by the algorithm, then constructing a crowd behavior model under the fire scene by combining with a social force model-based local collision avoidance algorithm, and finally providing fire escape guidance for each escape individual in the scene according to the model. The invention establishes a fire emergency evacuation simulation method with theoretical basis and guaranteed calculation precision, and develops verification application under various experimental conditions. The method has the characteristics of low computational complexity, high simulation effect fidelity and the like, and can support real-time emergency evacuation simulation of hundreds of escape individuals in practical application. The invention mainly comprises the following three aspects:
(1) the global motion planning algorithm based on the safety field shows different behavior modes in an emergency scene and a non-emergency scene, so that the real-time global path planning algorithm based on the non-uniform safety field is provided and realized. The algorithm assigns a risk level to each security mesh within the security farm to describe the risk condition within the current mesh. And then, the safety field calculates the optimal escape direction in each grid by using the danger level, so that navigation is provided for the escape individuals in the fire scene. According to the invention, through the design of diffusion potential energy and grid self-adaptive grade, a smoother individual escape path is generated, and the generation of oscillation phenomenon among individuals is effectively avoided.
(2) The method provides a local collision avoidance algorithm combined with a global path planning algorithm based on a safety field in order to improve the truth of fire emergency evacuation simulation. The algorithm is based on a social force model which can truly describe the movement criterion of people in a panic state, applies the social force model principle to each escape individual, and controls the collision among the escape individuals in a fire scene within a certain reasonable range, thereby simulating the escape condition of the individuals in the fire on higher precision.
(3) The method provides a crowd behavior model suitable for a fire evacuation scene. The smoke released by combustion is the first danger source in a fire disaster, and the model restrains the movement speed of an escape individual in the smoke by using a verified smoke concentration-escape speed formula. The model enables human-fire interaction to be possible, and provides a more real behavior model for the simulation of individual escape in a fire scene.
The method is based on the OpenGL framework, the calculation complexity is reduced as much as possible on the premise of ensuring the simulation reality degree, and the real-time emergency evacuation simulation of hundreds of individuals can be supported in practical application. .
Compared with the prior art, the invention has the advantages that:
1. the safety field designed by the invention can describe the dynamic information of a fire scene in real time by using the self-adaptive non-uniform safety grids, and calculate the optimal escape direction in each grid by using the danger level, thereby providing navigation for escape individuals in the fire scene. According to the method, through the design of diffusion potential energy and grid self-adaptive grade, a smoother individual escape path is generated, and the oscillation phenomenon among individuals is effectively avoided.
2. The local collision avoidance algorithm used by the invention is based on a social force model which can truly describe the movement criterion of people in a panic state, applies the social force model principle to each escape individual, and controls the collision among the escape individuals in a fire scene in a certain reasonable range, thereby simulating the escape condition of the individuals in the fire at higher precision.
3. The fire escape behavior model designed by the invention utilizes a verified smoke concentration-escape speed formula to restrain the movement speed of escape individuals in smoke. The model enables human-fire interaction to be possible, and provides a more real behavior model for the simulation of individual escape in a fire scene.
4. The calculation method used by the invention has theoretical basis, and the calculation complexity is reduced on the premise of ensuring the simulation truth of fire evacuation under the specific operation steps.
Drawings
FIG. 1 is a flow chart of a fire emergency evacuation simulation method;
FIG. 2 is a schematic view of an initial safety field;
FIG. 3 is a block diagram of a safety farm operation mechanism;
FIG. 4 is a flow chart of a global path planning algorithm based on a safety field;
FIG. 5 is a schematic view of a security mesh after splitting; (ii) a
FIG. 6 is a diagram showing the effect of single escape in an experiment;
FIG. 7 is a schematic diagram of interaction relationships of elements of a social force model;
FIG. 8 is a diagram showing the effect of 10 escape individuals moving in opposite directions;
fig. 9 is a schematic diagram of applying a fire escape behavior model to an individual escape process.
Detailed Description
Fig. 1 shows the general processing flow of the fire emergency evacuation simulation method, and the invention is further described with reference to other figures and embodiments.
The invention provides a fire emergency evacuation simulation method, which mainly comprises the following steps:
1. global motion planning algorithm based on safety field
The invention defines a non-uniform two-dimensional Security Field (SF) based on a global vector field by using the concept of Navigation Fields (SF).
a) Definition of Security field
First, the algorithm defines the security field SF as a set of unit vectors on a two-dimensional plane, i.e. SF: r2→S1，S1＝[0，2π]Wherein R is2Representing a two-dimensional Euclidean space, S1Representing the polar angle used to describe the vector.
The safety field SF is composed of several square safety grids. Each safety net grid is composed of a danger level, an adaptive level, diffusion potential energy and an escape direction, as shown in fig. 2.
Due to the high dynamics of the fire escape scene, the safety field describing the safety situation needs to read the dynamics in the field in real time and update each element of the safety field in time according to the dynamics, and the operation mechanism of each frame is shown in fig. 3.
As can be seen from the operation mechanism process diagram, each frame of the safety field needs to go through the steps of reading scene information, setting danger level, Astar routing, calculating adaptive level, calculating diffusion potential energy, updating danger level, calculating escape direction, judging and the like in the real-time updating process, and the specific operation method is detailed as follows.
Reading scene information: at each frame, the security field reads all the information in the experimental scene. These pieces of information include: the method comprises the following steps of obtaining position information of all individuals, static obstacles and dynamic obstacles in a scene, information of the movement speed, the direction and the target point of the individuals and the number of the individuals who escape, the number of safety grids in the scene and the self-adaptive level of the safety grids. This information will be used as input parameters to provide operational support for the next steps.
Setting a danger level: after successfully reading the information within the scene, the security farm sets a level of danger for all of the security grids it contains. It is noted that this step simply sets the initial risk level for the current frame for the security mesh, and that it is also possible to modify this risk level in the next step.
Astar way finding: after the setting of the danger levels of all safety grids in the safety field is finished, an Astar routing algorithm is utilized to generate a shortest escape path from a starting point to a target point of each escape individual in the scene, and an allowed walking belt of the shortest escape path is determined according to a certain rule. In particular, to ensure the safety of the escaper as much as possible, the safety ground requires that the escape route can only pass through a safety grid with a danger rating of 0. In the actual updating process, if the safety field finds that the current escape individual is in the walking belt allowed by the escape individual, the escape path does not need to be recalculated for the individual. In addition, for the escape individuals with the same shortest escape route (namely, the individuals with the same initial positions and the same target points), the concept of the travel-allowing belt also provides a wider motion space for the escape individuals, so that excessive collision among the individuals is avoided to a certain extent.
Calculating an adaptive grade: after the corresponding shortest escape path is calculated for each escape individual in the safety field, in order to save the calculation amount, the safety field only calculates the self-adaptive level of the safety mesh where each escape individual is located and the mesh in a certain range around the safety mesh by taking the individual as a unit. If the current original security mesh sgiAfter recalculation to di × di, it is split evenly to di2A plurality of security grids, each security grid being labeled as sgikjWherein j, k are integers and j, k ∈ [0, di-1 ]]，j，k∈N。
In particular, current algorithms only allow the original security mesh to be split once.
Calculating the diffusion potential energy: after the adaptive levels of the corresponding safety grids are calculated, the safety field calculates the diffusion potential energy of the safety grids with danger levels not being SAFE.
And updating the danger level: the purpose of this step is to have the hazard level within the current security mesh affected by the diffusion potential of its neighboring mesh. And after the safe field calculates the diffusion potential energy, recalculating the danger level of other grids in the safety grid neighborhood with the diffusion potential energy more than or equal to the diffusion threshold A.
Calculating the escape direction: after the corresponding danger levels of all the safety grids in the safety field are updated, the safety field calculates the escape direction in each safety grid. The escape direction is calculated by the danger level of the safety net lattice and the adjacent lattice. The invention requires that the escape direction should face the direction of the escape exit closest to the position of the escape exit as far as possible on the premise of ensuring safety, so that the escape individuals passing through the current safety net can escape from the fire scene as soon as possible.
b) Safety field calculation method
The global path planning algorithm based on the vector field (namely the safety field) provided by the invention is used for calculating the global planning speed v of each escape individualgThe process of (1). The speed reflects the movement speed and the movement direction of the escape individual based on the scene global consideration, and is a movement decision of the escape individual aiming at the final purpose of self escape. The algorithm block diagram is shown in fig. 4.
First, the algorithm targets the current original mesh (non-split security mesh) sgiRisk class sliTo calculate its adaptation level. Namely, traversing the safety grid sg of the current escape individual as the centeriA contiguous grid within a one-dimensional neighborhood. Algorithm set sgiIs a security mesh with which there is a common edge. As will be readily appreciated, there are four grids satisfying such a condition, each represented as sgi right，sgi up，sgi left，sgi downTheir corresponding hazard level is sli right，sli up，sli left，sli bottom. For sgiAnd assuming that the danger level of the grid where the escape individual is located is sl0Then its adaptation level diComprises the following steps:
the original security mesh is split into finer security meshes after the adaptive levels are calculated, as shown in fig. 5, it is obvious from the figure that the original security mesh is split into finer security meshes and exhibits different adaptive levels. For newly split safety grids sgikj，j，k∈[0，di-1]The algorithm specifies that its initial risk level is equal to the risk level of its corresponding original grid, and its calculation method is:
slijk＝sli
secondly, for the current security grid sgi(including split safety grids) diffusion potential EspiCorresponding danger class sliIs proportional to the square of the adaptive level diThe square of (a) is inversely proportional, and the calculation method is as follows:
it can be seen from the above formula that when the danger level of the current safety mesh is zero, the diffusion potential energy of the mesh is also zero, i.e. the mesh has no diffusion potential energy.
Current grid EspiIs mainly reflected in the risk level of its surrounding gridThe influence of (c). That is, a safety mesh with high diffusion potential can increase the danger level of its neighboring mesh, which is set by the present invention as follows:
subsequently, for the current security mesh, the global planning algorithm traverses its four neighboring security meshes in a counterclockwise direction and selects therefrom the neighboring security mesh with the lowest sum of the two risk levels:
wherein, sl isi' is the current safety grid sgiAfter traversing in the counterclockwise direction in four adjacent grids, the minimum sum of the risk levels in two adjacent directions. Then, the invention obtains sl according to the calculationiTo determine the respective two orthogonal directionsAnd
the algorithm then uses the calculated risk levels for the two orthogonal directionsAndthe difference is carried out according to the direction of the current frame, so that the safety grid sg at the current frame is obtained through calculationiBest escape direction αi：。
It is particularly noted that the diffusion potential energy also affects the escape direction of the safety net cells, i.e. it generates a "repulsion" action on the escape direction of the surrounding safety net cells, in other words, makes the escape direction of these cells deviate from the connection direction of the cells with the high diffusion potential energy:
wherein, αi adjacentDenotes sgi right，sgi up，sgi left，sgi downCorresponding escape direction αi right，αi up，αileft，αi downA collection of (a).
c) Calculation of escape individual global speed
Global planning velocity vgThe motion speed and the motion direction obtained by the escape individual based on the scene global consideration are motion decisions made by the escape individual for the final purpose of escape.
Firstly, the algorithm sets a uniform preference speed | V | for each escape individual, and the direction of the speed vector points to the target point of the escape individual from the starting point of the escape individual. I.e. p for the initial position0kA escaping individual akThe invention prefers the speed V to0 kThe method comprises the following steps:
secondly, each escape individual finds the safety grid where the escape individual is located by acquiring real-time information of the safety field. In particular, for the escaping individual akLet its current security grid be sgi。
Finally, the pair is in the safety grid sgiCurrent individual a ofkIn other words, the global programming speed v of the next frameg kComprises the following steps:
wherein, sl isiAnd αiFor safety grid sgiThe corresponding danger level and the escape direction. It can be seen that the global planning velocity v given by the algorithmgThe initial preference speed direction of the escape individual and the escape direction of the safety grid where the escape individual is located are subjected to compromise calculation, and the magnitude of the global planning speed is in inverse proportion to the danger level of the grid. The purpose of this is: the escape individual can avoid the danger source as much as possible while keeping the long-term plan (escape to the target point) of the escape individual, and the movement speed is reduced when the escape individual moves to a high-risk place. The operational effect is shown in fig. 6.
2. Local collision avoidance algorithm based on social force model
The global motion planning algorithm based on the safety field provides escape navigation of a global view angle for escape individuals in a scene, but when the number of the escape individuals is large, the algorithm cannot ensure that the escape individuals are prevented from colliding within a certain range. Therefore, the invention provides a low-computation-complexity local collision avoidance algorithm based on a Social force model (Social forces model) in pedestrian individual dynamics on the premise of ensuring certain computation accuracy by combining the mainstream local collision avoidance algorithm and the mainstream model at the present stage so as to achieve a more realistic simulation effect of fire emergency evacuation.
a) Construction of social force model the present invention simulates a social force model using repulsive forces between escaping individuals, between an individual and an obstacle, and a driving force for driving the individual to reach a desired speed, with a force model as shown in fig. 7. In an emergency situation, repulsive forces F are generated between the escapers to prevent the escapers from approaching each otherrep. Accordingly, the escaper also generates a driving force F directing it to the target pointatt. Specifically, let a denote the set of all escape individuals in the experimental scene, and O denote the set of all obstacles in the experimental scene. Then for the escaping individual akIn other words, the resultant force F (a) acting thereonk) Comprises the following steps:
wherein,
wherein tau is the reaction time of the escaping individual, mk、rkAnd pkRespectively represent escape individuals akMass, radius and current position of vkAndare respectively escape individuals akThe current movement speed and the desired speed.
Where μ is the sociality measure constant of the individual, γ is the personal space descent constant of the individual, nkjRepresentative of an individual akAnd ajThe normal vector of the line segment between.
Here, C is an obstacle metric constant of the individual, and D is an obstacle distance decreasing constant of the individual. dkoIndicating an individual a for escapekThe closest distance to the obstacle o. n iskoIs composed of an individual akA vector pointing to the nearest point on the obstacle o.
b) Calculation of local speed of escape individual
To reduce the amount of unnecessary computation, the present invention assumes that all individuals within a scene have faciesThe same quality, shape and size. Then for the escaping individual akThe invention can receive the driving force FattExpressed as:
wherein v iskFor escaping individual akThe speed of motion of the previous frame. This shows that the algorithm calculates the escape individual a in the safe place in each frame of calculationkGlobal planning velocity vg kThe local collision avoidance speed of the current frame is calculated by a Newton classical mechanics formula and is used as an input parameter of a local collision avoidance algorithmIn conclusion, for the escape individual akDuring the reaction time t of the reaction,this can be given by the following equation:
by aligning the local velocity vlThe algorithm utilizes a social force model suitable for a fire scene to provide a support of local dynamics for a plurality of escape individuals in the fire scene, and the demonstration effect is shown in fig. 8.
3. Construction of fire escape behavior model
a) Construction of fire escape behavior model based on smoke concentration
The invention samples the number of flame particles in a unit volume to calculate the approximate smoke concentration in a certain area. Because the number of flame particles in an experimental scene is large, in order to reduce the calculated amount, the invention only aims at each escape individual within a certain rangeThe flame particle number is sampled. The algorithm only calculates a certain radius n around the current escape individuali(the side length of one safety grid in the safety field) and approximately considering the sum of the numbers of the flame particles as the smoke concentration in the environment where the current escape individual is located. I.e. the current individual akSmoke concentration c effective for the surroundingskCan be expressed as:
wherein,indicating an individual a for escapekCurrent Position of (POS) (par)i) Indicating flame particles pariPar represents a flame particle satisfying the condition, ckIs the number of elements in the right set of equations. The invention stipulates in actual operation if ck>C, namely when the number of the sampled particles is larger than the upper limit C of the ion concentration, the algorithm automatically stops sampling the smoke concentration around the current individual, because excessive sampling not only consumes a large amount of computing resources, but also has no obvious significance on computing the movement speed of the escape individual.
The invention utilizes the smoke concentration-escape speed formula to calculate the motion speed scalar value of an individual, and then solves the escape speed equation of the individual to obtain the final escape speed of the individual, thereby providing final navigation for an escaper in a fire scene and achieving the purpose of establishing an escape behavior model.
Aiming at a large amount of smoke existing in a fire scene, the invention establishes a foundation for a fire escape behavior model by using a smoke concentration-escape speed formula provided by Frantzich and Nilsson. This equation can be expressed as:
v(Ks)＝μ+γKs
wherein, mu is 0.706m/s, gamma is-0.057 m2/s，KsIs extinctionThe Coefficient (expression Coefficient) represents the visibility information of the space around the escaping individual, v (K)s) The extinction coefficient of the escape individual is KsThe speed of motion in the environment of (a). Thus, for the escaping individual akThe algorithm may use the smoke concentration c around it, which has been calculatedkAnd extinction coefficient KsAnd establishing a mapping relation. The invention discovers the smoke concentration c through a large number of experimentskAnd extinction coefficient KsThe relationship between them is:
wherein, sigma is positive number and represents the current lethality of smoke, and sigma represents the current escape individual akDegree of panic. At this time, the smoke concentration-escape speed formula in the fire escape behavior model may be expressed as:
in practical calculation, the invention finds that 0.248 and 1 can satisfy most of fire emergency evacuation simulation scenes.
b) Solution of individual escape velocity equation
Due to the global programming speed vgWith local collision avoidance velocity vlThe invention respectively reflects the motion decision of the escape individual from different angles, and the invention utilizes a weighted summation mode to calculate the final escape speed v of the individual in the fire, and the calculation method comprises the following steps:
v＝λ·vg+(1-λ)·vl
wherein λ is the global programming speed vgAnd local velocity vlUnknown weighting parameters in between. The invention can calculate that the escape speed of each escape individual in a scene is high under a certain smoke concentrationSmall, the value of λ can be determined by solving the following equation:
the invention obtains an individual escape speed equation, namely the final escape speed v of the individual can be obtained, and the demonstration effect is shown in figure 9:
technical contents not described in detail in the present invention belong to the well-known techniques of those skilled in the art.
Although illustrative embodiments of the present invention have been described above to facilitate the understanding of the present invention by those skilled in the art, it should be understood that the present invention is not limited to the scope of the embodiments, and various changes may be made apparent to those skilled in the art as long as they are within the spirit and scope of the present invention as defined and defined by the appended claims, and all matters of the invention which utilize the inventive concepts are protected.
Claims (1)
1. A fire emergency evacuation simulation method is characterized by comprising the following steps: the method comprises the following three steps:
step (1), calculating the global speed of the escape individual by using a safety field: the two-dimensional vector field is represented as a plurality of arrows with lengths and directions on a plane, that is, the number of each area in a plane space can be represented by a vector to represent the characteristics of the area, such as speed, motion trend and the like, and based on the principle of the vector field, the safety field is defined as a unit vector set on the two-dimensional plane, namely SF: r2→S1，S1＝[0，2π]Wherein R is2Representing a two-dimensional Euclidean space, S1Representing the polar angle used for describing the vector and describing the dynamic situation in the fire scene by using the polar angle; then, by mapping the dynamically updated safety field into a global cost map, a global motion planning algorithm is designed to calculate the global speed of the escape individual in the global speed map, namely: giving the state Sk (t) of each escaping individual ak in the safety field at the moment t, and calculating the state Sk (t + tau) of the individual at the next moment by using the safety field so as to navigate the individual to the target point gk;
the social force model established based on the fluid dynamics can be used for describing the interaction between the moving pedestrian individuals and the interaction between the individuals and the boundary and the obstacle, the model is not only suitable for calculating the local dynamics of the crowd in a daily state, but also suitable for simulating the local dynamics of the crowd in panic in an emergency state, and the social force model is used for calculating the local speed of the escaping individuals: taking an escape individual as a unit, taking the global speed calculated in the step (1) as an input parameter of the step, and calculating the real-time stress condition of each individual by using a local collision avoidance algorithm based on a social force model through introducing the social force model based on the individual so as to calculate the corresponding local speed;
step (3), after calculating the global planning speed vg, the local collision avoidance speed vl and the smoke concentration of each escape individual within a certain range, calculating the motion speed scalar value of the individual by using a smoke concentration-escape speed formula, and obtaining the final escape speed of the individual by solving an individual escape speed equation, thereby establishing a fire escape behavior model for providing navigation for evacuees in a fire scene: the method comprises the following steps of (1) constructing a flame source particle system in a fire scene, sampling smoke concentration of flame particles, and calculating a final escape speed of each escape individual through a fire escape behavior model according to the global speed of the escape individual calculated in the step (1) and the local speed calculated in the step (2);
the global motion planning algorithm based on the safety field comprises the steps of calculating the safety field and calculating the global planning speed of an escape individual: aiming at each frame in the simulation process, reading dynamic information of a fire scene of the previous frame in advance by a path planning algorithm based on a safety field, then sequentially executing the safety field in the self-adaptive updating step (1) of danger level setting, Astar path finding, self-adaptive level calculation, diffusion potential energy calculation, danger level updating and escape direction calculation operation, and describing the dynamic information of the fire scene in real time; after the dynamic updating of each element value in the safety field is completed, the algorithm takes the escape individuals in the scene as units and calculates the global planning speed of each escape individual by combining the attributes of the individuals;
the social force model described in step (2) implies a modeled representation of the driving force, repulsive force, attractive force, pedestrian gross movement trends (1) the driving force is the most significant force in the model that determines the movement of a pedestrian individual towards a destination at a desired speed, the model assuming that if no external factors interfere with the movement of the current pedestrian individual α, the individual will move at speedTo a desired directionPerforming movement; the model passes through a defined "reaction time" tau due to the necessary deceleration or evasion processes during the actual movement of the individualαTo correct the actual speed of the individualWith desired speedDifference between them to achieve the desired speed(2) Repulsive force: the repulsive forces in the social force model come from two aspects: interaction between pedestrian individuals and obstacles; first of all, the first step is to,the moving direction of the individual pedestrian is influenced by other individuals of the pedestrian, and the individual pedestrian can expect the speed according to the individual pedestrianMaintain a certain distance from other individuals with the density of surrounding people, i.e., the pedestrian individual α generally feels more and more uncomfortable when approaching another unfamiliar pedestrian individual β, (3) appeal that the pedestrian individual α is sometimes attracted by other individuals or objects, and the model is assumed inHaving an individual or object therein having an attracting effect on the individual αUsing monotonically increasing potential functionsThe social force model specifies that α for the current pedestrian individual, the pedestrian individual in the visual field has larger influence on the current pedestrian individual than the individual outside the visual field, and the movement trend of other pedestrians only acts on the expected direction of the current individual, therefore, the local collision avoidance algorithm based on the social force model can calculate the local speed of each escape individual, the speed reflects the movement speed and the movement direction of the escape individual based on the peripheral information of the escape individual, and is the movement decision of the escape individual based on the instinct or even the subconscent, and when the social force model is applied to the larger-scale pedestrian group, the behavior characteristic of the escape individual is shown;
the fire escape behavior model in the step (3) utilizes a verified smoke concentration-escape speed formula to restrain the movement speed of the escape individual in smoke; the model enables human-fire interaction to be possible, and provides a more real behavior model for the simulation of individual escape in a fire scene.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410471493.7A CN104239636B (en) | 2014-09-16 | 2014-09-16 | Fire emergency evacuation simulation method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410471493.7A CN104239636B (en) | 2014-09-16 | 2014-09-16 | Fire emergency evacuation simulation method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104239636A CN104239636A (en) | 2014-12-24 |
| CN104239636B true CN104239636B (en) | 2017-05-03 |
Family
ID=52227690
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410471493.7A Active CN104239636B (en) | 2014-09-16 | 2014-09-16 | Fire emergency evacuation simulation method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104239636B (en) |
Families Citing this family (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3133550A1 (en) | 2015-08-20 | 2017-02-22 | Tata Consultancy Services Limited | Methods and systems for planning evacuation paths |
| CN105550500B (en) * | 2015-12-08 | 2017-12-15 | 吉林大学 | A kind of pedestrian's following behavior emulation mode based on social force |
| CN105335588B (en) * | 2015-12-14 | 2018-08-03 | 北京航空航天大学 | A kind of ground based on crowding factor crowd evacuation model |
| CN105740510B (en) * | 2016-01-22 | 2018-08-31 | 山东师范大学 | Evacuation crowd behaviour analogue system based on grid-density-relationship and its method |
| CN105550484A (en) * | 2016-02-26 | 2016-05-04 | 武汉大学 | Dynamic data driven modeling and simulation method for crowd evacuation and transportation behaviors in sudden-onset disaster |
| CN105716621B (en) * | 2016-04-08 | 2019-02-01 | 京信通信系统（中国）有限公司 | A kind of method and the network equipment of determining target best-effort path |
| CN105913157A (en) * | 2016-04-27 | 2016-08-31 | 熊运轩 | Fire-fighting emergency illumination and evacuation indication system intelligent escape logic vector searching algorithm |
| CN105955797A (en) * | 2016-04-28 | 2016-09-21 | 中山大学 | GPU-based crowd real-time simulation method adopting social force model |
| CN106096072B (en) * | 2016-05-17 | 2019-06-25 | 北京交通大学 | Dense crowd emulation mode based on intelligent body |
| CN106773717A (en) * | 2017-01-23 | 2017-05-31 | 复旦大学 | The efficient real-time control system of building emergency evacuation |
| CN107992036B (en) * | 2017-11-20 | 2021-03-16 | 汕头大学 | Method and device for planning vehicle access path in intelligent parking garage and storage medium |
| CN108733876B (en) * | 2018-04-04 | 2022-11-01 | 北京交通大学 | Fire evolution simulation method and fire evacuation comprehensive simulation method comprising same |
| CN109269517B (en) * | 2018-08-30 | 2019-07-26 | 中铁第四勘察设计院集团有限公司 | A kind of emergency for high-speed railway platform dredges emulation mode and system |
| CN109190325B (en) * | 2018-11-20 | 2019-05-28 | 山东师范大学 | Crowd evacuation Path Planning Simulation method based on the analysis of pedestrian's crowding |
| CN109783856B (en) * | 2018-12-12 | 2023-05-26 | 山东师范大学 | Emergency crowd evacuation simulation method and system based on group effect |
| CN109859558B (en) * | 2019-01-21 | 2020-12-29 | 北京科技大学 | Building fire virtual evacuation training method considering personnel physical influence |
| CN109655068A (en) * | 2019-01-31 | 2019-04-19 | 山东师范大学 | Path replanning method and device, system based on the event of fire that happens suddenly in building |
| CN109903490B (en) * | 2019-03-25 | 2020-12-01 | 哈尔滨工程大学 | Fire evacuation path calculation method and indication system |
| CN111144006B (en) * | 2019-12-27 | 2023-04-28 | 山东师范大学 | Crowd evacuation simulation method and system based on knowledge and emotion dual driving |
| CN111814346B (en) * | 2020-07-20 | 2022-04-12 | 北京大学 | Associated multi-parameter fire escape evaluation method for fire fighting system |
| CN111897850A (en) * | 2020-08-05 | 2020-11-06 | 湖南翰坤实业有限公司 | Method and system for generating escape route in personalized mode according to user portrait |
| CN112149305A (en) * | 2020-09-25 | 2020-12-29 | 中车长春轨道客车股份有限公司 | Railway vehicle fire personnel evacuation simulation method and system |
| CN112989614B (en) * | 2021-03-18 | 2022-08-05 | 武汉大学 | Emergency crowd evacuation simulation method based on time-space interaction process |
| CN114545932B (en) * | 2022-01-28 | 2023-06-06 | 北京交通大学 | Path correction method, device, equipment and readable storage medium |
| US11417106B1 (en) | 2022-02-04 | 2022-08-16 | King Abdulaziz University | Crowd evacuation system based on real time perception, simulation, and warning |
| CN114580195B (en) * | 2022-03-17 | 2024-05-10 | 山东交通学院 | Estimation method for expected and actual escape direction deflection angle under fire background of public place |
| CN117236147B (en) * | 2023-07-20 | 2024-07-02 | 中国地质大学（武汉） | High-density crowd evacuation simulation method, device and storage medium |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7188056B2 (en) * | 2002-09-09 | 2007-03-06 | Maia Institute | Method and apparatus of simulating movement of an autonomous entity through an environment |
| CN102693330A (en) * | 2011-03-25 | 2012-09-26 | 上海日浦信息技术有限公司 | Pedestrian evacuation simulation method based on extended BDI (base diffusion isolation) model |
| CN102609599B (en) * | 2012-04-11 | 2014-03-12 | 天津市市政工程设计研究院 | Method for designing emulational underground road alignment and transverse clear distance based on multiple intelligent agents |
| CN103279800B (en) * | 2013-01-10 | 2016-12-28 | 吉林大学 | Based on game theoretic large-scale crowd evacuation optimum escape route generating method |
| CN103995915B (en) * | 2014-03-21 | 2017-03-01 | 中山大学 | A kind of crowd evacuation emulation system based on compound potential energy field |
-
2014
- 2014-09-16 CN CN201410471493.7A patent/CN104239636B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN104239636A (en) | 2014-12-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104239636B (en) | Fire emergency evacuation simulation method | |
| CN110737968B (en) | Crowd trajectory prediction method and system based on deep convolutional long and short memory network | |
| Izquierdo et al. | Forecasting pedestrian evacuation times by using swarm intelligence | |
| Jain et al. | MVO-based path planning scheme with coordination of UAVs in 3-D environment | |
| CN109543285B (en) | Crowd evacuation simulation method and system integrating data driving and reinforcement learning | |
| CN110688749B (en) | Crowd evacuation simulation method and system | |
| CN110795833B (en) | Crowd evacuation simulation method, system, medium and equipment based on cat swarm algorithm | |
| Kim et al. | Analysis of evacuation simulation considering crowd density and the effect of a fallen person | |
| Chu et al. | SAFEgress: a flexible platform to study the effect of human and social behaviors on egress performance | |
| CN110132282A (en) | Unmanned plane paths planning method and device | |
| Karmakharm et al. | Agent-based Large Scale Simulation of Pedestrians With Adaptive Realistic Navigation Vector Fields. | |
| Clay et al. | Towards real-time crowd simulation under uncertainty using an agent-based model and an unscented Kalman filter | |
| US20190286766A1 (en) | Simulation device, simulation method, and storage medium | |
| Ni et al. | Agent-based evacuation in passenger ships using a goal-driven decision-making model | |
| WO2021258847A1 (en) | Driving decision-making method, device, and chip | |
| Ünal et al. | Generating emergency evacuation route directions based on crowd simulations with reinforcement learning | |
| Lee et al. | Online indoor fire evacuation system | |
| Kirik et al. | Turns of different angles and discrete-continuous pedestrian dynamics model | |
| Sirakoulis | Cellular automata for crowd dynamics | |
| Procházka et al. | Pedestrian modelling in netlogo | |
| Rodriguez et al. | Utilizing roadmaps in evacuation planning | |
| CN109871582A (en) | Knowledge based engineering crowd evacuation emulation method, system and medium in foreign environment | |
| Tsorvas et al. | Enhanced multi-parameterized cellular automaton model for crowd evacuation: The case of a university building | |
| Ibrahim et al. | Pythagorean theorem for shortest distance in CA based pedestrian simulation: a case study on the closed area | |
| KR101576608B1 (en) | Distance potential field calculating system and providing method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |