CN111639809A - Multi-agent evacuation simulation method and system based on leaders and panic emotions - Google Patents
Multi-agent evacuation simulation method and system based on leaders and panic emotions Download PDFInfo
- Publication number
- CN111639809A CN111639809A CN202010481785.4A CN202010481785A CN111639809A CN 111639809 A CN111639809 A CN 111639809A CN 202010481785 A CN202010481785 A CN 202010481785A CN 111639809 A CN111639809 A CN 111639809A
- Authority
- CN
- China
- Prior art keywords
- individual
- leader
- agent
- speed
- panic
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000004088 simulation Methods 0.000 title claims abstract description 19
- 230000008451 emotion Effects 0.000 title claims description 74
- 239000002245 particle Substances 0.000 claims abstract description 27
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 18
- 230000008569 process Effects 0.000 claims abstract description 16
- 230000003993 interaction Effects 0.000 claims abstract description 11
- 239000003795 chemical substances by application Substances 0.000 claims description 111
- 230000002996 emotional effect Effects 0.000 claims description 32
- 230000000007 visual effect Effects 0.000 claims description 28
- 230000036651 mood Effects 0.000 claims description 17
- 230000008447 perception Effects 0.000 claims description 14
- 230000009471 action Effects 0.000 claims description 13
- 208000015181 infectious disease Diseases 0.000 claims description 10
- 244000144980 herd Species 0.000 claims description 6
- 230000002238 attenuated effect Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000000276 sedentary effect Effects 0.000 claims description 4
- 239000000126 substance Substances 0.000 claims description 4
- 230000007613 environmental effect Effects 0.000 claims description 3
- 230000001186 cumulative effect Effects 0.000 claims description 2
- 238000001125 extrusion Methods 0.000 claims description 2
- 230000006870 function Effects 0.000 claims description 2
- 230000006399 behavior Effects 0.000 description 18
- 230000004888 barrier function Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000007480 spreading Effects 0.000 description 2
- 208000011597 CGF1 Diseases 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005315 distribution function Methods 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000004039 social cognition Effects 0.000 description 1
- 239000003381 stabilizer Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
- G06Q50/265—Personal security, identity or safety
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
- Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- General Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- Computer Security & Cryptography (AREA)
- Primary Health Care (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a multi-agent evacuation simulation method and system based on leaders and panic, and belongs to the technical field of crowd evacuation. The invention analyzes the attribute and the behavior influencing the evacuation of the individual and abstracts the evacuated individual into Agent. The interaction among different individuals and the interaction among the individuals and the environment are considered, the follower, the panic and the obstacle avoidance angle of the individuals in the crowd are researched, a crowd evacuation model based on multiple agents is established, the crowd evacuation process is simulated more truly, and the crowd evacuation under the emergency condition can be simulated more truly. The invention provides an improved particle swarm algorithm as an individual path selection strategy in an evacuation process. The field of view factor is considered, the evacuation scene in the building is divided into regions, the influence of limited field of view on individual path selection is researched, the global optimal individual selection strategy of the particles is optimized, and the improved particle swarm algorithm is more consistent with the actual situation of crowd evacuation in the building.
Description
Technical Field
The invention belongs to the technical field of crowd evacuation, and particularly relates to a multi-agent evacuation simulation method and system based on leaders and panic emotion.
Background
In recent years, the focus of research in crowd evacuation is an evacuation simulation model based on Agent technology (Agent). In this type of model, the behavior laws of the population are not preset, but the modeled footfalls are placed on individual persons. Individual persons are represented by "agents". By placing a large number of agents representing individuals in a virtual building space and studying their overall behavior, the behavior of the population in real conditions can be simulated.
In general, it is difficult, but not impossible, to simulate human behavior, and one of the more common approaches is to design an agent that can simulate human behavior, such agents operating in a scenario and being autonomous, cooperative, learning-capable, and adaptable; and has social ability, reaction ability and promptness. Autonomy means that agent movement may not be directly and continuously supervised.
Therefore, in order to research the crowd evacuation behavior and the crowd evacuation characteristics in the building and further reasonably and efficiently organize the crowd evacuation in an emergency, it is urgently needed to invent a building crowd evacuation simulation model which not only accords with the reality, but also comprehensively considers a plurality of factors.
Disclosure of Invention
Aiming at the defects of a crowd evacuation simulation model in emergency evacuation of crowd in a building in the prior art, the invention provides a multi-agent evacuation simulation method and a multi-agent evacuation simulation system based on leaders and panic emotion, aiming at more truly simulating the crowd evacuation process in emergency, and having important practical significance for crowd evacuation schemes in emergency and how to effectively organize crowd evacuation and reduce casualties in emergency.
To achieve the above object, according to a first aspect of the present invention, there is provided a multi-agent evacuation simulation method based on leaders and panic emotion, the method comprising the steps of:
s1, an initialization stage: establishing a continuous evacuation scene model of an indoor scene, and initializing common individuals into agents, wherein the agents comprise attributes in the aspects of position, speed, type, surrounding environment and panic and emotion, and the types comprise a leader, a conservative common individual, a steady common individual and a sensitive common individual;
s2, a sensing stage: each Agent carries out information interaction with other Agent behavior and environmental condition parameter values and transmits the obtained information to a decision stage;
the information acquired by the leader in the sensing stage is surrounding obstacles and whether a common individual follows;
the information acquired by the ordinary individual in the sensing stage is surrounding obstacles, whether a leader exists in the visual field range or not and the panic condition of other individuals in the visual field range;
s3, a decision stage: each Agent makes an evacuation decision according to the fastest evacuation principle by analyzing the self capacity and the environment condition according to the information acquired in the sensing stage;
the evacuation decision of the leader is as follows: moving towards the nearest outlet; secondly, if an ordinary person follows, the speed of the following person is matched; collision avoidance with other agents and obstacles is carried out in the process of traveling;
the evacuation decision of the common individual is as follows: if an outlet is found, directly moving towards the outlet; if the leader can be seen, the leader is followed to move; thirdly, the user can not see the exit or the leader and can select to move along with other common individuals in the visual field range; fourthly, people cannot be seen in the visual field, and one direction is randomly selected to move; collision avoidance with other agents and obstacles is carried out in the process of traveling;
s4, action stage: and each Agent responds to the evacuation decision in the decision stage, moves, and updates information to enter the sensing stage of the next cycle after moving.
Preferably, in the initialization phase, the attributes of the Agent are defined as follows:
wherein id represents the unique identification number of the Agent; pos (t) bits of Agent in continuous evacuation space modelPlacing;representing the actual velocity vector of the Agent; type represents the type of Agent, 0 represents leader, 1 represents conservative common individual, 2 represents stable common individual, and 3 represents sensitive common individual; reRepresents the view radius of the Agent; den (t) represents Agent as RePopulation density within a circular field of view of radius; v. ofmaxRepresenting the Agent maximum speed, obtained from the population density den (t); f (t) indicates whether the mood is infected, 0 indicates that the mood is not infected, 1 indicates that the mood is infected, and the leader is always 0, indicating that the mood of the leader is not affected by other people; e (t) is the emotional value of Agent, ehold(t) the bearable emotion value of the Agent is shown, the leader is 10000 all the time, and the emotion of the leader is not influenced by other people; pNThe emotion perception coefficient of the Agent is represented, the leader is always 0, and the emotion of the leader is represented to be not influenced by other people; dim denotes the emotional attenuation coefficient of Agent, panic (t) denotes the panic psychological value of Agent, parameters e (t), ehold(t)、PNDim and panic (t) are determined by Agent's type and F (t) together, t represents the time step.
Preferably, in the action phase, the speed update mode of Agent is as follows:
for the leader, the speed update includes the steps of: iteratively calculating the leader speed according to a particle swarm algorithm; adjusting the calculated leader speed according to the common individual condition of the following leader; updating the adjusted leader speed according to the surrounding obstacle condition;
for a normal individual, the speed update comprises the following steps: according to the particle swarm algorithm, iteratively calculating the speed of the common individual; calculating the panic emotion value of the common individual according to the emotion value, and adjusting the speed of the common individual according to the panic emotion value; and updating the adjusted speed of the common individual according to the surrounding obstacle condition.
Preferably, in the action phase, the speed updating method specifically includes the following steps:
for the leader, the speed update includes the steps of:
(i) iteratively calculating the leader speed according to an improved particle swarm algorithm, wherein the calculation formula is as follows;
(ii) adjusting the calculated leader speed according to the common individual condition of the following leader;
(iii) updating the adjusted leader speed according to the surrounding obstacle condition;
for a normal individual, the speed update comprises the following steps:
(i) according to the improved particle algorithm, the ordinary individual speed is calculated in an iterative mode;
there are five cases:
A. if the ordinary individual i can see nearby outlets in the visual field, the speed of the ordinary individual i is high
The formula is as follows:
B. when the ordinary individual i can not see the exit but can see the leader in the visual field range and if the fitness is fitnessl≥fitnessiThen, the velocity formula of the normal individual i is:
C. when the ordinary individual i can not see the exit but can see the leader in the visual field and the fitnessleader<fitnessiThen, the velocity formula of the normal individual i is modified as:
D. if the ordinary individual can not see both the exit and the leader but can see other ordinary individuals nearby, the speed update of the ordinary individual i is modified as follows:
E. if no ordinary individual can be seen in the visual field range of the ordinary individual, the speed updating formula of the ordinary individual is as follows:
wherein subscript l represents the leader; pbestl(t) representing a leader historical best location; x is the number ofl(t) representing the location of the leader; x is the number ofeixtIndicating the exit position; c. C1Weights representing a historical optimal solution portion of the individual; c. C2Weights representing a portion of the population-optimal solution; r is1And r2Is between [0, 1 ]]A random number in between; w is an inertia weight, so that the particles have the capability of keeping the motion inertia to balance the local and global searching capability of the particles; m islRepresents leader quality, ΩlRepresenting a leader neighborhood;representing the force between the leader and the individual j in the neighborhood; subscript i represents common individual i; pbesti(t) represents the historical best position of the average individual i; x is the number ofi(t) represents the position of the common individual i; m isiDenotes the i mass, Ω, of a common individualiRepresenting a neighborhood of a common individual i;representing the acting force between the common individual i and the individual j in the neighborhood; herd represents the number of slaves;representing the attraction between the common individuals i and j; x is the number ofrandRepresenting any one location in the environment;
(ii) calculating the panic emotion value of the common individual according to the emotion value, and adjusting the speed of the common individual according to the panic emotion value;
(iii) and updating the adjusted speed of the common individual according to the surrounding obstacle condition.
Preferably, the calculated leader speed is adjusted according to the general individual condition following him, and the formula is as follows:
wherein omegalA neighborhood of the representative leader is presented to,representing the velocity of the individual i in the neighborhood,representing the leader's own desired speed;
preferably, the adjusted Agent speed is updated according to the surrounding obstacle situation, specifically as follows:
① calculating the force f between a person and an obstacle based on the SFM modelio
fio={Aiexp[(ri-dio)/Bi]+kg(rij-dio)}nio+kg(ri-dio)(vi·tio)tio
Wherein A isiIs the psychological repulsive force coefficient, BiIs a distance coefficient; r isijIs the sum of the radii of the common individual i and the common individual j; r isiIs the model radius size of the common individual i; dioThe distance from the common individual i to the obstacle o; n isioA vector pointing to i for o; t is tioIs a tangential direction vector; k is the modulus of elasticity of the extrusion, g (x) is a piecewise function,virepresents the velocity of the normal individual i;
② taking into account the obstacle, the general individual velocity update formula isWherein the content of the first and second substances,represents the adjusted Agent velocity, miIs the weight of the ordinary individual i,representing the force between the average individual i and the average individual in the neighborhood.
Preferably, the panic emotion value of the common individual is calculated according to the emotion value, and the speed of the common individual is adjusted according to the panic emotion value, specifically as follows:
calculating an emotion value for an emotion uninfected Agent;
Wherein, c1、c2And c3Is a constant of proportionality, c4Is a constant, which is not 0 if the ordinary individual is still in the evacuation site; if the common individual escapes from the scene, the value is 0;representing a calculated sentiment value; q represents the neighborhood ΩiA panic cumulative value;represents the desired velocity of the average individual i; v. ofmaxRepresenting the maximum speed of the average individual i.
Preferably, the emotional value of the average individual is calculated using the P-Durupinar model of emotional infection.
Preferably, the emotion value of the common individual is calculated by adopting a PS-Durupinar emotion infection model considering emotion perception and emotion attenuation, and the emotion value is as follows:
(1) calculating the neighbor domain of the common individual;
(3) judging whether Q is larger than or equal to a panic threshold value T, if so, indicating that the sum of the panic of the neighbor domain received by the individual exceeds the self bearing threshold value, updating the individual to be in an emotional infection state, and updating the emotional value to be:
a. if the individual is conservative, e (t) rand (0, 0.4); b. if the individual is sedentary, then e (t) rand (0.4, 0.8); c. if the individual is susceptible, then e (t) rand (0.8, 1); otherwise, it means that the individual received an emotion value not reaching its threshold, so e (t) is 0;
and updating the emotion value by using the following formula: e.g. of the typei(t)=pi·ei(t)
(4) If the individual can see the exit or the leader within the visual field range, the individual emotion value is attenuated, and the emotion value of the individual is updated by the following formula; e.g. of the typei(t)=(1-dimi)ei(t)
(5) When e (t) < ehold(t), the individual is returned to an emotional uninfected state.
To achieve the above object, according to a second aspect of the present invention, there is provided a multi-agent evacuation simulation system based on leaders and panic emotion, the simulation system employing the simulation method as described in the first aspect.
Generally, by the above technical solution conceived by the present invention, the following beneficial effects can be obtained:
(1) the method analyzes and summarizes the attributes and behaviors influencing the evacuation of the individual, abstracts the evacuated individual into Agent, models the attributes, perception, decision and behaviors of the Agent, and establishes an individual Agent model. The interaction among different individuals and the interaction among the individuals and the environment are considered, the following leader behavior mode of the individuals in the crowd, the spreading of panic emotion in the crowd and the avoidance of pedestrians, pedestrians and barriers are respectively modeled, a crowd evacuation model based on multiple agents is established, the crowd evacuation process under the emergency condition is simulated more truly, and the crowd evacuation under the emergency condition can be simulated more truly.
(2) The invention provides an improved particle swarm algorithm as an individual path selection strategy in an evacuation process. The field of view factor is considered, the evacuation scene in the building is divided into regions, the influence of limited field of view on individual path selection is researched, and the global optimal individual selection strategy of the particles is optimized; for the panic psychology of an individual, quantitative description of panic factors is given, and a speed updating formula of an algorithm is optimized to reflect the influence of the panic factors on individual evacuation; and a social force model is introduced to solve the problem of interaction between different individuals and between the individuals and the barriers, so that the improved particle swarm algorithm is more in line with the actual situation of crowd evacuation in a building.
Drawings
FIG. 1 is a flow chart of a multi-agent evacuation simulation method based on leaders and panic emotions according to the present invention;
fig. 2 is a schematic diagram of an establishment process of an evacuation scene coordinate system according to the present invention;
fig. 3 is a crowd panic emotion propagation diagram of a building multi-Agent crowd evacuation model based on leaders and panic emotions provided by the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In addition, the technical features involved in the embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.
As shown in FIG. 1, the present invention provides a multi-agent evacuation simulation method based on leaders and panic emotions, comprising the steps of:
firstly, an initialization stage: including initialization of evacuation scenarios and pedestrian information.
1. Evacuation scenario modeling
The invention establishes a continuous evacuation scene model, pedestrians can move towards any direction, and the motion trail of the pedestrians is a continuous curve instead of a broken line.
The specific process of establishing the continuous evacuation scene model is as follows:
1) establishing a physical evacuation space model conforming to a real indoor evacuation scene, wherein the physical evacuation space model comprises: boundaries, free space, obstacles, and exits.
Boundary: simulating a real wall, wherein pedestrians cannot pass through the wall, and the moving range of the pedestrians is within the wall. Obstacle: there are a wide variety of obstacles in an evacuation scenario, for example: walls, uprights, tables and chairs, etc., which can interfere with the movement of pedestrians. Free space: inside the boundary, free space remains except where there are obstacles, and pedestrians can move randomly within the area. An evacuation outlet: target position for pedestrian evacuation.
2) And abstracting the physical evacuation space model into a continuous evacuation scene model.
The continuous evacuation scene model established by the embodiment is a two-dimensional plane, and can be extended to a three-dimensional plane. The left lower corner of the physical evacuation space is used as an origin, the longest boundary is used as the X-axis direction, the direction perpendicular to the X-axis is used as the Y-axis direction, the whole physical evacuation space is coordinated, the position of each pedestrian is determined by a unique horizontal and vertical coordinate (X, Y), the motion direction of each pedestrian is any one direction of [0 degrees and 360 degrees ], and the motion trail is any continuous curve. As shown in fig. 2, the coordinates of the pedestrian will change over time during the escape process. When meeting an obstacle, the pedestrian can change the moving direction according to the surrounding conditions.
2. Multi-agent
The multi-agent technology can be used for simulating the walking conditions of people, so that the simulated objects can appropriately avoid obstacles and select a certain path to walk, but appropriate parameter values can be set to simulate probabilistic deviation targets, and therefore the walking conditions of people in different psychological states under different conditions can be simulated.
The leader represents a role, and the individual playing the role satisfies the following conditions: familiarity with evacuation scenarios; mastering escape skills; can be identified by common individuals; the mood of which is not affected by others. For example, the scene is a station, and the leader is a worker; the scene is a market, and the leader is shopping guide and security guard.
The individual agents are the agents of action for people evacuation, so the attributes of each evacuated individual Agent are defined as follows:
wherein the content of the first and second substances,
id represents the unique identification number of the Agent; pos (t) represents the position of the Agent in the continuous evacuation space model; in this embodiment, pos (0) is initialized randomly in a free space region of an indoor scene.Representing the actual velocity vector of the Agent, including magnitude and direction; the present embodiment is satisfiedUnder the restriction of (a) to (b),the directions are random. type represents the type of Agent, 0 represents leader, 1 represents conservative common individual, 2 represents stable common individual, and 3 represents sensitive common individual; the present embodiment sets the total evacuation number and the four types of pedestrian ratios, and then randomly assigns the types under the condition that the set ratio limit is satisfied. ReRepresenting the view radius of the Agent, the visual angle range of the human is about +/-104 degrees; den (t) represents Agent as RePopulation density within a circular field of view of radius; calculated by the following formula.Wherein omegai(t) the other Agent set representing the view radius of the Agent at time step t. v. ofmaxRepresenting Agent maximum velocity, obtained from population density den (t)F (t) represents whether the emotion is infected, 0 represents that the emotion is not infected, 1 represents that the emotion is infected, the leader is always 0, the emotion of the leader cannot be influenced by other people, and common individuals F (0) of the three personalities are initialized randomly. e (t) is the emotional value of Agent, ehold(t) indicates the sustainable emotional value of Agent, PNExpressing the emotional perception coefficient of the Agent, dim expressing the emotional attenuation coefficient of the Agent, and panic (t) expressing the panic psychological value of the Agent. These parameters are initialized according to table 1 and table 2, depending on the type of Agent and whether the emotional flag f (t) is infected.
type | Ehold | PN | panic | dim |
Leader | 10000 | 0 | 0 | 0 |
Conserved form | According to F (0) | rand(0,03) | According to F (0) | rand[1,0.7] |
Weight stabilizer | According to F (0) | rand[0.3,06) | According to F (0) | rand(0.7,0.4) |
Sensitive type | According to F (0) | rand[0.6,1] | According to F (0) | rand[0.4,0] |
TABLE 1
F(0) | e(0) | ehold(0) | panic(0) |
0 | 0 | Equation 2 | 0 |
1 | 1 | Equation 2 | Equation 1 |
TABLE 2
panic=log-N(μq,σq 2) (1)
Ehold=log-N(μE,σE 2) (2)
Wherein log-N () is a lognormal distribution function. According to the mu of the individualTAndinitializing an individual emotion threshold T; according to the mu of the individualqAndan individual panic value q is calculated.
II, a sensing stage: and each Agent performs information interaction with other Agent behavior and environmental condition parameter values, and transmits the obtained information to a decision-making stage.
Distance constraint conditions are as follows: the constraint conditions of the distances between the Agent and other agents and between the Agent and the obstacle are as follows. Wherein the radius of view of the Agent is Re。
Visual angle constraint conditions: the constraint condition of the angles of the Agent, other agents and the barrier is that the included angle between the object point and the object to be dispersed is alpha, and alpha is more than or equal to-90 degrees and less than or equal to 90 degrees.
1. Perception phase of leader
The leader is familiar with the environment in the building, masters evacuation escape skills under emergency conditions, can complete self safe evacuation under emergency events, and can guide other individuals to evacuate. The invention sets the information that the leader knows the internal structure of the building, the exit and the like, and can be identified by common pedestrians.
The information that the leader needs to acquire in the perception phase is surrounding obstacles and whether someone follows.
2. Perception stage of the average individual
There are some more typical behaviors in the average individual: the following behaviors: when an emergency such as an emergency occurs, people may take the same action as surrounding individuals, i.e., the public mind, due to unfamiliarity with the environment and panic mind. Panic behavior: when an emergency occurs, evacuation personnel can generate certain psychological pressure, individuals can have an unhealthy behavior, and typical panic behavior characteristics such as blind actions of personnel, disordered crowd at bottlenecks and the like are generally caused. Therefore, the invention sets a common individual following leader behavior pattern. The information that the ordinary individual needs to acquire in the perception stage is surrounding obstacles, whether a leader exists in the visual field range or not, and other individuals in the visual field range.
Thirdly, a decision stage: and each Agent makes an evacuation decision according to the fastest evacuation principle by analyzing the self capacity and the environment condition according to the information acquired in the sensing stage.
1. Evacuation decision of the leader: moving towards the nearest outlet; if a person follows, the speed of the follower is matched. And secondly, collision avoidance with other agents and obstacles is required in the traveling process.
2. Evacuation decision of the general individual: firstly, if the pedestrian finds the exit, the pedestrian directly moves towards the exit; if the pedestrian can see the leader, the pedestrian will move along with the leader; thirdly, the pedestrian can not see the exit or the leader, and then the pedestrian can move along with other pedestrians in the visual field; fourthly, if people cannot be seen in the field of vision of the pedestrians, one direction is randomly selected to move. Collision avoidance with other agents and obstacles is needed in the process of traveling.
Fourthly, action stage: and each Agent responds to the evacuation decision in the decision stage, moves, and updates the information to enter the next cycle after moving.
The Agent takes the evacuation exit as a destination and continuously adjusts the speed and the direction of the Agent to move towards the exit. And in the process, the Agent can avoid collision with surrounding obstacles and pedestrians.
1. Action phase of leader
The updating of the leader speed comprises the following steps:
(i) and (5) iteratively calculating the leader speed according to the particle swarm algorithm.
v(t+1)=w.v(t)+c1r1(Pbest(t)-x(t))+c2r2(Gbest(t)-x(t))
Wherein, c1Representing the self-cognition coefficient, the weight of the individual historical optimal solution part; c. C2Representing a social cognition coefficient, and the weight of the optimal solution part of the population; r is1And r2Is between [0, 1 ]]A random number in between; w is an inertia weight, so that the particles have the capability of keeping the motion inertia to balance the local and global searching capability of the particles;
(ii) and adjusting the calculated leader speed according to the common individual condition of following the leader.
Since the leaders are familiar with the evacuation environment, the leaders move directly toward the nearby exits after evacuation begins. Meanwhile, the leader can lead the pedestrians to evacuate and can be influenced by the surrounding pedestrians. In order to achieve a good lead effect, the leader needs to adjust its speed according to the surrounding environment.
The speed of leader/is given by:
the above equation indicates that when there is a follower, the leader is in line with the slowest of them. If there is no follower around the leader, at its own desired speedAnd moving towards the target.
(iii) And updating the adjusted leader speed according to the surrounding obstacle condition.
And updating the position x (t +1) ═ x (t) + v (t) of the leader according to the updated speed and the current position of the leader by a particle swarm algorithm. After updating the position, the density and maximum movement speed are updated, and the next round (next time step) of the loop is entered (perception → decision → action).
2. Stage of action in a common individual
In the evacuation process, the decision and behavior of the pedestrian can be influenced by psychological factors. Due to differences in psychological characteristics, the level of panic in individuals may vary in sudden situations. Panic can affect pedestrian decisions and walking speed.
The ordinary individual speed update comprises the following steps:
(i) and (5) iteratively calculating the ordinary individual speed according to a particle swarm algorithm.
(ii) And calculating the panic emotion value of the common individual according to the emotion value, and adjusting the speed of the common individual according to the panic emotion value.
Calculating an emotion value
Calculation mode 1. for emotional uninfected Agent (P-Durupinar emotional infection model)
(1) Calculating the neighbor domain of the pedestrian;
(3) judging whether Q is larger than or equal to a panic threshold value T, if so, indicating that the sum of the panic of the neighbor domain received by the individual exceeds the self bearing threshold value, updating the individual to be in an emotional infection state, and updating the emotional value to be: a. if the individual is conservative, e (t) rand (0, 0.4); b. if the individual is sedentary, then e (t) rand (0.4, 0.8); c. if the individual is susceptible, then e (t) rand (0.8, 1); otherwise, it indicates that the individual received the mood value not reaching its threshold, so e (t) is 0.
Calculation mode 2 for the uninfected Agent (consider the emotional perception and emotional decay PS-Durupinar emotional infection model)
(1) Calculating the neighbor domain of the pedestrian;
(3) judging whether Q is larger than or equal to a panic threshold value T, if so, indicating that the sum of the panic of the neighbor domain received by the individual exceeds the self bearing threshold value, updating the individual to be in an emotional infection state, and updating the emotional value to be: a. if the individual is conservative, e (t) rand (0, 0.4); b. if the individual is sedentary, then e (t) rand (0.4, 0.8); c. if the individual is susceptible, then e (t) rand (0.8, 1); otherwise, it indicates that the individual received the mood value not reaching its threshold, so e (t) is 0. Updating e by using the following formula, and multiplying the e by the perception coefficient; e.g. of the typei(t+1)=pi·ei(t);
(4) If the individual can see the exit or the leader within the visual field range, the individual emotion value is attenuated, and the emotion value of the individual is updated by the following formula; e.g. of the typei(t+1)=(1-dimi)ei(t)
(5) When e < ehold(t), i.e. attenuation of e to less than ehold(t), then the individual reverts to an emotional uninfected state.
The model can take into account both psychological differences between individuals and the dynamic course of changes in individual emotional perception and emotional decay.
The second part reflects the influence of speed on the emotion of the pedestrian; in emergency situations, the pedestrian expects to escape from the scene at the maximum speed, but due to limitations, the actual speed of the pedestrian is difficult to reach its expected speed, thus creating a panic. The third part represents the absence of assistance due to isolation when no other pedestrians are seen in the field of view, due to unfamiliar evacuation environmentThe resulting panic mood. If other pedestrians can be seen, it is 0, otherwise it is 1. The fourth part is a constant which is not 0 if the pedestrian is still at the evacuation site; and 0 if the pedestrian escapes from the scene. c. C1、c2And c3Is a constant.
③ update speed under panicWherein the content of the first and second substances,represents the desired speed of pedestrian i; v. ofmRepresenting the maximum speed of the pedestrian i.
(iii) And updating the adjusted speed of the common individual according to the surrounding obstacle condition.
① calculating the force f between a person and an obstacle based on the SFM modeliw
fio={Aiexp[(ri-dio)/Bi]+kg(rij-dio)}nio+kg(ri-dio)(vi·tio)tio
Wherein A isiIs the psychological repulsive force coefficient, BiIs a distance coefficient. r isijIs the sum of the radii of the pedestrian i and the pedestrian j, riIs the model radius size of the pedestrian i. dioThe distance from the pedestrian i to the obstacle o; n isioA vector pointing to i for o; t is tioIs a tangential direction vector.
w∈ΩiIndicating that the pedestrian i is only affected by the obstacle w in the field of vision. Similarly, the acting force between the pedestrian i and the pedestrian jAndalso taking into account the view factor.
Similarly, after the velocity is updated, the position, the density, and the maximum moving velocity are updated in this order.
Preferably, the invention also provides an improved particle swarm algorithm for iterative computation of the leader and the common individual velocities.
For the leader, its speed and position are updated with the distance to the nearest exit as the global optimal Gbest:
for the average individual, five cases are distinguished:
A. if the pedestrian i can see the nearby exit in the visual field, the pedestrian faces the exit directly
Move, i.e. with the exit as global optimum. The velocity formula for pedestrian i is:
B. when the pedestrian i cannot see the exit but can see the leader in the visual field and if fitnessleader≥fitnessiI.e. the pedestrian i is located better (closer to the exit) than the leader and can know where the exit is by the leader, the pedestrian can be evacuated directly towards the exit instead of approaching the moving path blocking the leader towards the leader, the velocity formula of the pedestrian i is:
C. when the pedestrian i cannot see the exit but can see the leader in the visual field and the chatleader<fitnessiThen the pedestrian moves along with the leader, and the position of the leader is the global optimal solution of the particle in the particle swarm optimization. The velocity formula for pedestrian i at this time is modified as:
D. if the pedestrian cannot see both the exit and the leader, a case-by-case discussion is required. If the pedestrian can see other pedestrians nearby, the pedestrian can choose to follow the other pedestrians nearby, the following will is measured by the following coefficient herd, the greater the herd is, the stronger the following will is, and if the herd is 0, the pedestrian is not willing to follow. Introducing an attractive force between the coefficients herd and the pedestrian, wherein the speed update of the pedestrian i is corrected as follows:
E. if no other pedestrian can be seen in the pedestrian visual field, the global optimum in the particle swarm algorithm is not the exit or the leader but any position in the environment, and the speed updating formula of the pedestrian is as follows:
vi(t+1)=w·vi(t)+(xrand-xi(t))
fitness is as follows: the fitness is the distance between the particle and the outlet and is used for measuring the position of the particle, and the fitness is calculated according to the following formula: fitnessi=disi,exit。
In the whole invention, a Social Force Model (SFM for short) is adopted to describe the interaction between pedestrians and between pedestrians and obstacles, and the interaction describes the movement of the pedestrians from the angle of Force, so that the modeling is used for modeling collision of the pedestrians and obstacle avoidance.
The above formula shows that for the pedestrian i, the weight thereof is miWill face the target directionTo be provided withIs moved at a speed of (1). Meanwhile, the actual speed v of the device can be continuously adjusted in the processi(t),τiTo adjust the time. Reflecting the influence of the pedestrian on the moving object.
Acting force between pedestriansTwo acting forces exist among pedestrians, the first pedestrian i wants to keep a certain distance of psychological repulsive force with the pedestrian jThe second is the interaction force generated by the contact between the pedestrian i and the pedestrian jIn the case of emergency evacuation, people gather each other, and in the case of a large crowd density, collision between people is likely to occur.
Force f between person and obstacleio={Aiexp[(ri-dio)/Bi]+kg(rij-dio)}nio+kg(ri-dio)(vi·tio)tio. Acting force f of pedestrian and obstacleioIncluding the repulsive force between the pedestrian and the wall body and the contact force between the pedestrian and the wall body. f. ofioAnd fijSimilarly, the difference between the two is that the obstacle is stationary, the pedestrian is moving, and the pedestrian can be considered as a moving obstacle. Wherein r isiIs the radius of the pedestrian; dioI.e. the pedestrian to the obstacle. The distance of (d); n isioIs as follows. A vector pointing to i; t is tioIs a tangential direction vector.
Acting force f between pedestrians in the inventionijOn the basis of adding attraction among pedestriansTo describe the activities of people.Wherein: a. theattrRepresenting the attractive force intensity coefficient; b isattrIs a distance parameter; r isijIs the sum of the radii of the pedestrian i and the pedestrian j; r isiIs the model radius size of pedestrian i; n isijIs a direction unit vector directed by the pedestrian i to the pedestrian j.
The social force model considering the collision avoidance of people and obstacles and the improvement of people from many to many is as follows:
in fig. 3, circles represent emotionally uninfected individuals, triangles represent emotionally infected individuals, the evacuation environment is a space with the length of 60m and the width of 40m, as can be seen from fig. 3, initially, the number of emotionally affected individuals around the infected individuals gradually increases, the phenomenon of panic emotion spreading starts to occur by centering on the infected individuals, and in the position close to the exit, part of pedestrians can be changed from the emotionally affected individuals to the emotionally uninfected individuals, namely, the panic emotion of the pedestrians is attenuated, and the panic emotion of the pedestrians is relieved after the pedestrians see the exit, have evacuation targets and know that the pedestrians can escape.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and that any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A multi-agent evacuation simulation method based on leaders and panic emotions, the method comprising the steps of:
s1, an initialization stage: establishing a continuous evacuation scene model of an indoor scene, and initializing common individuals into agents, wherein the agents comprise attributes in the aspects of position, speed, type, surrounding environment and panic and emotion, and the types comprise a leader, a conservative common individual, a steady common individual and a sensitive common individual;
s2, a sensing stage: each Agent carries out information interaction with other Agent behavior and environmental condition parameter values and transmits the obtained information to a decision stage;
the information acquired by the leader in the sensing stage is surrounding obstacles and whether a common individual follows; the information acquired by the ordinary individual in the sensing stage is surrounding obstacles, whether a leader exists in the visual field range or not and the panic condition of other individuals in the visual field range;
s3, a decision stage: each Agent makes an evacuation decision according to the fastest evacuation principle by analyzing the self capacity and the environment condition according to the information acquired in the sensing stage;
the evacuation decision of the leader is as follows: moving towards the nearest outlet; secondly, if an ordinary person follows, the speed of the following person is matched; collision avoidance with other agents and obstacles is carried out in the process of traveling; the evacuation decision of the common individual is as follows: if an outlet is found, directly moving towards the outlet; if the leader can be seen, the leader is followed to move; thirdly, the user can not see the exit or the leader and can select to move along with other common individuals in the visual field range; fourthly, people cannot be seen in the visual field, and one direction is randomly selected to move; collision avoidance with other agents and obstacles is carried out in the process of traveling;
s4, action stage: and each Agent responds to the evacuation decision in the decision stage, moves, and updates information to enter the sensing stage of the next cycle after moving.
2. The method of claim 1, wherein in an initialization phase, the attributes of the Agent are defined as follows:
wherein id represents the unique identification number of the Agent; pos (t) bits of Agent in continuous evacuation space modelPlacing;representing the actual velocity vector of the Agent; type represents the type of Agent, 0 represents leader, 1 represents conservative common individual, 2 represents stable common individual, and 3 represents sensitive common individual; reRepresents the view radius of the Agent; den (t) represents Agent as RePopulation density within a circular field of view of radius; v. ofmaxRepresenting the Agent maximum speed, obtained from the population density den (t); f (t) indicates whether the mood is infected, 0 indicates that the mood is not infected, 1 indicates that the mood is infected, and the leader is always 0, indicating that the mood of the leader is not affected by other people; e (t) is the emotional value of Agent, ehold(t) the bearable emotion value of the Agent is shown, the leader is 10000 all the time, and the emotion of the leader is not influenced by other people; pNThe emotion perception coefficient of the Agent is represented, the leader is always 0, and the emotion of the leader is represented to be not influenced by other people; dim denotes the emotional attenuation coefficient of Agent, panic (t) denotes the panic psychological value of Agent, parameters e (t), ehold(t)、PNDim and panic (t) are determined by Agent's type and F (t) together, t represents the time step.
3. The method of claim 2, wherein the Agent speed update mode in the action phase is as follows:
for the leader, the speed update includes the steps of: iteratively calculating the leader speed according to a particle swarm algorithm; adjusting the calculated leader speed according to the common individual condition of the following leader; updating the adjusted leader speed according to the surrounding obstacle condition;
for a normal individual, the speed update comprises the following steps: according to the particle swarm algorithm, iteratively calculating the speed of the common individual; calculating the panic emotion value of the common individual according to the emotion value, and adjusting the speed of the common individual according to the panic emotion value; and updating the adjusted speed of the common individual according to the surrounding obstacle condition.
4. The method of claim 2, wherein the speed update during the action phase is as follows:
for the leader, the speed update includes the steps of:
(i) iteratively calculating the leader speed according to an improved particle swarm algorithm, wherein the calculation formula is as follows;
(i) adjusting the calculated leader speed according to the common individual condition of the following leader;
(iii) updating the adjusted leader speed according to the surrounding obstacle condition;
for a normal individual, the speed update comprises the following steps:
(i) according to the improved particle algorithm, the ordinary individual speed is calculated in an iterative mode;
there are five cases:
A. if the ordinary individual i can see nearby outlets in the visual field range, the velocity formula of the ordinary individual i is as follows:
B. when the ordinary individual i can not see the exit but can see the leader in the visual field range and if the fitness is fitnessl≥fitnessiThen, the velocity formula of the normal individual i is:
C. when the ordinary individual i can not see the exit but can see the leader in the visual field and the fitnessl<fitnessiThen, the velocity formula of the normal individual i is modified as:
D. if the ordinary individual can not see both the exit and the leader but can see other ordinary individuals nearby, the speed update of the ordinary individual i is modified as follows:
E. if no ordinary individual can be seen in the visual field range of the ordinary individual, the speed updating formula of the ordinary individual is as follows:
wherein subscript l represents the leader; pbestl(t) representing a leader historical best location; x is the number ofl(t) representing the location of the leader; x is the number ofeixtIndicating the exit position; c. C1Weights representing a historical optimal solution portion of the individual; c. C2Weights representing a portion of the population-optimal solution; r is1And r2Is between [0, 1 ]]A random number in between; w is the inertial weight; m islRepresents leader quality, ΩlRepresenting a leader neighborhood;representing the force between the leader and the individual j in the neighborhood; subscript i represents common individual i; pbesti(t) represents the historical best position of the average individual i; x is the number ofi(t) represents the position of the common individual i; m isiDenotes the i mass, Ω, of a common individualiRepresenting a neighborhood of a common individual i;representing the acting force between the common individual i and the individual j in the neighborhood; herd represents the number of slaves;representing the attraction between the common individuals i and j; x is the number ofrandRepresenting any one location in the environment;
(ii) calculating the panic emotion value of the common individual according to the emotion value, and adjusting the speed of the common individual according to the panic emotion value;
(iii) and updating the adjusted speed of the common individual according to the surrounding obstacle condition.
5. A method according to claim 3 or 4, characterized in that the calculated leader speed is adjusted according to the general individual situation following him, by the formula:
6. The method according to any of claims 3 to 4, characterized in that the adjusted Agent speed is updated according to surrounding obstacle situations, as follows:
① calculating the force f between a person and an obstacle based on the SFM modelio
fio={Aiexp[(ri-dio)/Bi]+kg(rij-dio)}nio+kg(ri-dio)(vi·tio)tio
Wherein A isiIs the psychological repulsive force coefficient, BiIs a distance coefficient; r isijIs the sum of the radii of the common individual i and the common individual j; r isiIs the model radius size of the common individual i; dioThe distance from the common individual i to the obstacle o; n isioA vector pointing to i for o; t is tioIs a tangential direction vector; k is the modulus of elasticity of the extrusion, g (x) is a piecewise function,virepresents the velocity of the normal individual i;
② taking into account the obstacle, the general individual velocity update formula isWherein the content of the first and second substances,represents the adjusted Agent velocity, miIs the weight of the ordinary individual i,representing the force between the average individual i and the average individual in the neighborhood.
7. The method according to any one of claims 3-6, wherein said calculating a panic emotion value for the average individual based on the emotion value, and adjusting the speed of the average individual based on the panic emotion value are as follows:
calculating an emotion value for an emotion uninfected Agent;
Wherein, c1、c2And c3Is a constant of proportionality, c4Is a constant, which is not 0 if the ordinary individual is still in the evacuation site; if the common individual escapes from the scene, the value is 0;representing a calculated sentiment value; q represents the neighborhood ΩiA panic cumulative value;represents the desired velocity of the average individual i; v. ofmaxRepresenting the maximum speed of the average individual i.
8. The method of claim 7, wherein the mood value of the average individual is calculated using a P-duropinar mood infection model.
9. The method of claim 7, wherein the emotion value of the general individual is calculated using a PS-duropinar emotional infection model that considers emotional perception and emotional decay, as follows:
(1) calculating the neighbor domain of the common individual;
(3) judging whether Q is larger than or equal to a panic threshold value T, if so, indicating that the sum of the panic of the neighbor domain received by the individual exceeds the self bearing threshold value, updating the individual to be in an emotional infection state, and updating the emotional value to be: a. if the individual is conservative, e (t) rand (0, 0.4); b. if the individual is sedentary, then e (t) rand (0.4, 0.8); c. if the individual is susceptible, then e (t) rand (0.8, 1); otherwise, it means that the individual received an emotion value not reaching its threshold, so e (t) is 0;
and updating the emotion value by using the following formula: e.g. of the typei(t)=pi·ei(t)
(4) If the individual can see the exit or the leader within the visual field range, the individual emotion value is attenuated, and the emotion value of the individual is updated by the following formula; e.g. of the typei(t)=(1-dimi)ei(t)
(5) When e (t) < ehold(t) when the individual is treatedAnd return to an emotional uninfected state.
10. A multi-agent evacuation simulation system based on leaders and panic emotion, characterized in that the simulation system employs a simulation method according to any of claims 1-9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010481785.4A CN111639809B (en) | 2020-05-29 | 2020-05-29 | Multi-agent evacuation simulation method and system based on leaders and panic emotion |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010481785.4A CN111639809B (en) | 2020-05-29 | 2020-05-29 | Multi-agent evacuation simulation method and system based on leaders and panic emotion |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111639809A true CN111639809A (en) | 2020-09-08 |
CN111639809B CN111639809B (en) | 2023-07-07 |
Family
ID=72332170
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010481785.4A Active CN111639809B (en) | 2020-05-29 | 2020-05-29 | Multi-agent evacuation simulation method and system based on leaders and panic emotion |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111639809B (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112183288A (en) * | 2020-09-22 | 2021-01-05 | 上海交通大学 | Multi-agent reinforcement learning method based on model |
CN112364529A (en) * | 2020-12-09 | 2021-02-12 | 长春理工大学 | Emergency evacuation scheme determination method and system for emergency evacuation scene |
CN112434958A (en) * | 2020-11-30 | 2021-03-02 | 上海应用技术大学 | Panic group emergency evacuation analysis method based on big data |
CN113177347A (en) * | 2021-05-10 | 2021-07-27 | 清华大学 | Crowd movement simulation method, device, equipment and storage medium |
CN113290567A (en) * | 2021-04-19 | 2021-08-24 | 山东师范大学 | Crowd evacuation guiding robot based on crowd emotion recognition |
CN113378382A (en) * | 2021-06-09 | 2021-09-10 | 哈尔滨工业大学 | Method for quantifying maximum squeezing force in dense crowd |
CN114545932A (en) * | 2022-01-28 | 2022-05-27 | 北京交通大学 | Path correction method, device and equipment and readable storage medium |
CN115359568A (en) * | 2022-08-24 | 2022-11-18 | 深圳职业技术学院 | Simulation method for pedestrian intelligent body movement and emergency evacuation and computer equipment |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107330147A (en) * | 2017-05-26 | 2017-11-07 | 北京交通大学 | A kind of subway station personnel emergency evacuation emulation mode and system |
CN107665282A (en) * | 2017-10-11 | 2018-02-06 | 山东师范大学 | Crowd evacuation emulation method and system based on isomery emotional appeal model |
CN108446469A (en) * | 2018-03-07 | 2018-08-24 | 山东师范大学 | A kind of group's behavior evacuation emulation method and device based on video drive |
CN109214065A (en) * | 2018-08-14 | 2019-01-15 | 山东师范大学 | The crowd evacuation emulation method and system of Q table are shared based on multi-Agent |
JP2019185074A (en) * | 2018-03-31 | 2019-10-24 | セコム株式会社 | Simulator, simulation method and action determination device |
CN111159854A (en) * | 2019-12-11 | 2020-05-15 | 华侨大学 | Method and device for building fire evacuation simulation model of nursing home |
-
2020
- 2020-05-29 CN CN202010481785.4A patent/CN111639809B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107330147A (en) * | 2017-05-26 | 2017-11-07 | 北京交通大学 | A kind of subway station personnel emergency evacuation emulation mode and system |
CN107665282A (en) * | 2017-10-11 | 2018-02-06 | 山东师范大学 | Crowd evacuation emulation method and system based on isomery emotional appeal model |
CN108446469A (en) * | 2018-03-07 | 2018-08-24 | 山东师范大学 | A kind of group's behavior evacuation emulation method and device based on video drive |
JP2019185074A (en) * | 2018-03-31 | 2019-10-24 | セコム株式会社 | Simulator, simulation method and action determination device |
CN109214065A (en) * | 2018-08-14 | 2019-01-15 | 山东师范大学 | The crowd evacuation emulation method and system of Q table are shared based on multi-Agent |
CN111159854A (en) * | 2019-12-11 | 2020-05-15 | 华侨大学 | Method and device for building fire evacuation simulation model of nursing home |
Non-Patent Citations (2)
Title |
---|
胡阿进: "突发事件下建筑物出口疏散能力模型及仿真研究" * |
鲍东宾等: "低能见度环境下多出口场馆人群疏散研究" * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112183288A (en) * | 2020-09-22 | 2021-01-05 | 上海交通大学 | Multi-agent reinforcement learning method based on model |
CN112434958A (en) * | 2020-11-30 | 2021-03-02 | 上海应用技术大学 | Panic group emergency evacuation analysis method based on big data |
CN112364529A (en) * | 2020-12-09 | 2021-02-12 | 长春理工大学 | Emergency evacuation scheme determination method and system for emergency evacuation scene |
CN113290567A (en) * | 2021-04-19 | 2021-08-24 | 山东师范大学 | Crowd evacuation guiding robot based on crowd emotion recognition |
CN113177347A (en) * | 2021-05-10 | 2021-07-27 | 清华大学 | Crowd movement simulation method, device, equipment and storage medium |
CN113378382A (en) * | 2021-06-09 | 2021-09-10 | 哈尔滨工业大学 | Method for quantifying maximum squeezing force in dense crowd |
CN114545932A (en) * | 2022-01-28 | 2022-05-27 | 北京交通大学 | Path correction method, device and equipment and readable storage medium |
CN115359568A (en) * | 2022-08-24 | 2022-11-18 | 深圳职业技术学院 | Simulation method for pedestrian intelligent body movement and emergency evacuation and computer equipment |
Also Published As
Publication number | Publication date |
---|---|
CN111639809B (en) | 2023-07-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111639809A (en) | Multi-agent evacuation simulation method and system based on leaders and panic emotions | |
Palechano et al. | Virtual crowds: Methods, simulation, and control | |
Pelechano et al. | Controlling individual agents in high-density crowd simulation | |
Pelechano et al. | Modeling crowd and trained leader behavior during building evacuation | |
Shao et al. | Autonomous pedestrians | |
Karamouzas et al. | A predictive collision avoidance model for pedestrian simulation | |
Kapadia et al. | Egocentric affordance fields in pedestrian steering | |
CN107480320B (en) | Crowd evacuation simulation method and system based on topological map and visual influence | |
US20090306946A1 (en) | Methods and systems for simulation and representation of agents in a high-density autonomous crowd | |
Sun et al. | Crowd evacuation simulation method combining the density field and social force model | |
Stüvel et al. | Torso crowds | |
Chu et al. | SAFEgress: a flexible platform to study the effect of human and social behaviors on egress performance | |
CN112417754A (en) | Crowd evacuation simulation method based on scene semantic information under complex indoor structure | |
Mao et al. | Modeling group structures with emotion in crowd evacuation | |
Zhou et al. | Route choice in the pedestrian evacuation: Microscopic formulation based on visual information | |
KR20140137068A (en) | Evacuation simulation system and providing method thereof | |
Zhang et al. | Knowledge-based crowd motion for the unfamiliar environment | |
Pelechano Gómez et al. | Comparison of crowd simulation for building evacuation and an alternative approach | |
Fang et al. | New agent-based egress model allowing for social relationships | |
Baig et al. | Realistic modeling of agents in crowd simulations | |
Cheng et al. | A modified particle swarm optimization-based human behavior modeling for emergency evacuation simulation system | |
Kamkarian et al. | Crowd evacuation for indoor public spaces using coulomb’s law | |
Ünal et al. | Generating emergency evacuation route directions based on crowd simulations with reinforcement learning | |
Rodriguez et al. | Utilizing roadmaps in evacuation planning | |
Chu et al. | Seismic evacuation simulation in a dynamic indoor environment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |