JP2014164540A - Evacuation action prediction system and evacuation action prediction program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an evacuation action prediction system taking account of relation among a plurality of evacuees.SOLUTION: The evacuation action prediction system can execute: evacuee position calculation processing that calculates position information of a certain evacuee for each time step such that the certain evacuee heads for a target evacuation place set to the certain evacuee on the basis of space data and human data; group integration processing that, on condition that evacuees belonging to a common belonging group enter a searchable range, with one evacuee defined as a reference point in the evacuee position calculation processing, integrates the evacuees to a group having a common target evacuation place; and group position calculation processing that calculates position information of a certain group for each time step such that the certain group heads for the target evacuation place set to the certain group on the basis of the space data and the human data.

Description

  The present invention relates to an evacuation behavior prediction system and an evacuation behavior prediction program for predicting an evacuation behavior of a victim when a disaster such as an earthquake, tsunami, flood damage, or fire occurs.

  Various disasters such as earthquakes, tsunamis, floods, and fires can be minimized by taking evacuation sites and routes in advance. Therefore, the applicant has proposed various evacuation behavior prediction systems (methods) shown in Patent Documents 1 and 2 and Non-Patent Document 1. With these evacuation behavior prediction systems (methods), it is possible to track the evacuation situation after a disaster, that is, the position information of the refugees over time. Therefore, it is possible to select an optimum condition by changing the conditions such as the evacuation place and the evacuation method and performing various evacuation behavior predictions.

JP 2006-107231 A Japanese Patent No. 4915550

Kakegawa Hidefumi, Oyama Takumi, “A simulation of crowd evacuation behavior at the block level during the tsunami run-up”, Coastal Engineering Papers, Vol. 53, 1341-1345, 2006

  By the way, the position calculation of the evacuees shown in these documents is calculated for each evacuee, and does not consider the relationship between the evacuees. When evacuating in the event of a disaster, it is conceivable to act together based on predetermined human relationships such as family, company, school, and neighborhood. An object of the present invention is to calculate the position of each refugee according to the actual evacuation situation in consideration of the relationship between such refugees.

In order to solve the above-described problems, the evacuation behavior prediction system according to the present invention is:
Storage means and control means,
The storage means
Disaster data that defines the extent of damage caused by a disaster;
Spatial data including evacuation site information and passage information,
For each evacuee, the affiliation group, the target evacuation location, and human data with the moving speed set are stored,
The control means can execute evacuee position calculation processing, group integration processing, and group position calculation processing,
The evacuee position calculation process calculates, for each evacuee, position information for each time step so as to go to a target evacuation site set for the evacuee based on the spatial data and the human data.
In the group integration process, in the evacuee position calculation process, an evacuee having a common group belongs to a common target on the condition that one refugee is within a searchable range defined as a reference point. Integrated into a group to the evacuation site,
The group position calculation process calculates, for each group, position information for each time step so as to go to a target evacuation site set for the group based on the spatial data and the human data. To do.

Furthermore, in the evacuation behavior prediction system according to the present invention,
In human data, a plurality of belonging groups are set for each refugee, and a priority is set for each refugee in the belonging group.
The group integration process is to integrate other evacuees who belong to a higher priority group into other evacuees and groups on the condition that they are within the searchable range. It is characterized by.

Furthermore, in the evacuation behavior prediction system according to the present invention,
Each target evacuation site has a safety level,
In the group integration process, the integrated group target evacuation site is installed in a target evacuation site with a high safety level among the target evacuation sites set for the refugees to be integrated.

  Furthermore, in the evacuation behavior prediction system according to the present invention, the disaster data changes over time.

Furthermore, in the evacuation behavior prediction system according to the present invention,
The control unit can execute disaster prediction processing for calculating disaster data based on at least disaster scenario data.

The evacuation behavior prediction program according to the present invention is
In the evacuation behavior prediction program executable by a computer based on various data stored in the storage means,
The storage means
Disaster data that defines the extent of damage caused by a disaster;
Spatial data including evacuation site information and passage information,
For each evacuee, the affiliation group, the target evacuation location, and human data with the moving speed set are stored,
For a certain refugee, based on the spatial data and the human data, an evacuee position calculation process for calculating the position information for each time step so as to go to the target evacuation place set for the refugee,
In the evacuee position calculation process, the evacuees having a common affiliation group are grouped to a common target evacuation site on condition that one refugee is within the searchable range defined by using one refugee as a reference point. Group integration processing to integrate,
Based on the spatial data and the human data, a group position calculation process for calculating position information for each time step so as to go to the target evacuation site set for the group is enabled. It is characterized by that.

  According to the evacuation behavior prediction system and the evacuation behavior prediction program according to the present invention, when an evacuee having a common affiliation group encounters during an evacuation behavior, that is, is located within any searchable range, it is integrated as a group. By moving to the target evacuation site set for the group, it is possible to calculate the position of the refugee according to the actual evacuation situation.

Diagram for explaining the configuration of the simulation model Conceptual diagram of space modeling Diagram for explaining selection rules for short-term movement targets for self-refugees Conceptual diagram about human visible range Conceptual diagram regarding the movement direction and position determination of evacuees Illustration for explaining walking speed considering crowd density The block diagram which shows the whole structure of embodiment of this invention The figure which shows the structural example of the human data of embodiment of this invention The flowchart which shows embodiment of this invention The figure for demonstrating the group integration process which concerns on embodiment of this invention (encounter of evacuees) The figure for demonstrating the group integration process which concerns on embodiment of this invention (a meeting of an evacuee and a group) The figure for demonstrating the group integration process which concerns on embodiment of this invention (encounter between groups)

  Now, an example of an evacuation simulation model, which is a prerequisite technology of the present invention, will be described. The evacuation simulation model deals with the interaction problem of the three elements of space, human and disaster area. For this reason, in the evacuation simulation model, it is important that these three elements are appropriately expressed in consideration of the relationship between the elements, and that the influence of each element on the evacuation behavior can be evaluated. There are the following two methods for modeling the space in the evacuation simulation model.

  (1) A method for modeling the geometric shape of a space as faithfully as possible. (2) A method of modeling space by network representation and mesh division. The method (1) is suitable for grasping the behavior of evacuees depending on the local state of the space. The method (2) is a method suitable for grasping the global transition of evacuees. It is necessary to decide how to model the space depending on the content to be evaluated by the evacuation behavior and how to model the person.

  The following methods can be used to model humans who are thought to have the most influence on the results obtained from the evacuation simulation model. (1) A method of modeling with emphasis on human individuals. (2) A method of modeling around a crowd. The method (1) is effective in evaluating the evacuation behavior in accordance with the mutual relationship between humans and local changes by giving variations to the characteristics of human individuals. When this modeling is adopted, it becomes effective when combined with the method (1) in the space modeling. On the other hand, the method (2) adopts the method (2) above for space modeling, and evaluates the evacuation behavior of a relatively large number of people globally using graph theory, network theory, or queuing theory. it can.

  When considering the impact of disaster propagation when predicting evacuation behavior, use the results of the disaster simulation model independent of the evacuation simulation model, and introduce disaster propagation to each space as an input condition in advance as a scenario A method can be considered. Another possible method is to incorporate a disaster simulation function into the evacuation simulation model itself and sequentially calculate the propagation of the disaster area in parallel with the evacuation action as time elapses after the fire breaks out.

  Explain the points to be considered when building an evacuation simulation model. The difference in the method of modeling each object of space, human beings, and disaster scope when constructing an evacuation simulation model depends on the contents of the guidelines for disaster prevention planning that should be emphasized using the model. Regarding the space modeling, in the case of a building, the position of the door, the wall, the surrounding people, etc. affect the evacuation behavior of individuals, so it is desirable that the model be able to handle the effects of these conditions appropriately. . In addition, considering that the evacuation trajectory is captured in an individual unit in a state close to reality, it is necessary to model the evacuee to freely move in the space.

  The point to consider from the aspect of evacuation behavior regarding human modeling is that refugees do not use only the evacuation route specified at the design stage, but select the route according to the situation according to the conditions of the space and disaster situation. It is to be. In addition, it is important for the progress of disasters to be able to capture evacuation behavior that takes into account the human impact of the disaster.

  The basic concept of modeling in the present invention will be described. In the present invention, an evacuation simulation model based on object orientation has been developed, and contrivances have been made with respect to modeling of three elements of space, human being, and disaster area, and human individual modeling.

First, modeling of three elements by object orientation will be described. Object orientation is a methodology for building a problem solving system (or model) using a computer. In the object-oriented method, several objects for solving the problem to be handled are found, the relationship between the objects is regarded as a network structure, and the problem is solved.

An object, which is a basic unit in object orientation, is defined by data indicating its own state and a specification (method) regarding how it behaves. In the problem-solving process, messages are sent between objects (messagesending), and the specifications corresponding to the messages sent to you are started to change your status or provide the data required by the destination. Is done.

  With regard to space modeling, we have modeled the assumption that humans can move freely in the space, using spaces such as rooms and corridors as basic units. In addition, in order to express the staying state that occurs near the door, the passage time of the door (the time required for opening and closing the door) can be taken into account. The propagation of the disaster area was modeled so that the results of the disaster simulation model could be handled as input conditions for the evacuation simulation.

  The reason for developing an evacuation simulation model based on object orientation is that it can be modeled by considering each of the three elements of space, human being, and disaster area as independent objects. This makes it possible to proceed with the development step by step for each model, and produces an advantage at the development stage as compared with the model development mode that emphasizes the algorithm. In addition, modification and improvement of the model can be performed for each object, and the extensibility of the entire model can be improved.

  Next, modeling of an individual person using object orientation will be described. In this model, human individuals are considered to be objects that have walking speed, evacuation action start time, occupied area on evacuation action, etc., which are considered important as characteristics in evacuation behavior, and these data can be set individually. .

  This human object can behave independently based on data defined for itself and data provided by other objects, even though the basic behavior for evacuation behavior is the same. Specifically, the human object can obtain the contents of the situation change due to the influence of the disaster range from the space, and can select the evacuation route independently according to the situation. To determine the evacuation direction of a human object, initially obtain data on the surrounding situation to determine the movement target for evacuation, receive attraction from the movement target and repulsion from obstacles, and calculate the vector of those forces. We decided to synthesize.

The features of the model of the present invention and the concept underlying the outline will be described with reference to FIGS. The evacuation simulation model 30 is basically composed of three models as shown in the explanatory diagram of FIG. The space model 31 is responsible for modeling a space in which humans can move within the building, and the human model 32 is responsible for modeling a human performing evacuation behavior.

  These two models form the basis of this model, and the simulator 35 manages both models. The disaster model 34 set in the disaster simulation model 33 plays a role of inputting the disaster propagation condition after the disaster has occurred into the space model 31 by the simulator. Each model shown here is basically composed of a plurality of objects.

  The space model 31 handles modeling of a space in which a human can move. Here, the model for the interior of the building will be described in detail. In a model for the inside of a building, a building is defined as an assembly of a plurality of levels, and a hierarchy is defined as an assembly of a plurality of rooms. Then, the connection information between layers is managed by stairs and the connection information between rooms is managed by a door. The space model 31 is not limited to the inside of a building, but can also be a block. In the model for the city block, it is possible to manage the same as the inside of the building by dividing the outdoors into spaces as appropriate and adopting signals as connection information.

  A room object that models a room, which is the smallest unit of the model, has a geometrical meaning surrounded by multiple walls as shown in the explanatory diagram of Fig. 2 for the purpose of capturing the evacuation trajectory in a form close to reality. Defined as a closed space. That is, in FIG. 2, when a space composed of a room 40 and corridors 41 and 42 is modeled, the room 40 is surrounded by four line segments (walls) 43 a to 43 d. The corridors 41 and 42 are partitioned by doors 44a to 44d and walls.

  In addition to information on walls from room objects, human objects include information on exits, doors, and guidance displays that are movement targets during evacuation, and information on other people in the room and disaster areas that are obstacles to evacuation. I can take it out. The human object can move relatively freely in the room based on the information.

  For the modeled door object, in addition to room connection information, a reference value for the door passage time can be defined individually. It is important to consider the passage time of the door particularly when an assisting device is used during the assisting action. When a human object passes through the door, the passage time of the door is determined from the reference value according to the relationship between the door width and the occupied area in the evacuation action, and it is decided to stop at the door during that time. It has been. While a person is using a door, other persons cannot enter the door, and can wait in the vicinity of the door, thereby enabling a representation of a staying state that occurs near the door.

Next, the human model 32 will be described. In this evacuation simulation model, people who can evacuate by themselves (self-refugee class) and people who cannot evacuate by themselves (helper class) from the viewpoint of differences in the characteristics of people evacuating in the building And three types of human beings (helper class) who help (or rescue) humans who cannot evacuate by themselves, for facility managers and fire brigade. Here, the class corresponds to the “type” definition of the object.

Next, determination of the movement target will be described. Self-refugee and caregiver class objects perform evacuation behavior while selecting a short-term movement target in order to satisfy a preset long-term movement target. The meanings of the long-term movement target and the short-term movement target are as follows: (a) The long-term movement target is a place where evacuation (movement) is finally desired. (B) The short-term movement target is a place where it is desired to reach (pass) for the time being in order to satisfy the long-term movement target. That is, while selecting the short-term movement target, the evacuation action is performed until the reached location coincides with the long-term movement month sign.

  “Evacuation site” is set as the long-term movement target of the evacuee object. The door is selected as the short-term movement target. In this model, based on the overall evacuation scenario for letting humans take evacuation behavior, the staircases inside and outside the building are set as final evacuation sites, and these are called “evacuation sites”. The door leading to “evacuation site” is represented as “exit”.

  FIG. 3 is an explanatory diagram showing a selection rule for a short-term movement target of a self-evacuated person. It is assumed that a self-refugee 50a exists in the room 40a, a self-refugee 50b exists in the room 40b, and a self-refugee 50c exists in the hallway 41. (1) If there is an exit in the space where you are, select the exit closest to you from the exit. In this example, the self-refugee 50c selects the exit 46.

  (2) If there is a guide light in the space where you are, select the door in the direction indicated by the guide light that is the closest to you among the guide lights. In this example, the self-evacuated person 50 a selects the door 44 g in the direction indicated by the guide light 45. (3) In other cases, the door closest to you is selected from the doors in the room where you are. In this example, the self-evacuated person 50b selects the door 44h that is the closest to the person among the doors in the room 40b where the person is located. In addition, the priority order which applies these rules is the order of said (1)-(3).

  FIG. 4 is a conceptual diagram regarding the human visible region. In FIG. 4, it is assumed that smoke is generated in the room 40c. The doors 44k, 44l, 44p, and 44r in the area 51 that cannot pass due to the disaster and the area 52 that is out of the sight of the refugee 50d are not selected. The evacuee 50d selects the doors 44m and 44n in the visible region as movement targets.

  Next, determination of the moving direction and moving position of the evacuees will be described with reference to FIG. FIG. 5 is a conceptual diagram of the moving direction and position determination of the evacuees 50e. As shown in FIG. 5, the movement direction after the short-term movement target is determined in the self-evacuation object and the assistant object is assumed to be a plurality of force vectors that are assumed to act psychologically on humans. Is determined by combining the vectors. R indicates the range reached at the walking speed per unit step, and Rx indicates the starting point of the movement position in the next step.

  As for the force acting on human beings, it was assumed that gravitational force was assumed from moving targets (doors and guide lights), and repulsive force was assumed from obstacles on evacuation behavior (walls, other human beings, and disaster areas). In the example of FIG. 8, the attractive force F (r4) acts on the evacuee 50e from the door 44. Also, repulsive forces F (r1) and F (r2) act from the walls, and repulsive forces F (r3) act from the other evacuees 50f. These forces are generally determined in inverse proportion to the distance from each object. In addition, an inertial force having a certain magnitude with respect to the moving direction was considered. The reason why the inertial force is taken into account is that the human movement is considered as a smooth movement close to reality.

  The starting point Rx of the next movement position in the simulation step is determined based on the direction of the vector ΣF obtained by combining these force vectors and the walking speed set in advance for each object. At this time, since the attractive force is set to be a relatively large value compared to the repulsive force, it is assumed that the situation where the attractive force and the repulsive force are balanced hardly occurs, but if the balanced force is balanced, it does not move. .

Next, the disaster model will be described. As described with reference to FIG. 1, this disaster model plays a role of introducing various disaster states such as fire, smoke due to fire, collapse due to an earthquake, flooding due to tsunami and flood damage into a spatial model by a simulator. Data to be introduced into the model is created using the results of disaster simulation independent of the evacuation simulation model. Regarding the effects of disaster on human objects, we considered the effects of obstructing human vision and moving away from the space in the space where the evacuation time limit was reached. It was assumed that humans who entered the space that reached the evacuation limit time would be unable to evacuate thereafter.

  FIG. 6 is an explanatory diagram illustrating an embodiment of the present invention, and FIG. 6A is an explanatory diagram illustrating an example of a walking speed of an evacuee considering a crowd density. In the evacuation behavior of an unspecified large number of people, the crowd density in the walking space may affect the walking speed. In this model, the horizontal walking speed was inversely proportional to the crowd density. In the calculation of the crowd density, the density around the target evacuee individual 50g is considered to be affected, and as shown in FIG. 6A, it exists in a semicircle with a radius of 3 m with respect to the traveling direction of each evacuee. Other evacuees (indicated by black circles like 50i) were considered as affecting the walking speed. Evacuees outside the semicircle with a radius of 3 m (indicated by white circles like 50h) are not considered for the effect on walking speed.

  FIG. 6B is an explanatory diagram of the walking speed on the stairs. The walking speed on the stairs was represented by the horizontal walking speed on the stairs in consideration of the relationship with the space modeling. As a reduction factor for the horizontal walking speed, 0.8 was adopted with reference to the actual measurement value. In FIG. 6B, the solid line is the horizontal walking speed (the upper limit is 1.0 m / sec), and the broken line is the horizontal walking speed for the staircase.

  FIG. 7 is a block diagram illustrating an example of an evacuation behavior prediction system that is an embodiment of the present invention. In FIG. 7, the evacuation behavior prediction system 1 includes an input unit 7, a control unit 15, and a data output unit 18. The input unit 7 inputs “disaster scenario data” 2, “time data” 4, “disaster data” 5, “spatial data” 6, and “human data” 7, respectively. The “disaster scenario data” 2 is information indicating initial conditions of various disasters, and includes the location of the disaster, the scale of the disaster, the weather conditions at the time of disaster, the conditions for disaster countermeasures, and the like. When the disaster is a fire, it is composed of the location of the fire, the heat generation rate (KW) of the fire source, the conditions of fire prevention equipment measures, the conditions of the opening, and the like. In the case of an earthquake or tsunami, it consists of the epicenter and magnitude of the earthquake.

  The “time data” 3 includes a disaster occurrence time, a time step interval for calculating the position information of the refugee in the simulation, and an evacuation start time when the refugee starts evacuation after the disaster occurs. The “disaster data” 5 is information separately calculated using the disaster prediction means 3 based on the “disaster scenario data” 2 described above, and is composed of the disaster situation for each simulation place. This disaster situation includes a danger level for each place. For example, the danger level is composed of 1 to 5 levels, the danger levels 1 and 2 are allowed to pass, the danger levels 3 and 4 are not allowed to pass, the danger level 5 is not allowed to pass, and the evacuees who were in the place are damaged (Death) can be set as appropriate. In the evacuation behavior prediction system according to the present embodiment, the evacuee to be simulated selects the target evacuation place according to the danger level of the place where he / she is located.

  In the case of a disaster such as an earthquake, the danger level for each location may not be changed over time, but for disasters such as tsunamis, floods, and fires that are likely to expand the scope of damage over time, change over time. That is, it is good also as giving a time history to the danger level for every place. In the present embodiment, the calculation of “disaster data” 5 by the disaster prediction unit 3 is performed outside the control unit 15. The disaster prediction unit 3 is evacuated by the control unit 15 described later. It may be performed together with behavior prediction. Furthermore, for calculation of “disaster data” 5, not only “disaster scenario data” 2 but also “spatial data” 6 that is a prediction target area of evacuee behavior may be used. It is possible to calculate accurate “damaged data” 5 in consideration of the topography of the prediction target area and the structure of the building.

  The “spatial data” 5 is information that defines the positional relationship, attributes, state, etc. of a block or a building. In other words, the information defines the action range of the refugee, and the refugee moves in the space defined by the “space data” 5. For example, when only the inside of a building is a target for evacuation behavior prediction, it is defined as a parameter relating to the dimensions, positional relationship, attributes, state, etc. of the space inside the building and the accompanying equipment or opening. Specifically, the floor number, floor height, room ID, room space attribute (type), room floor number, room coordinates (three-dimensional), number of openings (number, ID), number of doors (number, ID) ), Number of obstacles (number, ID), number of guidance displays (number, ID), opening ID, room ID with opening, opening coordinates (three-dimensional), door ID, space connection relationship, obstacle ID, obstacle coordinates (3D), room ID with obstacle, guidance display ID, room ID with guidance display, type of guidance display, coordinates of guidance display (three dimensions), contents of guidance display (type, ID) and the like.

  “Human data” 6 is information defined for each evacuee. FIG. 8 shows an example of “human data” 6. In this example, “human data” 6 defines an evacuee ID, an attribute group, an initial position, a target evacuation place, and a walking speed. In addition, information for determining the vertical relationship with other evacuees, such as the age and status of each evacuee, may be included. On the condition of this “human data” 6, the refugee moves from the initial position to the target evacuation site. In the present embodiment, in this “human data” 6, the affiliation group is information that defines the relationship between evacuees, and the relationships such as “family A”, “company A”, “school A”, etc. are defined. ing. In the evacuation behavior prediction of the present embodiment, when an evacuee having a common affiliation group encounters based on this affiliation group, it is integrated as a group and moved to a common target evacuation site. Furthermore, in this embodiment, each human data has a plurality of belonging groups defined by priority, and the evacuees form groups according to this priority.

  In “Human Data” 6, as described in FIG. 6B, the moving speed is a parameter that serves as a reference for the walking speed of the evacuees. Become. Further, when a group is formed by a plurality of evacuees as in this embodiment, it is conceivable that the lowest walking speed in the group is set as the group walking speed.

  The control unit 15 includes a “time data storage unit” 9, a “disaster data storage unit” 10, a “spatial data storage unit” 11, a “human data storage unit” 12, an “evacuee action position calculation means” 14, an “image display”. Means "16. The “time data storage unit” 11 stores and updates the parameters related to the time necessary for predicting the evacuation action, which are input in the “time data” 4. The “disaster data storage unit” 10 stores and updates the parameters related to smoke in the space considered as one of the obstacles of the evacuation action, which is input in the “disaster data” 5.

  The “spatial data storage unit” 11 stores / updates parameters relating to dimensions, positional relationships, attributes, states, and the like of the spaces inside the blocks and buildings and the accompanying facilities and openings, which are input in the “spatial data” 6. The “human data storage unit” 12 stores and updates the parameters specified in the “human data” 6 and specifying each evacuee. As described above, the “time data storage unit” 9, the “disaster data storage unit” 10, the “spatial data storage unit” 11, and the “human data storage unit” 12 function as storage means of the control unit 15.

The “evacuee action position calculation means” 14 is used to calculate the next evacuee's next time from the “time data” 9, “disaster data” 10, “spatial data” 11, and “human data” 12 stored in each storage unit. This is the part for calculating the position after the time step. The information regarding the position of the refugee calculated by the “evacuee action position calculation means” 14 is accumulated in other refugee data and the time history, the accumulation of the evacuation completion time and the number of evacuation completion persons, the number of people staying in a certain space. Is written in the data file 17. At the same time, the information calculated by the “evacuee action position calculation means” 14 is “spatial data” or “
Together with the “disaster data” etc., it is sent to the “image display means” 13 and displayed on a display unit such as a display. By providing such an image display means 13, it is possible to visually confirm the movement status of the evacuee's time history and the trajectory of the flow line of the evacuation action.

  The output unit 18 outputs the image display output unit 16 and the data file output unit 17. Information output by the image display output unit 16 includes evacuee movement status (time history / by room), the locus of evacuation behavior of the specific evacuees, and the like. In addition, as described above, the information output by the data file output unit 17 includes the evacuation completion time (by refugee), the cumulative number of evacuation completers (time history), the number of visitors (by time history and space), Number of people passing by {by time history, by door (stairs)}, etc. are included. As described above, the evacuation behavior prediction system of this embodiment determines the position of the refugee over time based on various conditions such as “time data” 4, “disaster data” 5, “spatial data” 6, and “human data” 7. By appropriately setting these conditions, it is possible to predict the actions of each refugee and investigate disaster prevention measures when a disaster occurs.

  FIG. 9 is a flowchart showing an embodiment of the present invention, and shows an evacuation behavior prediction process for each group in which refugees or a plurality of refugees are integrated. When the process is started, the initial position, the target evacuation place, the walking speed, etc. are extracted with reference to human data for the corresponding refugee (or group). In addition, spatial data at the position of the corresponding evacuees and disaster data (danger level) are also extracted (S11). Next, the movement direction of the corresponding refugee (group) is determined to reach the target evacuation place (S12). As described with reference to FIGS. 3 to 5, the moving direction of the evacuees is determined using the spatial data and the disaster area (danger data). Next, the position information of the refugee, that is, the position information after the time step interval (dt) is calculated based on the walking speed of the victim (S13). For movement of a group in which a plurality of evacuees are integrated, position information is calculated based on the lowest walking speed in the group.

  Here, the position information at time t is calculated for one evacuee and group, but the position information at time t is similarly calculated for other evacuees and groups. After calculating the position information at time t for all evacuees and groups, it is determined for each evacuee whether there are other evacuees within the searchable range defined with the evacuee as a reference (S14). . If it is determined that the group is present (S14: Yes), the group integration process (S15) is executed. On the other hand, if it is determined that there is not (S14: No), the group integration process S15 is skipped. In S16, the current time t is changed to the next time (t + dt), and the process returns to S11. By repeating such processing, the position of each evacuee and group, that is, the evacuation situation is calculated, and the information specified by the image display output unit 16 and the data file output unit 17 described with reference to FIG. It becomes possible to provide.

  Now, the details of the group integration process (S15) will be described. 10 to 12 are diagrams for explaining the group integration process, in which FIG. 10 is a form when refugees encounter each other, FIG. 11 is a form when refugees encounter a group, and FIG. These are figures which showed the form when a group encounters.

FIG. 10 is a diagram showing a form when evacuees encounter each other, and schematically shows a state after the calculation (S13) of the position information of each refugee described in the flowchart of FIG. . Here, the case where the refugee a and the refugee b encounter is considered. In the present embodiment, a circle having a predetermined distance (radius) centered on the refugee a is set as the searchable range L. Since the evacuee b is located within the searchable range L of the evacuee a, it is determined whether there is a common one with reference to the group to which both belong. When the group to which the evacuee a and the evacuee b belong is compared, it can be seen that the group A matches. The refugee a and the refugee b have a relationship such as a family, a company colleague, a school classmate, and the like, and are therefore integrated as a group A. On the other hand, when there is no group belonging to each other among the evacuees, each evacuee moves toward the target evacuation site set for each evacuee as before.

  In this embodiment, the circular searchable range L centered on the refugee is set in this way, but a visible range that takes into account the direction of movement of the refugee may be adopted as the searchable range. . Further, the searchable range may be such that the visible range of the refugee is limited in view of the terrain around the refugee and the state of the building (spatial data). Furthermore, it may be possible to limit the visible range according to the disaster situation (disaster data) such as fire smoke.

  Therefore, the moving direction of the evacuee a is set to the direction of the evacuee b until the evacuee b is reached. After the evacuee a and the evacuee b are integrated as the group A, the target evacuation place set for each refugee is changed to the target evacuation place as a group. The target evacuation site as a group is changed to one of the target evacuation sites that each refugee a and b has as a movement target. As the change rule, it is conceivable to select a distance from the current position to the target evacuation site, or to select based on an attribute in the human data, for example, an order relationship such as age or job title. Alternatively, when a safety level is set for each target evacuation site, a target evacuation site with a higher safety level may be selected. In addition, when a safety level is set for the target evacuation site, the target evacuation site may be selected in which a safety level that sufficiently covers the danger level is compared with the danger level at the current position. Or it is good also as selecting randomly, without using the attribute of human data, and the safety level of a target evacuation site.

  Thus, the refugee a and the refugee b are integrated into the group and start moving to the target evacuation site set in the group. Group movement is performed in the same manner as evacuees. The walking speed was set for the evacuees, but since the group moves with multiple evacuees, the group walking speed is set to the lowest walking speed among the group evacuees. The

  FIG. 11 is a diagram for explaining the group integration process (S15) for the form when the evacuees encounter a group. FIG. 11 shows a case where two groups A and C are located within the searchable range L of the evacuee c (the searchable range L is omitted in the figure). For evacuees c, belonging groups having priorities in the order of groups A, B, and C are set. Two groups A and C are located within the searchable range L of the refugee c. In this case, the refugee c starts moving so as to be integrated into the group A having a high priority. The integration is the same as the encounter between the refugees described with reference to FIG. 10, and the integrated group A starts moving toward the newly set target evacuation site. Thus, the evacuees are integrated into a higher priority group and moved to the target evacuation site.

  FIG. 12 is a diagram for explaining the group integration processing (S15) for the form when the groups encounter each other. FIG. 12 shows a case where group B and group C meet, that is, group B is located within the searchable range of group C (the searchable range L is omitted in the figure). In group C, three evacuees g, h, and i are integrated. Among these members, evacuee g has group B having a higher priority than group C to which it currently belongs. Accordingly, the evacuee g starts moving so as to leave the group C and integrate it into the group B. The integrated group B starts moving toward the newly set target evacuation site. On the other hand, in Group C, the target evacuation site may be changed as the evacuee g leaves, but the target evacuation site may be maintained. As described above, in the group integration processing of the present embodiment, when an evacuation action is initially performed with a colleague of the company, when the family is encountered, the evacuation action is performed with the family away from the colleague of the company. It is possible to calculate the position of the evacuees according to the evacuation situation.

  The evacuation behavior prediction system according to the present embodiment has been described above. With this evacuation behavior prediction system, the behavior (position) of each refugee over time after various disasters such as earthquakes, tsunamis, floods, and fires occur. Can be predicted. In addition, by making predictions (simulations) by changing various conditions, it is possible to detect problems at the time of disaster and to take measures for disaster prevention. Although the details of the evacuation behavior prediction system have been described here, the present invention can employ a computer-executable program as the object. By installing the program in a computer, the above-described evacuation behavior prediction system can be configured.

  Note that the present invention is not limited to these embodiments, and embodiments configured by appropriately combining the configurations of the respective embodiments also fall within the scope of the present invention.

DESCRIPTION OF SYMBOLS 1 ... Evacuation action prediction system, 2 ... Disaster scenario data, 3 ... Disaster prediction means, 4 ... Time data, 5 ... Disaster data, 6 ... Spatial data, 7 ... Human data, 8 ... Input part, 9 ... Time data storage part DESCRIPTION OF SYMBOLS 10 ... Disaster data storage part, 11 ... Spatial data storage part, 12 ... Human data storage part, 13 ... Image display means, 14 ... Evacuee action prediction means, 15 ... Control part, 16 ... Image display output part, 17 ... Data file output unit, 18 ... output unit

Claims (6)

Storage means and control means,
The storage means
Disaster data that defines the extent of damage caused by a disaster;
Spatial data including evacuation site information and passage information,
For each evacuee, the affiliation group, the target evacuation location, and human data with the moving speed set are stored,
The control means can execute evacuee position calculation processing, group integration processing, and group position calculation processing,
The evacuee position calculation process calculates, for each evacuee, position information for each time step so as to go to a target evacuation site set for the evacuee based on the spatial data and the human data.
In the group integration process, in the evacuee position calculation process, an evacuee having a common group belongs to a common target on the condition that one refugee is within a searchable range defined as a reference point. Integrated into a group to the evacuation site,
The group position calculation process calculates, for each group, position information for each time step so as to go to a target evacuation site set for the group based on the spatial data and the human data. Evacuation behavior prediction system. In human data, a plurality of belonging groups are set for each refugee, and a priority is set for each refugee in the belonging group.
The group integration process is to integrate other evacuees who belong to a higher priority group into other evacuees and groups on the condition that they are within the searchable range. The evacuation behavior prediction system according to claim 1. Each target evacuation site has a safety level,
The integrated group target evacuation site in the group integration process is set at a target evacuation site with a high safety level among the target evacuation sites set for the refugees to be integrated. The evacuation behavior prediction system according to claim 2. The evacuation behavior prediction system according to any one of claims 1 to 3, wherein the disaster data changes with time. The evacuation behavior prediction system according to any one of claims 1 to 4, wherein the control unit is capable of executing a disaster prediction process for calculating disaster data based on at least disaster scenario data. In the evacuation behavior prediction program executable by a computer based on various data stored in the storage means,
The storage means
Disaster data that defines the extent of damage caused by a disaster;
Spatial data including evacuation site information and passage information,
For each evacuee, the affiliation group, the target evacuation location, and human data with the moving speed set are stored,
For a certain refugee, based on the spatial data and the human data, an evacuee position calculation process for calculating the position information for each time step so as to go to the target evacuation place set for the refugee,
In the evacuee position calculation process, the evacuees having a common affiliation group are grouped to a common target evacuation site on condition that one refugee is within the searchable range defined by using one refugee as a reference point. Group integration processing to integrate,
Based on the spatial data and the human data, a group position calculation process for calculating position information for each time step so as to go to the target evacuation site set for the group is enabled. An evacuation behavior prediction program characterized by this.

Images