DE102009059892A1 - Device and method for the dynamic adaptation of simulators of person and object streams as a basis for a forecasting tool with human-machine interaction - Google Patents

Abstract

The present invention is based on a simulator with cellular state machines. A simulation of object flows, in particular flows of people, should be created as realistically as possible. In particular, it should be possible to react dynamically to changes in the situation. The invention is characterized in that a route selection, a destination selection and / or particle properties depending on an event, fixed in time or space or randomly controlled on the one hand by means of the control device (7) and the computer device (9) or on the other hand by means of the control device (7) and a man-machine interface (8) can be dynamically adapted.

Description

The present invention relates to a device according to the preamble of the main claim, and a method according to the preamble of the independent claim and uses according to the additional claims.

Wherever people or objects appear concentrated, mass-typical phenomena arise. Some of these phenomena endanger the safety of life and limb, such as when people are trampled to death at a rock concert. Other phenomena require appropriate steering measures to make processes efficient from a technical and economic point of view. These include evacuations of a site after a mass event, such as in a football stadium and its surroundings, or the steering of the road at peak hours.

Conventionally, there are some approaches, in particular to simulate passenger and car flows. The known approaches, however, do not include convincing methods to respond dynamically to changes in the situation.

For example, for streams of people when a fire breaks out or escape routes are opened. Previous object and personal current simulators are based on an initial situation, topology and parameter specification, which can not be varied during the simulation. Even the behavior of other people is considered only locally, generally at the perimeter of less than 10 meters. A dynamic adaptation of the person behavior and the simulation to interesting length scales larger than 10 meters practically does not take place, for example, jams are not recognized and thus can not be bypassed.

There are no known conventional solutions to the problem on this side.

A pedestrian simulator calculates the next position of a person in terms of time and space using mathematical methods and thus using fixed rules that are intended to mimic human behavioral patterns. Here, various models from the field of micro- or macro-modeling are used, ranging from methods based on partial differential equations to methods based on cellular automata.

In all known realizations, these methods are static. Methods relating to dynamic adaptation by interactive external intervention in the simulation, for example the possibility of opening doors at a certain time, or an adaptive behavior of the persons to intrinsic signals, that is to say artificial intelligence, for example a modification of the routing due to arising congestion, are not realized or provided in conventional object and pedestrian simulations. The prior art does not provide for dynamic adaptation in object and pedestrian simulators to outside intervention.

• – ein Zielmodell legt fest, wie sich Teilchen/Personen auf ein Ziel zu bewegen;
• – ein Modell zu Teilchen- oder Personenbewegungen legt fest, wie sich Teilchen/Personen untereinander verhalten;
• – ein Hindernismodell definiert, wie sich Teilchen/Personen um Hindernisse bewegen.

One frequently chosen approach to pedestrian simulations is devices and methods based on cellular state machines. In this case, an area, for example a street, is covered with a cell grid. For example, hexagonal grids or square cells are common. Each cell can occupy different occupancy states, such as filled with an obstacle, or occupied by a person, or empty. Such states are updated via rule sets or machines over time. The following submodels and their interaction contain the core ideas of this automaton:

• A target model determines how particles / people move towards a target;
• - a model of particle or person movements determines how particles / persons behave with each other;
• - An obstacle model defines how particles / people move around obstacles.

Proven here is an approach that mimics well-known mechanisms from the physics of electronics. In the formulation this is realized via potential fields.

Targets attract particles / persons as a positive charge attracts electrons. The strength of the potential field is determined in the prior art as a function of the Euclidean distance of the person / particle to the target.

Particles / people repel each other as electrons repel each other. The strength of the potential field is conventionally determined as a function of the Euclidean distance of the persons / particles from each other.

Obstacles repel particles / persons as a negative charge repels electrons. The strength of the potential field is conventionally determined as a function of the Euclidean distance of the person / particle to the obstacle.

It is an object of the present invention to provide an apparatus and a method for generating particles moving in a spatial area of the apparatus, in particular based on cellular state machines. It is supposed to be a simulation of object streams In particular, flows of people are created as realistically as possible. In particular, it should be able to react dynamically to changes in the situation. Based on the state of the art, an apparatus and an additional method are to be provided which overcomes the aforementioned conventional deficiency. It should result in a significantly improved overall behavior of particle flows, so a correct image of actual behavior. By means of devices and methods according to the invention, control pulses are to be generated to a control center for controlling building elements.

The object is achieved by a device according to the main claim and a method according to the independent claim.

The invention focuses on an apparatus and method for generating streams of particles. The invention relates to particle flows of any mobile particles. Such particles may be, for example, metal spheres. These particles may be, for example, persons, persons on means of transportation, such as bicycles or motor vehicles, or likewise, these particles may represent animals.

According to a first aspect, there is provided a device for generating motions of particles detected in a spatial area of the device by means of a first detection device, the area being covered by a cell grid and each cell being able to occupy different occupancy and total potential states by means of a computer and to update timing, wherein each cell is assigned a target potential, which determines how particles are attracted by a target, and an obstacle potential that determines how particles are repelled by an obstacle, and wherein each particle has a particle potential associated with it in which a total potential in a cell is composed of the values of the target potential and the obstacle potential in the cell and the particle potentials of particles detected by the first detection means in neighboring cells of the cell, and d Particles emanate from a respective start cell in each case from one cell to a neighboring cell with a lowest total potential change. The invention is characterized in that a routing, a destination selection and / or particle properties depending on an event, temporally or spatially fixed or randomly controlled on the one hand by means of the control device and the computer device or on the other hand dynamically adjusted by means of the control device and a man-machine interface ,

According to a second aspect, there is provided a method of generating particle streams, comprising the steps of providing a device having a spatial and cell gridded region, each cell occupying different occupancy and total potential states set by means of a driver and computer means Each cell is assigned a target potential that determines how particles are attracted to a target, and an obstacle potential is assigned that determines how particles are repelled by an obstacle, and where each particle has a particle potential, with a total potential in a cell the values of the target potential and the obstacle potential in the cell and the particle potentials of particles detected by a first detecting means in neighboring cells of the cell; Positioning particles at respective start cells, each particle changing from one cell to a neighboring cell having a lowest total potential; Detecting the positions of the particles by means of the first detection means; Updating the total potential states by means of the first detection device, the computer device and the drive device. The method is characterized in that a routing, a destination selection and / or particle properties depending on an event, temporally or spatially fixed or randomly controlled on the one hand by means of the control device and the computer device or on the other hand dynamically adjusted by means of the control device and a human-machine interface ,

According to a third and fourth aspect, an apparatus or method according to the present invention is used to simulate and / or control object streams, streams of people or animal movements.

A detection device may be an optical detection device, for example a camera.

Occupancy states can be: occupied or free of particles, obstacle, target or source.

Particles may be, for example, metal spheres. Particles may be, for example, electrically negatively charged metal spheres. Particles may be assumed to represent persons or animals. Particles thus generally represent objects.

A dynamic adaptation of the behavior of individual objects, in particular of virtual persons and thus of streams of people, is provided in order to be able to collect virtual experiences in an interactive training forecasting tool.

In accordance with the present invention, dynamic intervention in person and object simulation is provided and appropriate approaches for dynamically adapting the device and method to such procedures are claimed. In particular, methods for integrating dynamic response to external influences in a simulation are described. In accordance with the present invention, the development of training simulators for better planning and control of streams of people in infrastructures and mass events is thus made possible in principle. In accordance with the present invention, a human-machine interaction prediction tool may be provided.

• – von außen durch ein Ereignis gesteuert werden, und zwar insbesondere interaktiv über eine Mensch-Maschine-Schnittstelle beispielsweise durch ein Signal zur Evakuierung oder einem Öffnen von Türen;
• – innerhalb eines Modells, nach einem gewissen Schema – etwa einer bestimmten Zeit oder in einem bestimmten Bereich – injiziert werden, beispielsweise durch die Einfahrt eines Zuges oder eine Öffnung von Türen gemäß einem vorgegebenem Fahrplan;
• – zufallsgesteuert sein.

According to the present invention, it is possible to intervene dynamically in the simulation of person and object streams. According to the present invention, interventions can be done interactively via a man-machine interface, injected within a simulator according to a certain scheme, or random. Such an adaptation can

• Be controlled externally by an event, in particular interactively via a man-machine interface, for example by a signal for evacuation or opening of doors;
• - be injected within a model, according to a certain scheme - such as a certain time or in a specific area - for example by the entry of a train or opening of doors in accordance with a given timetable;
• - be random.

• – Adaptive Wegewahl, und zwar über eine Neuinitialisierung von Navigationsfeldern oder Neusetzen von Zwischenzielen;
• – Adaptive Zielwahl;
• – Adaptive Personeneigenschaften.

In order to make this meaningful, dynamic adjustments are made to three main elements of a model for a movement of particles or persons:

• Adaptive routing, via reinitialization of navigation fields or re-setting of intermediate destinations;
• - Adaptive targeting;
• - Adaptive person characteristics.

The present invention allows the realistic simulation of particle flows or streams of people while taking into account a dynamic adaptation of the particle flows or streams of people, as conventionally not realized.

Adaptive targeting

• – von außen durch ein Ereignis gesteuert werden, und zwar insbesondere interaktiv über eine Mensch-Maschine-Schnittstelle, beispielsweise durch ein Signal zur Evakuierung;
• – innerhalb des Modells, nach einem gewissen Schema – etwa einer gewissen Zeit, oder in einem bestimmten Bereich – injiziert werden, beispielsweise eine Einfahrt eines Zuges und eine Öffnung von Türen gemäß einem vorgegebenen Fahrplan;
• – Zufallsgesteuert sein.

Depending on the type of person-current simulator used, persons receive a fixed target at the beginning of the simulation. This strictly deterministic behavior is state of the art and neglects the fact that people tend, at least from the point of view of the observer, to spontaneously adapt their goals. For example, when a person spontaneously enters a shop on the way to the exit of a train station. Another possibility is when an escape is triggered. According to this present invention, this approach is broken and individuals can be assigned new goals. Such an assignment can again

• Be controlled from the outside by an event, in particular interactively via a man-machine interface, for example by a signal for evacuation;
• - be injected within the model, according to a certain scheme - for a certain time, or in a certain area - for example, a train entrance and an opening of doors according to a given timetable;
• - be random.

• – eine räumliche Triggerung. Beispielsweise betritt eine Person einen Bereich vor einem Laden, so wird ihr Ziel auf den Laden bzw. das Innere des Ladens umgestellt;
• – eine externe Triggerung. Wird beispielsweise ein Signal zur Evakuierung gegeben, so wird das Ziel beispielsweise auf den nächsten Notausgang umgestellt;
• – eine zeitliche Triggerung. Beispielsweise nachdem eine Person eine gewisse Zeit im Laden verweilt ist, schaltet diese auf ihr ursprüngliches Ziel um und verlässt somit wieder den Laden;
• – eine ereignisgesteuerte Triggerung. Beispielsweise fährt ein Zug ein, den eine Person noch erreichen möchte;
• – spontane, rein zufällige Triggerung. Beispielsweise fällt einer Person spontan ein, dass diese noch Besorgungen im Laden zu erledigen hat.

The change of destination can in turn be determined according to the characteristics of the individual or random. For example, not every person automatically enters a store at any time. Examples of triggering a new target are:

• - a spatial triggering. For example, a person enters an area in front of a store, so their destination is converted to the store or the interior of the store;
• - an external triggering. If, for example, a signal for evacuation is given, the destination is switched, for example, to the next emergency exit;
• - a temporal triggering. For example, after a person lingers in the store for a certain amount of time, they switch to their original destination and thus leave the store again;
• - event-triggered triggering. For example, a train arrives that a person still wants to reach;
• - spontaneous, purely random triggering. For example, a person spontaneously remembers that they still have errands to do in the store.

Adaptive adaptation of personal characteristics

• – von außen durch ein Ereignis ausgelöst werden – insbesondere interaktiv über eine Mensch-Maschine-Schnittstelle, beispielsweise durch ein Signal zur Evakuierung oder dem Öffnen von Türen; Bei Flucht wird eine Wunschgeschwindigkeit als Personeneigenschaft deutlich erhöht;
• – innerhalb des Modells, nach einem gewissen Schema – etwa einer bestimmten Zeit oder in einem bestimmten Bereich – injiziert werden. Beispielsweise ein Eintreten von Dunkelheit; Bei Eintreten von Dunkelheit wird im Allgemeinen aufgrund von Sichtbehinderung langsamer gelaufen;
• – zufallsgesteuert sein.

The virtual persons considered in the microscopic simulations simulations have different characteristics according to the prior art. These can be customized. Such an adaptation can

• - be triggered from the outside by an event - in particular interactively via a human-machine interface, for example by a signal for evacuation or the opening of doors; In flight, a desired speed as a person characteristic is significantly increased;
• - be injected within the model, according to a certain scheme - such as a certain time or within a certain range. For example, an onset of darkness; As darkness falls, walking is generally slower due to obstruction;
• - be random.

According to the present invention, a dynamic adaptation of simulators of streams of people or object streams to external or intrinsic, that is, in the course of the simulation resulting situation changes and signals.

• – eine dynamisch Adaptive Wegewahl virtueller Personen;
• – eine dynamisch Adaptive Zielwahl virtueller Personen;
• – eine dynamische Anpassung der Eigenschaften virtueller Personen.

According to the present invention, dynamic intervention in a simulation of passenger flows. Such interventions can be done interactively via a human-machine interface, injected within the simulator according to a certain scheme, or be random. In order to make this meaningful, the invention provides dynamic adaptation methods on three main elements of a model for the movement of particles or persons:

• - a dynamically adaptive routing of virtual people;
• - a dynamically adaptive targeting of virtual people;
• - a dynamic adaptation of the properties of virtual people.

The invention provides that the interventions take place interactively via a human-machine interface or are controlled intrinsically by the simulation. Adjustments can be made both deterministically and provided with certain probabilities.

According to the device according to the invention and according to the method according to the invention or the uses according to the invention, the following advantages result.

The invention enables interactive intervention in the simulation and thus for the first time the creation of a training simulator or a prognosis tool with human-machine interaction.

A dynamic adaptation to different circumstances corresponds to the human behavior much better than a purely static view. Thus, real flows of people can be depicted much closer to reality in the simulations. Typical statements about crowd levels, person densities, and flows are significantly more reliable using the proposed apparatus and method.

Thus, it is only through this approach with realistic movement patterns that a realistic calibration of essential parameters, especially in cellular approaches, possible. Furthermore, the invention allows better compliance with personal flow simulator guidelines, thanks to the explicit modeling closer to reality.

Further advantageous embodiments are claimed in conjunction with the subclaims.

According to an advantageous embodiment, the routing adjustment can be carried out by means of a reinitialization of total potentials by means of the drive device and the computer device.

In most pedestrian simulators, a person is given a path by means of a scalar navigation field, a so-called navigation potential. People move in each case in the direction of the minimum value of the navigation field, that is, along the steepest descent. The determination of the navigation field taking into account the necessary parameters, such as accessible space, goals and obstacles, is state of the art. In methods according to the prior art with the specification of a navigation field, the chosen path is fixed predetermined. The approach is also used for the simulation of object movements in general.

• – ereignisgesteuerte Initialisierung, beispielsweise nach Öffnen einer Tür.
• – zeitgesteuerte Initialisierung, das heißt wiederholte automatische Initialisierung in zeitlich konstanten oder variablen Intervallen, beispielsweise Initialisierung in jeden 10. diskreten Zeitschritt der Personstromsimulation bzw. Teilchenstromsimulation;
• – zufallsgesteuerte Initialisierung.

According to the navigation potential during the simulation is redetermined when a situation change occurs. This dynamic redefinition of navigation potential changes the routing for each object to fit the new situation. The following exemplary options are available for controlling the routing or reinitialization of the navigation potential:

• - event-driven initialization, for example after opening a door.
• Timed initialization, that is repeated automatic initialization at time-constant or variable intervals, for example initialization into every 10th discrete time step of the person flow simulation or particle flow simulation;
• - randomized initialization.

The dynamic adaptation takes place by taking into account a changed situation, for example opening a door. In accordance with the present invention, the navigation potential is also reinitialized at regular time intervals independently of situation changes injected in the outside. Then also changes in the simulation run, which are not controlled from the outside, can be considered.

According to a further advantageous embodiment, the total potential can be proportional to a density of particle quantities on accessible Because of being raised to a destination for cells on these paths. In particular, the invention proposes determining the navigation potential as a function of the density of particle quantities or quantities of people on the accessible routes to the destination. In this weighted re-initialisation, high density of people or particle densities increase the navigation potential. Particles follow the lowest potential, making paths that become jammed automatically less attractive.

Repeated reinitialization requires increased computation times, but leads to significantly more realistic person behavior.

According to a further advantageous embodiment, the routing adjustment can be performed by means of a replacement of intermediate destinations by means of the drive device and the computer device. In some pedestrian simulators, people are navigated by means of intermediate targets on graphs. This partly also in connection with a navigation field based navigation from intermediate to intermediate destination. According to the state of the art, at the beginning of the simulation, the persons receive a fixed path to the destination. However, this strictly deterministic behavior neglects the fact of dynamically adjusting the routing. Completely analogous to the routing when using a navigation field can also be used here a dynamic routing. According to the exemplary embodiment, it is provided that the intermediate destinations-injected by events, predetermined timings or coincidence-are re-selected as a function of the changed situation. The invention also provides that intermediate destinations are re-selected at regular intervals in order to respond to the situation change that has occurred as a result of the simulation run.

According to a further advantageous embodiment, a probability that an intermediate destination is chosen on a way to a final destination in the re-setting may decrease if there is a particle accumulation on the intermediate destination and / or with an increasing stowage strength. According to the embodiment, the routing is done depending on detected congestion. In particular, the likelihood that a certain intermediate destination will be chosen on the way to the final destination should decrease if there is a traffic jam on the intermediate destination. The probability should also depend on the stasis.

According to a further advantageous refinement, the destination selection adaptation can take place by means of a destination selection from a predefined set of destinations according to different viewpoints. The new one-touch dial may not only refer to a destination, such as persons in front of a store, but may also include the selection of a destination from a given set of destinations according to different viewpoints, such as in an escape the nearest exit is chosen. The choice may be deterministic, such as the nearest destination, or random.

According to a further advantageous embodiment, the particle properties may be the fixed speed at which particles move in a free path, or a preferred distance to other particles. The virtual persons contemplated in microscopic simulcast simulations have different characteristics in the prior art, such as different desired speeds that people attempt to maintain during simulation, that is, in each discrete time step of the simulation, or different preferred distances to other persons. In all passenger current simulators according to the prior art, these properties are fixed at the beginning of the simulation, a dynamic adjustment is not possible.

In reality, however, such adaptation occurs in certain situations. In a threatening situation that accompanies an escape, people deviate from their desired speed and move with the maximum possible speed, which also varies from person to person. Furthermore, persons in the event of an escape are quite willing to fall below the comfort distance to other people.

According to a further advantageous embodiment, the particle property of a velocity of a particle can be adjusted by scaling a maximum velocity by multiplying the maximum adaptation factor. The walking behavior of people and in particular the speed with which people walk has outstanding importance in the simulation of movements of persons. Therefore, in the following, the dynamic adaptation of the person speed to changes in the situation is considered more precisely. The invention provides in particular to adapt the desired speed of persons depending on the situation. Such situations can be, for example, an escape situation or a normal situation. The desired speed is the fixed speed with which people run at free choice, according to the prior art methods.

A particularly effective implementation is the scaling of the maximum speed v _{max of} a person. The maximum speed of persons varies from person to person and in reality is essentially normal. By the Scaling can dynamically provide corresponding variations in the desired speed _Wish = q _wish × v _max be realized, for example v _escape = q _escape × v _max .

The choice of factors q may depend on different parameters. It can be deterministic or subject to random and temporal, especially random, fluctuations. When determining q, it should be borne in mind that the speed of a person should be characterized and maintained on average by that person. Furthermore, frequent abrupt variations in q are unnatural. The invention provides that small q should be able to be designed as a function of parameters such as the person characteristic and of the running past. The exact design of such dependencies is still the subject of research.

According to a further advantageous embodiment, an event may be opening or closing a door, detecting or avoiding a jam, entering a store, blocking, evacuation, train entry or announcement of a change of track. These are conceivable scenarios in which dynamic adjustments are extremely important. Other events may include an occurrence of smoke or fire or an alarm.

According to a further advantageous embodiment, the timed adjustments can be performed in regular time steps.

According to a further advantageous embodiment, the spatially determined adjustments can be carried out as a function of the density of particle quantities on the accessible cells to the destination, with high particle densities increasing the total potential of the accessible cells to the target.

According to a further advantageous embodiment, real object movements can be detected by means of a second detection device for the initialization of positions of the particles, of start cells, targets and particle velocities.

According to a further advantageous embodiment, an evaluation device for evaluating the particle movement detected by means of the first detection device for the respective generation of control pulses to a control center can be provided.

According to a further advantageous embodiment, the control center can be provided for controlling building elements.

According to a further advantageous embodiment, building elements may be doors, windows, signs, loudspeakers, elevators, escalators and / or lights.

The present invention will be described in more detail by means of exemplary embodiments in conjunction with the figures. Show it:

1 a flowchart of a dynamic adaptation to simulation-internal situation changes in regular time steps;

2 a dynamic adaptation to external situation changes;

3 different time steps of a simulation of a person stream with dynamic adaptation using the example of opening a door or removing an obstacle;

4 an example of a dynamic and intelligent adaptation of the routing to different circumstances;

5 an example in which a simultaneous adaptive adjustment of goals and speed is realized;

6 an embodiment of a device according to the invention;

7 an embodiment of a method according to the invention.

1 shows a flowchart of a dynamic adaptation to simulation-internal situation changes in regular time steps. Along a time axis t, on the left side of the 1 is shown, the following steps are performed. With a step s1 a parameter selection takes place, which is static. With a step S2, an initialization takes place. With a step S3, a first simulation step is carried out. In a step S4, a second simulation step is carried out. With step S5, a kth simulation step is executed. A step S6 checks whether the situation has changed. If it is determined in a step S6 that the situation is not changed, it is returned to the step S2 without a change. If it is determined that the situation has changed, the parameter is returned from step S6 with a dynamic adaptation of the parameters in a step S7 to step S2. If it is determined in a step S6 that the time has expired, the simulation is ended with a step S8.

2 shows a dynamic adaptation to external situation changes. With a step S1, a static parameter selection takes place. With a step S2, an initialization takes place. With a step S3, a first simulation step is carried out. Thereafter, in step S4, it is determined whether there is an external intervention that is interactive, timed, or random. If such an intervention is present, a dynamic adaptation of the parameters ensues with a step S5. A re-initialization, ie an adaptation on the basis of step S5, ensues with a step S6. If such an intervention does not occur before or after step S6, a further simulation step follows in a step S7. This is followed by a decision in step S8 as to whether the simulation is to be ended or not. If the simulation does not end, step S3 is returned and a further simulation step is carried out. If the simulation is to end, the simulation is ended with a step S9.

3 shows different time steps of a simulation of a person stream with dynamic adaptation on the example of opening a door or removing an obstacle. There is a dynamic adjustment by taking into account a changed situation, here the opening of a door. In the pictures b and c the middle small obstacle is inactive, that means the door is open. The possibility of interactive intervention by the user in object and pedestrian simulations is of central importance. 3 shows such an intervention on the example of opening a door. Shown are different time steps of a simulation. The flow of people that originates in the person-generating source on the left, moves toward the target, which is spread over the entire right side. After opening the door in picture b and c, the persons or particles move through the door, as this is the shortest way. As soon as the door is closed, the particles or persons move around the wall. In picture d it can be clearly seen how also the persons who could not cross the door are forced to choose the detour around the obstacle.

4 shows an example of a dynamic and intelligent adaptation of the routing to different circumstances. By repeated weighted initialization of the goal-setting potential, the simulated persons can recognize congestion and deal with it accordingly. Repeated initialization places more emphasis on high personal density distances compared to middle and low population densities. In 4 the source from which the people pour is in the lower left corner, the target is in the lower right corner. Both are separated by an obstacle with only two openings. Due to the much smaller dimensions of the lower opening, there forms a jam. This is detected by means of the weighted flooding by the pedestrians and, when exceeding a certain size, they select a detour via the upper opening. 4 shows different time steps of a simulated pedestrian flow with intelligent dynamic adaptive intelligent routing. High density areas are avoided so that some people choose the detour via the above passage to the destination as soon as the congestion at the lower passage becomes too large. Every 10 time steps, the underlying potential was recalculated, with distances weighted differently depending on local population density. The navigation potential is also reinitialized at regular intervals independently of external changes in the situation. This also allows changes in the simulation run, which are not controlled from the outside, to be taken into account. The invention proposes, in particular, to determine the navigation potential as a function of the density of quantities of people or quantities of particles on the accessible routes to the destination. In this weighted reinitialization, high particle densities (densities) increase the navigation potential. People follow the lowest potential, so that traffic jams are automatically less attractive.

5 shows an example in which a simultaneous adaptive adjustment of goals and speed is realized. In each case different time steps of a simulation are shown. The persons move from the entrances and exits of one building to those of the other buildings. Due to the dimensioning, some traffic jams occur. In c, an event occurs, an intrinsic or extrinsic one that causes the person to escape. You automatically select the next emergency exit, these are the targets at the top and bottom, and move towards them. Since an evacuation situation is assumed, the speeds of the persons are adjusted at the same time. The sometimes different desired speeds - an inhomogeneous amount of people is considered - are multiplied by a constant factor. 5 shows different time steps of a passenger flow simulator with adaptation of goals and personal characteristics, which here is the speed, triggered by an external signal. In figures a-c, people move from one building to another. Between the time steps c and d, a dangerous situation occurs and the signal for evacuation is given. In the time steps d-f, the persons leave the premises at increased speed to the emergency exits at the upper and lower edge.

6 shows an embodiment of a device according to the invention.

The device I generates a movement of particles 3 which may be metal balls, for example.

The device I has a cell grid in a spatial area 5 on. Each cell is assigned a temporally variable total potential. particle 3 For example, metal spheres are initially placed on the cell grid 5 positioned. A number may be, for example, n = 50 beads. By means of a drive device 7 The cells can be assigned temporally variable total potential values. Each cell may for example be associated with an electromagnet whose magnetic force by means of the control device 7 is adjustable. The drive device 7 can set a respective potential by means of a current through an electromagnet. At a start time Ts are by means of the driving device 7 activates the potentials, the beads move starting from a respective start cell S on each other balls and obstacles H over to the target Z. At an end time Te all the beads can have reached their goals Z. For visualization and / or detection of the movement of the beads, a first detection device 1 , such as a camera. The information - these can be the directions of movement of particles 3 be - the first detection device 1 can in a computing device 9 are used to calculate respective particle potentials. The drive device 7 can by means of a man-machine interface 8th for example, to control device parameters. The information of the first detection device 1 can also be in an evaluation device 11 be rated. For example, a particle density in the cell grid 5 recorded and evaluated. The evaluation device 11 can send control signals to a control center 13 for controlling building elements 15 , such as doors or signs, spend. The device I can also be emulated, for example, by a computer. The device is particularly suitable for simulating flows of people, for example in buildings. The model of the device I according to the invention can be transferred to a computer with a corresponding model. Ie. The device I can also be simulated by a computer. Such an embodiment is also included within the scope of this application.

7 shows an embodiment of a method according to the invention.

In step S1, provision is made of a device having a spatial with a cell grid 5 coated area, each cell occupying different occupancy and total potential states by means of a drive device 7 and a computer device 9 each cell is assigned a target potential that defines how particles 3 are attracted by a target Z, and an obstacle potential is assigned, which defines as particles 3 be repelled by an obstacle H, and where each particle 3 associated with a particle potential, wherein a total potential in a cell is determined from the values of the target potential and the obstacle potential in the cell and the particle potentials of a first detection means 1 detected particles 3 in neighboring cells of the cell. With a step S2, a positioning of particles takes place 3 at respective start cells S, after which the particles 3 each change from one cell to a neighboring cell with the lowest total potential;
With a step S3, the positions of the particles are detected 3 by means of the first detection device 1 , With a step S4, the total potential states are updated by means of the first detection device 1 , the computing device 9 and the drive device 7 , With a step S5, a routing, a destination selection and / or particle properties depending on an event, temporally or spatially determined or randomly controlled on the one hand by means of the control device 7 and the computer device ( 9 ) or on the other hand by means of the drive device 7 and a human-machine interface 8th dynamically adjusted.

The invention has been explained with reference to a simulator based on cellular state machines. Transferring the methods to other approaches - such as highly individualized agent-based methods - is easily possible. In the simulation of streams of people with cellular automata, people with a fixed speed scheme move on the basis of potential considerations to their next cell and thus to their desired destination. This model is enhanced by introducing dynamic customizations at several, particularly relevant points.

Claims (32)

Device (I) for generating by means of a first detection device ( 1 ) detected movements of particles ( 3 ) in a spatial area of the device, the area having a cell grid ( 5 ) and each cell can occupy different occupancy and total potential states that can be determined by means of a computer device ( 9 ) and a drive device ( 7 ) and updated over time, each cell being assigned a target potential that specifies like particles ( 3 ) are attracted to a target (Z), and an obstacle potential is assigned that determines how particles ( 3 ) are repelled by an obstacle (H), and whereby each particle ( 3 ), wherein a total potential in a cell is determined from the values of the target potential and the obstacle potential in the cell and the particle potentials of the first detection device ( 1 ) detected particles ( 3 ) in neighboring cells of the cell, and particles ( 3 ) starting from a respective start cell (S) in each case by a cell in a neighboring cell with a lowest total potential, characterized in that a routing, a target choice and / or particle properties depending on an event, temporally or spatially fixed or randomly controlled on the one hand by means of the control device ( 7 ) and the computer device ( 9 ) or on the other hand by means of the control device ( 7 ) and a human-machine interface ( 8th ) dynamically adjusted. Device according to claim 1, characterized in that the routing selection by means of a reinitialization of total potentials by means of the drive device ( 7 ) and the computer device ( 9 ) is performed. Apparatus according to claim 2, characterized in that the total potential is increased in proportion to a density of amounts of particles in drivable ways to a target for cells in these ways. Apparatus according to claim 1, characterized in that the routing selection by means of a replacement of intermediate destinations by means of the drive device ( 7 ) and the computer device ( 9 ) is performed. Apparatus according to claim 4, characterized in that a likelihood that an intermediate destination is chosen on a way to a final destination in the re-setting, decreases when a particle congestion exists on the intermediate destination and / or with an increasing stowage strength. Device according to one of claims 1 to 5, characterized in that the destination selection adaptation takes place by means of a destination selection from a predetermined set of destinations according to different viewpoints. Apparatus according to any one of claims 1 to 6, characterized in that the particle properties are the fixed rate at which particles move in a free path or a preferred distance to other particles. Device according to one of claims 1 to 7, characterized in that the particle property of a velocity of a particle is adjusted by means of scaling a maximum velocity by multiplying the maximum velocity by an adaptation factor. Device according to one of the preceding claims 1 to 8, characterized in that the event, an opening or closing a door, a recognition or bypassing a jam, entering a shop, a barrier, an evacuation, a train entrance, an announcement of a track change, a Occurrence of smoke or fire or an alarm signal is. Device according to one of the preceding claims, characterized in that the timed adjustment is carried out in regular time steps. Device according to one of the preceding claims, characterized in that the spatially determined adaptation is carried out as a function of the density of particle quantities on the accessible cells to the target, wherein high particle densities increase the total potential of the accessible cells to the target. Device according to one of the preceding claims, characterized in that by means of a second detection means real object movements are detected for initialization of positions of the particles, of start cells, targets and particle velocities. Device according to one of the preceding claims, characterized by an evaluation device ( 11 ) for evaluation by means of the first detection device ( 1 ) detected particle movements and for the respective generation of control pulses to a control center ( 13 ). Device according to claim 13, characterized by the control center ( 13 ) for controlling building elements ( 15 ). Apparatus according to claim 14, characterized in that building elements ( 15 ) Doors, windows, signs, speakers, elevators, escalators and / or lights are. Method for generating particle streams, comprising the steps of providing a device with a spatial with a cell grid ( 5 covered area, each cell occupying different occupancy and total potential states, which are controlled by means of a control device ( 7 ) and a computer device ( 9 ), each cell being assigned a target potential that defines how particles ( 3 ) are attracted to a target (Z), and an obstacle potential is assigned that determines how particles ( 3 ) from an obstacle (H) repelled and each particle ( 3 ) is assigned a particle potential, wherein a total potential in a cell is determined by the values of the target potential and the obstacle potential in the cell and the particle potentials of by means of a first detection device ( 1 ) detected particles ( 3 ) in neighboring cells of the cell; - Positioning of particles ( 3 ) at respective start cells (S), after which the particles ( 3 ) each change from one cell to a neighboring cell with the lowest total potential; - Detecting the positions of the particles ( 3 ) by means of the first detection device ( 1 ); Updating the total potential states by means of the first detection device ( 1 ), the computing device ( 9 ) and the control device ( 7 ); characterized in that - a routing, a destination selection and / or particle properties depending on an event, temporally or spatially determined or randomly controlled on the one hand by means of the control device ( 7 ) and the computer device ( 9 ) or on the other hand by means of the control device ( 7 ) and a human-machine interface ( 8th ) dynamically adjusted. A method according to claim 16, characterized in that the routing adjustment by means of a reinitialization of total potentials by means of the control device ( 7 ) and the computer device ( 9 ) is performed. A method according to claim 17, characterized in that the total potential is increased in proportion to a density of amounts of particles in drivable ways to a target for cells in these ways. A method according to claim 16, characterized in that the routing adjustment by means of a replacement of intermediate destinations by means of the drive means ( 7 ) and the computer device ( 9 ) is performed. A method according to claim 19, characterized in that a likelihood that an intermediate destination is chosen on a way to a final destination in the re-setting, decreases when a particle congestion exists on the intermediate destination and / or with an increasing stowage strength. Method according to one of Claims 16 to 20, characterized in that the destination selection adaptation takes place by means of a destination selection from a predefined set of destinations according to different viewpoints. A method according to any one of claims 16 to 21, characterized in that the particle properties are the fixed rate at which particles move in a free path or a preferred distance to other particles. Method according to one of Claims 16 to 22, characterized in that the particle property of a velocity of a particle is adjusted by scaling a maximum velocity by multiplying the maximum velocity by an adaptation factor. Method according to one of the preceding claims 16 to 23, characterized in that the event, an opening or closing a door, a recognition or bypassing a jam, entering a shop, a barrier, an evacuation, a train entrance, an announcement of a track change, a Occurrence of smoke or fire or an alarm signal is. Method according to one of the preceding claims 16 to 24, characterized in that the timed adjustment is carried out in regular time steps. Method according to one of the preceding claims 16 to 25, characterized in that the spatially determined adaptation is carried out as a function of the density of amounts of particles on the accessible cells to the target, wherein high particle densities increase the total potential of the accessible cells to the target. Device according to one of the preceding claims, characterized in that by means of a second detection means real object movements are detected for initialization of positions of the particles, of start cells, targets and particle velocities. Method according to one of the preceding claims 16 to 27, characterized by an evaluation device ( 11 ) for evaluation by means of the first detection device ( 1 ) detected particle movements and for the respective generation of control pulses to a control center ( 13 ). Method according to claim 28, characterized by the control center ( 13 ) for controlling building elements ( 15 ). Method according to claim 29, characterized in that building elements ( 15 ) Doors, windows, signs, speakers, elevators, escalators and / or lights are. Use of a device according to one of claims 1 to 15 for the simulation and / or control of object streams, streams of people or animal movements. Use of a method according to one of claims 16 or 30 for the simulation and / or control of object streams, streams of people or animal movements.

Images