-For a comprehensive modeling of pedestrian dynamics in real-world scenarios the consideration of tactical level decisions in addition to operational ones is necessary. This paper presents a hybrid agent architecture employing a Floor Field approach at the operational level but granting agents an abstract representation of the simulated environment. The paper briefly presents the environmental model and hybrid agent architecture based on the floor field approach, then a sample practical application in a simple case study is also presented to show how it allows specifying abstract behavioural scripts for different groups of agents.
Index Terms—pedestrian simulation, agent-based modeling, floor-field model, hybrid agents, environments for multi-agent systems
The Floor Field approach to the modeling and simulation
of pedestrian dynamics, first introduced by [
system is “a first class-abstraction that provides the
surrounding conditions for agents to exist and that mediates both
the interaction among agents and the access to resources”.
An environment for agent-based systems can encompass both
abstractions and mechanisms, for instance regulating the
outcomes of agents’ chosen lines of action, as discussed by [
In order to support an automated production of the above elements and related data structures, a spatial representation of the area in which the simulation must take place, in the form of a CAD-like file, is required: on the other hand, this kind of map is generally produced when planning the construction of a building or available to managers of a premise. In order to allow algorithms to actually explore this representation and make sense of it, the designer is required to produce some A C
B
D GATE Hall-B
ENTRANCE 1 - EXIT 1
HALL ENTRANCE 2 - EXIT 2
E G
F H
A {lessonA}
C
B D form of annotation in it, as exemplified in Figure 1(a). In particular, the sub-areas in which the environment is divided into must be constrained by obstacles (in red in the figure), or passages, gateways to another sub-area (in cyan), and they must contain a specific block indicating a label that will be associated to the area. Both gateways and these label blocks are annotations that do not influence the walkability of the associated cells. Additional annotations represent start areas, in which pedestrian agents can be created (either initially or even at later stages of the simulation), end areas, final targets of movements in which pedestrian agents actually exit the simulation, and intermediate destinations (also associated to labels, not shown in the figure) that pedestrians must reach at a certain point of a more articulated movement plan.
The construction of such a plan requires the possibility to
explore and process a much simpler data structure, in
particular an abstract map in terms of a graph-like commonsense
representation of the environment, as discussed by [
Instead, for managing operational level tasks in the Floor Field approach, additional discrete grids containing gradientlike structures supporting agents’ navigation of the environment are necessary. Examples of these data structures are shown in Figure 1(c) and 1(d), respectively related to the static floor fields leading towards the gateway between the subareas labeled as “Hall” and “E” and the southern exit of the scenario. Once again, the annotated CAD-like representation of the environment supports the automated generation of these layers, by means of a simple cellular automaton whose description is omitted here for sake of space. Please notice that, however, we chose not to extend the diffusion of the static floor field associated to an area or marker to all the discrete representation of the environment, but to limit this operation to the sub-areas that are in direct connection to the target in the abstract commonsense representation. This, on one hand, simplifies the environment set up phase (especially considering relatively large environments, in which it would not be practically feasible to do this) and, on the other, is sufficient since the agents will be provided with a tactical level behavioural model allowing them to generate plans requiring the perception of these fields only in these adjacent sub-areas.
Additional layers are actually included in the agent environment to support the gathering of statistics about their dynamics, but also their interactions (in particular, the mutual perception of members of groups) and the management of conflicts (movement intentions are stored into one of these additional layers to support a simple identification and management of the conflicts by the environment itself).
it is clear that the information provided to agents’ perceptions
is sufficient to support basic operational level behaviour for
pedestrian agents. In fact, whenever an agent knows where
it’s headed (i.e. can perceive the static floor field associated structure, they can search for their goal in the abstract map
to that destination), the basic floor field model is sufficient to and in particular the intermediate steps leading towards this
allow the agent to achieve its own movement goal. However, goal. The result of this search operation (that employs state of
as mentioned in the previous section, the static floor fields the art graph search algorithms whose description is omitted
is not spread in the complete discrete representation of the for sake of space), leads to the construction of a plan in the
environment, so an agent must actually plan a course of action, form of a set of operating instructions (in the vein of [
to the previously described environment; in particular, it is a rough representation of a building including a set of lecture halls (labeled from “A” to “E”), all of which reachable from a central hall (actually including a surrounding corridor), with a northern and southern exits.
Within this scenario, exploiting the previously described tactical level extension, it is relatively simple to model different
The green agents, in other words, are already in the environment at the beginning of the simulation, in particular in lecture hall “A”, then they have to move to the sub-area in which “lessonE” takes place (i.e. lecture hall “E”) and finally they must leave the environment, half of them through the northern exit and the other half through the souther one. Yellow and white agents, instead, enter the simulation area respectively through the northern and southern entrances, then move to the sub-area where “lessonA” takes place (lecture hall “A”), then they have to move to “lessonH” (lecture hall “H”) and finally leave the environment, also splitting equally between northern and southern exits.It must be stressed the fact that the modeler is not forced to precisely and extensively define the path to be followed by the agents employing it, which will expand an abstract plan into a proper sequence of actionable operational level movements according to their tactical level knowledge.
The yellow agents’ plan, for instance, will be expanded into a sequence that is analogous to the one described in Figure 3. groups of pedestrian agents associated to different groups of Some screenshots of this simulation are shown in Figure 4. students, having different timetables. For instance we could In particular, in Figure 4(a) green agents are exiting hall define three groups (Green, Yellow and White agents) whose “A” and moving towards hall “E” (using both exits of hall tasks in the environment are the following: “A”), while yellow and white agents are moving to hall “A”.
Green [A → lessonE → Exit{0.5 North; 0.5 South}] In Figure 4(b) the three groups have reached the respective Yellow[EExnitt{ra0n.5ceNNoorrthth; 0.5→Southle}s]sonA → lessonH → tdheestiennavtiioronnsmanedn,t lawtehriloentihneFyigeullroew4(ca)n,dthewghriteeenggrorouupps lmeaovvees White [EExnitt{ra0n.5ceNSooruthth; 0.5→Soutlhe}s]sonA → lessonH → taorwearerdaschliencgtutrheeihrafill n“aHl”.leFsisnoanllyb,eifnorFeigleuarvein4g(dt)h,ebeonthvirgornomupesnt.
The fact that agents from the same group employ different paths to reach the same movement target may depend, on one hand, on the nature of the static floor field layer, but also on the fact that, at the tactical level, they may choose different intermediate steps (e.g. there are two gateways leading from the main hall to lecture hall “A”).
The paper has shown an extension of a floor field model to
encompass tactical level tasks and information. We introduced
a particular structure of agents’ environment allowing them to
perceive and act at the operational level, but also deliberate at
tactical level: environmental data structures are automatically
generated starting from an annotated CAD-like environment
description. The hybrid agent architecture, including a simple
reactive operational level able to trigger deliberation activities
of the tactical level whenever it is necessary, has also been
introduced. A sample application illustrating how this approach
(a) Green group leaves hall “A” while yellow and white ones move towards
it.
(b) All groups reached their intermediate targets.
(c) Green group leaves the environment, while yellow and white ones move (d) Yellow and white groups have almost completed their movement towards
towards hall “H”. hall “H”.
allows specifying simple behavioural scripts for relatively
complicated agent’s plans has finally been described. Future
works are aimed, on one hand, at supporting the possibility
to enrich the abstract commonsense spatial representation
for allowing tactical level reasoning about information like
estimated distances, level of crowdedness of visible areas or
passages, additional relevant information (e.g. a certain area is
a steep ramp or staircase, which would hinder the movement
of elderlies or persons on wheelchairs). Moreover, we are also
going to extend agents’ behavioural specification to allow the
coordination of group path planning and the definition of area
activities, in the vein of [