International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B4, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
and organization for evacuation simulation would be more
sounding (Yuan and Schneider 2010). This research trend is
also very helpful about the idea of introducing consistent basic
unit to represent the whole building area.
All these contributions by researchers have enlightened us to
develop an improved method to prepare navigation data for
evacuation simulation. In our opinion, the combination of a 3D
normalized spatial relationship model and a parallel
computational framework can answer the question of providing
accessible data for emergency evacuation simulation in large
buildings. Therefore, this paper will concentrate on the
introduction of these algorithms to the working procedure in
emergency communication data extraction.
To finish this task, this paper contains four parts including the
introduction part. In the second part, we come to analyse the
key features of inner-building space and formulate a normalized
spatial relationship model to extract communication
information; in the third part, we use a test area to prove our
model is workable; in the final part, we discuss the advantage
and disadvantage of our solution.
3. EVACUATION COMMUNICATION DATA
EXTRACTION FROM LARGE BUILDING
As mentioned in the first part, the striving needs of emergency
evacuation require large volume of simulation data. Evacuation
simulation needs two basic types of data: people behaviour data
and basic environment data(Vanclooster, De Maeyer et al.
2010). The people behaviour data could be generated from
social research or psychological research, and this is not the
focus of our research. Nevertheless, the related factors of the
building environment data draw our attention.
For the purpose of evacuation simulations is moving a group of
people from the dangerous area to the safe area, the spatial
distribution feature of evacuation building is the most important
element among its environment factors. Furthermore, the inner-
space feature of buildings is also crucial, since the artificial
structure mainly places its function area in the inner part of
building. Therefore, we should consider both the spatial
distribution and the inner-space feature of large buildings. Thus
the next part of the paper is about the accessible position
description of the inner space feature for large buildings and a
designed extraction solution for the communication data of
inner-building space.
3.1 Inner-Space Feature of Large Building
Due to the fact that buildings are artificial objects, the inner-
space of large buildings is defined before the structure is
physically constructed. This space is classified into many
functional parts, such as electrical supply part, heating-pipe part
or other parts, and all these parts must be accessible for
people(Martín, García et al. 2011). If the functional part is not
accessible, then the part cannot be maintained by workers.
Nevertheless, the ‘accessible’ meaning for skilled building
workers is quite different from normal pedestrians. This is
caused by the different mobility ability of these two groups of
people, for example a well-trained electrical worker could use
climbing tools like ropes to climb up a tall wall easily, and a
normal person cannot do this. For the emergency evacuation
simulation the mobility ability of people is restricted into the
minimum level, in other words the accessible area means the
area easily moved in and out on foot by normal people.
To restrict the accessible area for normal people, the limitation
for moving from one place to another place must be defined.
Otherwise, some un-acceptable moving condition would appear,
such as through one step people could jump directly from one
floor to the neighbour floors. The mobility limitation for people
in inner-building space has three types.
The first type of limitation is the step length limitation. Every
moving step from one position to the neighbour position must
not be over the average step length of normal people, otherwise
the moving request is rejected. The second type of limitation is
the step height limitation. Like the step length limitation, the
height difference between two neighbouring positions should
not be over normal people level. The third type of moving
restriction is the slope limitation. This type of limitation is
introduced to eliminate the case that two neighbouring position
are on the edge of a large slope. All these three types of
limitation will be carried out in the communication data
extraction for evacuation simulation.
The three types of movement limitation lead to the construction
of a special relationship model for the inner-space of large
buildings. The feature of spatial distribution for inner-building
space requires us to introduce several popular types of spatial
objects and their important relationships. In the next section, we
will focus on the explanation of these introduced spatial objects
and the related spatial relationships model.
3.1.1 Typical Spatial Object
From many aspects, the considering scope of candidates for
basic spatial object are comparative narrow. In detail, the
artificial feature of buildings determines that the inner-building
objects of the structure are in regular shape. This determines the
introduction of regular shape object (figure 2), and leads to the
choice of cuboid as the basic spatial object. We have evaluated
many types of polyhedron. At last we choose the cuboid,
because other types of polyhedrons have many disadvantages in
the evacuation simulations.
Accessible cell
Accessible cell representation
Physical representation
Figure 2. Diagram showing transformation from existing
building structure data to pedestrian accessible cell form
For example, if we use an octahedron with six rectangle faces
and two honeycomb faces as the basic unit, then each unit
normally has six neighbouring units, and this makes an obstacle
to simulate people choosing path in evacuation. This is because
people have been accustomed to eight direction choices for
moving to the neighbouring units. If they have six direction
choices, they will be confused.
Intern:
After the seti
an index to r
using the cu
for a specific
Besides the |
we should al
the existing
the evacuatic
to several c
compose the
information
evacuation c«
31.2 Nor!
Space
The triangles
spatial relatic
used to form
function is tr
model. Thes
space with o
building wo
simulation, fi
next step. Ne
out for being
we should al
inner-space
evacuation rc
defined by ar
aspects.
The two aspe
the possible i
this position i
question cou
triangles of tl
inner-space ai
properly dete:
analysing the
triangles beloi
The second q
accessible pos
section. This
and slope r
considering p
positions.
413 Man
The accessibl
Solution. And
in the forme
important det
process. First,
unit in our re
length and he
Square bottom
Whole area int
the accessible
Furthermore, t
Tegularize eve
simulated peor
