International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
) : ; & / !
&. Walking Path -HEPU
Fi) MM) Tm o s
e «Gm BE BX
37 UNKS
MN
T7 RO&DF |--
i gx
JE Sun W Suww v Biss marsh D
=
source A: Polygon 244
pestination B: Polygon 112
walking Path
FROM bolygon 244
TO polygon 231
CROSS link 2
: TO polygon 88
= CROSS link 1
tas 1 ETS potygon 143
Links
EE | Street Blocks
Selected Links
Selected Street Blocks
Figure 5: path between point A and point B
To know if there is any connectivity between any two polvgons,
say polygon 88 and polygon 231, just simply by making a
query of “polyl=88 and poly2=231" or “polyl=231 and
poly2=88". In this case, the record of link 2 is retricved.
Similarly, with the previous selected polygons, the result is
obtained. Figure 5 shows the result of using base map features
to find a walking path between two points. The pedestrian is
suggested fo start at polygon 244, polygon 231 and then cross
the link 2 to reach polygon 88, cross the link 1 to reach polygon
143 followed by crossing the link 5 to arrive polygon 112, the
destination. However, it is important to realize that land feature
configuration in most cases is not that ‘simple’ as illustrated in
the above example, but with a lot more complexities.
In Figure 6, there is a barrier between polygon 161 and 86 and
there are two links connecting polygon 231 and polygon 203. In
this application, barriers or traffic islands are not considered as
parts of the solution because they are inaccessible for
pedestrians. Therefore, these kinds of non-passable features are
checked and removed from the polygon list (Figure 3) during
the implementation of the algorithm. Furthermore, there are the
possibilities of more than one existing link connecting two
polygons (e.g. the two links between polygons 231 and 203).
With all these in mind, computation of the optimal walking path
follows the algorithm as summarized below:
Step I: Construct a line between the source and destination
points
Step 2: Get all intersection points on the line
Step 3: Overlay the start and end point of each line segment
with the street and construct a list with polygon ID
Check inadequate polygons: to see if any physical
dividers and traffic islands included in the above list. If
yes, remove them.
Step 5: Check 1f there are more than one links connecting two
polygons, if yes, compare the distance of each line
with the constructed line. Choose the closest onc.
Step 4:
60
Step 6: Query the link with provided polygon ID in the list to
obtain the connectivity relationships between polygons
Step 7: Display the result
ource A: Polygon 161
& Walking Path - HKPU
FROM poly
rt = yon 161
rs" \CRoss link 12
#2 UNKs {TO polygon 36
{TO BeTygun 88
F^ ROADF. CROSS link 2
Lj TO polygon 231
r""jcRoss link 19
^| jTO polygon 203
E
Moser. | TT TE
mee emm Links
Street Blocks
Barriers
Selected Links
Selected Street Blocks
Figure 6: Computation with the existence of barrier and more
than one link between polygons
4. DISCUSSION AND CONCLUSION
The study indicates that there is a potential to compute walking
path without the use of road centerline. Although the work
presented here lacks evaluation or examination, the results
should be considered as encouraging. It is understandable that
the accuracy required for walking path is relatively lower, when
compared with accuracv required for driving path. But one of
the important components of the future work should be included
the evaluation. part of the proposed algorithm to test the
reliability of the model. Besides, the algorithm should be
further developed and investigated with more walking path
features such as building blocks, stairs, subways, bridges. As
can be seen in Figure 5, pedestrians are allowed to move from
one polygon to another polygon (from polygon 244 to polygon
23] in this case) without crossing any links in between.
Therefore, some sophisticated assumptions or considerations
are needed to take into account in the algorithm. For this
particular case, the algorithm should consider the road width
and the existence of barriers etc. in order to determine if it is
suitable for pedestrians to cross. As a result, further analytical
and theoretical works and studies are required to explore the
feasibility of the proposed algorithm.
This paper introduces a new idea of using an independent of
any arc-node data model to find path for pedestrians. In fact, the
objective of transportation system is to improve individual
accessibility. However, most GIS-T focus on vehicle navigation,
individual accessibility is less common. This research could be
important and useful for those people who interested in this area.
sour
Destination B: Polygon 203
Interna
Chan 3
Traffic-
(2000) :
Dueker
Informa
Sharing
Goodch
and Ch:
Han Ge
Finding
Prohibit
Dimens
Horowit
Transpo
Plannin;
Kirkby :
Shortest
Data H
12-16, 1
Longley
Rhind I
Volume
NY: Wi
Miller I
Informai
Applicat
Sutton
Represer
Transpoi
Vuren V
in Path
Planning
Zhan F1
Algoritht
Transpoi
5.1 Ack
The worl
an asso
Polytechi
