etwork
d. This
t data.
nly the
ntenna
s. This
| major
d. Das
Nerden
nd den
igitalen
it mehr
ing von
n. Die
och die
rzeugt.
nd wird
om the
ce and
of the
variies
based
will be
mation
nt data.
id. The
he the
ase for
rocess
elevant
station
j
2. MICROCELL PLANNING
Due to the growth of the subscribers of the D2 network, the
resulting high traffic load and the use of low performance
mobile phones the network has to be adjusted continuously.
One possibility of expanding the capacity, that is to say
more subscribers are able to set up calls at the same time,
is the introduction of smaller cells with their own BTS. The
capacity (Erlang/km?) is proportional to the number of cells
within a specified area.
The growing number of cells (and BTS) leads to a problem
however. Each BTS uses a specified frequency for the
connection to and from the mobile. The number of possible
frequencies is limited however, therefore each frequency
has to be used as often as possible. On the other hand the
same frequencies sent from different BTS must not disturb
each other to avoid interference problems and dropped
calls. Therefore the knowledge of the topography is very
important.
Especially in the major cities the topographic data (100 m
by 100 m pixelsize) which has been used up to now is no
longer sufficient. In order to plan optimized coverage and
capacity precise digital data of the terrain height including
the location and the height of the buildings (city structure
data) are currently needed. The knowledge of city
structures is a presumption for the planning of further
antenna locations for the microcells. The height above
ground of these antennas will not project neighboring
buildings and the coverage performance of a microcell will
be within a radius of one kilometer.
For modelling the urban areas sophisticated fieldstrength
propagation models have been developed. Mannesmann
Mobilfunk is using the Urban-Micro-Model (Cichon et al.,
1993), which describes multipath wave propagation by
three components (vertical plane, transversal plane model
and multipath scattering model caused by reflections). The
computer-aided planning software bases on the use of
precise raster data.
3. DIGITAL CITY STRUCTURE
To evaluate the requirements of the building data different
procedures for providing data were initially tested and then
subsequently the accuracy of the resulting data sets. A
variety of different data sets were created from one testing
Site and others were aquired. These sets were compared
with a regard to their usability with the prediction model
currently used (Feistel, Baier 1995). This test resulted in
the following data requirements:
- combined dataset of both terrain and building heights
- pixelsize of 5x5 m?
* horizontal accuracy of about half pizelsize, vertical
accuracy *2m
- generation of all buildings with a larger size than 50 m?
and above ground height of more than 3 m
- generation of the buildings as boxes with flat roofs
(highest representative point)
- perpendicular rise of heights between ground and
building (height discontinuity)
- division of building blocks into several parts, if the
height differs by more than 3 m
183
Fig.1: Digital model of a city
Moreover the data has to be generated or transformed to
the same geocoded space as other data used at
Mannesmann, that means Gauss-Kruger coordinates with
Bessel ellipsoid and Potsdam datum, and they have to be
transferable to the common used format within the
company.
After establishing the prerequisites the first orders for
obtaining the 3-D building data for several German cities
were placed.
Fig. 2: Raster dataset of Berlin
4. HARD- AND SOFTWARE ENVIRONMENT
The planning of a cellular radio network is a very complex
business that cannot be done without the extensive use of
hard- and software. Within the headquarters planning
department a huge UNIX-based network has been
installed. Numerous SUN workstations are connected via
FDDI or Ethernet. Different backup media like Jukeboxes,
DAT and Exabyte are available. 2 Sparc Center 1000 for
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B2. Vienna 1996