Page 183 - Antennas for Base Stations in Wireless Communications

156 Chapter Four

4.9.1 Case Study
A case study of a real WCDMA network deployment was undertaken to
determine the capacity and coverage performance of an urban cellular
network when a limited set of sites are upgraded from three-sector to
21
six-sector sites. This is compared to a homogenous and regular hex-
agonal deployment, where a full network migration is performed. The
conclusion is that the performance gain shown in homogenous deploy-
ments is maintained in a real network.
The antennas used are characterized by a half-power beamwidth
scaled proportionally to the number of sectors per site, i.e., three- and
six-sector sites are configured with 65° and 33° half-power azimuth
beamwidth antennas, respectively. The vertical size of the antennas
is kept constant for all antennas, i.e., the directivity increases by 3 dB
for each halving of the azimuth half-power beamwidth. The resulting
capacities are equivalent to a capacity increase of 86% when doubling
the cells from three to six. Thus, the expected capacity increase per site
is about 1.8 times for doubling the number of cells per site.
Irregular deployment and nonuniform traffic distributions that char-
acterize a real network are suspected to have an impact on network
performance. A higher order sectorization case study addressing a real
network containing over a hundred sites in a major city serves as an
example of a possible and viable way for system improvement for capac-
ity increase. Three different radio network configurations are evaluated.
The first studied scenario (baseline) consists of the original network
deployment where all sites, even the higher order sectorization evalu-
ation sites, are equipped with three-sector sites. Each cell has a maxi-
mum downlink power of 20 W. The other two scenarios, H20 and H10,
are network configurations based upon five higher order sectorization
evaluation sites that are migrated to six-sector sites. In the H20 sce-
nario, the higher order sectorization sites are equipped with six cells
with 20 W downlink power each, whereas in the H10 sites the downlink
power is 10 W per cell. The total power per site is then 60 W in the latter
case, i.e., equal to the total power of the baseline three-sector site.
Figure 4.18 shows the speech capacity in the different network sce-
narios and for the three different cell groups. Since areas covered by
different cell groups are not the same in the different scenarios, it is
interesting to evaluate the capacity per site. The capacity increase rela-
tive to the reference scenario for the different cell groups are summa-
rized as follows. All active cells: +28% (H20) and +23% (H10); higher
order sectorization (H.O.S.) cells: +77% (H20) and +65% (H10); and
other cells: +4% (H20) and +5% (H10). The expected relative increase
of 1.8 times for a doubling of the number of sectors is thus confirmed
in the realistic network environment study. Higher order sectorization