Page 165 - Antennas for Base Stations in Wireless Communications

138 Chapter Four

4.5.1 Field Trial
A field trial of commercial two-branch and four-branch diversity implemen-
11
tations was conducted in a live GSM network. Two statistical measures
were recorded. The uplink signal strength, RxLev in dBm, was reported
as the maximum value taken over all two or four receive branches, i.e.,
the signals were not combined before RxLev was reported. This gives a
measure of the selection diversity gain. The active measurement range of
RxLev was bounded from below at –110 dBm, i.e., all measured values less
than –110 dBm were reported as RxLev = 0, which corresponds to –110
dBm. The uplink bit error rate, RxQual on a scale from 0 to 7, was mea-
sured after combining. A low value of RxQual corresponds to good signal
quality, whereas a high value of RxQual means poor signal quality.
The selection diversity gain was about 1 dB with four-branch diver-
sity compared to two-branch diversity, as measured by the improve-
ment in uplink RxLev for the best branch. This corresponds well with
simulated and measured results that show an improvement of 3–4 dB
in receiver sensitivity, from which an estimated power gain of 2.5–3
dB should be subtracted. Figure 4.5 shows a plot of the Cumulative
Distribution Function (CDF) of the uplink RxQual for measurement
reports for all mobile traffic in a cell with RxLev = 0, i.e., for cases when
the best branch has a received signal strength of –110 dBm or less. It
is clearly seen that four-branch diversity provides better quality than
two branches. For example, about 45% of the two-branch measurement
reports show RxQual < 4 compared to 66% of the four-branch measure-
ment reports, an improvement of almost 50%.

1
0.9
Probability (RXQUAL < abscissa) 0.7 4WRD 2WRD
0.8

06
0.5
0.4
0.3

0.2
0.1
0
0 1 2 3 4 5 6 7
RXQUAL
Figure 4.5 CDF of RxQual when RxLev equals 0 for a two-branch
(2WRD) and a four-branch receive diversity (4WRD) field trial for
all commercial traffic in a cell