Antenna diversity gain in the wireless local area network standards Hiperlan/2 and IEEE 802.11a

Abstract: The purpose of this thesis is to investigate the benefits present when
antenna diversity is used in the two Wireless Local Area Network (WLAN)
standards Hiperlan/2 and IEEE 802.11a. The standards represent the next
generation of high-speed extensions to a wired network by providing bit
rates up to 54 Mbps under optimal conditions. Typical application areas may
be airports, offices and exhibition halls.

Data transmission is done in the license free 5 GHz band, and the
transmission technique used is Orthogonal Frequency Division Multiplexing
(OFDM). This Multi Carrier Modulation technique was chosen due to its
satisfactory performance in highly dispersive environments.

The diversity technique investigated is called Antenna Diversity, which
means that the receiver is equipped with multiple antennas. If the antennas
are spaced apart at a properly chosen distance (or if different polarisation
directions are used), almost uncorrelated versions of the signal can be
obtained. These uncorrelated “branches” are combined in order to get a
stronger signal. The combining methods used are versions of Switch
Diversity, Selection Diversity and Maximum Ratio Combining (MRC). The
diversity gain is evaluated for various data rates that are common in the
two standards. The number of receiver antennas is varied between one and
eight.

The physical layers defined in Hiperlan/2 and IEEE 802.11a are almost
identical. Only minor differences occur, resulting in variations in the data
rates provided. A physical layer software simulator was implemented to
facilitate data rate simulations common to the two standards. The simulator
takes random data bits (which represent information sent from the MAC
layer), applies the physical layer functions and performs a base band
transmission over a time dispersive Rayleigh fading channel. A receiver,
equipped with multiple antennas, was also implemented. The output of the
receiver is an estimate of the transmitted bits after diversity algorithms
have been used.

The results of the simulations demonstrate that the coverage can be
significantly improved with the methods tested. Due to its low
implementation costs, Switch Diversity is the technique most likely used in
a real system. When this method is used with eight antennas, a gain of up to
2 dB can be achieved (this corresponds to an improvement of 20% of the
system range). This seems to be a small improvement, but the fact that the
switch algorithm only chooses between antennas, not sub carriers, has to be
considered. For Selection Diversity, the gain lies between 7-9 dB when eight
antennas are used (which means that the system range is improved with about
50%). The detection is done on sub carrier level rather than on antenna
level, and hence frequency selective dips can be avoided. This is also the
case in MRC, but this method performs even better. The measured gain for
eight antennas is 12-16 dB depending on the data rate chosen. This means
that the system range is improved with up to 100%.