Performance of MIMO Schemes in Radio-over-fibre-based Distributed Antenna System

The research presented in this thesis has focused on the use of MIMO wireless communications in a RoF-based DAS to improve wireless coverage and capacity performance in an indoor environment. The aim is to analyse the practical issues that cause throughput to deteriorate when commercial MIMO APs are used in a RoF-DAS, and also to verify that improved performance - lower error rates and higher capacities - can be achieved by a large physical separation between the RAUs when specific multi-antenna scheme algorithms are used.

The performance of an IEEE 802.11n MIMO-supported AP and IEEE 802.11g spatial-diversity-supported AP are investigated in a RoF-DAS when different fibre lengths are connecting the AP in the central unit to the RAUs, and when the RAUs are widely separated. The analysis indicates that for MIMO, the throughput drops rapidly due to severe ISI caused by differential delay when the fibre-length difference exceeds a certain distance, while for spatial diversity high throughputs can be maintained even at large fibre-length difference. Further, it was observed that largely separated RAU may lead to power imbalances and the throughput drops in specific wireless user's positions when the received power imbalance was above 12-15dB for MIMO-supported AP, while for spatial-diversity-supported AP the power imbalance does not affect the throughput.

The majority of previous works on RoF-DAS for improving MIMO systems were based on commercial products and the specific algorithms used within these products are unknown. An investigation was carried out at microwave frequency with SIMO algorithms in RoF-DAS uplink, MISO and MIMO algorithms in RoF-DAS downlink, and compared with the performance of a SISO system. This investigation was later extended to millimetre-wave frequency where larger bands of frequency are available enabling the possibility of wider bandwidth and higher data rates. The result shows significantly reduced error rate and modestly increased capacity for a wireless 1x2 SIMO uplink using MRC algorithm and 2x1 MISO downlink using Alamouti STBC algorithm. Also, error rate was reduced for a wireless 2x2 MIMO downlink using the zero-forcing algorithm while, most importantly, greatly increased capacity was achieved through the spatial multiplexing gain.