Abstract
With densification and heterogeneity being dominant themes for current third, fourth and future fifth generation networks, small cell base stations (SBSs) are expected to see widespread use and deployment. Using multiple antennas, such base stations can improve link reliability and increase spectral efficiency through diversity, beamforming and multiple input-multiple output (MIMO) spatial multiplexing. However, the implementation of such schemes comes at an increased complexity leading to an additional cost and power expenditure. This conflicts with simple design and growing energy efficiency (EE) demands now being placed on wireless networks. This aim of this thesis is therefore focused on the investigation and analysis of low complexity, multiple antenna techniques and approaches to enable energy efficient transmission and reception schemes applicable to small cell base stations. Particular attention is given to indoor installations. Consideration is given to existing low complexity transmit diversity schemes, applicable to third and and fourth generation wireless base station access technologies - Wideband Code Division Multiple Access (W-CDMA) and Long Term Evolution (LTE). By simulation and using appropriate channel models, anticipated transmit power reductions, using multiple antennas to maintain an equivalent quality of service (QoS) that would otherwise require a single antenna operating at a higher transmit power, are provided. Power consumption models (PCMs) are then used to show the total power consumed to realise such schemes. New expressions for overall power consumption are provided that account for the additional antennas and signal processing required to reduce the transmit power. A predicted energy saving of 15% for an LTE SBS employing 2 Tx, 1 Rx transmit diversity is shown. With increased indoor densification and the use of beamforming techniques seen as a predominant means to reduce interference, the work is extended using the expressions formulated, to consider a network of densely deployed SBSs employing downlink (DL) beamforming. Using a multiple input single output (MISO) configuration, appropriate path losses and interference based on the densification level are used to simulate realistic channel conditions in a simulated dense deployment. Based on this, DL weightings are obtained and used to optimise the transmission and interference properties and to minimise the expended energy in such networks. Exploitation of closely spaced antennas and channel correlation is considered to minimise the amount of channel state information (CSI) needed at the transmit beamformer. Using previously derived total power expressions a 26% increase in SBS EE over an omni-directional single input-single output (SISO) configuration is shown. Importantly it is shown that under certain circumstances there is no need for transmit CSI on a per antenna basis and a reduction in CSI of 50% and 75% is shown to be viable in a 2 and 4 element beamformer respectively. With the move to 5G, using large scale antenna systems and high bandwidth spectrum, SBS can theoretically achieve the anticipated future data bandwidth demand of 10,000 fold in the next 20 years. Emphasis is given to exploit small cell distances to simplify SBS design particularly considering indoor installations such as user deployed SBSs and dense indoor networks. Theoretical results based on link budget analysis are compared with the previously developed indoor dense SBS model using recently researched millimetre wave (mmWave) channel propagation conditions concentrating on diverse bands (28GHz and 72GHz) of the spectrum. Investigation of the performance of low complexity approaches using a minimal number of antennas at the base station and the user equipment (UE) is provided. Using the appropriate power consumption models and state-of-the-art sub-component power usage the total power consumption and energy efficiency of such systems are determined. Significantly, the work shows that despite LSAS and massive MIMO schemes being dominant themes in current research, the use of single omni-directional SBSs with suitable forward error correction (FEC) provides a low complexity and energy efficient alternative for the indoor SBS. With mmWave being typified by eroded link margin and non-line of sight (NLOS) communication, and as an alternative to using large scale antenna systems (LSAS) for small cells, the use of combined spatial and temporal processing with a minimum number of antenna elements is given consideration. Based around a spread spectrum physical layer, using transmit and receive beamforming with Rake combining. A detailed analysis of circuit power due to signal processing is provided. The new air interface is compared to a 4G OFDM based system and is shown to be more power efficient and significantly the SE/EE trade-off is improved by an order of magnitude i.e. a 10-fold improvement is seen.