Abstract/Summary

Fifth generation (5G) wireless communication systems promise increased capacity, increased data rate, enhanced reliability, reduced latency, improved energy efficiency, improved spectrum efficiency and reduced interference, and massive multiple-input multiple-output (MIMO) has been identified as a driving technique in achieving this. In massive MIMO, the base station is equipped with hundreds of antennas to service tens of terminals in the same time-frequency resource. But there are several challenges associated with massive MIMO that prevent the achievement of these benefits, and these include channel estimation, pilot contamination and radio frequency (RF) impairments, etc. The main focus of this thesis is on the use of continuous-time statespace models to identify the dynamics of massive MIMO wireless channels, i.e. channel estimation. Two identification models, namely the continuous-time integer-order state-space and continuous-time fractional-order state-space identification models are considered when identifying the massive MIMO frequency-selective wireless channels. These models are designed based on the multiple-input multiple-output output-error state space (MOESP) algorithm, a subspace system identification algorithm that has been proven to successfully identify the dynamics of a system. Through simulations it is shown that with increase in model order, the continuous-time integer-order state-space model is able to model the massive MIMO channels with increased accuracy. The performance of the continuous-time fractionalorder state-space model is also studied for different fractional-order values, and its performance is then compared with that of the continuous-time integer-order state-space model. Having identified the dynamics of the massive MIMO system, equalizers are then designed to help combat the effects of inter-symbol interference (ISI) caused by the massive multiple-input multiple-output (MIMO) frequency-selective wireless channels. We propose the use of state-space models for channel equalization. The minimum mean square error – decision feedback equalizer (MMSE-DFE) is the equalizer of choice in addressing the ISI and is built based on the continuous-time integer-order state-space and continuous-time fractional-order state-space identified models.