ABSTRACT:

Millimeter Wave (mmWave) wireless communications is considered an enabling technology to allow fifth generation (5G) to achieve data rates in the order of giga-bits-per-seconds for future mobile applications. The main objective of this thesis is to design a low complexity hybrid beamforming system taking into consideration the corresponding mmWave hardware constraints.

In this thesis, hybrid precoders and combiners are reviewed and newly modified implantations are proposed. Also, spatial characteristics of the channel model in multiple input multiple output (MIMO) fully-connected structure are studied to be used in system model. The design problem is presented as a sparse reconstruction and solved using Orthogonally Based Matching Pursuit (OBMP) algorithm. Maximum Likelihood (ML) fixed weight beamforming is used along with OBMP for direction of departure estimation. The hybrid precoder design is simplified by combining the design of ML beamforming, orthogonal RF codebook and the beamspace full digital precoder.

Performance of modified hybrid precoding-combining design is presented by evaluating spectral and energy efficiency. System performance is investigated by varying the following system parameters, number of data streams, radio frequency (RF) chains, RF codebook layers, signal to noise ratio (SNR) and power consumption values. Simulation results shows near optimal efficiency (above 90%) compared to unconstrained full digital optimal beamforming for higher SNR and RF chains. The results are compared with variable hardware complexity and it is shown that with the modified designs an efficient decrease of iterative process is achieved. Also, less matrix inversion calculation is needed compared to the state-of-art approach.

Moreover, the effect of using uniform planner array (UPA) antenna at the receiver is investigated. The thesis shows that beamspace acquisition with ML fixed weight beamforming is an optimum implementation to achieve similar to optimum mmWave full digital precoder.