Robust Hybrid Receive Beamforming For MmWave Massive MIMO Systems

The modern 5G cellular communication networks and beyond require large-scale antenna systems to improve gain in order to face the weaknesses in the suggested millimeter-wave(mmWave)signals and to meet the higher capacity demand.These systems implement digital beamforming,which requires a radio frequency(RF)chain connected to each antenna.As such,the increase in the number of antennas and RF chains results in high energy requirements and computational complexity.Consequently,hybrid analog and digital(HAD)beamforming has recently been receiving considerable and welldeserved attention for the practical implementation of large-scale antenna systems.Compared to fully-digital beamforming,partially-connected HAD beamforming can significantly reduce the hardware cost,complexity,and power consumption.In this work,new hybrid beamforming design approaches have been proposed to achieve the goals of interference suppression,robustness against mismatches,and reducing computational complexity using robust adaptive beamforming techniques and deep-learning.The main contributions of this dissertation are given as follows:Firstly,to reduce the peak sidelobe level(SLL)of the beam pattern and steer the nulls in the desired directions while reducing the power consumption,hardware cost,and complexity,a novel partial-connected hybrid analog-digital receive beamformer based on an atom search optimization(ASO)algorithm is proposed.ASO is a recently invented metaheuristic optimization algorithm imitating the physical movements of atoms as described in molecular dynamics simulation,with the advantage of having few parameters to tune.The desired optimization weights are accomplished by controlling the amplitudesonly of the digital beamformer’s vector along with the phase shifts of the analog beamformer’s vectors.Further,to demonstrate the effectiveness of the proposed beamformer,a comparison with other state-of-the-art metaheuristic algorithms through several scenarios has been conducted,in terms of peak SLL reduction,null depth,and convergence rates.Simulation results show that the proposed algorithm has always outperformed other types of beamformers.Secondly,to mitigate interference and increase robustness against direction-ofarrival(DOA)mismatch,a novel robust HAD beamforming receiver with a partiallyconnected structure is proposed.It is based on methods of an improved bat algorithm(I-BA)and robust adaptive beamformers(RABs)in the digital domain.Since most of the RAB methods are sensitive to the DOA mismatch and depend on the complex weights which lead to an expensive receiver,the I-BA is proposed.In the analog part,analog phase alignment by linear searching(APALS)with sufficiently fine grid points is employed to optimize the analog beamforming matrix.The performance of the proposed I-BA is assessed using MATLAB simulation and compared with the BA and particle swarm optimization(PSO)algorithms.It shows better performance in terms of convergence speed,stability,and convergence rate.Besides,the proposed hybrid I-BAAPALS approach showed better performance compared to other proposed robust hybrid techniques,i.e.,diagonal loading(DL)APALS(DL-APALS)and DL spatial matched filter(SMF)APALS(DL-SMF-APALS).Finally,a framework for integrating deep learning into robust HAD beamforming in uplink mmWave MU-MIMO systems has been developed.To ensure robustness,a diagonal loading method based on a spatial matched filter(SMF)is used to scale an identity matrix and add it to the covariance matrix.Then,using a metaheuristic PSO algorithm,phase-only digital BF weights for null steering are constructed to ensure lower complexity and a simple design of the deep neural network(DNN).As a function fitting problem,DNN is used to solve the null steering problem.We will avoid the computational complexity of diagonal loading and metaheuristic techniques by using deep learning.According to simulation results,the proposed DNN model performs similarly to PSO.