| Massive multiple-input multiple-output(MIMO)is one of the important technologies in the 5G communication system,which improves the spectral efficiency and link reliability by deploying hundreds of antenna elements at the base station.The current massive MIMO systems mostly use the centralized processing architecture,which results in a large amount of data aggregation.Therefore,the central processing units suffer from extremely high bus bandwidth and computational complexity.On the other hand,the performance of massive MIMO depends on the quality of channel state information(CSI).However,in practice,CSI is seldom perfect due to many issues such as inaccurate channel estimation,quantization of CSI,or feedback error.Thus,the imperfection of CSI must be considered to increase the robustness of the system.Facing the challenges in centralized architecture and imperfect CSI,we conducts a series of researches on massive MIMO distributed signal processing schemes and robust transmission methods.Firstly,we study the massive MIMO uplink communication system.Reviewing the classical linear equalization algorithm,the bit error rate(BER)performance of zero forcing,minimum mean-square error and maximal ratio combining is compared.Two distributed iterative algorithms,unidirectional-chain equalization algorithm and bidirectional-chain equalization algorithm,are proposed for the distributed daisy-chain architecture.The simulation results show that the proposed algorithm can approximate the performance of the centralized algorithm and effectively reduce the bus bandwidth.It is obvious that the convergence speed of the bidirectional-chain equalization algorithm is higher than that of the unidirectional-chain equalization algorithm.In order to further reduce the computational complexity,a recursive-based non-iterative distributed equalization algorithm is proposed.Simulation results indicate that its BER performance is the same as that of the centralized algorithm.Secondly,we explore the distributed precoding algorithm for massive MIMO downlink communication system.After reviewing several traditional precoding algorithms,according to the distributed daisy-chain architecture system model,the unidirectional-chain precoding algorithm and bidirectional-chain precoding algorithm are respectively established.In order to further reduce the computational complexity,a recursive-based non-iterative distributed precoding algorithm is proposed.The simulation results indicate that the proposed algorithms can achieve the same performance as the centralized algorithm,and effectively reduce the traffic bandwidth and computational complexity.Finally,we study the massive MIMO robust transmission method.Firstly,based on the MMSE criterion,we propose the worst-case robust channel estimation problem,which is transformed into a semidefinite program problem.Next,for the uncertainty region defined by the matrix norm,the original problem is reduced to a power allocation problem,and an iterative algorithm is proposed to solve it.Closed-form solutions are also provided for the uncertainty region defined by a specific matrix norm.Then,based on the statistical CSI,the worst case and the outage probability respectively,robust precoding schemes are provided to maximize the received signal-to-noise ratio.The simulation results show that the robust precoding scheme based on statistical CSI and the worst case guarantees the average performance and worst-case performance of the system respectively,and the robust precoding scheme based on the outage probability significantly improves the outage performance of the system. |
