Research On High-efficiency Mm-Wave Transceiver Design Optimization

With the development of the communication systems,mobile communication's requirement of the transmission rate and bandwidth keeps increasing.However,due to the bandwidth shortage in the low and medium frequency spectrum,new spectrum must be investigated in order to realize reliable,high-speed and wide-bandwidth transmission.Hence,millimeter-wave(mm Wave)becomes a key technology in 5G.Hybrid analog-digital(A/D)transceivers designed for millimeter wave(mm Wave)systems have received substantial research attention due to the benefit of their lower cost and modest energy consumption compared to their fully-digital counterparts.However,in future mm Wave systems,there still exist some problems,great overhead of channel estimation and computation of high-dimensional precoding matrices,severe pathloss experienced by electromagnetic waves in the mm Wave band and heavy inter-user interference caused by the increasing number of users.In order to overcome the first problem,we employ the two-timescale scheme where for hybrid A/D transceivers,we update the high-dimensional analog beamforming matrices(long-term variables)over a long timescale based on the channel statistics and update the low-dimensional digital processing matrices(short-term variables)over a short timescale based on the low-dimensional effective instantaneous channel matrices.To implement this scheme,stochastic optimization is used to update the long-term variables while conventional methods or deterministic optimization is used to update the short-term variables.For second problem,we use relay to obtain an adequate coverage area.And for third problem,we utilize the nonlinear Tomlinson-Harashima precoding(THP)technique to further reduce the inter-user interference.Specifically,we consider the following two scenarios.First,we conceive a full-duplex(FD)mm Wave MIMO multiple-relay systems with the objective of improving the coverage area and the sum rate attained.With the aid of the cut set bound,we are able to formulate the optimization problem to two different-timescale problems.For the long-timescale problem,we develop an efficient analog beamforming algorithm based on the stochastic optimization theory while for the short-timescale problem,we propose a digital beamforming algorithm based on the theory of penalty dual decomposition.This two-timescale hybrid beamforming design can outperform the conventional beamformers both in terms of requiring a lower CSI-estimation overhead and a higher sum rate in the face of outdated CSIs.Second,we conceive a THP based nonlinear joint design for the downlink of multiuser MIMO mm Wave system.In this scenario,the users are close to each other and thus the inter-user interference is severe.The THP structure subtracts the interference at the base station before transmitting the signal to the users.In order to enhance the reliability of the system,we aim to minimize the mean square error(MSE)of the system under channel uncertainties subject both to realistic transmit power constraint and to the unit modulus constraint imposed on the elements of the analog beamforming matrices.In this paper,we propose a single-timescale joint design algorithm and a two-timescale joint design algorithm.The two-timescale algorithm can be seen as an extension of the single-timescale algorithm,since their updating steps of the digital processing matrices are the same.Besides,we determine the cancellation order for the THP structure based on the lower bound of the MSE.The simulation results illustrate that our proposed algorithm can outperform its linear counterpart.