Hybrid Precoding For Dynamic Subarrays In Mm Wave Massive MIMO Systems

The fifth-generation mobile communication network system adopts a large amount of unutilized bandwidth in millimeter wave(mmWave)band.MmWave greatly alleviates the tight spectrum resources,but brings serious path loss in the propagation.This problem can be solved by deploying hundreds or even thousands of antennas at the transmitter,which can achieve higher array gain,spatial freedom and accurate space division multiplexing.However,the high hardware cost,hardware complexity,and power consumption of the massive antenna arrays are also the problems to be solved.For mmWave massive multiple-input multiple-output(MIMO)systems,the hybrid analog/digital precoding is an efficient transceiver.The fully-connected architecture is widely adopted in hybrid precoding systems.In this architecture,each radio frequency(RF)chain is connected to all the antennas with phase shifters(PSs)and RF adders.Therefore,both the hardware cost and power consumption are high since the numbers of PSs and RF adders increase linearly with the increase of antennas.To address these challenges,a hybrid precoding in partial-connected architecture is used.The partial-connected architecture can greatly reduce the number of PSs and eliminate the need for the RF adders.Thus,the hardware cost and power consumption are effectively reduced.The partial-connected architecture can be divided into two categories:the fixed subarray and the dynamic subarray.In the fixed subarray architecture,each RF chain is connected to a fixed antenna subset;meanwhile,each antenna is only connected to a single RF chain.In the dynamic subarray architecture,each RF chain is connected to a subset of antennas which is adaptively partitioned according to the channel information.First,a hybrid analog/digital precoding algorithm based on the maximum mutual information of transmitted signal is proposed for the massive single user MIMO(SU-MIMO)system.Then,based on this hybrid precoding algorithm,a corresponding antenna subarray partitioning algorithm is designed.The performance advantage of the proposed algorithm is proved by simulation.Subsequently,a downlink hybrid analog/digital precoding algorithm based on the dynamic subarray architecture for mmWave massive multi-user MIMO(MU-MIMO)systems is proposed.Different from the existing schemes,the base station(BS)firstly selects the multi-user set by calculating the analog effective channel based on the best beam.Then,the antenna subarray partitioning algorithm allocates each antenna element to RF chain according to the maximum increment of the signal to the interference noise ratio(SINR).Finally,the hybrid precoding is optimized for the dynamic subarray architecture.By calculating SINRs on the analog effective channels of the selected users,the antenna subarray partitioning algorithm can greatly reduce the computation complexity and the size of search space.Moreover,it can also guarantee the user fairness since each antennaelement is allocated to acquire the maximum SINR increment of all selected users.Extensive simulation results demonstrate that both the energy efficiency and sum rate of the proposed solution obviously outperform those of the fixed subarrays.Compared with the fully-connected architecture,the proposed solution obtains higher energy efficiency with slight loss of sum rate.The research in this paper proves that dynamic subarrays can achieve a compromise between the performance and the hardware complexity for mmWave massive MIMO systems.