Massive 3D MIMO Hybrid Precoding

Massive MIMO(Multiple-Input Multiple-Output)technologys are able to provide considerable array gain for millimeter wave communication system,which has been attracted extensive attention.In Massive MIMO technologies,3D MIMO technology can exploit the elevation domain information in order to meet people's demand for higher data rates.However,with the continuous expansion of MIMO system scale,the huge hardware complexity and power consumption make the traditional precoding architectures impractical to implement.In recent years,in order to save the cost and achieve satisfactory performance at the sametime,the hybrid precoding architectures are widely investigated for millimeter wave communication system.Unfortunately,most researches assume that the hybrid precoding architectures are implemented with infinite or high-resolution phase shifters,which is difficult to realized in engineering.Based on this fact,this paper proposes three low-complexcity hybrid precoding algorithms based on low-resolution phase shifters.In this paper,a hybrid precoding structure based on inverters and adaptive connection network is proposed,which uses inverter to realize the function of 1-bit resolution phase shifter.The adaptive connection network helps each RF chain to connect the optimal antennas,and then a hybrid precoding algorithm based on the structure is proposed.In the analog precoding,firstly,grouping the transmit antennas and designing the analog precoding matrix are jointly conducted to match with the subchannels of the different users.Then the digital ZF(Zero Forcing)precoding matrix is designed to eliminate inter-user interference based on the equivalent channel matrix,and the simulation results verify the superiority of the algorithm.Secondly,this paper proposes a hybrid precoding for multiuser 3D MIMO systems with 1-bit resolution phase shifters,in which each user is equipped with one antenna.For the millimeter wave 3D MIMO channel can be written in the form of Kronecker product,the analog precoding is precessed in horizontal domain and vertical domain simultaneously.Then the digital MMSE(Minimum Mean Square Error)precoding is designed to eliminate inter-user interference based on the equivalent channel matrix.Through complexity analysis,the algorithm complexity and hardware complexity are reduced.Finally,in order to improve the performance of the system furtherly,a hybrid precoding for multi-user 3D MIMO systems with 2-bit resolution phase shifters is proposed,in which each user is equipped with multi-antennas.Simulation results show that the algorithm reduces the complexity and is able to achieve higher performance.