The Research On Channel Modeling And Precoding For 3D Massive MIMO Systems

As one of the core contents of 5G(the Fifth Generation),massive MIMO(Multiple-Input Multiple-Output)has been widely concerned and studied at home and abroad for its high data transmission rate and reliability.In massive MIMO technologies,the 3D MIMO technology that can exploit the elevation domain information has become an important research hotspot in wireless communication for its ability to increase degrees of freedom of the channel and better meet the increasing demand of wireless data in 5G transmission.At present,most of the channel models are only for 2D space,which to a certain extent,limits the development of 3D MIMO technology.Therefore,the construction of 3D channel model is of vital importance.Moreover,the existing 3D MIMO precoding techniques are not mature and mostly concentrated in the azimuth direction only,which lead to high calculation complexity and poor system performance.Hence,new precoding schemes are urgently needed to fully exploit the potential of 3D MIMO systems.Firstly,based on the 2D SCM(Spatial Channel Model)proposed by 3GPP,the construction of 3D SCM is completed,which includes the derivation of channel response,antenna gains and channel coefficients.What's more,the system-level simulation is made on the performance gains of 3D MIMO over the traditional MIMO systems.Secondly,based on the 3D SCM constructed,this thesis studies the 3D MIMO precoding methods.A novel 3D MIMO precoding optimization algorithm is proposed,based on the Kronecker product structure of 3D channel.By utilizing the channel's characteristics,the objective function is derived.And then the precoding matrix can be calculated easily by selecting the maximum value of the modulus.The simulation results verify the advantages in both complexity and system performance.Finally,the concept of tensor analysis is introduced into 3D MIMO precoding scheme.From the new point of view,the precoding matrix is solved.By utilizing the decomposition characteristics of tensors and the properties of multidimensional translation invariant array,the global array processing problem can be transformed into a series of subarray precoding problems,which greatly reduces the computational complexity.The simulation results verify the feasibility of this scheme.In addition,the idea of high-dimentional precoding strategy based on the maximum SINR is proposed,and the derivation process is given by using the properties of tensors.