Study On Precoding And Interference Alignment In Multiuser MIMO System

Multiple-input multiple-output (MIMO) technology can improve the capacity ofwireless communication systems effectively without any addition of bandwidth andtransmission power. So the technology becomes a hot topic in the research of currentwireless communication. With the change of the need in the wireless communication,MIMO has been applied into multi-user communication system. In the multi-userMIMO communication system, multiuser interference deteriorates the communicationquality and limits the system capacity severely. As an effective method to enhancewireless communication performance, the precoding is applied to the suppression ofmulti-user interference, and the pretreatment applied to the transmission data at thetransmitter is used to eliminate the multi-user interference. Recently, a novel methodnamed as interference alignment, has been proposed. Through using the idea ofoverlapping interference space, the interference alignment achieves the spatial isolationbetween the desired signal and the interference signal, so as to improve the channelcapacity and provide a new idea for suppression of multiuser interference.Starting with the channel model of multi-user MIMO system, the thesis studied onthe multi-user interference cancellation techniques. Then the thesis focused on theprecoding under the multi-user MIMO broadcast channel and distributed interferencealignment (DIA) under multi-user MIMO interference channel. The specific work is asfollows:①The thesis did a systematic study on the precoding under the multi-user MIMOsystem respectively from the linear precoding and non-linear precoding. Thecomparative analyses show that the linear precoding has less computational complexityand is more suitable for the actual wireless communication system, although thecapacity in the system with linear precoding is less than that with non-linear precoding.②Analysis demonstrated that the traditional block diagonalization algorithm hashigh computational complexity, because of applying the singular value decompositionto multi-user channel for solving the null space without interference. To reduce thecomputational complexity and ensure the performance, an improved algorithm wasproposed. Channel block diagonalization was fast finished by combining matrixpseudo-inverse and orthogonal decomposition instead of the complex singular valuedecomposition, and LDLHdecomposition was applied to optimize the equivalent channel. The simulation results indicate that the proposed algorithm ensures the systemBER and reduces the computational complexity. Compared to traditional blockdiagonalization algorithm, the proposed algorithm obtains almost same system capacityperformance under the same antenna configuration.③Research on the optimized constraints criteria of DIA and do someperformance analysis. Based on the condition of maximized sum rate, the rankconstraint interference alignment algorithm was accomplished, which combines fullrank of the useful signal subspace and minimum rank of interference signal subspace.The thesis analyzed the impact of different antenna configurations on DIA performanceincluding minimum remaining interference algorithm and maximum SINR algorithm.Simulation results show that the rank constraint method obtains better system capacityperformance, when the number of antennas and users meet the condition of K <min (M,N).