Research On Robust Interference Alignment Algorithms With Uncertain CSI In MIMO Interference Channels

In a multi-user wireless communication system,interference seriously exist in the desired signal at the receiving end.Interference alignment,as an effective interference management technique,enables interference from multiple transmitters to be compressed together and eliminated at the receiving end,which can dramaticlly improve the system capacity.However,most of the interference alignment schemes at present require the transmitters to have local or even global ideal channel state information,but this cannot be satisfied for a variety of reasons in the real communication system,and thus the interference cannot be completely aligned,the system performance also cannot achieve as we expected.Therefore,this paper will design a robust interference alignment algorithm in MIMO interference channels to overcome the problem of performance degradation due to error and limited feedback CSI.The main work is as follows:A robust interference alignment algorithm based on maximizing Rayleigh entropy is proposed for the problem of high sensitivity to the interference due to the error CSI in the K-user MIMO interference channel.Different from traditional MAX-SINR algorithm which independently design interference arrangement matrix without considering the correlation among data flows,interference suppression matrixes at the receiver are obtained through maximizing the signal Rayleigh entropy and take its unitary matrix form for the optimal interference suppression matrix in order to improve the user rate in this thesis.Moreover,in order to achieve the optimal power allocation among data streams,the water power algorithm is used to allocate more power to better sub-channels.In addition,the convergence,freedom and spectral efficiency of this algorithm are analyzed theoretically.The simulation results show that the proposed algorithm can effectively improve the performance of the system in different case of CSI.A novel limited feedback interference alignment scheme is proposed in interfering MIMO-MAC.In order to overcome the loss of spectral efficiency in low SNR,the precoder matrix is designed through maximizing the target signal to the interference leakage and noise ratio,and interference suppression matrix is designed by maximizing the signal to interference plus noise ratio for each data stream.In addition,bit allocation schemes of traditional limited feedback are based on the mean of interference leakage,which are not optimal and make the total number of the effective bits less than the bits provided by system.Thus,a novel bit allocation scheme is provided in this thesis which can improve bit utilize and decrease the impact of quantization error.Theoretical analysis and simulation experiments both show that this algorithm can effectively reduce the impact of quantization error on system performance and significantly improve bit utilization of the system,so that the performance of limited feedback CSI can be improved.