Precoding Algorithm In Multi-User MIMO System

With the rapid development of information society, the new mobile communication system has a higher requirement for information transmission rate and quality of service. Multi-Input Multi-Output (MIMO) system could make full use of space resource, and increase the channel capacity and spectral efficiency without increasing the transmission power and signal bandwidth, which therefore has been widely concerned. Nevertheless, single-user MIMO system is not proper for the requirement of modern communications. Thus, resource allocation and transmission plan concerning multi-user MIMO systems has become a hot topic. A major problem in multi-user MIMO systems is how to remove the interferences between users. We can use precoding through utilizing the channel state information in multi-user MIMO systems, such as to decrease or eliminate the interferences between users.Based on MIMO channel models, the principle and performance advantage of MIMO systems is studied in the thesis and we focus on the capacity of channels of MIMO systems. For MIMO system using space time code we proposed a new channel estimation algorithm. When the correlations of transmission sequences satisfy specific conditions, through simple signal processing, a self-correlation matrix with specific structure could be obtained at the receiver. Then the matrix could be used in channel estimation through simple computation without matrix decomposition, which has low computation complexity. And then, the model is further extended to multi-user MIMO systems and we focus on the block diagonalization(BD) precoding algorithm and the precoding algorithm based on signal-to-interference-and-noise ratio. Also, the capacity of the multi-user MIMO systems, which uses the BD algorithm, is given. In the end, we proposed several improved precoding algorithm, include BD based on equity of different user which improve the symbol-error-rate of the weak user, BD with antennas selection which achieves higher capacity performance in specific multi-user MIMO systems, precoding algorithm for channels with spatial correlation which improves the symbol-error-rate of the whole system. In addition, we compare the computation complexity and bit error ratio between the system using precoding and the system which directly sent the original parallel signal without precoding. Finally, we show the correctness of all the proposed precoding algorithms through simulations.