Abstract: In multi-user MIMO (Multiple Input Multiple Output) system, the users transmit the Channel State Information (CSI) to the base station through feedback. Then the system could achieve the multiplexing gain through precoding or user scheduling with the received CSI at the base station. Concerning that the traditional way of feedback takes up too many uplink resources, a limited feedback MIMO precoding algorithm was proposed based on Grassmannian codebook. Only a few bits were required to be the feedback to the base station because the new approach quantized the channel matrix and transmitted the sequence number of the quantized channel matrix according to the codebook, which reduced the feedback load considerably. The simulation results of the Bit Error Rate (BER) and throughput rate show that the new approach maintains the performance of the system. Furthermore, the system throughput is illustrated on the condition that the estimated error and quantized error of the channel matrix exist. The analytic expression with the throughput, variance of error and transmitted power will benefit the channel estimation process.

Key words: Multiple Input Multiple Output (MIMO), limited feedback, Channel State Information (CSI), channel matrix error, throughput

摘要： 多用户多输入多输出(MIMO)系统中，用户可以通过反馈的方式把信道状态信息(CSI)传递给基站，基站利用该CSI进行预编码或者用户调度，可以有效地提取复用增益。但是传统的反馈方式，反馈量太大，占用过多上行资源。针对收发都是多天线条件的多用户MIMO系统，提出一种基于Grassmannian码本的有限反馈预编码方法，通过对用户信道矩阵进行量化，每个用户仅仅需要反馈一个序号而不是信道矩阵，从而大大减少了反馈量。通过对系统误码率(BER)和吞吐率的仿真，表明了该方法能保证系统性能。同时，由于信道估计存在误差，量化过程存在误差，所以，分析了信道矩阵误差对于新方法性能的影响，得到了系统吞吐率同信道矩阵误差的方差，以及基站发射功率三者的关系，并通过仿真进行验证，从而为评估信道估计和量化过程的优劣提供了理论依据。

关键词: 多输入多输出, 有限反馈, 信道状态信息, 信道矩阵误差, 吞吐率