中文摘要：

在通信系统中，信道估计的准确与否直接影响到接收端符号的检测性能，对MIMO系统更是如此。本文首先介绍隐含序列的信道估计方法，并分析此类方法存在的问题，包括信道估计性能受发送端符号数据的影响以及受未知直流偏移影响。然后，重点研究基于ZCZ（Zero Correlation Zone）训练序列的DDST（Data—De—pendent Superimposed Training）方案，此方案可以实现多天线系统的多径信道和直流偏移联合估计。最后，给出理论分析和仿真结果，二者均表明平衡ZCZ训练序列不但能分离直流偏移估计和信道估计，而且能使二者的估计均方差性能同时达到最优。

英文摘要：

In communication systems, data detection performance depends on channel estimation, especially for MIMO systems. In this paper, the channel estimation method based on implicit training sequences was first introduced, to- gether with its disadvantages including negative effect of data symbols and unknown DC-offset on channel estimation performance. Then, the Data-Dependent Superimposed Training （DDST） method based on the ZCZ（Zero Correlation Zone）training sequences was researched in detail, in which the problems of MIMO multipath channel and DC-offset estimation were resolved. At last, analytical and simulation results were given. It is shown that balanced ZCZ training sequences can not only make DC-offset and multiple channel estimates decoupled, but also minimizes the mean square error performance of channel and DC-offset estimation simultaneously.