无线MIMO技术被认为是实现未来高速宽带无线Internet接入网的关键技术之一,在第三代（3G）乃至三代以后（B3G）的移动通信系统中有着广阔的应用前景。本文在他人研究工作的基础上，首先对无线MIMO系统的信道模型、信道容量、信道估计算法三个主要方面涉及到的有关理论进行了进一步的研究，并通过计算机进行了仿真分析。要实现一个MIMO系统，在接收端除了需要低复杂度高性能的检测算法外，还需要对MIMO信道进行精确的估计。由于MIMO信道的参数远多于传统单输入单输出（SISO）的信道，因此信道估计的复杂度明显增加，给接收机的实现带来了困难。 其次分析了收发两端的多天线单元之间存在相关性、同时存在直射分量（LOS）和散射分量的传播环境、己知和未知信道的状态信息等因素对MIMO信道容量的影响，给出了有关仿真结果。得到了容量的具体计算公式，并做了物理解释。 另外在研究无线信道特点的基础上，分析了MIMO信道的特征，建立了数学模型。在研究信道估计原理的基础上，分析了常用的MIMO信道估计算法，提出了最大似然一滑动平均估计和最大似然一维纳滤波估计两种改进算法，仿真分析、比较了这两种信道估计算法的性能。最后对频率选择性MIMO信道的信道估计技术和导频设计问题进行了分析讨论。

It is believed that the wireless MIMO technology will be one of the key ones that realize the high-speed broadband wireless Internet access networks in the future and has wide application prospect in the third generation（3G）or beyond third generation (B3G) mobile communications. It has mainly made further research on the related theories about three main aspects：channel models，channel capacity, channel estimation and implementation algorithms on the basis of others’ research work. on the one hand, the important results acquired by others were tested and verified, on the other hand , a number of new useful experimental results were obtained by a number of computer-aided simulations. In the first , to develop a MIMO system , exact channel estimation is necessary . MIMO channel estimation approaches are more complex than SISO channel, this leads to the complexity of the receiver . In the next the paper it was particularly analyzed the impact of the correlation between antenna arrays both sides of transmit and receive , the coexistence of line of sight(LOS) component and scatter power component in the radio propagation environments , and known and unknown channel state information on the capacity of MIMO channel . The related simulation results are also given. The formulae for computing the capacity , and the physical explanations about these formulae are also given in the thesis. In addition, the paper analyses the wireless channel , and expounds the MIMO channel characteristic , and then models the MIMO channel. It introduces the principle of channel estimation , and the paper gives two advised algorithms: ML-MA(Moving Average Filter) and ML-Wiener. Improves and simulates our schemes , and compares their performance. In the end discusses MIMO channel estimation in frequency-selective fading, and investigates the design of training sequence．