MIMO-OFDM通信系统是未来通信的必然要求，其高容量、高质量和高移动性对接收端准确的恢复出原始数据提出了挑战性的要求,因此，其信道估计对该系统有着至关重要的作用，而复杂多变的移动通信信道则使信道估计面临着较大的挑战，本文针对以上问题，重点研究了MIMO-OFDM系统的信道估计技术。 本文首先讨论了无线信道的衰落特性，对信道估计的准确性和用户移动速度之间的关系行了仿真分析，分析了MIMO-OFDM的结构和技术原理及基于导频的信道估计算法，对梳状、块状导频的性能及基于STBC的 MMSE、LS信道估计算法进行了仿真分析，研究了降低MMSE信道估计算法复杂度的方法，设计了一种低复杂度的MMSE信道估计算法，仿真结果表明，该算法的MSE和BER与文献中的MMSE信道估计算法接近，但复杂度有所降低。另外研究了基于子载波分组的信道估计算法，在进行信道估计时，根据导频插入的特点，考虑保留能量较大的采样点，从能量较小的采样点获取噪声信息，再经过处理而消除部分噪声，提出一种改进的抑制噪声的信道估计算法，仿真结果表明在低信噪比的情况下，该算法的MSE比文献中的基于子载波算法有很大程度的降低。

MIMO-OFDM is the inevitable requirement of future communication, its high-capacity, high-quality and high-mobility are a big challenge for receiving antenna to recover the original data accurately, while the complicated mobile communication channels give rise to the difficulty of channel estimation. To solve above problems, this paper focuses on the channel estimation of the MIMO-OFDM system. The author elaborates the fading character of the wireless channel, and gives some simulation results of relationship between moving speed and accuracy of channel estimation. The principle and structure of the MIMO-OFDM is discussed, channel estimation technique based on pilot was analyzed, and the performance of block-type and comb-type pilot arrangement and the LS and MMSE channel estimation algorithm based on STBC is emulated, and an algorithm of reducing complexity for channel estimation is studied and designed. The simulation result shows that its performance of MSE and BER is similar to that of the MMSE channel estimation algorithm mentioned in the literature with lower complexity. In addition, the algorithm based on sub carrier grouping is discussed. According to the characteristics of the pilot insertion, considering high energy of sampling points, noise information is obtained from low energy of sampling points and though series of the processing an improved noise-suppressing algorithm is proposed with some of the noise is eliminating. The result of the simulation showed that the MSE of the improved noise-suppressing algorithm is lower than that of the sub carrier grouping algorithm mentioned in the literatures when SNR is low. Key words: Channel Estimation; LS; MMSE; sub carrier grouping; noise suppression