MIMO-OFDM技术非常适合高速率通信业务的要求，是下一代移动通信的核心技术。但是MIMO-OFDM对时间和频率同步误差比较敏感，时间的不同步会导致符号间干扰，而频率不同步则会导致载波间干扰，两者均会导致系统性能的下降。因此准确的时间同步和频率同步对MIMO-OFDM系统来说是非常重要的。 本文主要研究了MIMO-OFDM的同步技术。首先，分析了OFDM同步处理的分类和同步误差对系统性能的影响，比较了不同的OFDM同步方法。其次，重点研究了MIMO-OFDM系统的同步方法，由于Schenk提出的使用训练序列的同步算法存在峰值平台，导致其定时同步不准确，因此提出一种改进的同步算法，使用具有良好自相关特性的CAZAC序列作为训练序列，采用TSTD发射分集方式，有效提高了同步准确率和频偏估计性能，并通过仿真验证了改进算法的性能。最后，介绍了MIMO-OFDM仿真实现的模型和流程图，在不同的频偏和移动速度下对改进算法的频偏估计性能进行了验证。 仿真结果表明由于改进算法利用了CAZAC序列良好的自相关特性，定时同步函数峰值尖锐，定时同步准确，在信噪比大于5dB时，定时同步准确概率能够达到97%以上。在低信噪比条件下，改进算法的频偏估计的MSE值与理论MSE值基本接近，在信噪比较高时频偏估计性能与原有算法基本相同。在同等条件下，定时同步和频偏估计的性能较原有算法的有所提高。改进算法在频偏分别为0.1个子载波间隔，0.2个子载波间隔和频偏为10kHz时具有良好的频偏估计性能，当终端的移动速度分别为0km/h、60km/h和120km/h时，改进算法也能够实现良好的频偏估计。

MIMO-OFDM technology is suitable for requirements of high-speed communications as a new generation of mobile communication core technology. However, MIMO-OFDM system is very sensitive to synchronization error. Timing error will lead to inter-symbol interference and frequency offset can introduce the inter-carrier interference, both results in the decline of system performance. So accurate timing synchronization and frequency offset of MIMO-OFDM system are very important. This thesis focuses on the synchronization algorithms in MIMO-OFDM systems. In the thesis the influence of timing offset and carrier frequency offset on OFDM system is analyzed first， and different synchronization algorithms are introduced. Then the synchronization algorithms are mainly researched, because of Schenk’s algorithm has a peak value plateau, its timing synchronization is inaccurate, a modified algorithm that use the CAZAC sequence and TSTD are developed. Finally the frame and flow chart of computer simulation are introduced, the simulation are carried out to validate the performance of modified algorithm in different conditions. The results of computer simulation demonstrated that the modified algorithm has correct timing synchronization because of using the auto correlation characteristic of CAZAC sequence; the right probability of timing synchronization can achieve 97% when SNR is more than 5dB. In lower SNR, the MSE value of modified algorithm is close to the theoretical value; their performance is comparative in higher SNR. In the same condition, the modified algorithm enhanced the performance of timing synchronization and frequency offset. when the frequency offset is 0.1, 0.2 of sub-carrier spacing, and 10kHz, and UE speed is 0km/h, 60km/h, and 120km/h, the modified algorithm can estimate the frequency offset accurately.