Research And Implementation Of Channel Estimation In MIMO-OFDM Systems

Nowadays, people are calling more and more for the convenience of mobile communication and thus motivating the rapid development of mobile technologies, among which the'Beyond the Third Generation'mobile system is under consideration. In order to carry out high data rate as well as multi-media transmission, advanced methodologies are under research. Orthogonal frequency division method is now attracting the focus of the whole world with its high spectrum efficiency, ease of implementation (via FFT), and low receiver complexity (single-tape equalization), etc. Meanwhile, multiple input multiple output (MIMO) method is also a good choice. It extensively enhances the spectrum efficiency without extra resources and transmitter power. So, the combination of these two technologies, say the MIMO-OFDM system, is a good point for research and implementation.Among the many key factors of the MIMO-OFDM system, channel estimation is with no doubt of high significance. Coherent detection, with better performance in low SNR, can be carried out through channel estimation.In order to give a better description about the theory of channel estimation, the channel characteristics are first introduced in this thesis. Detailed explanation is given on the fading features as well as the mathematical models of the wireless channel in the mobile environment, following which are the basic principles of MIMO and OFDM technologies.The keystone of this thesis is the channel estimation in the MIMO-OFDM system. Because of its complexity and confusion to understand, some channel estimation methods in the single antenna OFDM system are described as the fundamental. Then, we focus on the channel estimation method based on the maximum mean square error (MMSE) principle, which is a prospective technology in the MIMO-OFDM system. However, due to its a very high rank matrix contrary in the process of computing, the basic MMSE algorithm is too complex to realize in the hardware platform. Some simplified algorithms are essential. The Significant Tap Catching (STC) method is a good choice. As the simulation shows, it distinctly reduces the workload of computation with little extra consumption in SNR. Finally, it is carried out in a hardware platform using FPGA chips.At the end of the thesis, circs for the most advanced channel estimation technology is briefly introduced.