MIMO-OFDM System Channel Estimation Technology

MIMO-OFDM technology is currently a hot research field of communication. It can significantly increase the channel capacity and can be effective against multi-path fading, making it a promising solution for wireless channel to achieve high-speed data transmission. It has a very broad research and development prospect. The whole performance of MIMO-OFDM system is directly affected by the accuracy of channel state information, so it is of great significance to research MIMO-OFDM system channel estimation.In this paper, OFDM, MIMO and MIMO-OFDM system are studied. The training sequence based channel estimation and the pilot symbol based channel estimation are discussed deeply. In the training sequence based channel estimation method, the LS time-domain channel estimation and the LS frequency-domain channel estimation are discussed. Then the improved algorithms of them are given. The results of the experiment show that the system performance can be improved by the improved algorithm. In the pilot symbol based channel estimation method, the EM based LMMSE channel estimation and the two-dimensional pilot symbol based MMSE channel estimation are discussed. By researching the Gauss interpolation and the Cubic interpolation, the estimation performance of complete channel response in MIMO-OFDM system based on interpolation is analyzed. The EM algorithm can change multiple input multiple output (MIMO) channel estimation problem into a series of single input single output (SISO) channel estimation. This transformation can avoid large-scale matrix operation so as to simplify the algorithm and improve the performance of the system and the accuracy of the estimation. In order to enhance the performance of the EM based LMMSE channel estimation, the initial value of the iterative EM algorithm is improved through designing the training sequences. The improved performance of the system is verified by several simulation experiments.