Researchof Channelestimation Based On MIMO-OFDM System

In wireless communications, high data transfer rate and high spectrum utilization have always been the pursuit of the goal of the system. Orthogonal frequency division multiplexing (OFDM), as a multi-carrier transmission solution, has been widely utilized in high speed wireless data transmission systems, with high spectral efficiency and resistance to multipath fading. Multiple input multiple output (MIMO), as a major breakthrough in the field of communications and signal, in the case of no increasing the transmission bandwidth and power, can greatly increase channel capacity and spectrum efficiency by establishing parallel transmission channel between the transmitter and receiver. Therefore, the organic combination of OFDM and MIMO, can improve system capacity and transmission rate to a large extent and obtain strong reliability. In MIMO-OFDM, because of accurate channel state information playing a vital role on the demodulation and decoding, the channel estimation is one of the key technologies of MIMO-OFDM system. Taking the MIMO-OFDM as background, the method of channel estimation is researched.Firstly, the thesis describes the development process and present situation of mobile communications technology and MIMO-OFDM, and analyzes the propagation characteristics of radio channel and the simulation model. The space-time coding and spatial multiplexing technology are researched.Secondly, LS, MMSE, LMMSE criterions are researched, and simulation shows the LMMSE criterion is the best. Next, superimposed training-based channel estimation is researched, in the iterative decision algorithm, the estimation accuracy is poor for ignoring the residual error and noise when judging the valid information sequence by initial channel estimation and superimposed training. This paper presents a improvement measure and simulation results show the accuracy of the improved algorithm has been improved.Finally, the constant, linear, Gaussian and DFT interpolations are compared by theoretical analysis and computer simulation, which explains the DFT interpolation is the best. In view of only zeros processing without filtering and denoising in DFT interpolation, a joint estimation method combining DFT interpolation and LMS adaptive filtering for filtering is proposed. And simulation analysis is taken in the end.