Research On Channel Estimation In MIMO-OFDM System

Multiple-Input Multiple-Output (MIMO) becoming an attractive technology which can dramatically boost channel capacity and spectral efficiency; while Orthogonal Frequency Division Multiplexing (OFDM) can cope with multi-path fading. So the combination of MIMO and OFDM, MIMO-OFDM which can make the two technologies benefit each other is preferred among the wireless transmitting technologies of the next generation physical layer. Hence, the technologies of MIMO-OFDM system become the hot spot in the field of wireless communication. As a key technology in MIMO-OFDM system, channel estimation provide the other technologies such as precoding, signal detetion of system with necessary channel state information(CSI) and directly influences the performance of the receiver, so it gets more attention.This dissertation focuses on channel estimation in MIMO-OFDM systems. All the research of the dissertation is based on analysis of a variety of MIMO channel models, system framework and capacities for MIMO-OFDM systems. A series of researches on channel estimation are carried out in different environment. The work is briefly stated as follows:1. Time invariant channel estimation is studied. The performance of Maximum Likelihood(ML) algorithm is not well caused by its sensitivity to noise on the Rayleigh flat fading channel. So a joint iterative ML channel estimation and symbol detection algorithm is proposed to further improve the performance. The algorithm associates with the Minimal Mean Square Error(MMSE) detection algorithm which can restrain the noise. Simulation results show that, compared to traditional ML algorithm, fewer pilots are required, and then more data resource can be made the best use of.2. Time varying channel estimation based on linear model is studied. An interpolation-based MIMO-OFDM time varying channel estimation algorithm is proposed to improve the performance and deal with the symbol delay. Firstly, part of time domain channel response parameters in the current OFDM symbol are estimated, and then the other time domain channel response can be recovered by interpolation associated with the estimated ones. Theoretical analysis and simulation show that the algorithm gets better performance with the pilots in groups, and the symbol delay can be avoided.3. The characteristics of LTE system combined with channel estimation in MIMO-OFDM system are discussed. A new reference symbol mapping scheme is proposed to improve LTE downlink channel estimation performance in the time varying environment. The proposed scheme is based on the concept of inserting null sub-carriers in the sub-carriers nearby reference symbols (i.e., pilot symbols). This helps mitigate the effect of inter-carrier interference (ICI) at pilot sub-carriers. In addition, the proposed scheme allows the ICI gain parameters to be estimated via a simple estimator. And the proposed scheme achieves excellent performance improvements over the Release 8 standard one in the time varying environment.