Research On Channel Estimation And Precoding Of MIMO-OFDM System

MIMO-OFDM system inherit high spectral eiffciency of OFDM technology,bandwidth scalability, resistance to multipath fading, etc., as well as the advantages ofMIMO technology which improves data transmission capacity. MIMO-OFDM is afocus for the next research generation of wireless communication systems. Channelestimation and multi-cell precoding technology are key technologies of MIMO-OFDMsystem. Therefore,this paper focuses on the problems faced by these two technologies,and gives the corresponding solution.Firstly, the article describes the basic principles of channel estimation.,andanalysises traditional channel estimation methods, such as LS estimator,LMMSEestimator. This paper compares the advantages and disadvantages of the various channelestimation techniques.The feasibility and problems of DFT-based channel estimationalgorithm are analyzed. Due to the presence of virtual subcarriers，leading todiscontinuity of the pilot, as well as sample rate mismatch problem and energy leakage,the performance of traditional DFT methods decline. This paper, based on the existenceof the problem, proposes an improved algorithm based on the DFT channel estimation.MATLAB simulations proves the validity of the proposed method.Secondly, Inter cell precoding does not consider multi-user transmission,as well asequipped multi-antenna of user.We propose,based on the existence of the problem, alow complexity precoding algorithm for multi-cell scene，which makes Ml use ofmulti-user transmission and user’s multi-antenna collaboration processing capabilitiesand overcome disadvantages of BD method that ignore noise and high computation ofSVD decomposition.The proposed method can gains effectively improvement of thesystem throughput. It is well known that precoding requires perfect knowledge ofchannel state information (CSI) at the transmitter, and in practice the perfect CSI maynot be available because of channel estimation error. To solve the problem, we present aprecoding method which is robust to channel estimation error based on the previouslyproposed method and show that the proposed algorithm is robust to channel estimationerror through simulations.