Study On Channel Estimation Methods In OFDM Systems

Orthogonal frequency division multiplexing (OFDM) is a popular modulation techniquefor wireless digital communications. This thesis will provide an overview of the OFDMsystems and various approaches of channel estimation which different constraints assumed onthe channel and data.Due to the dispersive wireless channel OFDM transmitter can e ectively combatmultipath channel impairments that can cause deep fades in some subchannels. However,OFDM receiver requires accurate and computationally e cient channel state informationwhen coherent detection is involved. It needs a good method estimating the channel and oneof the most used methods of wireless communication for OFDM Systems are the least square(LS) and minimum mean-squared error (MMSE) Channel Estimation methods. HoweverMMSE requires a large number of processing operations. Improve the performance of LS andMMSE channel estimation DFT-based channel estimation derived.In order to avoid the high complexity of the existing method, advanced channelestimation method semi-blind also is modified in this thesis. At the receiver, we calculate theautocorrelation matrix of the received signal, and the Q matrix is achieved by the subspacedecomposition method, then the initial channel estimation can be obtained. Due to theexistence of noise, the initial channel estimation is not ideal enough.The modified algorithm optimizes the Q matrix through the initial channel estimation.The channel estimation error caused by the noise is estimated by the initial channelestimation, and eliminated in the Q matrix, thus the performance of the channel estimation isimproved. Aiming to get a better estimation performance, the modified estimation uses the Qmatrix and the optimized Q matrix to optimize the channel knowledge obtained frompilot-aided estimation synchronously. In this thesis, study is carried out, based on themathematical analysis and simulations in MATLAB, to modify channel estimation methodsfor OFDM system environment.