On Channel Estimation Techniques For MIMO-OFDM Systems

Orthogonal Frequency Division Multiplex (OFDM) and Multiple Input Multiple Output (MIMO) are widely recognized as a revolutionary technology for providing higher data rate in the future broadband wireless communication system. Obviously, accurate channel state information is essential to reliable information recovery. Channel estimation technique has different levels of impact to synchronization and sub-carrier allocation. Therefore, the accurate estimation of the channel state information is one of the key issues for MIMO-OFDM system. There are mainly two channel estimation methods: pilot-aided channel estimation and blind channel estimation. This thesis mainly concerns only with the pilot-aided channel estimation techniques for MIMO-OFDM system.First, pilot symbol assisted modulation (PSAM) channel estimation is investigated for SISO-OFDM systems, mainly focusing on the block-type and comb-type channel estimation. A comparative analysis on popular pilot assisted estimation algorithms is provided, including LS, LMMSE pilot estimation algorithms, followed by three interpolation algorithms, i.e. linear interpolation, Gaussion interpolation and cubic-spline interpolation. It is shown by Matlab simulation and analysis that, with the increase of signal-to-noise ratio, the cubic-spline interpolation algorithm outperforms other two interpolation algorithms.Next, comb pilot symbol assisted modulation (PSAM) channel estimation for MIMO-OFDM system is investigated. For the channel estimation algorithm of pilot location, in order to reduce the high complexity of LMMSE algorithm, a simplified method based on the singular value decomposition is given. Then the channel estimation algorithm based on the interpolation of the data location, including linear interpolation, Gaussion interpolation, cubic-spline interpolation algorithms and DFT channel estimation method, is discussed. On this basis, by neglecting secondary sampling in the time-domain to eliminate noise interference, the traditional DFT channel estimation method is improved, followed by Matlab-based computer simulation and discussion.