Research On Channel Estimation Algorithm In MIMO-OFDM Systems

OFDM can divide a frequency selective channel into a number of parallel flat fadingchannels. This technology achieves high spectral sffiency,and overcome the adverse ofmulti-path effects. MIMO systems,employing multiple transmitting antennas and multiplereceiving antennas,can not only multiple channel capacity,but also multiple the spectralefficiency without increasing the bandwidth and transmit power. MIMO-OFDM,combiningOFDM and MIMO, will achieve high rate and large capacity in limited bandwidth.MIMO-OFDM technology has become one of the core technology of the next mobilecommunications. This paper focuses on this key technology of MIMO-OFDM system channelestimation, in-depth study channel estimation algorithm based on pilot.First，this paper briefly describes the research background and development status ofMIMO-OFDM technology. the radio channel characteristics and the model was presented, themultipath delay, Doppler spread, Rayleigh fading and Rician fading was associated discussion.Then characteristic of communication and principle of OFDM,MIMO,MIMO-OFDM systemare discussed.Second, the research of the MIMO-OFDM systems based on one-dimensional pliotchannel estimation are discussed. Then derived The method of pilot insertion in MIMO-OFDMsystem according to the method of pilot insertion in OFDM. Which were inserted orthogonaldomain or frequency domain orthogonal pilots at the time or frequency axis,give thesimulation analysis and comparison of Least Squares (LS), Minimum Mean SquareError(MMSE) and improved algorithm at pilot position,then introduced the first ordercubic-spline interpolation and transform domain-based interpolation algorithm into thefrequency domain orthogonal pilot,compare the performance of their respective advantagesand disadvantages by matlab simulation.Finally, for the shortage of channel estimation based on one-dimensional pliot, theresearch of channel estimation based on two-dimensional pliot channel estimation arediscussed. The distribution structure of two-dimensional pliot are discussed， in the case ofinserting rectangular pilot pattern,introduce Wiener filter filterinterpolation and its improvedWiener filter interpolation algorithm, simulation results show that thealgorithm is superiorone-dimensional interpolation algorithm. As interpolation algorithm will bring ininterpolation noise, In this paper, add the noise-reducing filter under DFT/IDFT transform based on interpolation algorithm for removing the interpolation noise effectively. Andverified by simulation that the algorithm has better estimation performance.