Channel Estimation Algorithm With Special Training Symbols For MIMO-OFDM Systems

Li Li,Wang Ke,Han Li
DOI: https://doi.org/10.3963/j.issn.2095-3844.2008.06.032
2008-01-01
Abstract:The paper proposes a new time domain least square channel estimation using special training sequences for MIMO-OFDM systems.In this algorithm,using a constant envelope of exponential sequence by inserting zeros as training symbols,the training symbols has characteristics of desirable PAPR,which eliminate the influence of intersymbol interference(ISI).The simulation results show that the constant envelope training symbols is more advantage than stochastic training symbols,the algorithm using constant envelope training symbols performs the algorithm using stochastic training symbols with respect to the mean square error(MSE) of the estimated channel by 15 dB if the signal-to-noise ratio(SNR) is 20 dB;When using special symbols as training sequences,the proposed algorithm is also by more advantage of the conventional algorithm and the proposed algorithm outperforms conventional algorithm with respect to the MSE by 10 dB if the SNR is 25 dB.
What problem does this paper attempt to address?