基于为MIMO-OFDM系统设计了一种训练序列,并给出了相应的信道估计算法.提出使用PN序列代替传统OFDM系统中的循环前缀作为训练序列,并且不同发射天线选择相互正交的PN序列,利用PN序列的自相关性和互相关性,采用时域相关的方法估计多信道.与传统采用频域导频的信道估计器相比,该方法由于无需矩阵求逆,能够大大降低计算复杂度,同时PN序列作为OFDM符号的一部分可以实时估计信道,具有跟踪时变信道的能力.仿真结果表明该估计是无偏的,而且均方估计误差极小.
A new training sequence is designed for MIMO-OFDM systems, and the corresponding channel estimation algorithm is proposed. The method uses PN sequence as training sequence instead of cyclic prefix, and the PN sequences are orthogonal at different transmit antennas. Due to the autocorrelation and cross-correlation of PN sequences, the channel estimation was acquried by correlation in the time-domain. Compare to traditional estimators that used pilots in the frequency-domain, the proposed estimator need not matrix inversion that can reduce computational complexly. Because PN sequence is a part of OFDM symbol, the channels can be real-time estimated and time varying channels can be tracked. The computer simulation results show that the proposed estimator is unbiased and has significantly smaller MSE.