Research On Equalization Techniques For MIMO-OFDM Systems

As a combination of MIMO and OFDM techniques, MIMO-OFDM can not only increase the transmission rate of wireless communications systems by times, but also effectively use the multipath fading to suppress the interference and noise. The MIMO equalization and detection techniques have been drawing much attention because of their key position in MIMO-OFDM systems.In this work, various equalization algorithms for MIMO-OFDM are studied first, and then an improved layer threshold sphere decoding (SD) algorithm is proposed. Compared with the original SD, the improved SD algorithm shows lower complexity but the same performance. However, the MIMO equalization performance is found to be much influenced by channel estimation errors. Thus, the preamble symbols aided channel estimation and noise variance estimation algorithms are studied next. A new estimation method based on QR decomposition is proposed. The new method multiplies the LS channel estimation vectors by a unitary matrix which is obtained from the QR decomposition of a partial frequency domain transfer matrix, and then the channel estimation noises are separated out. The new method is free of virtual subcarriers and channel variations, and hence can provide more reliable channel state information for MIMO-OFDM equalization in different channels.