Transmitter precoding for interference mitigation in closed-loop MIMO OFDM systems

This thesis contributes to transmitter precoding design in multiple-input multiple-output (MIMO) orthogonal frequency-division multiplexing (OFDM) systems for reducing intercarrier interference (ICI), and mitigating the degradation due to antenna correlations. Transmitter-based techniques usually require the complete channel state information at the transmitter (CSIT), which is difficult to obtain in practical applications. The complete channel information, including frequency offset and channel gains, is not readily estimated at the transmitter or feedback capability from the receiver is limited. Therefore, precoding with partial CSIT is first considered. An important property of the ICI matrix is derived and non-linear Tomlinson-Harashima precoding (THP) with only partial CSIT, not including frequency offset, is developed. To further reduce the feedback requirements, linear and non-linear limited-feedback precoders are proposed. Our results significantly reduce the BER increase due to frequency offsets in single-user and multiuser MIMO OFDM.;The proposed schemes have good performance, low feedback requirements and low complexity, which are desirable for interference reduction in the future wireless systems.;Precoders are also developed to mitigate the impact of transmit-antenna and path correlations for MIMO OFDM systems. A pairwise error probability (PEP) optimal precoder is derived, which requires only covariance information at the transmitter, and significantly reduces feedback requirements. Furthermore, mean (first-order statistics) -feedback SNR-maximizing precoders are designed for a general MIMO channel model in MIMO OFDM with estimation errors and feedback delay. The quality of mean feedback will be degraded due to estimation errors, and it is more sensitive to the channel time variations and feedback delay than covariance information. In contrast, covariance precoding may become less effective when mean feedback is accurate. Au adaptive dual-mode precoder is thus proposed, in which either new mean-feedback precoding or covariance precoding is adaptively chosen at the receiver according to the channel conditions. The intuition is confirmed that if mean feedback is sufficiently accurate, it improves the system performance. Our proposed precoders, both the mean-feedback precoder and the dual-mode precoder, reduce the error rate. Adaptive precoding outperforms both mean-feedback precoding and covariance precoding applied individually.