Abstract
Orthogonal frequency-division multiplexing (OFDM) has been adopted in various wireless communication standards due to its high spectrum efficiency and robustness against frequency selective fading channels. Meanwhile, multiple-input multiple-output (MIMO) transmission that utilizes multiple antennas at both the transmitter and receiver sides has been shown effective to enhance the system performance and/or throughput. Combining OFDM with MIMO transmission results in a new system called MIMO OFDM. Despite the inviting features, MIMO OFDM requires superior synchronization as well as channel estimation to restore the orthogonality among subcarriers and to support MIMO detection. Challengingly, most of the conventional synchronization and channel estimation methods for OFDM cannot be directly applied to MIMO OFDM without major modifications. Furthermore, synchronization and channel estimation depend on each other, yet they are mostly investigated separately. In this dissertation, we develop joint synchronization and channel estimation schemes for both OFDM and MIMO OFDM systems. For OFDM systems, we first develop a joint scheme based on a special training sequence structure. To ensure superior performance, we derive an optimal threshold and a suboptimal threshold for the proposed joint scheme. The optimal threshold clearly reveals how the threshold should react to the channel statistics, the signal-to-noise ratio (SNR), and other design parameters. On the other hand, the suboptimal threshold that can be calculated without pre-simulation or prior information on the channel statistics is shown a practical substitute for the optimal threshold. The proposed joint scheme with optimized thresholds improves both synchronization and channel estimation performance as compared with a related work with pre-simulated thresholds, especially when the channel’s first tap is rather insignificant and/or the SNR is relatively low. To integrate a carrier frequency offset (CFO) acquisition scheme with the proposed joint scheme, while preserving the original training sequence structure, we construct a partially geometric sequence that has an ideal periodic auto-correlation function. Accordingly, we develop an acquisition scheme that exploits the partially geometric feature to extend the CFO estimation range. For MIMO OFDM systems, we first develop a novel arrangement of the training sequences, which skillfully eliminates the mutual interference among the simultaneously transmitted training signals of the active antennas. Based on the training sequence arrangement, we generalize our joint scheme for OFDM into MIMO scenario and further present a majority vote refinement (MVR) of the timing estimates for different transmit-receive links. The proposed MVR fully exploits the features of our training sequence arrangement and the MIMO diversity to improve synchronization as well as channel estimation. It is shown that MVR not only enhances the robustness of the proposed joint scheme but also contributes significant performance gain at lower SNRs. The performance of the proposed joint approach is also analyzed with respect to MVR and threshold selection, and the theoretical results agree well with the simulation results over several channel models. To provide a trade-off between performance and complexity, we also develop low-complexity joint schemes for both OFDM and MIMO OFDM systems using maximal length sequence based training structures.