Abstract
In this thesis, a novel preamble structure of two time-domain training signals is proposed for time synchronization, carrier frequency offset (CFO) estimation, and channel estimation in orthogonal frequency-division multiplexing (OFDM)-based two-way amplify-and-forward (AF) relay systems. The first training signal consists of a specific complex sequence with a nice property that the Hadamard product of the sequence and its complex-conjugate circular-shift version (excluding the zero-shift case) is a near-perfect sequence. The second training signal is a sequence whose frequency-domain counterpart is formed by placing two partial Zadoff-Chu sequences proposed recently on subcarriers in an interleaved manner, where these two sequences are separable in both of the frequency and time domains. Using the proposed preamble for one terminal of the two-way system and a similar preamble with the first training signal circularly shifted appropriately for the other terminal, we can conduct CFO-immune coarse time synchronization for both terminals based on auto-correlation and cross-correlation in the first half of the training period, and then perform CFO estimation, fine time synchronization, and channel estimation in the second half. As compared to existing related works, the proposed approach achieves better performance with a little increase in the computational complexity.