Timing and frequency synchronization for orthogonal frequency division multiple-access systems

Reliable timing and frequency synchronization is a major requirement in orthogonal frequency division multiple-access (OFDMA) systems as synchronization errors can result in inter-symbol-interference (ISI) and inter-carrier-interference (ICI) in the received signal, which severely degrade system performance. Thus, the objective of the proposed research is to develop, analyze, and prototype timing and frequency synchronization techniques for downlink (DL) and uplink (UL) OFDMA transmissions. For synchronization in DL OFDMA transmissions, we have developed conditions to select appropriate Zadoff-Chu sequences as training symbols that allow timing synchronization in the presence of large carrier frequency offsets (CFO). Using the proposed training block, we have designed training signal detection, timing synchronization, and integer CFO estimation algorithms. We have further proposed the training signal design for frequency synchronization in DL coordinated multi-point (COMP) transmissions, in which a user has to synchronize to multiple base-stations at the same time. In this respect, a frequency synchronization algorithm for DL COMP receiver has been designed using the proposed training signal along with its hardware implementation to analyze real-time performance. For frequency synchronization in UL OFDMA transmissions, we have proposed a null sub-carrier-based CFO estimation algorithm, and carried out its identifiability and acquisition range analysis. The proposed algorithm supports both sub-band and generalized sub-carrier allocations. We have also designed a CFO estimation and compensation algorithm for UL single-carrier frequency division multiple-access (SC-FDMA) transmissions. The proposed algorithm is based on parallel factor analysis and supports interleaved sub-carrier allocation. In addition, it guarantees the identifiability of CFO estimation and allows the system to operate on full load. Detailed simulations results have been provided along with discussions on computational requirements, which reveal that the proposed algorithms provide significant improvements in performance and efficiency compared to state of the art schemes in the literature.Ph.D