Tile-based iterative carrier frequency estimation for OFDMA uplink transmissions

This paper examines the carrier frequency offset (CFO) estimation problem in the orthogonal frequency division multiple access (OFDMA) uplink systems, which is very challenging due to the presence of multiple CFOs. The existing solutions to this problem are either too complex to implement or not flexible in subcarrier allocation. To solve these problems, this paper proposes a tile-structure based iterative multi-CFO estimation technique. The proposed method is developed based on a special training sequence with repetitive structure. The inherent multi-user interference (MUI) compression provided by the tile structure allows us to utilize the repetitive property of the training sequence to jointly estimate the CFOs in the frequency domain with low complexity. Combining the CFO estimation with an interference cancellation scheme and performing iteratively, the algorithm achieves high estimation accuracy and fast convergence. The proposed algorithm is suitable for any subcarrier assignment schemes. In addition, as compared with other existing time domain based algorithms, which achieve the Cramer Rao Bound (CRB) at the price of unaffordable complexity, it closely matches their performance with over 70% computational saving.