An Efficient Reduced-Complexity Two-Stage Differential Sliding Correlation Approach for OFDM Synchronization in the AWGN Channel

In this paper, we propose a new scheme for data- aided time and frequency synchronization for OFDM systems, based on a single-symbol preamble. The preamble, of useful length $2^{m}-2$, is composed of two consecutive identical m-sequences (with length $2^{m-1}-1$ each). This preamble is extended by a cyclic prefix of convenient length. This stucture is adequate for a two-stage synchronization scheme, namely a reduced complexity coarse synchronization stage, followed by a finer synchronization one. The first stage, based on Cox and Schmidl-like sliding correlation, determines a reduced uncertainty interval over which the fine stage is carried. The second stage is indeed based on a differential correlation, which is more complex compared to the first stage. The combined use of m-sequences and differential correlation offers an almost perfect peak of the computed metric at the preamble start. To assess the performance degradation occasioned by the reduction of complexity characterizing the proposed two-stage approach, we also consider the brute force single-stage approach, where differential correlation is exclusively used. As a byproduct of our two-stage approach, the fractional frequency offset is estimated and its performance is assessed and compared for both two- stage and one-stage approaches. The brute force approach outperfoms all the considered benchmarks. Compared to the reduced complexity scheme, the brute force one provides similar performance, at the expense of a significant complexity overload. Only for SNR lower than $-5$ dB, the brute force scheme presents a slight enhancement with respect to the reduced complexity one. The simulation results show that the proposed method gives better performance than any other considered estimator. Although our technique is expected to be well suited to multipath channels, thanks to the underlying properties of m-sequences, in this paper we focus on the Additive White Gaussian Noise channel.