Skip to Main Content
In this paper, we analytically study the performance of a recently proposed efficient reduced complexity time synchronization approach for orthogonal frequency division multiplexing systems. This method uses a preamble of two identical parts and proceeds in two stages. In the first stage, the repetitive structure of the preamble is exploited to provide the coarse time estimate respecting the algorithm of Cox and Schmidl. In the second stage, a fine metric, based on differential correlation, is carried over a reduced time window centered on the coarse estimate. We here study the performance of the fine stage, assuming a successful coarse stage whereby the fine search window is centered on the correct frame start. We approximate the fine metric by a Gaussian distribution to derive a closed form expression of the frame start correct detection probability. To this end, a statistical characterization of the fine metric is achieved by its mean and variance computation. Simulations are used to validate the results of the analysis. Indeed, the evaluated rate of correct detection perfectly concords with the theoretical probability in both additive white Gaussian noise and multipath channels.