Improved Label Design for Timing Synchronization in OFDM Systems Against Multi-Path Uncertainty

Timing synchronization (TS) is vital for orthogonal frequency division multiplexing (OFDM) systems, which makes the discrete Fourier transform (DFT) window start at intersymbol-interference (ISI)-free region. However, multi-path uncertainty in wireless communication scenarios degrades TS correctness. To alleviate this degradation, we propose a learning-based method enhanced by improving design of training label. In proposed method, classic crosscorrelator extracts initial feature benefiting following machine learning. Wherein, network architecture unfolds one cross-correlation process. Against uncertainty, novel label designed representing ISI-free region and especially highlighting its approximate midpoint. Therein, closer to boundary smaller values are set, expecting locate maximum output with high probability. Then, guarantee correctness labeling, exploit priori information line-of-sight (LOS) form LOS-aided labeling. Numerical results confirm that, effectively enhances learner against uncertainty.