Reduced Complexity Initial Synchronization for 5G NR Multibeam LEO-Based Non-Terrestrial Networks

Auteurs

Kumar Meshram A., Kumar S., Querol J., Andrenacci S., Chatzinotas S.

Référence

IEEE Open Journal of the Communications Society, vol. 6, pp. 1528-1551, 2025

Description

This paper presents a computationally efficient technique to mitigate the overall Carrier Frequency Offset (CFO) impairment introduced by the User Equipment (UE) crystal oscillator imperfection and the Doppler effect introduced by the satellite and UE movement. We assume steerable multibeam regenerative Low Earth Orbit (LEO) satellite-based 5G New Radio (NR) Non-Terrestrial Networks (NTN) under trajectory uncertainty and beam-pointing errors. The UE and the satellite do not require any prior information on the satellite ephemeris or UE location to achieve initial synchronization. The novelty of our method lies in exploiting the instantaneous satellite state vector for Doppler Pre-Compensation (DPC) to each beam of the LEO satellite relative to its Beam Center (BC). The UE then performs post-compensation to address the residual frequency offset by employing aggregated 5G New Radio (NR) Primary Synchronization Signal (PSS), followed by PSS detection, offering two-fold search space reduction during initial synchronization. We provide practical design parameters for the proposed algorithm to ensure that UE can efficiently perform initial synchronization by evaluating approximate bounds on the spot beam radius along with the No-Benefit Region (NBR) bounds of DPC. We conducted extensive simulations to assess the performance of the S- and Ka-bands under the NTN channel model and validated it with the corresponding analytical expressions conditioned on satellite trajectory uncertainty and beam-pointing error. Our method ensures that the probability of PSS detection remains above 90% while maintaining a false alarm rate of 1% at a Signal-to-Noise Ratio (SNR) as low as −6 (dB) if the standard deviation of the satellite trajectory and beam-pointing errors are within 5 (km) and 0.05°, respectively.

Lien

doi:10.1109/OJCOMS.2025.3543625