Beam Scheduling and Multi-Scenario Initial Access Evaluation for 5G NR Non-Terrestrial Networks

Abstract

This paper presents an integrated framework for beam scheduling optimization and initial access evaluation in 5G new radio (NR) non-terrestrial networks (NTN). A mixed integer linear program (MILP) is formulated to minimize the synchronization signal block (SSB) sweep period while enforcing full cell coverage, hardware beam limits, and spatial separation among active beams. The optimal solution is shown to match the unconstrained round-robin lower bound. The resulting schedule is used as the basis for a system-level simulation of a single LEO satellite flyover, modeling six terminal access scenarios that combine two UE antenna classes, directional phased-array and handheld user terminals (HUT), with three beam alignment strategies, namely perfect pointing, ephemeris-errored, and twoprobe spatial search. A time-varying non-homogeneous Poisson process with a Gaussian envelope captures realistic peakhour traffic over a $\mathbf{1 2 0}$-second overpass. Results show that terminal antenna gain is the primary determinant of access reliability: directional UEs achieve near-perfect access success rates (> 99.9%), while HUT terminals are coverage-limited at approximately 76%. Spatial separation is validated through perUE SINR analysis, confirming less than 0.05 dB interference degradation for over 95% of accepted users. Critically, the twoprobe search strategy, although beneficial for link-level robustness, introduces significant queue congestion and access latency, with the normalized queue load peaking at 18× for HUT-Search, revealing a fundamental system-level trade-off between signaling diversity and resource contention.