Scalable OTFS-Based Multi-LEO Downlink Access Using Doppler-Block Resource Allocation

Abstract

Future non-terrestrial networks (NTNs) are expected to rely on dense low Earth orbit (LEO) satellite constellations capable of providing simultaneous coverage to a common user terminal. While Orthogonal Time Frequency Space (OTFS) modulation has emerged as a promising waveform for high-mobility satellite communications, most existing studies focus on conventional single-satellite links. This paper proposes a scalable multi-user OTFS downlink framework that exploits residual Doppler differences between simultaneously visible satellites to enable Doppler-domain resource allocation and user separation at the receiver. Dedicated Doppler blocks, pilot-assisted channel estimation, synchronization, and low-complexity linear minimum mean-square error (LMMSE) detection are jointly employed to recover multiple superimposed OTFS downlink signals. Simulation results for a four-user scenario under NTN-inspired channel conditions demonstrate reliable user separation and symbol recovery with end-to-end symbol error rates below $10^{-4}$ at high signal-to-noise ratios. The practical feasibility of the proposed framework is further validated through a software-defined radio (SDR) experiment involving two simultaneous OTFS transmissions and an RFSoC-based channel emulator. Near error-free detection is achieved for both satellite links, confirming the effectiveness of the proposed architecture under realistic hardware conditions. The presented framework provides a scalable solution for future multi-satellite OTFS-based NTN communications.