In low Earth orbit (LEO) satellite communications, the link quality fluctuates drastically during a satellite pass, exhibiting a wide dynamic range from the horizon to the zenith. Moreover, the high relative velocity induces severe Doppler shifts. While orthogonal time frequency space (OTFS) modulation effectively resolves the doubly-selective fading, its spectral efficiency is fundamentally bounded by the Nyquist limit. To break this bottleneck while adapting to dynamic channel variations, this paper proposes a LEO satellite-assisted flexible faster-than-Nyquist (FFTN)-OTFS (LEO-FFTN-OTFS) scheme. Conventional fixed-parameter FTN signaling suffers from severe inter-symbol interference at low elevation angles or spectral inefficiency at the zenith. To overcome this, a low-complexity Look-Up Table (LUT) mechanism is designed to adaptively optimize the time-domain compression factor based on the instantaneous signal-to-noise ratio. At the receiver, a linear minimum mean-square error (LMMSE) detector is formulated to suppress the colored noise and structured interference with minimal computational overhead. Besides, a rigorous theoretical framework is established incorporating 3GPP Tapped Delay Line (TDL) channel models to derive analytical expressions for effective throughput, energy efficiency, and bit error rate (BER) bounds.Simulation results demonstrate that the proposed adaptive scheme eliminates the irreducible error floor inherent in aggressive static FTN configurations at low SNRs, and maximizes the effective throughput across the entire elevation trajectory, achieving a superior trade-off between spectral efficiency and transmission reliability.

翻译：暂无翻译