Abstract
Direct Device-to-Satellite (D2S) communications promise global connectivity to unmodified user equipment (UE), extending coverage beyond terrestrial networks. Realizing this promise is fundamentally challenging: severe path loss and limited UE transmit power push uplink SNRs far below terrestrial norms, while suitable spectrum remains scarce. Together, these constraints impose a spectral-efficiency (SE) bottleneck, and under such conditions the efficiency of the UE power amplifier becomes critical, jointly governing transmit power and battery life. To improve UE-side power efficiency, 3GPP has adopted Discrete Fourier Transform-spread OFDM (DFT-s-OFDM) as an optional uplink waveform, exploiting its substantially lower Peak-to-Average Power Ratio (PAPR) relative to OFDM. To break the SE bottleneck, we show that aggressive non-orthogonal transmission, in which the number of concurrent users exceeds the number of receive antennas by more than 2×, can unlock substantial capacity gains that remain entirely unexploited. Realising these gains, however, requires receiver architectures that, to the best of our knowledge, have not yet been developed. DFT-s-OFDM intensifies the difficulty: the DFT spreading couples signal components across subcarriers, inflating the effective dimensionality of the detection problem. We address both challenges with a novel receiver design that jointly exploits the SE gains of aggressive non-orthogonal transmission and the power-efficiency benefits of DFT-s-OFDM. Simulations under realistic channel-estimation errors and high-mobility Doppler show that the proposed scheme achieves 2× the SE of baseline, surpasses recent nonlinear MIMO receivers by 40% at 15% of their complexity, and reduces PAPR by up to 6 dB relative to DFT-s-OFDM MIMO and 11 dB relative to OFDM.