OTFS-SDMA for Massive Grant-Free Random Access in LEO Satellite Internet of Things

Low earth orbit (LEO) satellite-based Internet of Things (IoT) has great potential to provide seamless global coverage, but the large propagation delay and severe Doppler shift in terrestrial-satellite link (TSL) will become the most challenging problem. To handle these challenges and facilitate massive grant-free random access, we propose an orthogonal time frequency space-based scramble-division multiple access (OTFS-SDMA) scheme, where the scrambling technique is used to tackle the correlated TSL channels between neighboring devices. At the receiver, we first propose a user activity detection (UAD) method based on capturing the dominant line-of-sight (LoS) path, without relying on the assumption of a static TSL. To facilitate accurate channel estimation (CE) against severe Doppler shifts, we exploit prior information about satellite velocity to detect the angles of arrival (AoAs) of active devices with the two-dimensional multiple signal classification (2D-MUSIC) algorithm, and further estimate the Doppler shifts. Building on the Doppler estimation, the orthogonal matching pursuit (OMP) algorithm is used to estimate the sparse TSL channel in the time-delay (TD) domain. In accordance with the OTFS-SDMA scheme, we propose a cross-domain elementary signal estimator (CD-ESE) for multi-user detection (MUD). In the CD-ESE MUD structure, both bit-level and symbol-level scrambling sequences help to distinguish neighboring active devices with correlated TSL channels, and the channel decoder works in conjunction with the CD-ESE to enhance MUD accuracy. Simulation results are provided to demonstrate the superior performance of the proposed OTFS-SDMA scheme over the state-of-the-art solutions.