Angular Correlation-Aware Grant-Free Detection in Multi-Satellite Cooperative Networks

Recent advancements in satellite communication have positioned multi-satellite cooperation as a promising paradigm for next-generation communication networks. In terrestrial-satellite links, the dominance of line-of-sight (LoS) propagation induces strong angular correlations between the positions of user terminals and their corresponding angles of arrival (AoA) at satellite receivers. To exploit this angular correlation, this work proposes an approach to enhance grant-free random access detection in multi-satellite-enabled Internet of Things systems. Specifically, a multi-satellite terminal angular relationship (MSTAR) model is proposed to capture inter-satellite angular correlations, thereby improving spatial diversity utilization in signal detection. Based on this model, a multi-satellite cooperative message-passing and expectation-maximization (MSC-MP-EM) algorithm is developed for joint active user detection (AUD) and channel estimation (CE). The algorithm iteratively alternates between two components: LoS-MP, which incorporates Bayesian inference with LoS priors for initial AUD and CE, and MSA-EM, which refines AoA estimation by exploiting the angular correlations modeled by MSTAR. To evaluate the theoretical performance limits of the proposed scheme, the Cramér–Rao bound is derived as a benchmark for AoA estimation accuracy. Theoretical analysis and simulation results demonstrate that the proposed algorithm achieves superior performance over benchmarks in terms of AUD and CE accuracy in multi-satellite cooperative networks.