Abstract
This letter explores cooperative channel estimation for space-MIMO OFDM uplinks with multiple low-Earth-orbit (LEO) satellites. By leveraging spatial orthogonality across satellites , we mitigate pilot contamination in massive MIMO multiuser scenarios. To reduce inter-satellite link (ISL) feedback overhead, we propose a sensing-matrix-based scheme, where satellites forward only selected antenna signals to the master node (MN), preserving dominant signal energy. Simulation results show that cooperative satellites improve spatial decorrelation by 55% and reduce normalized mean square error (NMSE) by 83% compared to non-cooperative random pilot assignment. The proposed compression scheme effectively reduces ISL payload while maintaining channel estimation performance.