This thesis proposes to exploit this angular structure through beam-space clustering: partitioning the set of visible satellites into groups such that within each group, the spatial signatures are sufficiently separable for joint spatial decoding (e.g., via MMSE or successive interference cancellation), while different groups are multiplexed across orthogonal time–frequency resources. The goal is to find a sweet spot between full TDMA (no interference but low spectral efficiency) and full spatial multiplexing (high spectral efficiency only if all signatures are well separated).