Abstract
Abstract
Spatial multiplexing can increase the rate of entanglement distribution over free-space optical (FSO) channels, but atmospheric turbulence introduces modal crosstalk and unequal loss among spatial modes. We propose adaptive spatial multiplexing (ASM), which uses channel state information (CSI) for each coherence block and singular value decomposition (SVD)-based transmit–receive processing to adapt the spatial basis and the number of Bell pairs distributed in parallel. For a passive, polarization-independent channel with at most one photon in each selected singular input mode, vacuum in every unused input mode, and a portwise receiver that preserves polarization qubits in singly occupied output modes and maps all other occupations to erasures, matched SVD processing reduces the collected multiphoton channel to parallel qubit erasure channels with transmissivities given by the squared singular values of the transmission matrix. The resulting one-way hashing rates are additive and nonnegative. Therefore, when the minimum per-stream rate is the only constraint on selection, the aggregate achievable rate is maximized exactly by selecting every parallel channel that meets it. We study ASM over a simulated four-mode turbulent FSO channel ensemble with ideal, polarization-preserving mode processors. At the nominal minimum-rate requirement, ASM outperforms receiver-only combining in aggregate rate, per-stream rate, and priority-stream non-outage probability. The results show that ASM turns spatial multiplexing into a resource-aware strategy for entanglement distribution, balancing overall throughput with stream reliability and photon efficiency. Its aggregate-rate advantage over receiver-only processing persists under the CSI-aging conditions considered.