Space communications is the link layer of everything in orbit and beyond, and the supplied research front is now split between an operational, twenty-year-old RF backbone and a set of largely simulation-stage upgrades (satellite-to-phone multiple access, learned random access, semantic compression, quantum key distribution) whose gains are asserted more often than quantified.
Summary
Space communications covers every path data takes to and from spacecraft: ground station networks talking to satellites and deep space probes, satellite-to-satellite relay, and increasingly satellite-to-handset links. NASA consolidated its own networks into the SCaN (Space Communications and Navigation) programme in May 2006 and now treats it as the backbone supporting the International Space Station through to deep space science ref. That is the mature part of the field: large radio-frequency apertures, licensed spectrum, hardened electronics, and a small number of very expensive ground assets. The fragility of that layer is visible in the same source set: damage to the 70-metre Deep Space Station 14 antenna at Goldstone was classified as a Type A mishap on the basis of total repair cost ref.
The research front sits above that backbone and attacks four different bottlenecks. First, capacity and access: direct-to-cell (DTC) service, connecting ordinary phones to satellites, forces a rethink of multiple access, and non-orthogonal multiple access (NOMA) is being proposed to raise spectral efficiency and user scale over conventional orthogonal schemes. Second, protocol behaviour in low Earth orbit: long propagation delays, large Doppler shifts and many simultaneous access attempts break conventional contention-based random access, and learned collision classifiers are being proposed to fix it under 3GPP-compliant LEO settings. Third, bandwidth economy: semantic transmission schemes that send only the most important modality and generate the other at the receiver, targeted at synchronised audiovisual services over constrained satellite links. Fourth, security and new physical layers: quantum satellite communication (QSC), plus free-space optics in the 8 to 12 micron mid-infrared atmospheric window, supported by component work on modulators and single-photon detectors.
The parameters that decide each of these are physical and unforgiving: link budget against atmospheric loss and turbulence, beam pointing and tracking accuracy, payload mass, power and thermal budget, clock synchronisation, spectrum availability, and radiation tolerance of the RF front end. The QSC review names precisely this list as the reason large-scale deployment has not happened, alongside scalability and integration with terrestrial infrastructure. On the hardware side, radiation effects in gallium nitride power amplifiers are still handled with approximate, trial-and-error models rather than physical understanding, which is the stated motivation for a $450,000 NSF award to build a predictive design framework.
Evidence base
- NASA established the SCaN programme in May 2006 to unify its networks, and it now supports the ISS through to deep space science and exploration ref.
- NASA classified the damage to the 70-metre Deep Space Station 14 antenna at Goldstone as a Type A mishap based on total cost of damages, investigation concluded 5 June 2026 ref.
- A 5 May 2026 review lists atmospheric loss, beam pointing and tracking, payload constraints, synchronisation, scalability and terrestrial integration as the bottlenecks still preventing large-scale quantum satellite communication, with Micius as the representative milestone.
- A deep learning random access framework for LEO satellite communications reports higher access success probability and lower delay than conventional schemes, evaluated in simulation under 3GPP-compliant LEO settings (3 June 2026).
- A mid-infrared upconversion detector achieved 80% conversion efficiency, 37% overall detection efficiency, noise equivalent power of 1.8x10^-17 W/Hz^1/2 and photon-number resolving up to 9 photons at 3 microns (4 June 2026).
- An NSF award of $450,000 dated 4 August 2026 funds physics-based radiation transport modelling of AlGaN/GaN HEMTs because existing radiation-damage prediction for satellite and deep-space power amplifiers relies on approximate, trial-and-error models.
Open questions
- What is the actual quantified gain of NOMA over orthogonal access in a direct-to-cell link budget, including receiver complexity and imperfect successive interference cancellation, rather than a case-study comparison?
- Do the learned random access gains under 3GPP-compliant LEO simulation survive real Doppler, real preamble collisions and on-board inference constraints in orbit?
- Can semantic and generative transmission schemes be certified for operational satellite services, given that the received content is partly synthesised rather than transmitted?
- Can mid-infrared plasmonic modulators reach the modulation speed and depth needed for the 8 to 12 micron free-space window, and can that hardware be space-qualified for radiation and thermal environments?
Assessment drafted 2026-08-31 from up to 11 KB sources using the technology-scorecard framework; scores are a draft read pending review.