Lunar network constellation

A layered design: frozen elliptical orbits for south-polar comms and navigation, a mid-altitude mesh for global reach, and halo-orbit relays that keep the far side connected to Earth. Edit any layer and the coverage maps recompute.

–

T+0.0 h

Coverage over one sidereal month

% of time

020406080100%

Loading…

Summary — ≥1 satellite in view

Global

–

South pole <−85°

–

Far side

–

S-pole worst gap

–

City sites

Area averages are cos(latitude)-weighted. Visibility uses a smooth sphere plus the elevation mask; local terrain (crater walls at the poles) will cut real coverage, which is why rim towers matter.

Moon → Earth trunk: link sizing

Reference anchors

NASA LLCD (LADEE, 2013): 622 Mbps optical downlink from lunar orbit, ~0.5 W laser, 10 cm space telescope, 1550 nm PPM; ground receiver used an array of small (~40 cm) telescopes. Default optical inputs reproduce this.
LRO Ka-band: ~100 Mbps class science downlink from lunar orbit (~0.75 m high-gain antenna) to an ~18 m ground dish.
Artemis II O2O: Orion's optical terminal, designed for up to ~260 Mbps from lunar distance.

Earth segment

Optical ground stations spread in longitude and climate. Clouds block 1550 nm, so you need site diversity: 3 longitude bands × 2–3 uncorrelated weather sites each keeps one station with Moon in view and clear sky most of the time.
RF backup. Ka/S-band through DSN-class 34 m antennas or commercial ground networks for command, safing and bad-weather periods; lower rate, near-all-weather.
Relay via GEO? A GEO optical terminal sits above weather and can see the Moon ~always, then drops to Earth over short, weather-diverse links. It trades extra hardware and a second hop for availability; worth it once trunk demand passes a few Gbps.
Storage absorbs gaps. With DTN store-and-forward, a 10 Gbps optical trunk that is up 70% of the time can still carry 7 Gbps average if relays have enough onboard storage.

Design rationale

01Why layers

No single orbit does everything. The Moon is small (radius 1,737 km), so low orbits see a little at a time; the poles never see Earth well; the far side never sees it at all.

ELFO: long dwell and multi-satellite geometry over the south pole, where the first cities and water ice are.
Walker mesh: global reach for rovers, pits and far-side sites.
NRHO / L2 halo: high, Earth-visible nodes that bridge the far side to Earth.
Direct-to-Earth: nearside sites simply point at Earth, which never sets.

02South polar ELFO

Elliptical Lunar Frozen Orbits (a ≈ 6,142 km, e ≈ 0.6, i ≈ 57.7°, ω = 90°) put periapsis over the north and let the satellite hang near apoapsis (~8,100 km altitude) over the south for most of each 12 h orbit. Four satellites in two planes give continuous south-polar coverage and, for much of the time, the four-in-view geometry a position fix needs. This is the ESA Moonlight / NASA LCRNS approach.

A mirrored north ELFO pair (ω = 270°) is on by default because Peary Rim is a north-polar city site; drop it to see how much the mesh alone provides there.

03Getting to Earth

The 9:2 southern NRHO (6.56 d period, perilune ~3,000 km over the north, apolune ~70,000 km over the south) never goes behind the Moon as seen from Earth and sees the south pole for days at a time. An Earth–Moon L2 halo (~15 d period, Az ≈ 12,500 km) sits ~50–75,000 km behind the far side, always in view of both Earth and the far side, like China's Queqiao.

These carry the high-rate optical trunks home. Mesh satellites reach them through crosslinks.

04LunaNet and DTN

Build to the LunaNet interoperability framework so that NASA, ESA, JAXA and commercial nodes interoperate: common link layers, a network time and navigation service, and Delay/Disruption-Tolerant Networking (Bundle Protocol). Bundles are stored at each node and forwarded when the next link comes up, so outages, occultations and cloudy ground stations cost delay, not data.

05Optical vs RF

Optical for crosslinks and the Earth trunk: tiny beams (µrad) give Gbps from watts and small telescopes, need no spectrum licence and leak almost nothing to radio telescopes. Downsides: precise pointing, acquisition time, and clouds on the Earth end.

RF for users and backup: S-band for surface-to-orbit and the navigation signal, Ka-band (around 23 / 26 GHz) for higher-rate proximity and trunk backup. Wide beams are forgiving and work with simple surface terminals.

06Protecting the radio-quiet far side

The far side is the only place near Earth shielded from human radio noise, and Daedalus is planned as a radio observatory. ITU-R RA.479 and the Radio Regulations (Art. 22) set aside the Shielded Zone of the Moon for radio astronomy, so relays that transmit RF over the far side work against that.

Optical-only links to far-side terminals and observatory backhaul.
Scheduled RF silence over the observatory, enforced in spacecraft software.
Narrow spot beams, with nulls toward the array; no transmissions in observatory bands.

07Orbit stability

Lunar gravity is lumpy: mascons make most low orbits decay or crash within months. "Frozen" inclinations (~27°, 50°, 76°, 86°) and ELFOs keep eccentricity and ω nearly constant, but Earth's pull and mascons still demand regular station keeping. NRHO and halo orbits are unstable and need small, frequent trim burns (of order a few m/s per year for Gateway's NRHO). Here, Keplerian layers use two-body propagation and the relays use CR3BP periodic orbits, both idealised.

08Build-out phases

Phase 1: 2 relays (NRHO + L2 halo) + 1 south ELFO. Far side reaches Earth; the south pole gets intermittent comms. Toggle layers to see this.
Phase 2: grow to 4–6 ELFO satellites for continuous south-polar comms and PNT (≥4 in view).
Phase 3: global Walker mesh plus surface networks around each city.

09The surface last mile

The horizon is close: √(2Rh) gives ~2.4 km for a 1.7 m observer and ~18.6 km for a 100 m tower, so two such towers link over ~37 km at most. Surface networks therefore use towers on crater rims (which also catch sunlight), point-to-point optical or mm-wave hops, and fiber laid along the highways between cities. Orbiting layers handle the gaps and mobile users.