A pressurized, shirtsleeves lunar vehicle engineered for 30–60 days of autonomous surface operations. Not a car — a mobile habitat, integrated into the Aegis orbital–surface logistics chain.
Open rovers solve early sortie missions. Aegis solves permanent lunar presence — eliminating the single greatest bottleneck in lunar exploration: EVA fatigue.
Full life support at 56.5 kPa / 34% O₂, the NASA exploration atmosphere. Crew works, sleeps, eats, and plans without suits. Productivity multiplied 4–6× over EVA-only concepts.
700 kg of water in deployable bladders, with food, tanks and racks arranged around a Zone C huddle, gives about 7 cm of water on five faces. Cuts an August-1972-class SPE to an emergency exposure; 10 cm would need about 1.0 t of water.
Planning table, science bench with glove box, full tool storage, maintenance bench, medical station with telemedicine. The rover is a workspace, not just transport.
Integrated with Aegis Station (orbit), Short-Hoppers, Long-Haulers, Tankers, and WOK prospecting rovers. Closed-loop mobility and water-logistics from orbit to surface.
Manual, semi-autonomous, or fully autonomous. Return-to-base and autonomous scouting modes. Forward-looking LIDAR terrain classification for path planning.
Fold-down medical bench, full CMO kit (diagnostics, surgery, IV), AED with cardiac monitoring, telemedicine link. Pressurized refuge during EVA incidents.
Every dimension derives from a single constraint: 2.1 m interior cabin height for standing room over 30–60 day missions. The rest follows through structural, thermal, and stability logic.
Forward is operational, aft is personal. A continuous 0.75 m aisle on the port side is the emergency egress path, stretcher route, and primary circulation spine. Dust stays at the door.
| Zone | Name | Length | Area | Functions |
|---|---|---|---|---|
| A | Forward Command | 1.6 m | 5.4 m² | Pilot/co-pilot stations, nav displays, forward windows, hand controllers |
| B | Central Workspace | 2.4 m | 8.2 m² | Planning table, galley, tool storage, science bench, comm station |
| C | Crew Quarters | 2.2 m | 7.5 m² | 3 berths (2 lower + 1 upper), hygiene closet, medical bench, lockers |
| D | Systems Bay | 1.2 m | 4.1 m² | ECLSS racks, PDU, spares, floor access panels |
| E | Airlock Vestibule | 0.6 m | 2.0 m² | Suitports (×2), dust management, HEPA, exterior hatch |
The baseline EVA interface uses suitports rather than a traditional airlock. The dust-covered suit never enters the cabin — crew enters from inside through a sealed hull port, seals, and detaches. On return, the suit docks and crew exits inside. This eliminates ~95% of dust ingress and saves 5–10 kg of atmosphere gas per 60-day mission. Ingress/egress takes 10–15 minutes vs. 30–45 minutes for a full airlock cycle. The vestibule can revert to traditional airlock mode if suitport seals fail.
Private sleeping berths (2.0 × 0.70 × 0.85 m) with sound-attenuating curtains, individual LED reading lights, and USB-C ports. Circadian lighting system transitions from 5000K blue-enriched (day) through 3000K warm (evening) to 1800K amber (night) on a 24-hour schedule. Zone C targets ≤45 dBA for sleep quality. No shower — body cleaning uses no-rinse wipes consistent with ISS and submarine practice. Exercise via EVA activity plus resistance bands and isometric protocols.
At polar sun angles a horizontal radiator is nearly the ideal radiator: the sun grazes it at 6°, loading it by only ~17 W/m² when clean. Rev B therefore carries an unshaded white roof radiator sized to dusty end-of-mission emissivity, and deletes the Rev A roof array tier that sat above it: a full-coverage array 0.4 m up blocks about 70% of the radiator's view of space and would not shade it at 6° anyway (the shadow needs a ~3.8 m overhang). Power when parked comes from the deployable wings; a sun-tracking mast array for driving is an open item. The roof plate remains the ASI node power and thermal interface shared with the hopper.
NMC/NCA chemistries were explicitly rejected. Thermal runaway at 140–170°C with violent oxygen release, fire, and HF gas is unsurvivable in a sealed ~66 m³ cabin. LFP (LiFePO₄) is stable above 250°C with slow, self-limiting failure and no fire. The mass penalty — ~926 kg pack vs. ~600 kg for NMC — is accepted. The battery sits underfloor at 0.75–0.95 m height, pulling the CG low.
The Aegis-Class Rover flight software is built on NASA's core Flight System (cFS): the cFE kernel, the standard NASA apps, and seven Aegis mission apps for thermal, power, life support, radiation, communications, navigation, and the autonomy executive. All hardware access goes through a hardware abstraction layer; a physics simulator stands behind it today, and the hardware side of that layer is a documented porting boundary. The mission apps are the same source in the simulation and on hardware. The simulator's clock is the mission's: the image runs the world at a pace the ground sets, in lockstep with the flight software's scheduler, so a 37-minute leg is driven in under four minutes and two runs agree to the tick.
Thermal, power management, life-support atmosphere, radiation, communications monitoring, navigation, and an autonomy executive that owns the vehicle mode, on the core Flight System. Every command is length- and range-checked before it is acted on; every app reports its state as events the ground can read.
Limits are watched by the stock cFS Limit Checker and answered by Stored Command sequences: shelter deploy, heaters off, scrubber to maximum, non-critical load shed. Each response ends in a defined vehicle mode; the radiation shelter mode is cleared only by crew or ground, and recovery re-arms the engine so a returning fault fires again. Four more watchpoints compare a mechanism with its command: a heater, a battery contactor, the shelter and the fuel cell, each answered through the same sequences.
A waypoint catalog as a flight table, a guidance state machine, and a heading-and-speed controller that commands body-frame velocity through the abstraction layer, which a six-wheel chassis realises: counter-phase steering on a 6 m minimum radius, the speed bounded by traction on the slope and by the motors' power. A destination is offered only if the battery can get the rover there and back with reserve, directly or via one recharge stop at a LUNET node. The rover stops and holds if its pose sensor is lost.
The cFS image runs against a physics simulator (radiator and cabin zones, four battery modules behind contactors, wings and a fuel cell, cabin atmosphere, a six-wheel chassis on procedural terrain, LIDAR, radiation). Each fault is injected into the simulated plant, never into the software, and traced from sensor to Limit Checker to Stored Command to actuator to mode change. Every leg is recorded as a tape of telemetry fields that replays and draws. Standalone tests gate every change; the live image is the final check.
A browser console on a CCSDS-to-WebSocket bridge: summary, pilot display, thermal, power, life support, and the event log. Every value shown is telemetry and every control is a command the flight software implements. Scripted scenarios run the four fault classes with console-driven recovery between them.
The hardware abstraction layer is a documented contract: sensor and actuator structures with units and frames, return codes, timing budgets, and a compiling hardware backend template with a porting guide. Filling it in is the next layer of evidence, and it needs a chassis and avionics partner.
One 4.95 km leg, driven at ten times real time in November 2026, with four faults injected into the simulated vehicle on the way. A zone heater sticks on against its command: the Limit Checker catches the disagreement after three samples, a stored sequence turns the heaters off, the executive enters THERM_SAFE and the leg goes on. A battery module's contactor trips: the non-critical loads are shed and the executive enters LOADSHED; the operator closes the module, restores the loads and re-arms. The pose sensor fails: the navigation app stops the vehicle after ten failed reads and aborts the leg, and a second GOTO finishes it when the sensor returns. A wheel locks: the speed halves and the currents rise, and nothing on board watches for it yet. Every element in the frames below is a telemetry field from the running software.
The software proves the mode logic, the fault-response path, the navigation gate, and the ground loop. It stops at the abstraction layer. These are the systems it does not cover.
Guidance ends at a body-frame velocity command that the chassis realises. Wheel and steering control are the chassis's own, modelled in the simulator and read back as six stations, not flight software. Path planning, obstacle avoidance, and localization are not implemented; the LIDAR fan and SLAM health are displayed to the operator, not acted on; a locked wheel has no on-board response.
The flight software commands the CO₂ scrubber speed, the power-distribution switches, the battery contactors, the wings and the fuel cell. Oxygen generation, pressure control, water, and the thermal-loop pumps have no commands behind the abstraction layer yet, and a scrubber that stops scrubbing is answered by the high-CO₂ response, which assumes it works. A load shed does not stop a leg, and a lost pose sensor leaves the last pose on the telemetry unflagged.
The communications app monitors the link. There is no radio control, no protocol stack, no encryption, and no command authentication; the console is protected only by binding to the operator's machine.
No device drivers, no target other than Linux, no state that survives a processor restart, no redundancy, and no verification package beyond the tests in the repository. Flight software and vehicle architecture remain with Aegis Station.
The crew module is designed to mate to a third-party chassis from the maturing open-architecture lunar rover market. The reference configuration below represents the operational envelope and interface requirements the crew module brings to a candidate platform; specific chassis selection and final parameter values follow once a platform partnership is established around this envelope.
The reference platform provides six independently driven and steered wheels on three axles — no driveshafts, differentials, or transfer cases. In-hub motors with harmonic-drive reduction are the baseline expectation, keeping mechanical complexity and dust-exposed mechanisms off the chassis. A candidate platform need not match this mechanism, but must hold the crew module within the envelope the figures below describe.
The reference suspension is a semi-active double-wishbone at all six stations: titanium control arms, Ni-Ti superelastic coil springs (fatigue-immune across −170°C to +120°C), and electromechanical ball-screw dampers providing passive damping (back-EMF), semi-active rate control, and active ride-height adjustment (±100 mm) — with no working fluids anywhere. This fixes the ride-quality envelope the crew module assumes: ~1.1 Hz ride frequency (lunar-gravity corrected) for automobile-like crew comfort, electronic anti-roll via differential damper control, and active cabin leveling on cross-slopes up to ±10°. A candidate platform is free to meet this differently, but must keep the crew module within that comfort and leveling envelope.
The reference drive system uses six independent motors for per-wheel torque vectoring, slip detection and limiting (5–25% threshold by mode), electronic stability control, and regenerative ABS — across Highway (5% slip, equal torque), Off-Road (15% slip, adaptive), Rock Crawl (25% slip, per-wheel override), and Emergency (limits relaxed) modes. The governing constraint the crew module imposes is traction, not motor power: with a traction coefficient of 0.5 and rolling resistance of 0.15–0.25, the slope limit is ~21° on firm and ~15° on soft regolith regardless of mass, because traction and slope demand both scale with weight. What the 10.5 t Rev B mass does change is ground pressure (12 kPa on 0.4 m wheels), speed at a given power, and energy per kilometre (1.04 kWh/km, 77 km per charge on firm ground, 46 km soft). A candidate platform must deliver stable, slip-managed torque control within the traction margin and meet the ≤7 kPa ground-pressure requirement at Rev B mass.
The figures here represent design-baseline expectations consistent with currently demonstrated lunar mobility platforms and serve as the integration starting point for partner discussions.
Defense-in-depth for the most energetic system on a crewed pressurized vehicle. Every tier is independent — each layer assumes the previous has failed.
LFP inherent stability (olivine cathode stable >250°C, no oxygen release). CID disconnects on overpressure. PTC limits current >80–90°C. Ceramic-coated separator maintains integrity beyond polymer melt point.
Sealed enclosure contains worst-case multi-cell vent (2–5 bar). Dedicated vent line to hull exterior — not cabin. Ceramic fiber thermal barriers limit adjacent module surface to <60°C during 200°C transient. Pyrotechnic contactor for permanent irreversible disconnect.
BMS monitors every cell at 1 Hz normal / 10 Hz anomaly. Graduated response: Advisory (log + monitor) → Caution (reduce rate, alert crew) → Warning (disconnect module, close thermal valves) → Emergency (fire pyro, auto-vent, isolate bay, master alarm).
ECLSS gas sensors (CO, HF, electrolyte vapor, H₂). Battery bays under negative pressure during vent. Fixed fire suppression (Novec 1230 clean agent). Crew SCBA at forward/aft stations, 30-second donning time.
Carry sensors, drills, and ISRU payloads. Cargo module hauls processing equipment to sites identified by WOK prospecting rovers.
Haul equipment, deploy robotics, serve as mobile command post. Train-style module configuration allows mission-specific arrangements.
Move crews safely between outposts, sites, domes, and landers. Hopper collar mate at the aft dock; no EVA for crew transfer.
Pressurized visibility modules optional. Science bench with sealed glove box for regolith sample handling without cabin contamination.