CubeRover

NASA/Glenn Benson. Public domain (NASA / US government work).

Overview

A rover small enough to ride as one payload among several on a commercial lander has to give up almost everything a flagship rover carries in reserve: no separate cruise stage power, no spare mass for redundant strings, and a delivery contract that prices the trip by the kilogram rather than by the mission. Astrobotic Technology of Pittsburgh answered that constraint by treating rover mobility itself as the product. CubeRover applies the CubeSat unit convention to surface mobility: one U is a 10 x 10 x 10 cm payload volume carrying 1 kg, and the vehicle is offered in 2U, 4U and 6U sizes that scale volume and payload capacity while reusing the same power, thermal, structural, avionics and software systems [1]. Astrobotic sells it as a service rather than as hardware, alongside the Peregrine and Griffin landers that are meant to carry it to the surface [8].

Development was funded through NASA Small Business Innovation Research and a $2 million Tipping Point award made in September 2019 [3], [5]. No CubeRover has yet flown; Iris, the Carnegie Mellon student rover that rode Peregrine Mission One in January 2024, shares the CubeRover design lineage but is a separate vehicle, and Peregrine’s own propulsion failure 9 minutes after separation ended that mission before any lunar surface operations could be attempted [8].

Where CubeRover sits among small wheeled rovers

The class CubeRover competes in is defined less by any single design choice than by the absence of margin. A JPL survey of current robotic and autonomous surface systems places small commercial rovers such as CubeRover alongside the CLPS payload manifest as one of several converging efforts to fly cheap, small mobility rather than large flagship rovers [7]. The nearest published points of comparison are other micro-rovers built on the same scaling logic: MoonRanger, an 18 kg four-wheeled rover developed at Carnegie Mellon for lunar polar prospecting, whose thermal design work shows how tight the survivable band gets at this mass class, with a viable operating range of only -10 to 35 degrees C and permanently shadowed region excursions limited to about one hour by camera and sun sensor cold limits rather than by the rover’s own avionics [11]. CubeRover’s four independently actuated wheels with grousers sit at the opposite end of the wheel-design spectrum from a limbed platform like ATHLETE, whose 850 kg mass and harmonic-drive joints let it accept 20 to 45 kPa ground pressure because it can walk out of soft soil; a rigid, mass-constrained rover like CubeRover has no such option and must instead get its mobility margin from the wheel and grouser geometry alone [10].

That is also why the wheel is the one subsystem CubeRover has iterated in public. Discrete element simulation of small lunar rover wheels shows why grouser geometry is load bearing rather than cosmetic at this scale: reduced gravity weakens inter-particle friction in the regolith, so a wheel selected for adequate slope performance in 1 g can lose climbing ability under lunar gravity, and self-propelled slip can rise sharply even on level ground [9]. The grouser shape JAXA selected for the LEV-1 hopping rover’s own single wheel, chosen by resistive-force-theory modeling and confirmed against its January 2024 lunar operation, is one data point for how small-rover wheel design converges on similar answers under similar mass and gravity constraints [12]. Against that backdrop, Astrobotic’s own account of opening a dedicated regolith mobility lab holding about 19,000 kg of simulant, and of an outside group’s compliant wheel outperforming CubeRover’s own wheels in side-by-side testing there, is consistent with a vehicle still in the middle of that iteration rather than at a settled design [6].

Every vehicle number below comes from Astrobotic’s own specification material [1], [2]. Where the manufacturer’s documents disagree, both values are given. Test results are from NASA and manufacturer accounts of the same NASA KSC campaigns [3], [5]. No peer-reviewed engineering description of the platform has been published.

Specifications

Parameter	2U	4U	6U
Rover mass	4.6 kg	8 kg	10.6 kg
Payload capacity	up to 2 kg	up to 4 kg	up to 6 kg
Guaranteed payload envelope	20 x 10 x 10 cm	20 x 20 x 10 cm	30 x 20 x 10 cm
Additional internal volume	2100 cm3	3900 cm3	5700 cm3
Additional external volume	910 cm3	1768 cm3	2625 cm3
Payload energy allocation	100 Wh+	150 Wh+	200 Wh+

Manufacturer figures, from [1].

Parameters common to all three sizes:

Parameter	Value	Source
Payload power bus	28 V DC	manufacturer [1]
Payload thermal environment	-20 to +60 degrees C
Payload wired interface	RS-422
Payload data allocation	10 kbps per kilogram of payload
Payload wireless standard	802.11n WLAN
Payload data storage	32 Gb or more
Nominal speed	4 cm/s
Top speed	10 cm/s	manufacturer [2]
Slope limit	30 degrees
Obstacle diameter	nearly 15 cm

The two speeds come from different manufacturer documents and are not related to each other in either: 4 cm/s is the nominal speed on the datasheet [1], 10 cm/s the top speed on the product page [2].

Mission profile

No CubeRover has been assigned to a flight. The datasheet states support for several lander configurations, and a design mission of 8 Earth days [1], covering multiple kilometers in a single lunar day [2].

Iris, flown on Peregrine Mission One in January 2024 [1], originated in the same Carnegie Mellon CubeRover work but is a separate vehicle, and never reached the lunar surface because the lander itself failed in cislunar space [8].

Mobility

Four aluminum wheels, each independently actuated, with dust covers and seals [2]. There is no rocker or bogie in the published description, so the vehicle is a rigid four-wheel skid-steer platform, which is consistent with the point-turn testing performed [5]. That rigidity puts the whole mobility margin on the wheel and grouser design, in the way small lunar rover wheel studies generally find: performance is sensitive to slip, sinkage and grouser geometry at once, and a wheel adequate at one gravity level is not guaranteed adequate at another [9].

Mobility characterization was run in the Granular Mechanics and Regolith Operations laboratory regolith bin at NASA KSC Swamp Works, which holds about 120 tons of lunar regolith simulant [3]. Over 150 mobility tests were performed with 11 sets of wheels [5], a count NASA’s own account of the same campaign corroborates independently of the manufacturer [4]. Drawbar pull, slope and point-turn data were taken from the rover’s own sensors. Some wheel sets climbed 30 degree slopes and others turned in deep regolith, and two sets exceeded the test expectations; a second campaign on 30 June 2022 used a mass-offloaded vehicle in the same bin to represent mobility at lunar gravity [5]. A separate Astrobotic regolith lab has since tested at least one outside-developed wheel against CubeRover’s own, with the manufacturer describing the outside wheel as outperforming everything tested to that point, which places the current wheel set as one iteration among several rather than final flight hardware [6].

The 30 degree slope figure quoted as a vehicle capability [2] is the best result across 11 wheel sets rather than a qualified vehicle limit [5].

Power and energy

The vehicle is solar powered. Array power and battery capacity are not published; what is published is the energy made available to the payload, 100 Wh or more on a 2U and 200 Wh or more on a 6U [1].

The baseline vehicle does not survive the lunar night: mission duration is 8 Earth days [1]. Night survival is offered as a customization, and a separate Tipping Point contract funds wireless charging intended to let CubeRover endure the night [2].

Thermal

The thermal design is not published beyond the service it provides. All CubeRover sizes share the same thermal system [1], and the payload thermal environment provided is -20 to +60 degrees C. No radiator area, insulation stackup or vehicle operating temperature range is published. A comparably sized lunar micro-rover, MoonRanger, needed three levels of thermal modeling to hold a much narrower band and still limited its permanently shadowed region excursions to about an hour, which gives a sense of how little thermal margin a rover this small is likely to carry even where CubeRover’s own numbers are not public [11].

Compute and avionics

All CubeRover sizes share the same avionics and software systems [1]. No processor part, memory, radiation tolerance approach or software framework is published.

Autonomy

Navigation is teleoperated with a visual-inertial system, using wide angle cameras front and rear [2]. An onboard safeguard returns the rover automatically to its last known position. No stereo baseline is described, so depth would have to come from motion.

Communications

The published payload wireless standard is 802.11n WLAN, the payload data allocation is 10 kbps per kilogram of payload, and onboard storage is 32 Gb or more [1]. No link budget, radio part or direct-to-Earth capability is stated.

Payload and instruments

CubeRover carries no instruments of its own; the payload is the customer’s. Top and bottom mounted payloads are both supported [1]. Non-standard services are stated to be available on request.

Modes of operation

The datasheet describes the platform as customizable to mission needs, including lunar night survival and missions to craters and permanently shadowed regions [1], but no deployment sequence or operating mode set is given.

Technologies developed

The transferable product is the interface standard rather than the vehicle: a CubeSat-style unit convention for surface mobility, with a published specification fixing envelope, 28 V bus, RS-422, WiFi, data rate per kilogram and thermal environment, so that a payload developer sizes against a standard rather than against a bespoke rover [1]. The second is the wheel test dataset from NASA KSC: drawbar pull, slope and point-turn measurements across 11 wheel sets in 120 tons of simulant, at both Earth gravity and mass-offloaded to represent lunar gravity [3], [5].

What is not established

No CubeRover has flown, so every number above is a manufacturer specification or a ground test result, not flight data. The wheel and grouser geometry actually used has not been disclosed, and the manufacturer’s own regolith lab has since tested at least one outside wheel design that reportedly outperformed CubeRover’s, so the vehicle described here should be read as one point in an ongoing iteration [6]. Mass, top speed and payload capacity figures disagree between manufacturer documents from different years, and no peer-reviewed engineering paper on the platform exists to arbitrate between them [1], [2]. Thermal design, avionics, and the processor and software stack are described only as “shared across sizes,” with no released values to check that claim against.