CubeRover

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

Overview

A rover that rides as one payload among several on a commercial lander gives up most of what a flagship rover keeps in reserve: spare mass for redundant strings, a dedicated cruise stage, and a delivery contract priced by the mission. Astrobotic Technology of Pittsburgh made rover mobility 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 share one structural, power, thermal, avionics and software design [2]. Delivery is sold as a service at 4.5 million US dollars per kilogram of payload, with 0.5 W of continuous power and 10 kbps per kilogram included [3].

No CubeRover has flown. The vendor’s own 2021 poster self-assessed the vehicle at TRL 4 to 5, with vibration, thermal vacuum and EMI testing still planned [2]. The vendor’s 2023 datasheet still states that no CubeRover has operated on the lunar surface [1]. The ground evidence is one set of regolith simulant mobility campaigns at NASA Kennedy [4], [5], and the performance figures in vendor material are specifications, not qualification results [1], [3].

Iris, the Carnegie Mellon student rover on Peregrine Mission One in January 2024, shares the lineage but is a separate vehicle [8]. Peregrine lost its landing capability when a helium pressure control valve failed and the oxidizer tank ruptured about 92 minutes after separation, so no surface operations were attempted [8].

Design space and where CubeRover sits

The small lunar rover class has converged on a few choices. The vendor offers 4.6 to 10.6 kg vehicles, solar powered, with teleoperated navigation [1], [3]; MoonRanger at 18 kg uses the same mass-limited approach [11]. A JPL survey of robotic and autonomous surface systems places small commercial rovers beside the CLPS manifest as one of several efforts to fly cheap, small mobility [7]. Earlier JPL micro-rover work, from Sojourner to nano-rovers, set the template for what a rover of this size can carry [26].

The flown or near-flown comparison points differ in what they have demonstrated:

• MoonRanger, an 18 kg four-wheeled Carnegie Mellon rover for polar prospecting, has a published thermal model giving a surface operating band of -10 to 35 degrees C and about one hour of roving in permanently shadowed regions, limited by the cold soak limits of its lasers and sun sensors; the paper is modeling only, with no thermal vacuum data [11].
• SORA-Q, the transformable nano rover from JAXA, TOMY, Sony and Doshisha University, was deployed from the SLIM lander [28].
• LEV-1, JAXA’s hopping rover, drove its single grouser wheel on the Moon after SLIM landed in January 2024, which is the only lunar surface wheel result for a rover in this size range in the held record [12].
• The three CADRE rovers, suitcase-sized and solar powered, shipped in February 2025 to Intuitive Machines for the IM-3 lander; the held source gives no results [25].
• ATHLETE sits at the other end of the wheel design spectrum. At 850 kg it can accept 20 to 45 kPa ground pressure because it can walk out of soft soil; the paper gives design values and no mobility measurements [10]. A rigid small rover has no such option.

Early lunar practice sets the baseline. Wheeled vehicles worked on Apollo and Luna soil when ground contact pressure stayed below about 7 to 10 kPa, and surface roughness rather than power capped speed [18]. The Apollo rover wire mesh wheel was sized with Bekker analysis to a 4.5 cm nominal deflection, and the analysis overpredicted the wheel: meeting the 25 degree climb requirement implied about 59.5 percent mean slip [29]. Single-wheel tests of the flight wheel in simulant found pull coefficient and efficiency insensitive to speed and wheel load, with soil strength the controlling variable, and gave Apollo slope limits of about 19 degrees in nominal soil and 23 degrees in firm soil; the tests were at 1 g in an air-filled bin [30]. The Apollo-era wheel test plan for the dual-mode rover specified load-deflection and soil tests before qualification but is a plan, not a set of results [19].

Wheel and terramechanics evidence

The wheel is the subsystem that CubeRover has iterated, and the literature on small lunar wheels shows why. Semi-empirical wheel-soil theory descends from dimensionless analysis of pneumatic tire data on terrestrial soils, with soil strength measured by cone index [17]. The lunar numbers that feed any model are a friction angle near 49 degrees and cohesion near 1.6 kPa over the top 60 cm [23], within a simulant requirement envelope of cohesion 0.1 to 1 kPa and friction angle 30 to 50 degrees [22]. A review of Apollo experience concluded that the Bekker equations failed for small rovers in one sixth gravity [18].

Discrete element simulation of a 250 mm lugged wheel for a lunar micro rover found that reduced gravity weakens inter-particle friction, raises self-propelled slip to 26.3 percent on level lunar terrain, and leaves the selected wheel unable to climb 25 degrees; the lunar-gravity results have no reduced-gravity measurement behind them [9]. A continuum SPH model of a 137 kg six-wheel Yutu rover found that larger rim diameter and greater width both lower the slip needed to climb a slope, taller grousers help at a cost in torque, and doubling grouser count from 20 to 40 changes almost nothing; it is a preprint validated only against 1 g single-wheel data [15].

Small skid-steer rovers have a second problem that a slope test does not show: sinkage during point turns. Single-wheel tests in GRC-1 simulant found that wider, larger wheels turned best, that 45 degree diagonal grousers improved turning without added mass but raised sinkage by about 30 percent, and that drawbar pull did not change significantly [13]. A comparison of a 200 mm grouser wheel in Toyoura sand and FJS-1 simulant across slip ratios of 0 to 50 percent and loads of 5 to 70 N shows that the choice of test soil changes the result, and it used reduced vertical load, not reduced gravity [14]. Field data from a 21.63 kg rover in Tohoku silica sand showed that slip on a grousered wheel must be computed against an effective radius that includes grouser height, or driving appears as negative slip [16].

JAXA’s LEV-1 grouser shape was derived with resistive force theory, tested in silica sand and DEM, and confirmed qualitatively by about two hours of wheel-driven operation on the Moon after seven hops [12]. Commercial wheel development is also marketing-stage: a hyper-deformable wheel for the FLEX rover is described only by requirements, not measurements [24]. Against this background, Astrobotic’s 2021 announcement of a regolith lab holding about 19,000 kg of GRC-1 simulant, and of an outside group’s compliant wheels that “perform better than any wheels we have tested to date”, records a wheel set still under iteration; the claim has no metric or baseline behind it [6].

Excavation rovers at Kennedy share the bin and the soil. RASSOR 2.0, a 66 kg excavator, selected 17 inch wheels over tracks and smaller wheels in 1 g tilt table tests in BP-1 simulant at a single 25 degree angle with three attempts per configuration [20]. Its successor IPEx raised a common wheel, arm and drum actuator design to TRL 5 with a four-test qualification sequence that included dirty vacuum, all at 1 g [21]. CubeRover was driven alongside RASSOR to rehearse driving into and out of a dug trench [5].

Test campaigns

In December 2020, CubeRover ran more than 150 mobility tests with eleven sets of wheels in the 120 ton regolith bin of the Granular Mechanics and Regolith Operations laboratory at NASA Kennedy, under a 2 million dollar Tipping Point contract [5]. NASA’s account of the same days reports dune crossing, point turns and trench wall ascent and descent, with sensor, motor and power behavior observed in dust [4]. Drawbar pull, slope and point-turn data were collected [5]. Two of the eleven wheel sets exceeded expectations, and some sets climbed 30 degree slopes while others turned in deep regolith [5].

The 30 degree slope and 15 cm obstacle figures on the product page are stated for the product line [3]. The test report supports 30 degrees only for some wheel sets [5]. A further campaign on 30 June 2022 used the same bin with a stripped mobility chassis; no published results from it are held.

No values for drawbar pull, slip or dust ingress appear in either account, and the tests were at 1 g in a bin [4], [5]. A flight qualified CubeRover planned for 2022 was not delivered [4].

Specifications

All values are manufacturer figures. They disagree between documents.

Parameter	2U	4U	6U
Rover mass, 2023 datasheet	4.6 kg	8 kg	10.6 kg
Rover mass, 2021 poster	3 to 4 kg	6 to 8 kg	10 to 12 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+

Mass and envelope rows are from the datasheet [1] except the poster row [2].

Parameter	Value	Source
Payload power bus	28 V DC	[1]
Continuous payload power	0.5 W per kg, 10 W peak	[2]
Payload thermal environment	-20 to +60 degrees C (datasheet), +10 to +30 degrees C (poster)	[1], [2]
Payload wired interface	RS-422	[1]
Payload wireless	802.11n WLAN	[1]
Payload data allocation	10 kbps per kg	[1]
Payload data storage	32 Gb or more	[1]
Nominal speed	4 cm/s	[1]
Top speed	10 cm/s	[3]
Slope and obstacle	30 degrees, nearly 15 cm	[3]
Mission duration	8 Earth days	[1]

NASA’s account gives a rover mass around 9 pounds, which matches neither vendor figure [4].

Subsystems

Mobility. Four independently actuated wheels, with dust covers and seals [3]. The poster credits the motors with flight qualification because they are used on the Perseverance rover and the batteries because they will fly on SLS; that is inherited heritage, not qualification of this vehicle [2]. No suspension is described in the held material.

Power. Solar. Array power and battery capacity are not published. The poster does not claim or deny lunar night survival, and the datasheet mission duration of 8 Earth days implies none [1], [2]. For comparison, NASA’s crewed Lunar Terrain Vehicle thermal work starts from a night of at least 85 hours [27].

Thermal. The chassis is wrapped in multi-layer insulation with a top-mounted radiator as the thermal ground [2]. No radiator area or vehicle temperature range is published.

Compute and navigation. A custom single board computer handles command and data handling, image processing, motor control and power, and front and rear 5 megapixel monocular sensors provide navigation [2]. Navigation is visual-inertial teleoperation with an automatic return to the last known good location on off-nominal conditions [3].

Communications. The link to the lander is 802.11n WiFi at 10 kbps per kilogram [1], [2]. No direct-to-Earth capability, radio part or link budget is stated.

What is not established

No CubeRover has flown, so every vehicle number is a vendor specification or a ground test result [1], [2]. Wheel and grouser geometry for the eleven wheel sets, and for the wheel the vendor intends to fly, are not published, and no test report gives drawbar pull, slip or sinkage values [5], [6]. The 30 degree and 15 cm capabilities are best-case wheel results generalized to the product [3], [5]. Vibration, thermal vacuum and EMI qualification were future work in 2021 and no result is held [2]. Mass, speed and thermal environment disagree between vendor documents of different years, and no peer-reviewed engineering description of the platform exists to arbitrate [1], [2], [3]. The reduced-gravity behavior of small grousered wheels rests on simulation and 1 g bench tests, and the one lunar surface wheel result, LEV-1, is a single hopping rover with qualitative confirmation [9], [12]. Thermal design, avionics and software are described only as shared across sizes. Power sizing and lunar night capability are not stated.