The Small Frontier: Trends Toward Miniaturization and the Future of Planetary Surface Rovers

Abstract

The robotic exploration of space began only five decades ago, and yet in the intervening years, a wide and diverse ecosystem of robotic explorers has been developed for this purpose. Such devices have greatly benefited from miniaturization trends and the increased availability of high-quality commercial off-the-shelf (COTS) components. This review outlines the specific taxonomic distinction between planetary surface rovers and other robotic space exploration vehicles, such as orbiters and landers. Additionally, arguments are made to standardize the classification of planetary rovers by mass into categories similar to those used for orbital satellites. Discussions about recent noteworthy trends toward the miniaturization of planetary rovers are also included, as well as a compilation of previous planetary rovers. This analysis compiles relevant metrics such as the mass, the distance traveled, and the locomotion or actuation technique for previous planetary rovers. Additional details are also examined about archetypal rovers that were chosen as representatives of specific small-scale rover classes. Finally, potential future trends for miniature planetary surface rovers are examined by way of comparison to similar miniaturized orbital robotic explorers known as CubeSats. Based on the existing relationship between CubeSats and their Earth-based simulation equivalents, CanSats, the importance of a potential Earth-based analog for miniature rovers is identified. This research establishes such a device, coining the new term ‘CanBot’ to refer to pathfinding systems that are deployed terrestrially to help develop future planetary surface exploration robots. Establishing this explicit genre of robotic vehicle is intended to provide a unified means for categorizing and encouraging the development of future small-scale rovers.

1. Introduction

The nature of space exploration has fundamentally changed in the last five decades. Nearly fifty years ago, the initial U.S. manned lunar program concluded with Apollo 17. Six successful manned landings have taken place, and twelve human beings have strolled across the surface of Earth’s only natural satellite. By 1973, almost USD 25 billion had been spent to complete this endeavor, which would be equivalent to around USD 257 billion today [1]. The U.S. had planned to return human explorers to the Moon in 2024 with a more sustainable program, known as Artemis, which would allow for longer-term occupation on the surface and in lunar orbit than the original Apollo missions. While both programs aimed to put boots on the Moon, architects of the Artemis program sought to engage the private sector for much of the required technology [2]. NASA has already awarded contracts to private space companies SpaceX and Blue Origin to develop the manned landers for Artemis III and V, respectively [3]. These fixed-price contracts and the general competition between the companies are intended to drastically reduce the cost of future manned moon landings. Privatization of space exploration has also been expanded to robotic missions. NASA’s Commercial Lunar Payload Services (CLPS) has partnered with a number of private companies to deliver NASA payloads to the moon at an unprecedented reduction in cost [4]. By lowering costs to explore planetary bodies, such as the moon, there has been a noted increase in mission cadence. Unlike fifty years ago, small-scale research organizations and private companies are now able to design, build, and deploy planetary exploration missions in this new era of space exploration.

The following material examines the trends of miniaturization in planetary surface rovers and predicts the need to establish terrestrially-based simulation tools to support future developmental pipelines. The second section of this research describes the taxonomy of planetary surface exploring robots, establishing the difference between orbiters, landers, and rovers. Planetary rovers are then further categorized by size in the following subsection to establish mass-based classes. The third section discusses specific case study rovers as examples of their mass class in addition to categorizing all major historical planetary rovers by year, mission, mass, distance traveled, and suspension or mobility technique. Additional commentary concerning the established relationship between CubeSats and CanSats has been included in the fourth section. This analysis concludes that an additional terrestrial simulation tool for mobile plenary rovers already exists and coins the term ‘CanBot’ to refer to such devices. The final sections discuss the compiled information and conclude the work.

2. Historical Planetary Surface Exploration Robots

When considering the exploration of space and other planetary bodies beyond the Earth, historically, robotic systems have largely preceded manned human exploration. This was true in the 1960s, when the former Union of Soviet Socialist Republics (USSR) and the United States of America (USA) competed to first crash—and then eventually softly land—robotic vehicles on the surface of Earth’s Moon. The USSR’s Luna 2 probe, featured in Figure 1a, demonstrated the first successful crash landing in September 1959, while the US Ranger 7 robotic mission, depicted in Figure 1b, captured detailed images of the lunar surface in February 1965 before crashing into the Moon. Additional Ranger probes targeted potential crewed landing sites, including Ranger 8 in the Sea of Tranquility, where Apollo 11 would eventually land with the first humans to step foot on the Moon [5].

In addition to mapping the lunar surface, robotic missions were also critical in testing orbital procedures and the process of landing softly on the Moon. The Surveyor probes were America’s first trial run in landing on another planetary body. In 1965, Surveyor 1 landed off target but still demonstrated, for the first time, that soft landings could be completed successfully. Robotic landers—from Surveyor 2, which crashed unexpectedly, to Surveyors 3, 5, 6, and 7, which were able to land gently on the Moon’s surface—provided invaluable information about how to land and about surface conditions on the Moon. By the end of the decade, the USA would successfully place two human astronauts on the surface of the Moon, but this was only made possible by the lessons learned from launching and crashing these earlier robotic vehicles on Earth’s only natural satellite [5].

2.1. Planetary Surface Robot Taxonomy

Robotic systems such as the Ranger and Surveyor probes were important trailblazers in the quest to both better understand the Moon and put the first human-occupied boots onto this stellar body. However, even at this early stage of robotic surface exploration, there were some clear taxonomic differences between the various robotic systems being developed to explore the lunar surface. Robotic devices such as the US Lunar Orbiter 1 (Figure 2a), which successfully entered lunar orbit in 1966, had significantly different science objectives than the Surveyor 1 probe (Figure 2b) previously mentioned. While the Surveyor missions were intended to physically land on the surface of the Moon and interact with the local regolith, the Lunar Orbiters primarily focused on collecting images of as much of the lunar surface as possible [5].

Fundamentally, the Lunar Orbiter demonstrated a macroscopic investigative approach where the entire moon was the target of inquiry, whereas the Surveyor missions focused on the relatively small area of the moon in which the probe had landed. This dichotomy presents the first two major taxonomic categories for planetary surface-exploring robots. In the first case, Lunar Orbiter 1 is an example of an orbiter. As the name suggests, these devices are intended to maintain a high orbit around a planetary body or the Moon. Their science objectives and useful lifespans are characterized by a large physical separation between themselves and the surface of the investigative target. Although in some cases orbiters may be directed to fly through the upper atmospheres of target planets, and their end-of-life (EOL) procedures typically involve crashing into a planetary body, orbiters do not generally interact with the surface of a planet as part of the scientific mission.

By contrast, landers, such as the Surveyor probes, are characterized by conducting planetary science that requires interacting with the surface of the target planet or Moon. Landers, therefore, include additional components and capabilities to enter a planet’s atmosphere and perform some form of controlled ‘landing’ on the planetary body. For example, the Surveyor landers could be distinguished from the Lunar orbiters by the addition of landing legs and scientific instruments specifically intended to investigate the composition of the lunar surface, such as an alpha-scattering experiment. This scientific instrument had to be physically conveyed to the surface of the Moon using the lander’s robotic arm, and had to remain in contact with the lunar surface to complete an analysis of the chemical composition of the regolith [9]. Historically, landers that have followed the initial Surveyor probes have employed similar mission profiles, where the scientific objectives have required physical contact and interaction with a planetary surface.

In addition to robotic landers, an additional taxonomic group of planetary surface robots also persists. Where landers are primarily immobile after completing their landing, planetary rovers are characterized by completing science objectives that require mobility. Much like landers, rovers are equipped with tools to closely investigate the surface of their target planet or moon, which often involve physically touching and interacting with surface regolith. But unlike landers, which are primarily constrained to investigate what can be seen and reached from a stationary position, rovers can actively move across the surface of an alien world to find new locations of scientific interest. There are some limited examples of landers demonstrating short relocating “hops” after their initial landing, but such translations have historically been only a few meters in distance, such as with Surveyor 6. To be a true rover, such as the USSR’s Lunokhod 1 (Figure 2c), significant surface mobility must be demonstrated and required to complete the mission’s objectives. Where Surveyor 6 demonstrated a single 2.4 m hop [9], Lunokhod 1 traversed 10.5 km as the first lunar rover, and Lunokhod 2 nearly quadrupled this distance after traveling 37 km in 1970 and 1973, respectively [11]. All three taxonomic classes of planetary exploration robots are shown in Figure 2.

2.2. Planetary Rover Sizing

Planetary rovers come in a diverse range of sizes and shapes. Beginning in 1997 with the Sojourner rover, the trend with NASA rovers has been to build bigger and more capable machines. Sojourner landed on Mars in 1997 with a mass of 15.5 kg, the Mars Exploration Rovers (MER) landed on Mars in 2004 with a mass of 174 kg, and the Mars Science Laboratory (MSL) landed on Mars in 2011 with a mass of 899 kg [12]. However, with the recent success of much smaller and non-traditional mobile vehicles such as SORA-Q or even the Ingenuity Rotorcraft, a new paradigm of small-scale rovers may be dawning. This emerging paradigm mirrors previous developments in microsatellites, such as CubeSats, which will be discussed later in this work. And just as microsatellites have changed the landscape of orbital science, micro-rovers may provide a new way to explore planetary bodies that do not require complex and high-cost missions.

Various categorization scales could be employed to discern between the wide spectrum of sizes available in planetary rovers. In their analysis of planetary rovers, Ellery establishes “mini-rovers” from 50 to 100 kg, “micro-rovers” from 10 to 50 kg, and “nano-rovers” from 5 to 10 kg. However, this “arbitrary” scale, as Ellery called it, would group the MER and MSL in the same category of “macro-rovers” that are greater than 100 kg in mass [11]. If, instead, the typical size classifications employed for orbital satellites are transferred to the planetary rover field, then MER could be considered a “mini-rover” as it was between 100 and 500 kg, while MSL would be classified as a “macro-rover” because it was between 500 and 1000 kg [13]. This latter classification scale will be adopted for the remainder of this research because of its perceived improvement in separating historical planetary roving vehicles.

Table 1. Robotic space probe.

Name	Ellery (Rovers) [11]	Soyer (Satellites) [13]
Large *	Na	>1000 kg †
Macro/Medium	>100 kg	500 kg–1000 kg
Mini	50 kg–100 kg	100 kg–500 kg
Micro	10 kg–50 kg	10 kg–100 kg
Nano	5 kg–10 kg	1 kg–10 kg
Pico	Na	100 g–1 kg
Fempto	Na	<100 g

* Text color indicates named size discretizations; † Soyer’s mass discretizations used in following tables.

compares the two mass-based classification systems, with each classification assigned a unique color code for later reference.

While macro and mini-scale rovers, such as MER and MSL, have shown great promise in furthering planetary science and exploration, this work will focus primarily on micro and smaller-scale rovers. The importance of these small-scale planetary explorers exists in a number of pertinent areas. Foremost, smaller rovers are cheaper to fly and operate; they have, therefore, been historically utilized for highly experimental and ground-breaking work. A historically significant example of this characteristic was the Sojourner rover, which was primarily a technology demonstration that was added to the primary Pathfinder mission. Arguably, Sojourner (Figure 3a) represented a secondary mission payload, as the rover could have failed, and the primary Pathfinder mission would still have moved forward. Representing the original micro-rover, Sojourner was able to demonstrate that wheeled mobility was not only possible on Mars but could be a valuable asset to exploration. Following the success of Sojourner in 1997, nearly every major nationally developed planetary rover that has come after has employed the same springless rocker-bogie suspension system. This low-cost rover undoubtedly had a major and long-lasting impact on planetary exploration.

Nearly two decades later, another ‘technology demonstration’ may be the leading edge of a new planetary exploration paradigm. Similar to Sojourner, the Ingenuity rotorcraft, depicted in Figure 3b, was added to the Mars 2020 (Perseverance) mission as a non-critical secondary payload. And similar to Sojourner, Ingenuity has significantly exceeded expectations in its capabilities and potential utility to planetary surface exploration. While initially intended only to demonstrate the feasibility of powered flight on Mars, the miniature rotorcraft completed a total of 72 flights and traversed a distance of approximately 17 km before suffering a fatal crash [14]. This marked success has led to interest in future multirotor exploration robots, such as the flagship Dragonfly mission, which will attempt to land a nuclear-powered rotorcraft on the surface of Saturn’s largest moon, Titan [15].

Secondary mission payloads such as Ingenuity and Sojourner represent a unique low-cost, high-risk, high-reward exploration mentality. While these missions make use of experimental technologies and approaches that have a higher chance of failure, their relatively small size and low launch cost allow them to be added onto larger missions at a generally low incremental cost. Perhaps this characteristic of being low-cost and potentially high-reward has led to the significant shrinking of many modern privately developed rovers.

Figure 3. Examples of small-scale experimental rovers, including (a) Sojourner Rover [16], and (b) Ingenuity Rotorcraft [17].

Figure 3. Examples of small-scale experimental rovers, including (a) Sojourner Rover [16], and (b) Ingenuity Rotorcraft [17].

3. The Miniaturization of Planetary Rovers

Notably, dedicated rover missions such as MER, MSL, and Mars 2020 continue to result in progressively larger robotic vehicles. However, a more modern trend, especially in privately funded spacecraft, has been the miniaturization of space exploration vehicles. CubeSats represent an obvious example of this trend for orbital vehicles, but this trend has also held true for robotic planetary surface exploration. What is also significant is the fact that, with the exception of the previously listed dedicated rover missions, all other modern rovers have been part of a larger mission. In many cases, these rovers represented noncritical secondary mission payloads, which can be added to larger missions because of the rover’s diminutive size. Some relevant examples of these small-scale rovers have been described in the following section, and most historical robotic roving vehicles have been compiled in

Table 2. Compilation of relevant historical planetary surface rovers.

Name	Year	Mass (kg)	Outcome	Deployed from
Lunokhod 1	1970	756 [11]	Success	Luna 17
Prop-M (1)	1971	4.5	Failure	Mars 2
Prop-M (2)	1971	4.5 [11]	Unknown	Mars 3
Lunokhod 2	1973	840 [11]	Success	Luna 21
Sojourner	1997	10.5 [11]	Success	Pathfinder
PLUTO	2003	0.86 [18]	Failure	Beagle 2
MER A	2004	174 [11]	Success	Na
MER B	2004	174 [11]	Success	Na
MINERVA	2005	0.591 [19]	Failure	Hayabusa 1
MSL	2012	899 [12]	Success	Na
Yutu 1	2013	140 [20]	Success	Chang’e 3
MINERVA II-1A	2018	1.1 [21]	Success	Hayabusa 2
MINERVA II-1B	2018	1.1 [21]	Failure	Hayabusa 2
Yutu 2	2019	140 [22]	Success	Chang’e 4
Pragyan 1	2019	27 [23]	Failure	Chandrayaan-2
Mars 2020	2021	1025 [24]	Success	Na
Ingenuity	2021	1.8 [25]	Success	Mars 2020
Zhurong	2021	240 [26]	Success	Tianwen-1
Pragyan 2	2023	27 [23]	Success	Chandrayaan-3
Rashid	2023	10 [27]	Failure	Hakuto-R Mission 1
Sora-Q (1)	2023	0.25 [27]	Failure	Hakuto-R Mission 1
COLMENA	2024	0.06 [28]	Failure	Peregrine Mission One
Iris	2024	2 [29]	Failure	Peregrine Mission One
LEV 1	2024	2.1 [30]	Success	SLIM
LEV2 (Sora-Q 2)	2024	0.25 [30]	Success	SLIM
Jinchan	2024	5 [31]	Success	Chang’e 6
YAOKI	2025	0.5 [32]	Failure	IM-2 (Athena)
Tenacious	2025	5 [33]	Failure	Resilience (Hakuto Mission 2)

Text color indicates mass discretizations as described by Soyer in Table 1.

. Entries in Table 2 have been color-coded, as in Table 1, to indicate the size of the rover based on pre-existing satellite size classification.

3.1. Prop-M: The Original Nano-Rover

Starting as early as the 1970s, scientists recognized the potential for low-cost mobile robotic platforms to conduct planetary exploration. The first nano-rover to land on another planet was Prop-M, which was developed by the Soviet Union’s space agency. The rover had a mass of approximately 4.5 kg and was developed in just 1.5 years. A six-segment boom was responsible for placing the rover on the surface of Mars, following the landing of its mothership lander. The rover was powered, communicated, and sent telemetry data through a 15-meter umbilical cable, which limited the rover’s mobility [34]. It moved using two rotating skis that allowed the rover to travel at a slow speed of up to 1 m per hour while consuming approximately 5 watts of power [35,36]. The Prop-M rover was equipped with two instruments, namely, a densitometer and a penetrometer [34]. The densitometer would have been used to measure the soil’s density, while the penetrometer was intended to measure its bearing strength [37]. The rover was also equipped with a rudimentary obstacle avoidance system that used tactile sensors to detect obstacles in its path [35]. The Prop-M rover was launched on four missions to Mars: Mars 2 and Mars 3, and later, the Mars-6 and 7 landers. However, none of these rovers was entirely successful [37]. The Mars 2 and 6 landers crashed on Mars, and the Mars 3 lander lost communication with Earth shortly after landing. Mars 7 missed Mars entirely. While it remains unknown if the Prop-M rover associated with the Mars 3 mission was ever successfully deployed, this tiny rover was a pioneering effort in planetary exploration. It was the first rover to be designed to walk on another planet, and it paved the way for future rovers [36].

3.2. Iris: A More Modern Nano-Rover

In addition to participating in NASA’s CLPS program, the private space company Astrobotic has also proposed a small-scale planetary rover format called the CubeRover. Clearly inspired by the successful CubeSat format, Astrobotic’s CubeRover also follows a modular and COTS framework where scientific payloads can be readily integrated into existing hardware solutions to accelerate development and deployment. Current documentation for CubeRovers suggests that they can be developed in a variety of standard sizes, ranging from a 2-unit or 2U system weighing 5 kg to a 6U system weighing 10 kg. Also notable are the CubeRovers’ proposed mobility mechanisms, which reject the typical rocker-bogie technique in favor of a fixed-axis design. This approach utilizes four rigidly affixed wheels with independent drive capabilities to facilitate skid steering [38].

The first CubeRover to be developed to flight-worthiness was the Iris rover, which was intended to be transported to the Moon as part of Peregrine Mission One [29]. Although Astrobotic’s first attempt to launch this CubeRover failed with Peregrine Mission One, the Iris rover still represented a significant attempt at developing a nano-rover. Significantly, their continuing efforts to propel a low-cost and modular CubeSat-like architecture for planetary rovers have the potential to reduce the barrier to entry for future small-scale rovers. Because Iris weighed only two kilograms, it falls under the same size classification as Prop-M, but is a considerably more recent example of a nano-rover, as the mission launched and failed in 2023 [39].

3.3. SORA-Q: A Pico-Rover

In 2024, the Japan Aerospace Exploration Agency (JAXA) successfully soft-landed the Smart Lander for Investigating the Moon or SLIM lander. This mission included a number of small micro-rovers known as lunar excursion vehicles (LEVs). Both rovers were able to successfully deploy from the SLIM lander while the primary lander was descending toward the lunar surface. LEV-1 was intended to hop on the Moon, while LEV-2, known as SORA-Q, collected and transmitted images of the lunar surface and the SLIM lander [40]. The LEV-2 micro-rover weighed only 250 g [30] and consisted of an 8-cm-diameter sphere before deployment of the rover’s wheels and camera [41]. In addition to taking photographs, SORA-Q was also able to traverse the lunar surface using two hemispherical wheels, having been designed to engage both loosely packed terrain and inclined slopes of no more than 30 degrees [42].

3.4. COLMENA: A Femto-Rover

As another passenger on the doomed Peregrine Mission One, the tiny COLMENA rovers never had a chance to be deployed on the moon either. However, these miniature rovers consisted of flat 10-cm pucks, each equipped with two wheels, with a total weight of only 60 grams a piece. Because of this diminutive scale, five identical rovers were included on the lander, and they were intended to be “tossed” onto the lunar surface using a simple mechanical catapult. Upon deployment, the rovers were going to be networked together to act as a swarm or “hive” as implied by their name [28]. Although this first attempt to demonstrate mobility with Femto-Rovers failed with the parent mission, the individual COLMENA rovers represent the smallest known planetary rovers ever to be developed.

3.5. Small-Scale Actuation and Locomotion

The bourgeoning development of small-scale and low-cost planetary rovers has allowed for significant locomotion and actuation diversification in recent years. Following its original deployment in 1997, the rocker-bogie suspension technique has become a mainstay for planetary rovers, especially those developed by major national organizations. Table 3 categorizes the mobility techniques employed by previous historical planetary rovers, indicating a notable dominance of the rocker-bogie system from the late 1990s to the early 2020s. The prevailing use of the rocker-bogie system may in part be attributed to its relative simplicity and robustness, as no compliant springs or complex actively driven mechanisms are required to achieve consistent stability of the mobility system. However, another significant driving force behind the continual use of this mobility technique may simply be the continuing flight heritage associated with each additional mission that uses the rocker-bogie system. The rocker-bogie system has long been established as a ‘flight-proven’ approach, which significantly reduces the risk to the large flagship missions that employ it.

While the use of safer flight-proven technologies will likely persist for expensive flagship missions developed by national agencies, Table 3 also reveals a new modern trend. Starting around 2023 and continuing to the present, there has been a notable increase in the diversification of mobility and actuation techniques for planetary surface rovers. These largely privately developed and small-scale rovers have abandoned the rocker-bogie standard to investigate new, unproven forms of mobility. In some cases, like the Rashid rover developed by the Mohammed Bin Rashid Space Centre in the United Arab Emirates, completely novel forms of mobility and wheel actuation have been developed. Rashid presents the first known lunar deployment of an actively articulated suspension system, where the height and position of the wheels were driven by a single motor connected to a series of gears and electromagnetic clutches. This unique mechanism was capable of demonstrating 15 degrees of freedom (DOF) to improve the stability and maneuverability of the wheeled chassis [43]. Unfortunately, the effectiveness of this mobility technique remains unknown, as the Rashid rover was not able to be deployed when the Hakuto-R M1 lander carrying it crashed into the Moon.

Table 3. Planetary surface rover suspension or mobility techniques.

Name	Wheel Count	Suspension/Mobility	Distance Traveled (m)
Lunokhod 1	8	Torsion Bar	11,000 [11]
Prop-M (1)	na	Rotating Skis	Not Deployed
Prop-M (2)	na	Rotating Skis	Unknown [11]
Lunokhod 2	8	Torsion Bar	37,000 [11]
Sojourner	6	Rocker-Bogie	106 [11]
PLUTO	na	Internal Hammer [18]	Not Deployed
MER A	6	Rocker-Bogie	7730 [11]
MER B	6	Rocker-Bogie	45,161 [44]
MINERVA	na	Rotating Torquer	No Surface Operation [19]
MSL	6	Rocker-Bogie	35,270 † [45]
Yutu 1	6	Rocker-Bogie	100 [20]
MINERVA II-1A	na	Rotating Torquer [21]	Unclear, non zero
MINERVA II-1B	na	Rotating Torquer [21]	Unclear, non zero
Yutu 2	6	Rocker-Bogie ‡	1613 [46]
Pragyan 1	6	Rocker-Bogie	Not Deployed [47]
Mars 2020	6	Rocker-Bogie [12]	35,480 † [48]
Ingenuity	na	Coaxial Helicopter	16,971 [49]
Zhurong	6	Rocker-Bogie * [50]	1900 [51]
Pragyan 2	6	Rocker-Bogie [23]	103 [52]
Rashid	4	Articulated Suspension [43]	Not Deployed
Sora-Q (1)	2	Eccentric Wheels [42]	Not Deployed
COLMENA	2	Fixed Wheels ‡	Not Deployed
Iris	4	Fixed Wheels [38]	Not Deployed
LEV 1	1	Hopping Pad/Wheel [53]	Unclear, non zero
LEV2 (Sora-Q 2)	2	Eccentric Wheels [42]	Unclear, non zero
Jinchan	4	Fixed Wheels ‡ [31]	Unclear, non zero
YAOKI	2	Fixed Coaxial Wheels ‡  [32]	Not Deployed
Tenacious	4	Fixed Wheel ‡ [54]	Not Deployed

* Partially active suspension system; † Still operational, distance traveled as of 3 July 2025; ‡ suspension system or mobility solution assumed based on images or renderings; Text color indicates mass discretizations as described by Soyer in Table 1.

Table 3. Planetary surface rover suspension or mobility techniques.

Name	Wheel Count	Suspension/Mobility	Distance Traveled (m)
Lunokhod 1	8	Torsion Bar	11,000 [11]
Prop-M (1)	na	Rotating Skis	Not Deployed
Prop-M (2)	na	Rotating Skis	Unknown [11]
Lunokhod 2	8	Torsion Bar	37,000 [11]
Sojourner	6	Rocker-Bogie	106 [11]
PLUTO	na	Internal Hammer [18]	Not Deployed
MER A	6	Rocker-Bogie	7730 [11]
MER B	6	Rocker-Bogie	45,161 [44]
MINERVA	na	Rotating Torquer	No Surface Operation [19]
MSL	6	Rocker-Bogie	35,270 † [45]
Yutu 1	6	Rocker-Bogie	100 [20]
MINERVA II-1A	na	Rotating Torquer [21]	Unclear, non zero
MINERVA II-1B	na	Rotating Torquer [21]	Unclear, non zero
Yutu 2	6	Rocker-Bogie ‡	1613 [46]
Pragyan 1	6	Rocker-Bogie	Not Deployed [47]
Mars 2020	6	Rocker-Bogie [12]	35,480 † [48]
Ingenuity	na	Coaxial Helicopter	16,971 [49]
Zhurong	6	Rocker-Bogie * [50]	1900 [51]
Pragyan 2	6	Rocker-Bogie [23]	103 [52]
Rashid	4	Articulated Suspension [43]	Not Deployed
Sora-Q (1)	2	Eccentric Wheels [42]	Not Deployed
COLMENA	2	Fixed Wheels ‡	Not Deployed
Iris	4	Fixed Wheels [38]	Not Deployed
LEV 1	1	Hopping Pad/Wheel [53]	Unclear, non zero
LEV2 (Sora-Q 2)	2	Eccentric Wheels [42]	Unclear, non zero
Jinchan	4	Fixed Wheels ‡ [31]	Unclear, non zero
YAOKI	2	Fixed Coaxial Wheels ‡  [32]	Not Deployed
Tenacious	4	Fixed Wheel ‡ [54]	Not Deployed

* Partially active suspension system; † Still operational, distance traveled as of 3 July 2025; ‡ suspension system or mobility solution assumed based on images or renderings; Text color indicates mass discretizations as described by Soyer in Table 1.

Along with Rashid, many recent nano-scale or smaller rovers have attempted to demonstrate unique mobility and actuation approaches that have never been proven in space environments. LEV1, associated with JAXA’s Smart Lander for Investigating Moon (SLIM), employed a single wheel in conjunction with a unique ‘hopping pad’. This system allowed the small-scale rover to control its azimuth orientation with its singular wheel, while exploring the lunar surface by hopping in the low gravity [53]. The second lunar excursion vehicle, LEV2, employed a unique eccentric wheel configuration. The successful mission completed by this rover, which was also known as SORA-Q, has been previously discussed. However, the unique mobility system is of note here, as it represents a mechanically simplified approach. Instead of the six-wheel configuration implemented by a typical rocker-bogie technique, two hemispherical wheels were developed that featured a small radial offset, which allowed the rover to drive or crawl across the lunar surface [42].

The relative mechanical simplicity of SORA-Q’s mobility technique is also reflected in many recent rovers developed in the last few years. Although significant technical details are still forthcoming for many of these vehicles, images and renderings of rovers like Iris, Jinchan, and Tenacious all appear to show rovers that maintain no significant suspension systems at all. Initial analysis of these rovers suggests that they have wheels that are firmly fixed to their drive motors or other locomotive actuators. Although fixed wheel skid steering is not a new concept for mobile robots on Earth, the zero-radius turning capabilities associated with the rocker-bogie system have overwhelmingly been leveraged in space applications, until very recently.

Significantly, these unique robotic rovers did not develop in a vacuum, but instead represent years of research and development that culminated in a carefully engineered system. As there are still a limited number of these systems to investigate, previously developed trends in miniature satellites will be elaborated upon as a potential template for future developmental trends in miniature rovers.

4. Future Developmental Pipeline

The future of developing miniature planetary rovers may follow similar trends that have allowed miniature satellites like CubeSats to grow in popularity and scientific merit. Notably, CubeSats, the most popular format of microsatellite, were developed concurrently with CanSats, which have become important Earth-based simulation aids for orbital CubeSats. This relationship may have contributed to the success of CubeSats, and a similar technological relationship may also be necessary to support the future of miniature planetary rovers.

4.1. CubeSats and CanSats

Much like SORA-Q, Iris, and COLMENA, the first interplanetary CubeSats were added to a larger mission as secondary technology demonstrations. The Mars Cube One (MarCO) mission included two six-unit CubeSats that each had a mass of approximately 13.5 kg, affectionately referred to as WALL-E (MarCO-A) and EVE (MarCO-B). These vehicles were launched with the InSight lander, which was the primary mission headed to Mars. The MarCO microsatellites accompanied the InSight lander all the way to Mars, where they provided a more rapid means for relaying InSight’s entry, descent, and landing (EDL) data to ground control operators back on Earth than would have been available with the standard Deep Space Network (DSN). Again, while not the primary purpose of the InSight mission, MarCO provided a unique capability thanks to a low-cost design approach known as the CubeSat [55].

The design standards that are now widely employed by CubeSats have a surprisingly simple and arbitrary origin. As early as 1995, a graduate engineering program was developing microsatellites at Stanford University. These efforts were largely directed by Professor Bob Twiggs, who later partnered with California Polytechnic’s Jordi Puig-Suari to develop an even smaller pico-satellite design. Although the initial launch opportunity for the pico-satellite was canceled, the investigation into pico-satellites as a means for education and scientific discovery continued. In order to power the pico-satellite design, a four-inch square area needed to be covered in solar cells; this necessity first inspired the idea of a square-shaped satellite. In order to develop an initial engineering model, plastic 4-inch display cases for Beanie Babies were sourced. However, concerned with the mishap of the Mars Climate Orbiter, which failed to reach Mars because of an improper conversion between English and metric units, the 4-inch cube was converted to a 10-cm cube, which was approximately the same general size. As 10 cubic centimeters of water have a mass of 1 kg, the standard was set: CubeSats would be 10 cm cubed and have a maximum mass of 1 kg, all because of some solar cells and a Beanie Baby display case [56].

While the CubeSat standard was under development, Professor Twiggs went to the 1998 University Space Systems Symposium (USSS). Accounts from this conference claim that Professor Twiggs suggested that a satellite could be developed to fit within a standard aluminum soda can [57]. As far as the analyzed records can show, this was the first public announcement of the CanSat concept. The first CubeSats launched into orbit in 2003 [56], while some of the first CanSats were tested in terrestrial launches in 1999 following the USSS [58]. It remains unclear whether Professor Twiggs truly intended for CanSats to be orbital vehicles, but the 1999 ARLISS (A Rocket Launch for International Student Satellites) event partnered with U.S.-based high-powered amateur rocket builders to launch CanSats made by both Japanese and American students to altitudes of 3.6 km [57]. In the following years, additional suborbital CanSat launches took place, with the ARLISS event largely focusing on education and providing an important opportunity to test systems in conditions that could be largely analogous to a real orbital launch. In particular, with the addition of parachutes, the CanSats launched on a ballistic trajectory had flight times of 10–15 min, which is similar in length to a single horizon-to-horizon pass of a small satellite in low Earth orbit (LEO) [58].

Figure 4. Historical levels of nanosatellite launches [59].

Figure 4. Historical levels of nanosatellite launches [59].

Notably, in the early days of ARLISS, a new angle on the CanSat flight paradigm was proposed, with the first “ComeBack” competitions taking place in 2001. In this event, the CanSats were equipped with parafoils or fixed wings and designed to fly back to a designated landing area. Later developments on this concept included “run-back” devices that employed simple wheels to be able to drive the CanSat along the ground and back to the targeted landing zone [57]. However, in more modern times, these types of self-propelled CanSats now appear to be the exception, not the rule. A previous review of both scientific and competition CanSats suggests that a large percentage of these vehicles employ passive parachute recovery systems for use after deployment from a rocket or balloon [60]. This flight profile is further supported by various competition documents and manuals for building CanSats that describe rocket launches followed by parachute deployments [57,61,62].

CubeSats have also evolved over the intervening years into wildly successful research tools. According to the Nanosats Database, as of 1 January 2024, some 2323 nanosatellites have been launched, with 396 of those launches taking place in 2023 alone [59]. Figure 4 further illustrates this exceptional increase in launch cadence for small-scale satellites over the last two decades. The exponential growth in CubeSat launches indicates the utility of these vehicles for space-based science. From the Planetary Society’s LightSail 2, which successfully demonstrated small-scale solar sailing in Earth orbit [63], to NASA’s MarCO CubeSats, which successfully demonstrated interplanetary travel and communication [55], to JAXA’s OMOTENASHI CubeSat, which would have attempted the first CubeSat landing on the Moon [64], these miniature vehicles are attempting complex and meaningful scientific objectives. While not all missions are successful, such as OMOTENASHI, which failed to control its attitude after deployment, these low-cost platforms are able to try science that would be too risky on the larger and higher-cost flagship missions.

CanSats also continue to be relevant, mainly as educational tools, but also as important pathfinding vehicles that can be used to test space-oriented or even space-bound systems [60]. However, some of the CanSat’s most valuable assets are also among its more significant shortcomings. CanSats were developed during the same time as the now much better-known CubeSat, and this parallel development could imply that CanSats are nothing more than a failed nanosatellite standard. While some attempts have been made to establish a cylindrical nanosatellite convention known as a “TubeSat” [65], the popularity and wide acceptance of CubeSats have well overshadowed these efforts. Failing as an orbital satellite standard still placed CanSats in the important territory of educational aids, and, again, as early starting points for the space technology pipeline. However, it is also clear that CanSats were always meant to be oriented toward orbital satellite technologies, and this naturally limits their wider application. In Figure 5, a general technological progression has been depicted whereby systems developed for CanSats may graduate to use in CubeSats and eventually to large-scale orbiters.

4.2. From CanSats to CanBots

In an effort to provide a distinct name and classification for Earth-based prototype robotic vehicles that are intended to aid in the creation of future miniature planetary rovers, this research coins the term “CanBot”. CanBots represent an expansion upon the existing CanSat concept. While CanSats have provided a robust and accessible means for simulating orbital satellite technologies in space-oriented projects, these devices are not widely deployed for simulating robotic planetary surface explorers. With the increased number of small-scale planetary rovers and the dawning of a possible modular nano-rover format, a robust development pipeline for these robots is required. There can be little doubt that the success of nano- and pico-sized satellites, such as CubeSats, has been facilitated by CanSat architectures. CubeSats have now traveled to other planets, and CubeRovers may soon be widely deployed to alien worlds as well. Therefore, the need for a new paradigm of simulated planetary surface exploration robots is now clear.

As previously described, there has been a history of non-traditional CanSats that demonstrate self-motivation and even long-duration terrestrial operation. Such vehicles are generally categorized as those that drive on the ground using wheels [66,67,68], or those that use some form of fixed aerodynamic surface to produce lift and, therefore, controlled flight [69,70,71,72,73]. Such vehicles have been classified by their own researchers and by a previous review as special types of aerial and terrestrial CanSats [60]. However, in this research, efforts will be made to reclassify these devices as a new robotic vehicle type known as the CanBot.

In addition to CanSats demonstrating surface mobility, there are also a number of small-scale test models for planetary rovers that could already be described broadly as CanBots. One such system could be the Wandering Observer of Lunar Features (WOLF) rover developed by Creamer Et al., which was created with both CubeSat and the initial CubeRover format in mind. Their first engineering prototype did not contain space-grade hardware and was intended only to validate the general parameters of the rover through Earth-based testing regimens. Therefore, this first model of their rover could broadly be categorized as a CanBot [74]. Another broad example of an existing CanBot, which has yet to be called by this name, could arguably be the Earth-based prototypes for the forthcoming CubeRover design that the Astrobotic company has been developing. To validate the general technology readiness level (TRL) of the CubeRover design, a number of initial engineering models were created and tested. In particular, versions of the CubeRover were used for mobility testing, featuring vehicles with a significant reduction in mass to simulate the lower lunar gravity. These vehicles would never be able to be launched into space, as core components were removed, and were, therefore, solely Earth-based test models. Acting as learning and pathfinding versions of future planetary rovers could qualify these robots as CanBots as well [75].

Notably, simulating planetary rovers on Earth as pathfinders for future space-bound systems is not a new concept. The MERs benefited from investigations done with the Field-Integrated Design and Operations (FIDO) rover. This initial vehicle was never intended to go to space, but instead operated solely on Earth to test various aspects of the future planetary rover mission that would later prove to be very successful on Mars. The FIDO rover was also modular, including hardware, electronics, sensors, and actuators that were cross-compatible or based on other previous test rovers in the “FIDO Family”. This “set of common module design resources” allowed for “rapidly [prototyping] very different robotic systems” that directly contributed to the success of the MERs [76]. Similar Earth-based rovers have been developed for MSL, such as Scarecrow, which features the same drivetrain as the Mars-bound rover, but was drastically reduced in weight to be analogous to the MSL on Mars when testing with the Scarecrow on Earth [77]. Again, while not explicitly enunciated by their makers, these terrestrial-based simulation rovers could be considered CanBots.

By way of comparison to CanSats, CanBots are also largely research tools intended for simulating space-oriented circumstances. However, instead of attempting to simulate orbital satellites, as CanSats do, the proposed CanBot vehicle is intended to simulate planetary surface rovers. Where a CanSat simulates planetary orbiters, CanBots simulate planetary rovers. This distinction is necessary as orbiters and rovers are significantly different vehicles. Moreover, simulating this kind of surface mobility with a CanSat confuses the purpose of CanSats, as they remain largely oriented toward orbital simulations. The need to separate CanBots from CanSats is clear, just as the previously described operational differences between orbiters and rovers are clear. CanSats should remain a significant milestone on the technology pipeline for developing orbital micro- and nanosatellites such as CubeSats. But just as small-scale satellites have benefited from supporting technologies such as CanSats, the development of future small-scale planetary rovers may also greatly benefit from a distinct prototyping platform, such as the proposed CanBot methodology. This principle is illustrated in Figure 6, where systems developed for CanBots may eventually be able to be deployed as CubeRovers or other small-scale rovers, where they could be matured for eventual application in larger-scale planetary rovers.

5. Discussion

Technologies continue to be reduced in size and improved in capability, which, in turn, has empowered unprecedented applications and advanced uses. Robotic planetary exploration represents one such unprecedented application. Improved and miniaturized technologies have allowed the gargantuan satellites of the 20th century, which possessed only limited scientific capabilities, to give way to the tiny CubeSats of the 21st century, which continue to redefine the plethora of scientific studies that may be performed from orbit. However, CubeSats did not develop alone, and their Earth-based counterparts, known as CanSats, continue to be important pathfinders and teaching tools that support further development in Microsatellites. And just as miniaturization has fundamentally altered the orbital satellite technologies, trends of miniaturization have begun to become apparent in current planetary surface rovers. From nano-rovers that weigh only a few kilograms to pico-rovers that weigh less than a kilogram, a new generation of roving robots is beginning to be introduced to alien worlds. Unlike rovers that have come before, these miniature systems are small, cheap, and accordingly more likely to employ risky cutting-edge technologies. Such robotic vehicles must be encouraged to advance human understanding of other planets and the general advancement of robust mechatronic systems.

6. Conclusions

To support a robust developmental pipeline for future small-scale planetary rovers, this research summarizes the major historical context for planetary surface rovers in addition to describing modern trends in miniaturized mobile vehicles, including their mobility and actuation techniques. Additionally, this work emphasizes the potential importance of CanSats in elevating the acceptance of CubeSats and identifies that similar Earth-based testing devices for miniature planetary rovers already exist. Furthermore, this research coins a new term, “CanBot”, to refer to these Earth-based test rovers in an attempt to provide a consistent mechanism to reference these devices and encourage their continued development.

There is little doubt that miniature planetary rovers will continue to be developed, and if previous trends in miniature orbital satellites are any indicator, the future of planetary surface exploration will also be in these small-scale rovers. Understanding and acknowledging such trends, in addition to engaging supportive technologies, such as CanBots, may help ease the transition into a new paradigm of space exploration.