Exploration is, was, and always will be a risky and expensive business. In order to reduce risk, we send rovers to other planets instead of people. Still, nothing will completely defray the risks of going first. As a result, planetary robots have to be over-engineered, remotely controlled, and are rarely used to their fullest potential.
A perfect example of this is the Mars Science Laboratory, Curiosity. Curiosity is an absolute wonder; its unprecedented size forced its designers to ascend great heights, with elegant wheels, an arm so loaded with equipment it can’t pick itself up in Earth gravity, and the dramatic Sky Crane maneuver. Despite those heroics, it took Curiosity almost three years to drive a mere ten kilometers. Part of that was the justifiable cautiousness of driving a billion-dollar rover. But part of it was also that they encountered terrain that was unexpectedly difficult, and it tore up their wheels. Imagine how much more they could have accomplished if they hadn’t suffered that setback.
What they need is a sidekick that can roam out and explore, whose loss wouldn’t cripple the mission. This symbiotic scout would take on the risk, ranging out quickly to cover lots of terrain and make sure it’s safe for its mothership. This scout will even be able to spot interesting features for its mothership to explore. Curiosity could have used this sort of scout when it was planning to examine the Glenelg site. Glenelg is a feature that Curiosity went completely out of its way to visit, and the scientists didn’t know for months if it would be worth the trek. Imagine if they could have known in hours.
So now, for the first time, meet a CubeRover.
CubeRover is the specification for a new kind of explorer. A CubeRover isn’t a headline rover; instead, it will ride with one. In order to not impose, CubeRovers must fit in a tiny, 30 centimeter cube and weigh in at no more than 10 kg with its own payload. These CubeRovers couldn’t carry heavy duty instruments; like the CubeSats that are squeezed into major satellite launches, these tiny rovers’ advantage is how easy they are to add to a mission.
As part of developing this specification, the team has built the very first CubeRover, Tetramorph. Tetramorph, named for its unique, unfolding deployment system, is a marvel of engineering and miniaturization. It borrows in many ways from the design of Andy, this team’s own headline rover. With a similar four-motor, pivoting-axle drive train and wheels with straight grousers, Tetramorph could be mistaken for a smaller version of Andy.
But that’s where the similarities end. In order to save space, each of Tetramorph’s wheels is packed with electronics and batteries; it also loses the large and heavy solar panels, instead recharging from its mothership. The biggest difference, though, is the conditions under which they’ll operate. Andy behaves like those headline rovers: it is the centerpiece of the mission, and will be at least partially driven from Earth. Tetramorph will range out for up to a kilometer in any direction before returning for a recharge - and do so autonomously.
Autonomy is crucial to the function of Tetramorph. Even on targets close enough to drive in real time, like the Moon, staying in radio communication would be difficult. It would limit how much foresight Tetramorph could give its symbiote, and it would draw too much power. It’s an exciting time in space robotics: because rad-hardened computers lag so far behind terrestrial ones, we’re just now getting to the point where serious autonomy - like what’s developed here at Carnegie Mellon - is viable.
The CubeRover spec was invented at the end of the summer of 2014. Responding to a then-recent NASA request for proposals, Carnegie Mellon Professors William “Red” Whittaker and Uland Wong envisioned a rover light and small enough to be added to existing missions. Development started shortly afterwards with the fall semester, lead by Carnegie Mellon graduate student Jay Jasper.
While it’s important to know the box Tetramorph must fit in, that was only the beginning of the story. In order to be rugged and mobile enough to scout, the team decided that the CubeRover would have to expand from its shipping configuration. Of all the ways to do that - some as exotic as wheels that expanded like umbrellas - the Tetramorph seemed most strong and efficient. Now, armed with an overall structure, the team set about designing the first prototype in all its detail. It quickly became apparent how cramped the design would be. Space was so tight that even the choice of computer board in the chassis affected the choice of motor in the wheels. It took ninety design revisions before the team was ready to fabricate.
At the same time, the software team was hard at work. Tetramorph faces a harder autonomy problem than most robots. While autonomous driving is never easy, it’s even harder in the featureless environments often found on other planets. Coupled with size, weight and power (SWaP) limitations, it took months of design to find algorithms and techniques capable of sustaining the mission.
Members of the team worked through their Thanksgiving break to finish building Tetramorph’s complicated wheels. Each one - composed of forty-nine pieces of aluminum - had to be painstakingly assembled before it was ready to be integrated into the rest of the rover. With machining, laser-cutting, and a small amount of 3D printing, the many components of Tetramorph were rapidly built by the team. Of course, not everything was made at home - the team also benefited from the generous donation of motors by Maxon Motor and motor-controllers by Trinamic Motion Control. It all came together just in time for the end of the semester; the team presented the working, mechanical prototype.
While the heroics of the fall semester made the vision into a mechanical prototype, Tetramorph was still not totally functional. Its next milestone would be a functioning, teleoperated robot. Over the course of the spring semester, even though Andy took precedence, Tetramorph was improved to the point that it’s now a functioning prototype
Autonomy is the next prototype milestone. The software team has done fantastic ground work, and the next step is to fully implement it.
Then, Tetramorph will be ready to change how we explore.