Meet the Next Generation of Planetary Rovers

Want to go scuba diving on the Saturnian moon Enceladus? Get in line. When NASA announced the discovery of a subsurface ocean the size of Lake Superior on the tiny moon it inspired a new rush of speculation about how we might land a rover on Enceladus' alluring surface.

“There’s such a large amount of interest in this small body,” Luciano Iess, lead author of the Endeladus study, told Motherboard last week. “You could deploy a rover that could penetrate through this rather thick ice layer with heat. You can melt the ice and then by gravity, the submarine would get pushed down and would, sooner or later, end up in the ocean. It’s complicated, but that’s the target.”

But Enceladus is only one of many such complicated targets positively begging to be explored. Roving around the Jovian moon Europa—which boasts a much larger subsurface ocean than Enceladus—has been on the to-do list for decades (and NASA might finally be getting around to it). Saturn's biggest moon, Titan, is another popular candidate. Though the Huygens probe pulled off a perfect landing in Titan's Shangri-la section in 2004, it was only able to send a few hours' worth of data back to Earth. Imagine how much more insight we could get into Titan's atmosphere and methane-rich oceans with a next-gen rover.

But as the list of extraterrestrial destination spirals upwards, so too does the need for cheaper, smarter, and more versatile rovers. This is especially true when it comes to more exotic destinations in the outer solar system. If we're seriously going to pursue Europa, Titan, Enceladus, and other bodies like them, we need to think beyond the “lab-on-wheels” setup. Bulky, expensive, and vulnerable, this design will have to make way for the extraterrestrial explorers of the 21st Century.

This is not to argue that the traditional rover model—a hefty science lab piggybacked on all-terrain wheels—should be phased out. After all, this design has racked up significant milestones on the moon and Mars. Curiosity has delved into the planet's aquatic history, while Opportunity is still chugging along discovering Martian balls at ten years old, far longer than NASA expected. And while China's Yutu rover is turning out to be a total drama queen, getting a mobile robot back on the lunar surface after 37 years is an accomplishment in itself.

But what will these 'rovers of the future' look like? We reached out to Adrian K. Agogino, a researcher at the NASA's Intelligent Systems Division, for answers. Together with senior robotics researcher Vytas SunSpiral, Agogino has been working on a new rover design based on the concept of tensegrity—where tension meets integrity. Nicknamed the “Super Bot Ball,” this bundle of rods and cables looks like a high-tech tumbleweed, but it has the potential to leave traditional rovers in the Martian dust.

Motherboard: When most people think of rovers for planetary exploration, they think of the hefty Mars rovers. The Super Bot Ball concept seems to explode this model. What kind of challenges can a tensegrity-based explorer overcome that a traditional rover cannot?

When we look at traditional wheeled rovers, we are looking at a very delicate tool. It can get stuck, it can tip over, it can crash into things. Even more advanced concepts such as powerful rigid walking robots can fail over relatively mundane tasks, such as maxing out on torque because it is crawling over some small rocks at just the wrong angles.

We saw a tensegrity structure as something very robust and compliant. Something that will not get into a lot of trouble and something that can get out of a lot of trouble. Tensegrity structures are really good at absorbing impact forces because of their inherently compliant design. We are using this property to design a robot that can survive landing from orbit. Unlike traditional robots which need landing systems like airbags, this quality is inherent in the design of the tensegrity robot and thus it has this level robustness to collisions as it explores another planet.

With a traditional rover one has to be very careful about collisions or, worse, falling or rolling down a steep slope or cliff. Since planetary scientists are often interested in the exposed rock found on cliffs and mountains, these are exactly the regions they would like to send the rover to. With a tensegrity robot which can survive landing from orbit, these risks are greatly decreased, enabling a much more aggressive exploration strategy.

On your NASA project page, you describe tensegrities as "counter-intuitive tension structures with no rigid connections." Could you expand on the inner workings of these explorers? How big do you expect them to be?

A basic tensegrity structure is essentially made of cables and rods. The key thing is that the rods do not connect rigidly to each other. They only connect through cables. The counter-intuitive part is how something connected with cables can have rigidity as a whole.

The answer is the tension network. There is an interplay between the compression in the rods and the tension in the cables that allow the system as a whole to have a variable degree of rigidity. The tension network allows contact forces to diffuse through the whole structure, which allows for a system wide participation in absorbing the stress of impacts.

For the actual tensegrity robot the big issue is actuation: how do we change the shape of the tensegrity. Our current approach is to reel the cables in and out on a spool contained in an end-cap of the rod, though this could change in the future. Our current size goal is for each tensegrity rod to be about 1.5 meters in length.

Where are you in the development process?

We are currently working with a tensegrity ball robot that has a few cables that can be actuated. This is helping test our basic mechanical properties and control. We are also developing a new actuated tensegrity rod. The new rods will lead to our second-generation prototype, which should have significantly improved capabilities and efficient rolling locomotion. Within six months, we plan to put six of these rods together to create a tensegrity ball. This ball should have significant capabilities including rolling.

Motherboard: The Super Bot Ball has been billed as ideal for the exploration of Titan. Why the focus on that moon specifically? Could the model deployed on other planets and moons, even with extreme landscapes like Europa?

Adrian K. Agogino: Tensegrity robots can be helpful most anywhere, though Titan has an interesting use case. Long ago I did the calculation that in principle a person (in a protective suit) could be dropped from 500 kilometers above the surface of Titan and survive the landing without any parachute due to Titan's thick atmosphere and low gravity (note that entering the upper atmosphere from orbit and not burning up is an entirely different issue that needs to be resolved through heat shields).

This person would hit pretty hard, and a rover landing at the same speed would almost certainly break. However, a tensegrity robot could potentially land at this speed and be just fine. In fact we have performed drop tests from buildings that tend to confirm this.

Titan's other interesting property is it is actively changing, so things like the edges of lake shores likely change location depending on season. Yet due to its atmosphere details about the surface are difficult to monitor from orbit. Such an environment is exciting to science, but very problematic to navigate in a traditional rover. A tensegrity robot, though, should be robust enough to handle unexpected environments.

For other destinations such as Europa, a tensegrity robot would need some sort of primary landing system, but could still be significantly more economical than traditional landing systems, since it would not need the final part of the landing system such as airbags or the Curiosity rover's sky crane. In addition, once on the surface a tensegrity robot would retain all of its robustness and compliance advantages.

Motherboard: A lot of our readers (myself included) are interested in how robotic planetary exploration will evolve over the coming decades. I'd love to get truly sci-fi with you. Other than tensegrity models, what are some other concepts that have the potential to radically change planetary exploration? Do you think traditional rovers will continue to be deployed, or will they be edged out for lighter, more durable models?

Adrian K. Agogino: The big trend that we are seeing in robotics today, of which tensegrity robots are an example, is the emergence of "soft machines". Traditional rigidly constructed robots are really good for high speed repetition of known tasks, such as factory automation, which is where the technology matured. But it turns out that they are very poor choices for interacting with dynamic natural environments—their rigidity means that any miscalculation about the external world can cause significant damage to themselves or those around them.

As such, as robots are increasingly being used outside the factory or laboratory, the technology is moving towards increased compliance and even incorporating soft deformable materials. There are many challenges to overcome, but this trend will continue into the future and tomorrow's robots will move more like animals in their ability to be soft and strong and adapt to their environment.