Building robots for “Zero Mass” space exploration

Sending 1 kilogram to Mars will set you back roughly $2.4 million, judging by the cost of the Perseverance mission. If you want to pack up supplies and gear for every conceivable contingency, you’re going to need a lot of those kilograms.

But what if you skipped almost all that weight and only took a do-it-all Swiss Army knife instead? That’s exactly what scientists at NASA Ames Research Center and Stanford University are testing with robots, algorithms, and highly advanced building materials.

Zero mass exploration

“The concept of zero mass exploration is rooted in self-replicating machines, an engineering concept John von Neumann conceived in the 1940s”, says Kenneth C. Cheung, a NASA Ames researcher. He was involved in the new study published recently in Science Robotics covering self-reprogrammable metamaterials—materials that do not exist in nature and have the ability to change their configuration on their own. “It’s the idea that an engineering system can not only replicate, but sustain itself in the environment,” he adds.

Based on this concept, Robert A. Freitas Jr. in the 1980s proposed a self-replicating interstellar spacecraft called the Von Neumann probe that would visit a nearby star system, find resources to build a copy of itself, and send this copy to another star system. Rinse and repeat.

“The technology of reprogrammable metamaterials [has] advanced to the point where we can start thinking about things like that. It can’t make everything we need yet, but it can make a really big chunk of what we need,” says Christine E. Gregg, a NASA Ames researcher and the lead author of the study.

Building blocks for space

One of the key problems with Von Neumann probes was that taking elements found in the soil on alien worlds and processing them into actual engineering components was resource-intensive and required huge amounts of energy. The NASA Ames team solved that with using prefabricated "voxels”—standardized reconfigurable building blocks.

The system derives its operating principles from the way nature works on a very fundamental level. “Think how biology, one of the most scalable systems we have ever seen, builds stuff," says Gregg. “It does that with building blocks. There are on the order of 20 amino acids which your body uses to make proteins to make 200 different types of cells and then combines trillions of those cells to make organs as complex as my hair and my eyes. We are using the same strategy,” she adds.

To demo this technology, they built a set of 256 of those blocks—extremely strong 3D structures made with a carbon-fiber-reinforced polymer called StattechNN-40CF. Each block had fastening interfaces on every side that could be used to reversibly attach them to other blocks and form a strong truss structure.

A 3×3 truss structure made with these voxels had an average failure load of 900 Newtons, which means it could hold over 90 kilograms despite being incredibly light itself (its density is just 0.0103 grams per cubic centimeter). “We took these voxels out in backpacks and built a boat, a shelter, a bridge you could walk on. The backpacks weighed around 18 kilograms. Without technology like that, you wouldn’t even think about fitting a boat and a bridge in a backpack,” says Cheung. “But the big thing about this study is that we implemented this reconfigurable system autonomously with robots,” he adds.

Robotic building teams

In a lab experiment, three robots used all 256 voxels to assemble a shelter in four and a half days. The first robotic team member was a cargo handler that transported the voxels from the supply area to the right place on the structure under construction. Once there, the cargo robot handed the voxels over to a crane robot that placed them exactly where they needed to be. Finally, a fastening robot moving inside the structure attached each new voxel to the structure.

The robots oriented themselves exclusively using internal reference frames—they basically counted the voxels they stepped on. This meant no vision, no lidar, no advanced sensors or control systems. “In this demo, the structure was preplanned. The preplanning resolves robots bumping into each other or whether the structure is stable as it is being built. We can do that automatically,” says Cheung. “But we also tried models like finite state automata, which is how ants build their colonies. This way we could solve high-level problems like finding something and building an enclosure around it,” he claims.

The system is scalable in the sense that, with more voxels, structures can be made larger; with more robotic teams, they can be assembled faster. The first real-world application the team aims at is building towers on the Moon.

Towers on the Moon

The towers are needed because the landing site for the Artemis 3, a mission intended to bring human astronauts back to the silver globe, is near the Moon’s south pole. “The Sun angle there is low, so to maximize the amount of sunlight, you should put the solar panels as high as possible. You can’t bounce radio waves off the atmosphere because there is no atmosphere, so you need line of sight for communications. The antennas, too, must be as high as possible,” says Cheung.

The height of the towers in this location must be over 100 meters to get the job done, and pulling that off with current deployment systems would be very difficult. So, the team is now focused on demonstrating how their building blocks and robots could be used in building communication and solar towers on the Moon. “Our next papers are exactly about that. They are coming out in March,” says Gregg.

Science Robotics, 2024.  DOI: 10.1126/scirobotics.adi2746

Jacek Krywko is a science and technology writer based in Olsztyn, Poland. He covers space exploration and artificial intelligence research.