Swarming for Success

Someone has to invent the future. And NASA wants a piece of the action. The NASA Institute for Advanced Concepts (NIAC) was created to encourage scientists and engineers to think outside the box, to come up with ideas just this side of science fiction. Their hope is that some of these ideas will pan out, and provide the agency with technologies it can use 20, 30, or 40 years down the road.

Penny Boston and Diana Northup taking a pH or ORP measurement in the Ragu passage (Cueva de las Sardinas. Tabasco, Mexico).
Credit: Kenneth Ingham, 2001

Earlier this year, NIAC funded Dr. Penelope Boston and Dr. Steven Dubowsky to continue working on a NIAC project to develop "hopping microbots" capable of exploring hazardous terrain, including underground caves. If the project pans out, hopping microbots may some day be sent to search for life below the surface of Mars.

Boston is the director of the Cave and Karst Studies Program and an associate professor at New Mexico Tech in Socorro, New Mexico. Dubowsky is the director of the MIT Field and Space Robotics Laboratory at MIT, in Cambridge, Massachusetts. He is known in part for his research into artificial muscles.

Astrobiology Magazine interviewed Boston shortly after she and Dubowsky received their Phase II NIAC grant. This is the second and final part of a two-part interview. In Part I Boston described the hopping microbot concept. In this part, she talks about the swarm-like behavior of large groups of microbots that enables them to perform sophisticated research tasks.

Astrobiology Magazine (AM): You and Dr. Dubowsky are working on a project to develop microbots, miniature robots that could be used to explore subsurface caves on Mars. One of the technologies you’re focusing on is having a group of these microbots act in concert as a swarm. What would this swarm behavior look like?

credit:Render by R.D.Gus Frederick
Close view of Ball Bot interior. Click image for larger view.
Image Credit: Render by R.D.Gus Frederick

Penny Boston (PB): We imagine them having a number of different algorithms. So there would be a deployment strategy, where they would be spreading themselves out, seeking, and keeping contact with each other. And then when a particular unit finds something that falls within an area of interest – for example, emissions of chemically reduced gases, like sulfides or methane, that might have biological significance – then it would switch to some kind of ensemble mode. It would put out a call and the entities that are in the immediate vicinity that have the talents that it needs to further investigate that finding would switch to the mode where they would make their way over. Then we would have a concentration of all of the available sensing and analytical modes to bring to bear on this particular interesting find.

AM: You said that these microbots would operate with simple rules. I’m wondering, how simple could you get and still do something useful?

PB: Well, apparently it’s very simple. I’m no roboticist myself, but one of the things that captured me about Rodney Brooks’s work back in late 80s into the 90s with insect robots was how simple a set of relational rules from one unit to another could produce extremely complex and capable behaviors. That comes directly from biology. An individual ant has a very limited palette of behaviors with which it relates to its fellows and its home nest and so forth. And yet their behavior, as they meet the exigencies of their environment, is extremely complex. As an ensemble they’re very, very capable. So even though the basic rules of the microbots are very simple, the emergent behavior of the system, which is directly modeled after the way communal insects behave, is extremely effective.

credit:Render by R.D.Gus Frederick
Ballbot, or Planetary MicroBots. Click image for larger view.
Image Credit: Render by R.D.Gus Frederick

AM: You’d also need a set of simple triggers to get the behaviors started. You mentioned methane detection as one. What are some of the others?

PB: The detection of any chemically reduced gases that are associated with life on Earth – that would certainly be one indicator. Other indicators would probably have to be processed in a way that would allow pattern recognition. And pattern recognition software is being worked on actively. One of the things that attracts our attention as human investigators is when we find anomalous deposits of soft materials in the caves that we work in. We work with all sorts of bizarre substances in caves that are the product of biological activity interacting with the mineral and rock environment in the subsurface. Those tend to be high-contrast against background.

I can imagine that being a fairly simple algorithm when applied to a robotic device. Even though we come at them as human investigators with our big brains, what triggers us to start on the process of investigation of something is really a simple set of rules, which we then take to great lengths of complexity as we analyze them. But we’re envisioning these guys as doing step-one recon, to call attention to something that’s worthy of future investigation. Even if one had a human base on Mars, these guys could be used to do the initial recon in difficult terrain.

They could also be used as a tool for planetary protection. The ability to render something like this aseptic is much greater than trying to render a human investigator aseptic. So we could send them places that are of potential biologic sensitivity to do the initial testing. I can even imagine using something like this as a permanent interface. If one is concerned about keeping an actual find of martian life (or potential martian life) protected, then keeping it one step removed from human investigators on site seems like another potential application. We think that this is a very generalizable approach, not only for life detection on Mars, but also for getting into difficult terrain on worlds that are not exobiological targets.

credit:Render by R.D.Gus Frederick
Camouflage BallBots Baseball-Sized. Click image for larger view.
Image Credit: Render by R.D.Gus Frederick

AM: You mentioned that when you see gooey or spongy stuff in a cave, that’s a pretty good indication that biology’s involved. But when geologists look at ancient rocks for evidence of biology, the older the rocks, the harder it is to tell whether what they’re looking at is of biological or of chemical origin. Do you think this microbot strategy is better at finding extant life than ancient life, or do you think it could be equally well adapted to both?

PB: I think it can be adapted to both. That’s based partly on the unique nature of biosignatures in caves, which are very different from surface outcrops. If one is looking for fossil material (for example, microbial fossils) in outcrops, that material has undergone a great deal of alteration from surface weathering. On Earth it’s undergone a great deal of other types of alteration as well; the extent of that on Mars is still debatable. But in any case, if something is surface exposed, it’s going to have been in the harsh martian surface conditions for significant period of time. What we find in Earth caves is that, because they are protected from surface weathering, the preservation of biological activity, even when it is lithified and has become fossil remains, is exquisite.

We have been studying many different kinds of materials and structures in caves where the organisms clearly were self-lithifying. They were in very mineralizing environments – analogous to what goes on in Yellowstone, where the high-temperature mats are undergoing a great deal of lithification as they’re growing. We find the same thing in caves, and in caves we then have protection of that material from surface influences. So once it is formed, it’s there for you. The preservation of microbial fossils is exquisite and easily visible. The biofabrics and biotextures are untouched.

People tend to think that caves are very short-lived. Some of them are. But a lot of the deep, ancient, pristine caves that we work in, which are much more analogous to what we would expect to find on Mars – being in a very arid environment, and therefore not altered by a great deal of fluid flow – here we see things that are probably millions of years old without any kind of subsequent alteration. So caves are treasure houses for protecting the materials that actually once lived in them. We’re getting real expertise in being able to detect those formations with human investigators. What we’re learning from that, I believe, is quite translatable to the robotic units that we’re anticipating.

credit:Render by R.D.Gus Frederick
Ballbot, or Planetary MicroBots. Click image for larger view.
Image Credit: Render by R.D.Gus Frederick

So I guess the answer is this: If the only traces of ancient life on Mars are in very deep pores, in sediments that are very, very deep in the planet, then these microbots are not going to uncover them. But neither are surface missions. So this is another way of looking at environments that may have preferentially housed organisms in the past, and may have preserved those materials much better than surface materials.

AM: All right. Now for the sci-fi question. Do you ever worry that some behavior among these entities might emerge that causes them to go off and do their own thing, or attack their home base?

PB: Well, I suppose it’s possible they could go awry. But because the strategy employs a limited number of basic interactions between them, even though the emergent behavior is complex, I think it’s quite steerable. And since we’re not going to mount them with any weapons capability – they’re going to be very benign – the worst that I can imagine happening is mission failure, where they decide to troop off somewhere else and do something else. But we’ve had mission failures before. We’ll have them in the future. I guess that’s a possibility.

AM: It would certainly make for an interesting story.

PB: Yes it would: Microbots Claim Mars for Selves!

AM: Are there any near-term applications of this technology?

PB: We’re very interested in their near-term use. Even on Earth there are potential spin-offs from this – being able to have little units that could get into difficult or small or hazardous places. We’re thinking about mine and cave rescue; we’re thinking about getting into human structures that have collapsed as a result of natural or man-made disasters. There are a number of non-extraterrestrial applications for these that might push the development of them sooner and also be a valuable contribution to problems that we have here on Earth.

Related Web Pages

Slime Team Website
The Mysterious Lives of Caves
Biomes, Inc.
Cave Slime
Touch-Sensitive Cyborg
Machines That Make-Out
Walking Tall Like a Human
Remote Control Insects
Unstoppable Creepy Crawlers