Robot Builds Ramp by Randomly Flinging 3,600 Toothpicks

With amorphous building materials like toothpicks and foam, robots build structures to climb over obstacles

3 min read
Robot Builds Ramp by Randomly Flinging 3,600 Toothpicks

One way of making a simple robot more capable is to give it the capacity to modify its environment. We've seen this in practice in the last year or two with robots that have the capacity to create tools, build buildings, and even manufacture other robots. This concept can be taken even farther, though, with robots that can construct large structures out of amorphous materials like glue, foam, and toothpicks.

Robots are fairly decent at using prefabricated materials to build things, but when you get out into an unstructured environment (whether it’s somewhere like a forest or a city after a major disaster), it doesn't really make sense to bring anything prefabricated, because you have no idea what you’re going to need. What makes more sense is to bring along building materials that can be adapted into whatever you want on-site, which means large amounts of stuff that you can then build up into exactly what you need.

Researchers from Harvard University and Worcester Polytechnic Institute have taken inspiration from animals like weaver birds, termites, and beavers, and developed robots capable of using cheap materials to build large structures. Let’s start with the beavers, ‘cause they’re the cutest. Beavers (and weaver birds) build things by sticking together large amounts of, uh, sticks. This new robot can do something very similar with prefabricated sticks (toothpicks) and glue. At this point, the robot in question has a deposition mechanism with which it just “flings" individual toothpicks (a technical term, apparently) after adding glue to them. While it has approximately zero control over placing the toothpicks into any sort of arrangement that would make structural sense, the sheer number of toothpicks plus a generous helping of glue means that eventually, the bot can build ramps or anything else that on some level consists of a random pile of wood and glue.

Termites, on the other hand, build structures out of mud, without any underlying framework. Robots can do something similar with urethane casting foam. By successively depositing layers of liquid that puff up as they dry, a robot can build ramps or any anything else that on some level consists of of a random blob of foam. Yes, there’s a theme here, and it’s not a sophisticated one, but part of the point is that it doesn’t have to be sophisticated: with cheap materials and patient robots, you can make structures that are perfectly functional, even if they’re not particularly efficient (or pretty).

Here’s a video of both of the robots in action:

You might have noticed that the toothpicks and glue weren't just piling up on the ground, but were actually sticking to the vertical wall a little bit, too. The researchers suggest that it might be possible to build things like arches and bridges with the toothpick-flinging method, once they figure out how to get the targeting down to a little more accurate than the aforementioned random piles.

Future research will also make the ramp building both autonomous and adaptive, meaning that the robots will be able to look at an obstacle, figure out on their own what sort of structure they need to create to get over it, and then build that structure by themselves. This may involve getting multiple robots to work together in a swarm, perhaps combining different materials and mechanisms to come up with the most efficient ways of surmounting every barrier we try and stop them with.

"Materials and Mechanisms for Amorphous Robotic Construction," by Nils Napp, Olive R. Rappoli, Jessica M. Wu, and Radhika Nagpal from Harvard University and Worcester Polytechnic Institute, was presented today at the 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems in Vilamoura, Portugal.

[ Harvard SSR ]

The Conversation (0)

How the U.S. Army Is Turning Robots Into Team Players

Engineers battle the limits of deep learning for battlefield bots

11 min read
Robot with threads near a fallen branch

RoMan, the Army Research Laboratory's robotic manipulator, considers the best way to grasp and move a tree branch at the Adelphi Laboratory Center, in Maryland.

Evan Ackerman

This article is part of our special report on AI, “The Great AI Reckoning.

"I should probably not be standing this close," I think to myself, as the robot slowly approaches a large tree branch on the floor in front of me. It's not the size of the branch that makes me nervous—it's that the robot is operating autonomously, and that while I know what it's supposed to do, I'm not entirely sure what it will do. If everything works the way the roboticists at the U.S. Army Research Laboratory (ARL) in Adelphi, Md., expect, the robot will identify the branch, grasp it, and drag it out of the way. These folks know what they're doing, but I've spent enough time around robots that I take a small step backwards anyway.

The robot, named RoMan, for Robotic Manipulator, is about the size of a large lawn mower, with a tracked base that helps it handle most kinds of terrain. At the front, it has a squat torso equipped with cameras and depth sensors, as well as a pair of arms that were harvested from a prototype disaster-response robot originally developed at NASA's Jet Propulsion Laboratory for a DARPA robotics competition. RoMan's job today is roadway clearing, a multistep task that ARL wants the robot to complete as autonomously as possible. Instead of instructing the robot to grasp specific objects in specific ways and move them to specific places, the operators tell RoMan to "go clear a path." It's then up to the robot to make all the decisions necessary to achieve that objective.

Keep Reading ↓ Show less