Termite mounds are impressive structures. Not only can they climb many feet above the ground, but they have built-in temperature control and are beneficial to other creatures in their habitats. Perhaps most impressively, termites build without blueprints or a centralized plan.
This fact inspired Justin Werfel, a researcher at Harvard's Wyss Institute for Biologically Inspired Engineering, to look at how termites create structures. These insects became the model for a new approach to building with swarms of robots. "We learned about these colonies of huge numbers of tiny insects that together build these very large-scale, complicated structures," Werfel told TreeHugger. "Now, how could we build and program a swarm of robots so that they build things for us?"
Instead of programming the robots with a sophisticated set of coordinated step-by-step instructions, Werfel's team gave the robots a simple set of rules to follow. In other words, they didn't give the bots blueprints.
The robots are equipped with sensors, so they can tell where to stop, turn around and how to place a brick. They build along a grid without communicating with each other, instead knowing to place their blocks in an open spot or move along if a spot is filled. The bots also follow traffic rules, that tell them where other bots are and to not bump into each other.
The key to this approach is that the low-level rules are defined by the user's end result. Werfel's team demonstrated the feasibility of the approach in a paper published in the February 14 issue of Science. They tested the idea with three real robots, and with a much larger swarm of simulated robots. "In an ideal world, where everything goes right, you can really build with any number of robots and of any size," said Werfel.
There are a couple of advantages to this programming approach. The bots make for a very resilient team; if one bot breaks down the other can build without it and complete the same structure. The inverse is also true, more bots can be added without re-tooling the others and making it easy to scale up.
Werfel says that situations where it's desirable to use robots instead of people can usually be characterized by "the three D's"—dirty, dangerous or dull. The team's long-term vision is to be able to develop building systems that can be deployed in places where it would be difficult to send humans.
"So, for instance, if we want to build underwater, or to build a base on Mars, those are situations where it's very dangerous to send people," said Werfel. "But imagine if you could send a team of robots ahead to build a habitat first and have astronauts show up later, that gives you a great head start."