Hundreds of thousands of ants can walk along narrow lanes in and out of their nest, and
the flow doesn’t even stop when two ants bump into each other.
They all just keep on moving.
Scientists have noticed this, and they think we can learn a thing or two from these tiny
travelers -- and lots of other animals that do similar things.
Colonies of termites, herds of zebras, flocks of birds, schools of fish -- they all exhibit
some kind of swarming behavior, where a group of animals collectively acts as one big thing.
And they can do great things together, like perfectly synchronizing their movements, or
building huge mounds.
Wouldn’t it be great if we could get machines to act like that?
The science of swarming behavior is inspiring scientists to build robots that in the future
might be able to help with everything from building construction to search-and-rescue missions.
The thing that makes swarming behavior so perfect for robotics is that in a swarm, no
one member does anything that’s too complicated.
An animal is just following a few simple rules, like staying the same distance from all of
its neighbors.
That means you don’t need to make the robot super fancy, and you don’t need to program
each one to tell it exactly what to do.
Instead, you just give a bunch of robots the same basic rules, and because of how those
rules play out in large numbers, the group will self-organize, and figure out how to
do whatever complicated thing you want them to do.
This is already a reality in today’s robotics.
In 2014, we told you about some researchers at Harvard who made over a thousand robots
that could arrange themselves into almost any pattern or shape the scientists wanted.
The scientists never told any individual robot where to go.
Instead, they just gave each one of them the same simple rules to follow: like, measure
how far you are from your neighbors, or find an outer edge of your robot swarm, and move
along that edge.
By doing those things over and over, the robots figured out exactly where to go.
Those same Harvard engineers also took some inspiration from termites to make robots that
could build pyramids, castles, and other structures out of foam blocks.
In this case, they borrowed a strategy that termites use, known as stigmergy -- a method
of indirectly communicating with each other to reach a common goal.
When humans work on huge construction projects, we need checklists and blueprints and chains
of command.
And all that involves a lot of direct communication.
But termites build by paying attention to tiny clues left over by fellow termites in
their environment.
When they make mud balls, they add in some pheromones, which tells other termites where
to build.
Each termite is really doing its own thing, but this indirect form of communication allows
them to coordinate their actions.
At Harvard, researchers used a similar idea to design robots that could place blocks based
on what the structure looked like at the moment.
So one robot could put its block somewhere that indicated where the robots behind it
should put their own blocks down.
So, they weren’t just blindly building because of how they were programmed.
Instead, the robots could adapt on the fly, even when the researchers tried to mess with
them by moving blocks that robots had previously put down.
Each robot placed its block based on how the block that was put down before it was oriented.
These robot swarms are so far confined to labs, but the idea is to eventually have them
work for us, and solve real-world problems.
Some variation of these stigmergic robots might be able to build things in dangerous
places, like disaster areas -- or even on Mars.
So soon, there might be robot swarms all over the place.
...which brings us back to traffic.
Scientists have found that car traffic would flow a lot more smoothly if cars acted more
like members of swarms, with everyone following the same simple rules -- like staying the
same distance from your neighbors and letting nearby cars know what you’re doing.
Humans, you might have noticed, are pretty bad at this.
Some people don’t even communicate well enough to use their turn signal and they sit
there at the green light ‘cause, I don’t know, what are you doing?
Playing Angry Birds?
But we could use swarm-based technology to teach self-driving cars to stay the same distance
from cars around them, or we could design them so that they periodically let other cars
on the road know what they’re doing.
Like everything with autonomous vehicles, it’s all still in the testing and development phase.
But the natural logic of swarms might someday be our ticket to less gridlock.
Thanks for watching this episode of SciShow, which was brought to you by our patrons on Patreon.
If you want to help support this show, just go to patreon.com/scishow.
And don’t forget to go to youtube.com/scishow and subscribe!