Do Not Be Afraid Of This Swarm of 1,000 Robots
Harvard scientists have created 1,000 interacting robots forming the biggest robot swarm ever.
Not impressed? The biggest interacting swarm until now had consisted of only a couple hundred robots. This is incredibly hard to pull off since the algorithms and software required to control so many bots at once is fiendishly difficult.
If you’re imaging hundreds of walking interacting terminator-style robots then I admire your imagination but you’re way off. We’re not talking about Asimo or even that little black rolling beeping robot on the death star. Anything like that is incredibly expensive and time consuming to build. We have to work out the basics of interaction before we attempt that.These robots were cheap and easy to assemble ($14 USD) and tiny (penny sized). They had no legs of course, not even wheels. They’re each supported by three stiff toothpick-like legs and they move by vibrating and skittering across the floor.
These bots are a balance then between the relatively sophisticated swarm behaviors expensive robots could achieve and the uselessly simplistic ones that totally lame bots could pull off. Granted, these robots are a little closer to the lame end of the spectrum. Still they’ve made some interesting developments.
Instructions to all the robots come via an over-head infrared device. This, of course, is incredibly convenient considering the number of robots they’re dealing with. Once given their kill commands…I mean their instructions…4 of the robots are designated as “seeds” and they move to specific locations that the other robots will use as a two-dimensional coordinate system. Other robots along the edge of the swarm then move until they determine they are at the boundary of the shape they are forming or behind another robot. They communicate their final position to each other so everyone knows where the open spots are. Once this process is complete they’ve been able to form specific shapes like the letter K, a sea star, or a Darwin Fish (I made that last one up).
Ok, making letters and other shapes doesn’t sound overly impressive but remember, we’re dealing with coordinating a vast number of machines here, more than ever before. The potential end-game for this type of technology makes me giggle sometimes.
What can this lead to? Well, eventually they may be able to form much more complicated shapes like the letter W. Seriously, what if they could form a much bigger robot that can shape-shift for specific tasks like moving more efficiently (rolling or swimming) or operating machinery or performing any of the innumerable task robots will be performing in the future. Similar algorithms could be used to control the thousands or million of future autonomous cars. A truly exciting possibility is the potential of programmable matter. This matter can form almost any 3D object or tool or consumer item we might desire from a chair to a cell phone to an entire house. The ultimate expression of this is what’s been called a utility fog.
Ready for some mind blowing? Utility fog is an active, polymorphic material consisting of a mega-swarm of foglets. Billions or trillions or more of them. Imagine cell-sized supercomputers with a dozen tiny extendable arms for latching onto partner devices. Clearly some form of self-replicating molecular nanotechnology will be required to create these. Such nano-scale devices should be able to reproduce almost anything (except food) down to the color, texture, optical and electrical, even biological…all quickly and easily and by voice command. Whole rooms or even buildings could change or disappear on a whim. What about talking to a friend on the other side of the planet. The foglets could potentially re-create him or her next to you on the couch and mimic their every move thousands of miles away almost instantly and indistinguishably. Creepy? maybe but what if you truly couldn’t tell the difference? (think of it as Skype video version 50). I’m not even going to mention the other possibilities you’re thinking about right now. Finally, (roll with me on this one) imagine if we could eventually replace our bodies and minds with these foglets.
Sure this all sounds way over the top and maybe impossible to you or even repellent. I whole hardheartedly disagree. From what I can tell there’s nothing in physics or materials science that would prevent most if not all of this from eventually happening. When? It’s so hard to tell but that’s besides the point really. Whether it’s a matter of decades or far far longer consider this; if we ever achieve this most impressive foglet-swarm end-game, people may still remember the foundation laid with those goofy 1,000 self-organizing robots that formed the letter K.
Swarm Video: https://www.youtube.com/watch?v=JmyTJSYw77g#t=21
Image Credits: Michael Rubenstein, Harvard University
Having read this I believed it was very informative. I appreciate
you taking the time and energy to put this informative article together.
I once again find myself spending way too much time both reading
and leaving comments. But so what, it was still worth it!
Some skepticism about this work is warranted. Even though the work is done at Harvard, technically the kilobots are at ‘science fair-level’. There is nothing challenging about the ‘robots’: very basic hardware with infrared communication. They move very slow so the assembly takes over half a day.
The physical kilobots add nothing to the depth of the research, as it is much better to simulate swarm behavior in computer models. The only legitimate purpose is to inspire middle school kids, or to bamboozle the technically illiterate out of grant money.
The ‘Grasp’ project at Upenn makes a flying swarm. which is much cooler-looking, but equally shallow: https://www.grasp.upenn.edu/
The scientific depth of all of this ‘swarm’ work is rather disappointing. Studying self-assembly and optimal design algorithms is very interesting. In fact, its my career. Decades of research, however, has brought us surprisingly little fundamental understanding. In my day job I design such algorithms for use in Integrated Circuit layout design for chips with billions (1,000,000,0000+) of components. This problem is very similar and practically relevant. The state-of-the-art in this field has barely moved the past 20 years.
[…] AND TECHNOLOGY * Robot swarms! * Ebola Patient Zero has been identified * Tick bite can make you allergic to mammalian […]
Are we not already a bunch of self-organizing foglets ?