The past five to ten years have seen the birth of microbotics. A whole range of components that are vital for building robots, such as actuators, motors or batteries, became available in micro-scale only fairly recently. Finally enthusiasts got what they needed to put their own systems together, and the whole field benefited from their work. But there are obvious limitations to scaling down robots full of sensors, motors, and other mechanisms. That is, unless you make the machines extremely simple, which is exactly what Ron Pelrine of SRI International has done. His work on levitated microrobots may have powerful implications for robotics, and is likely to bring us a step closer to fast, precise and affordable robotic systems comprising thousands, if not millions of microrobots.

The idea is remarkably simple. The robots are stripped of pretty much everything, including sensors, actuation systems and power source. The off-loaded components are incorporated into a complex, off-board control system. What is left are just simple clusters of magnets that can be customized with appendages designed for particular purposes. The robots hover in the air, controlled by the magnetic field generated by circuits below the work surface. This approach has several advantages over self-contained microrobots. These advantages may turn out to be significant enough for the idea to become a paradigm-shifter.

First of all, using a magnetic field to have your robots float over the work surface solves one of the most important problems common to all mechanical systems. Levitated microbots have practically zero wear, so they can operate for much longer than their friction-prone counterparts. This greatly reduces the likelihood of physical damage occurring, and there are relatively few other mishaps that could potentially befall a simple bundle of magnets. There just isn't much that can go wrong in such an extremely simplified system.

Second, as could be expected of robots powered by the same technology that propels super-fast trains, such micro machines are remarkably nimble. They are, most likely, the fastest robots out there in terms of relative speed. The ones pictured in the videos below measure from 0.1 to one centimeter across, and they can sprint at a stunning 217 body lengths per second (a cheetah can run at 18 body lengths per second).

What's more, these tiny robots can be controlled with great precision, with movement repeatability potentially reaching about 40 nanometers. This is especially important if you think about complex systems comprised of thousands of microbots, each of them busy performing its tiny bit for the greater good. Precision is the prerequisite for these robots, in order not to step into each other's way.

However, the biggest advantage of this model is that it makes use of materials that are already available widely and cheaply. Suddenly, all you need to build your own micro-scale factory is some off-the-shelf circuitry and a few magnets (though a PhD in robotics would probably help a great deal, too). Abundant evidence shows that research projects stand to gain from the involvement of a wider community - and there certainly is no shortage of robotics enthusiasts. Low entry barriers mean that we may soon find full micro-scale factories operating in the backyards of robotics fans, hobbyists and professionals alike.

And what could such factories be good for? Dr. Perline mentions, among other things, medical applications, such as cell printing and tissue growth. He can also see the robots used in rapid prototyping of items with embedded electronics. Since scaling such systems down would be relatively easy to achieve, for the first time microrobots could be used to assemble materials at the microstructural level. This opens a whole new world of possibilities in terms of material properties.

But there's more. Perline even goes as far as to suggest that the systems could replicate themselves, at least partially. Of course these are just predictions, but it's hard not to see the potential.