Slashdot Mirror
Swarms of Microrobots Over Europe?
Roland Piquepaille writes "In 'Mini robots to undertake major tasks?,' IST Results describes a EU-funded project which allowed to build several kinds of microrobots in the last three years. These robots are very small (about 1.5 cm by 3 cm), have limited on-board intelligence and are wirelessly controlled by a central robot control system. A follow-on project has already started, with an even more ambitious goal: deploy 'real' swarms of up to 1,000 robot clients. Such robot swarms are expected to perform 'a variety of applications, including micro assembly, biological, medical or cleaning tasks.' Read more for additional details, pictures and references about this follow-on project not described by the article mentioned above."
Well such an idea sounds reasonable enough. In fact in Richard Feynman's "plenty of room at the bottom" famous speech, he describes something similar: building small machines that are then used to build even smaller machines, until finally you have atomic-scale machines. This speech is considered by many to be the "original idea" for nanotechnology.
So why don't we have nanobots yet? Well it turns out its a little more complicated than that. The basic problem is that designs for large-scale robots do not work at smaller scales. You can take macroscopic engineering principles and scale them up or down to a point, but eventually they break down. The design of a 200ft long bridge is not just a 4X scale version of a 50ft bridge, after all.
If you read Drexler's technical book on the subject (Nanosystems) he goes into detail on how various properties (strength, elasticity, conductivity) scale down to the nano realm. Some of them scale favorably, whereas others do not. Thus nano-scale robots will not merely be "small versions" of macro robots. For instance the viscosity of a liquid becomes much more important than gravity, at small scales (whereas at large scales dealing with inertia and gravity are important).
My point is that robots cannot simply build exact (but smaller) copies of themselves. The half-sized robots will be useless within a generation or two, and will require new designs, optimized for that size. (Added to that, robot designs that are self-replicating are not trivial to begin with, at any size-scale!)