Slashdot Mirror
Swarms of Solar-Powered Microbots On the Way
Mike writes to tell us that Inhabitat has an interesting article, complete with some pretty pictures, about a new solar-powered swarm robot that could be used to collect data and aid in surveillance. "These mini-robots are quite revolutionary, considering that they contain all that's necessary to collect data and relay it back using one single circuit board. In the past single-chip robots have presented significant design and manufacturing challenges due in part to the use of solder as an adhesive. These new microbots use conductive adhesive to attach the components to a double-sided flexible printed circuit board using surface mount technology. The circuit is then folded into thirds and wrapped around the ASIC (application-specific integrated circuit). On top, a solar cell generates power for the robot and delivers 3.6 V to the unit, which is enough for it to walk. Locomotion is achieved via three vibrating legs, while a fourth horizontal vibrating leg is used as a touch sensor."
I have a solar calculator that works just fine under lights at night (with the dead battery removed). Why couldn't something this small draw enough power from overhead fluro's?
Me failed English...
FreeBSD over Linux. If my comments seem odd, this may explain...
They're already programmed at a particularly high level: they're swarm robots. Yout don;t give them instructions, you give them goals. Why would you want memory on board anyway when you can just broadcast it back to a central storage device? These are the sensor and manipulator portions of the swarm, they only NEED to sense and manipulate.
These things look like something right out of a science-fiction movie. I wonder how expensive they are to produce? They look light enough that you could literally spray them from a passing plane to gather intel on a suspicious site.
The ultimate way to stop any physical machine nowadays is to cut off its power supply. The ability to configure swarm-based robots that use their own powers is a new milestone at creating a potentially unstoppable force.
Grey-goo requirements checklist:
1. Decentralized: check
2. Self-sustaining: check
3. Adaptable: not yet but can be potentially achieved with sufficiently complex programming
4. Self-replicating: not yet, our last bastion of hope
What a good thing it is that robots can't fsck... yet.
Somebody remembers that Gibson novel where exactly these things were made of nano components, and therefore so light that they could fly (suspend would be the more accurate word)?
OK, I 'll give them ten years to get there ...
Every time I hear the "grey goo" debate I ponder why some people love so much to hypothesize about the potential future without taking as much at a single glimpse at the past and present. The grey goo hypothesis states that, as a result of some technological advancement, there will be a matter that is able to function and replicate by consuming available background resources until all resources are consumed and turned into more grey goo. This position utterly fails to realize that there already is matter that does exactly that. It's called life.
Life already functions in the most optimal way possible at consuming energy and replicating more of its own kind. But the "grey goo" scenario doesn't happen due to a simple natural law of diminishing returns -- as more and more grey goo (or, in our case, life) is produced, the less and less marginal advantage is there at producing more of the same kind. Identical species (or in simpler cases, where there are no "individual specimens", identical biomasses (e.g. mold, grass) first spreads out by consuming the most readily available resource, but as its numbers grow and resources dwindle, it gains less and less marginal advantage at consuming more resources, and becomes its own competitor more than a co-operator. The fact that during the billions of years that life existed on Earth, Earth has not turned into a uniform mass of a single biomatter, utterly destroys the "grey goo" hypothesis.
On top of that, there is this "adaptability" thing. As grey goo spreads more and more, and becomes its own competitor, some strains of grey goo (lets call it blue goo), through trial and error, will function better when instead of cooperating with other grey goo, exploits it, for example by consuming grey goo directly rather than consuming what the grey goo consumes. In turn, the grey goo will now have to modify its behavior to not only consume and replicate but also to defend itself against blue goo. Then we get yellow goo which likewise will consume blue goo. Then we get some violet goo which adapts to the blue-yellow goo rivalry by, for example, becoming poisonous go yellow goo to consume while offering habitat protection to blue goo in exchange for some released energy from what the blue goo consumed.
This sequence goes on and on, until we get an ultimate form (brown goo) that sufficiently adapts to consciously exploit other forms of goo for its own needs, build constructed habitats for itself, wage wars on its own kind, and occasionally debate on whether all brown goo specimens originally evolved from humble grey goo, or whether they were created by some Divine Heavenly Goo instead.
Something else to fly up my nose while riding my bike.
Locomotion is achieved via three vibrating legs, while a fourth horizontal vibrating leg is used as a touch sensor.
I understand that there are quite a few videos about on the 'net showing the use of a horizontal vibrating "leg" as a touch sensor. Not that I frequent such sites, you understand.
Wake me up when I'm able to program these in a high level language with a decent amount of memory onboard.
You see, this is the beauty of having a swarm -- you don't need any individual to have a lot of brainpower. We see the same thing in ants (and Red Sox fans), one ant has an IQ barely higher than that of a carrot but is programmed with one simple set of instructions. When millions of creatures with different simple sets of instructions end up bumping into each other and interacting, some extremely complex and "intelligent" behaviors can emerge -- ants build underground cities with temperature control and hydroponic gardens. They keep slaves and livestock. Ants wage large scale war. This is pretty impressive for a creature with only about 100 neurons each, until you realize how little each ant has to do, and that all of the higher level function comes from the millions of minute interactions between individuals, which is often personified as the "hive mind".
To the haters: You can't win. If you mod me down, I shall become more powerful than you could possibly imagine
Bring the cost down to near nothing, make them self replicating, then foist them off to agriculture. Since they use nice clean energy, they can replace tractors in planting and harvesting. In between those activities, they can tend the crop. Enough of these little dudes can monitor individual plants for disease, then treat or remove affected plants. Monitor and regulate moisture for maximum effect at each plant. Heck, they could even pollinate plants since the honeybee population has been devastated in recent years.
If I had a zillion little microbots or nanobots, I could find a LOT of better uses than spying on my neighbor. My neighbors are pretty damned homely anyway, I don't WANT to watch them doing whatever they do when I can't see them!
"Windows is like the faint smell of piss in a subway: it's there, and there's nothing you can do about it." - Charlie Br
Hizook.com has some cool videos showing the micro robots moving around. Worth checking out.
http://www.hizook.com/blog/2009/08/29/i-swarm-micro-robots-realized-impressive-full-system-integration
I don't see where bearings could give any potential advantage anyways:
- Wheels are a great big joke for transportation outside of paved or relatively smooth surfaces - look at offroad vehicles, they're horrendously inefficient in every way and there's a reason they're not small. To really get around in nature with wheels would take something like a rock crawler, a necessarily large, extremely complicated and inefficient vehicle that requires a lot of skill to use successfully.
- Fuel-burning engines are impractical at very small scales and are relatively maintenance-intensive. ICEs are pretty inefficient and turbines are only practical are fairly large scales.
- And finally the bearings themselves are relatively fragile and maintenance-intensive.
The best chassis a self-replicating autonomous robot could hope for would be a biological body that can heal itself, doesn't need lubrication systems and has limbs for transportation. Maybe a self-replicating nanobot that acts as a pathogen and 'roots' raccoon bodies (agile, opposable thumbs, pretty well-armed between the claws, teeth and other pathogens they're carrying) would be a successful one. Maybe a life cycle where the bot grows like a plant, producing a sweet infected fruit that the raccoon eats, where it infects the brain and grows more "seeds" in the digestive system would work (although it would still need to reproduce with other raccoons, perhaps also passing the bot along as an STD, or the raccoons that aren't attracted to the fruit would have a selective advantage).
A good, feasible compromise would be an insect body facsimile - it could be actuated with hydraulic systems, linear electric motors or artificial muscles, all of which are fairly robust, and it could potentially allow the robot to get itself around much better than a wheeled chassis. Imagine a grasshopper-like body that has the option of crawling or taking a huge leap.
"When information is power, privacy is freedom" - Jah-Wren Ryel