From some fertile minds at the University of Pennsylvania’s General Robotics, Automation, Sensing and Perception (GRASP) Lab comes the meanest looking quad-rotor micro air vehicle I’ve ever seen.
DARPA, as well as all the services, are funding research into Micro Air Vehicles, drones small enough to fit in the palm of your hand fitted with sensors that can penetrate deep inside buildings, tunnels, caves and urban canyons searching out the enemy. The Brits use the small 6-inch WASP MAV in Afghanistan.
The next step is to arm the little suckers. A networked MAV swarm carrying small explosive charges in kamikaze attacks could take out vehicles, missile batteries, radar installations or other soft targets like people. A swarm of these quad-rotor MAVs coming your way would definitely give you pause.
– Greg Grant
{ 25 comments… read them below or add one }
imagine the applications….
Now that's just plain bad-ass robotics.
… rise of the machines.
That is brutal…
It is cool, but … all the lights in fixed positions let the copter calculate it's exact position and angle at any moment. So we just need the bad guys to install a grid system of lighting ahead of time.
I'm kidding, but getting accurate position and angle data is a huge problem for MAVs in GPS denied environments. This is a cool lab demo, but unlikely to scale to the real world.
Never taken a robotoics course eh? What about infered sensors we used them to calcluate small distances in the maze robot contest.
GPS is for the bigger outside would but u dont need GPS to calcuatle your distance form an object how would that even work. The object your calculating your distace to isnt on no GPS plot?
Gyroscope lets u know your angle no? I don't get your comment at all.
you may have taken a robotics course but it was definately not a high level one. dead reckoning is a very large problem in robotics. In this video the robot has A) a prestored map of the room and the 'window' location and B) it knows it's own exact coordinates based upon its camera's view of the stationary lamps in the room. Without these two factors this robot would not have been able to perform as it did. In the future this may not be the case, but it is the case now.
I'm confused with the Velcro bit.. in order for it to land on something, someone would need to first sneak into enemy territory with their top secret mission of… "Planting tactical Velcro onto enemy hide outs"
Which we all know wont happen..
That is what the private sector is for , plausible deniability
Velcro your kidding right, u just reverse the props from lift to push and the force would keep u against the wall.
Sound is the downfall if anything. I don't think they used velcro. Just a target to land on.
The guy speaking in the video himself said "By placing velcro on the bottom of the quad-rotor and on a target the quad-rotor can perch on a surface"
the video demonstrated a proof of concept of the vehicle being able to position itself correctly for perching, something that is hard to do. To do the actual perching without velco one could use other systems, such as this:
http://www.wired.com/dangerroom/2010/04/pentagon-…
Waiting for the day we can just land one of these guys on someone's head and detonate a buckshot blast through the eye socket at night. Baddies, keep your windows closed through the hot desert summer!
This thing with the size of a table, some weapons and fielded in large numbers could own the battlefield. Ahh, the sci-fi future..
Hmm, so please clarify. When it buzzes through a window, it's calculating not hitting the frame ITSELF, and it's not just being flown there like a very controllable model plane? I'm guessing yes or there'd be no announcement.
There is talk of fixed reference points. This machine is trained in one environment, and could not perform the same tricks elsewhere? I think this is a rather fundamental point. If it came across a similar obstacle out of the lab could it maneuver without outside input?
Seems to be some dispute about how it;s operating.
A 5.7x28mm mounted underneath would be nice, and maybe reusable.
It could possibly use an artificial intelligence system to assist it in making on-site decisions in navigation. I was personally acquainted with one of the scientist at DARPA and he was doing artificial intelligence research for them as an outside consultant before he went to DARPA. He is a sharp, sharp guy.
Before anything else these little toys will be an up and close assissan's calling card. There is no reason they couldn't be made to drop a granade, release an agent, or fire small break-off missles.
i think the room is a problem but the this thing is a great targeting devise if you can fly by wire to the bad guy in the field it can be used to get over the hill and look right in to bennie boys eyes just in time to deliver that package
If memory serves me correctly, the Germans during WW II developed a little vehicle about half the size of a golf cart, packed it with explosives, and tried to guide it to its target by radio control. Fortunately for the allies, the radio proved too fragile, the signals too uncertain, the controls too unreliable, and the vehicle too prone to get stuck to do the Germans much good. These vehicles sound like a quantum leap ahead of those WW II vehicles.
@ james,
i believe this is what you're thinking of, though it was not radio comtrolled but wire guided and command detonated by wire.
http://en.wikipedia.org/wiki/Goliath_tracked_mine
Read Dean Ing's Book "Butcher Bird" And see how it could work
It's a prop driven, noisey, flying chainsaw that could be taken out by a mesquito net. I rest my case.. ( It still looks cool and flys fancy though ) :-)
sensors mounted on the quads and rotors work very smoothly, still one has to wonder if anyone would care for more help?
These will be cheap enough that clones will be flying into American compounds and Humvees in no time. The Lord of War will get around.