Drone quadracopters throw and catch inverted pendulum All those stories you've read lately about drones invading your privacy by photographing you from afar? Kids' stuff … now that we have the evidence below of drones throwing and catching an inverted pendulum. That trick has been accomplished by boffins at ETH Zurich’s Institute for Dynamic Systems and Control, where a “Flying …
The biggest worry won't be 'a' drone hovering around your bedroom window, it will be the 1000's flying through the air controlled by one handed hormonal teenagers with the likelihood that the battery will run down and the thing will drop out of the sky and hit you on the head.
We're is my hard hat....
No, I'm being serious. Not the one in the video, that's probably beyond my budget. I just happened to be shopping for one of these and came across this story, and figured that The Reg comment section is one of the better places to ask for advice. Can someone please post some UK links? Thanks.
"Are you serious? you really can't find them for yourself?
Or are you asking for some good reviews"
Yeah, I'm looking for some decent sites where I can find advice & purchase the parts. Yes, I can use Google and I tried that before putting the question in here. The problem is there are TOO MANY links, and rather than spending 1/2 hour trying to sift through them and try to work out what would be good value vs. what would be just cheap rubbish. I know I'm being a bit lazy by trying to call on the community to share their experience, but Larry Wall taught me that Laziness is a good thing, and I took him literally and applied that to everything in life.
Perhaps I phrased the question incorrectly. Perhaps some people need to stop taking things literally, and look for meaning in a question.
Try this, a review of a low-cost Quadropter (about $40 USD, available in the UK too), followed by a discussion thread. Its rivals are mentioned in the thread. It seems well suited for beginners.
Syma X1
http://www.rcgroups.com/forums/showthread.php?t=1797545
I have a conventional Syma helicopter, S105G (more stable than it is agile, so suitable for people learning in normal sized rooms), and I can't believe the number of times it has bounced of walls and ceilings with no real damage... since it was around £20 all-in, I was prepared to be generous. The battery has required soldering back onto the PCB- not hard- after my mate swatted it from the air cos we were buzzing him with it.
This reminds me of an article I read in New scientist a long time back. Some researchers were doing a “similar” experiment wherein they had a pendulum pivoted at its base on the back of a motorised wheeled truck and they were devising an algorithm to program the truck to keep the pendulum balanced by moving backwards and forwards.
Problem was they didn’t have the processing power to solve the necessary simultaneous equations in real-time so they employed a fuzzy logic approach. This looked at angular displacement, velocity and direction to compute a move backwards/forwards quickly/slowly solution to keep the pendulum upright.
The big advantage was it was (i) easier to compute (ii) cope with different length/mass pendulums within – quite wide IIRC – ranges without alteration of the algorithm. Whereas as described in the article when you do the “actual” calculations you have to allow for every parameter change.
Of course it was a real Heath Robinson affair with data and power cables trailing from it. I’d bet they’d be amazed to see that vid though.
For those of us who studied back in the day, there were things called "analogue computers" which could piss all over digital computers for some task - simultaneous equations being one. What takes a lot of computing power when done digitally is simple mathematics in the real world.
"BAE's "Gorgon Stare", for example, can track multiple objects within a 4KM radius beneath a drone."
Wikipedia says:
"Gorgon Stare is a video capture technology developed by the United States military. "
and
"Gorgon Stare is being developed and tested on the MQ-9 Reaper at Eglin Air Force Base, Florida. with the 53rd Wing. These sensors pods have been in development since 2009 by the Air Force’s Big Safari group and Sierra Nevada Corp.[9]"
Where's BAE in that?
My first thought was how can these things be weaponized. Depending on how much weight these things can lift, speed, battery life, miniaturization, AI etc these things could eventually render meat based infantry obsolete.
If a state can afford to cheaply mass produce 1000s (or millions?) of such drones armed with blades (like the manhacks in halflife2), or even explosives then they effectively have a form of cheap robotic infantry without having to develop difficult navigation and obstacle avoidance AI for walking or wheel based robots.
Question is for the price of 1 human grunt how many manhacks could be afforded? Given they can be stored for nothing (unlike meat soldiers) I imagine a lot. Defensively they could be scattered around an area to sit idle like landmines, the difference being that unlike landmines they can move in to attack a target. Or they could be used offensively by sending in swarms of them to attack bases at night. I don't see how current ground infantry would be able to survive in an environment swarming with these things. It would be hard enough to see them let alone hit them.
These kind of drones will also be a big problem for counter-terrorism. How is any leader going to be able to do any outdoors speech in the future when an assassin could just release a dozen kamikaze drones 2 miles away?
I can imagine there will soon be an international ban on development of such drones for warfare, because typically when there is a game-changing weapon that is cheap and dangerous the big states ban it. For example using cheap lasers to blind soldiers or pilots is banned, but you are perfectly allowed to blow them up with expensive missiles!
It's not clear from the article but I'm sure these drones aren't autonomous, there must be a motion capture system out-of-band which is doing all the processing and drone control, hence the white sheets to give good contrast for the system's cameras.
Being able to do the same thing with autonomous drones must be much harder just due to their viewpoints, even if the processing and drone comms could all be done on-board. I guess that'll all be do-able in the near future though.
Still very impressive nonetheless.
You're right - definitely not autonomous. Not only the white sheets but there are two reflective markers on the pole, making its position and orientation easy to calculate without intensive image processing. There will be cameras placed at different viewpoints to enable this to be done in three dimensions.
The copters also have two markers on them to allow their positions and orientations to be calculated in the same way. The cameras have LEDS round the lens to produce very bright reflections from the markers, and image processing occurs inside the cameras so there's no video to be analysed at all - just a stream of co-ordinates. This keeps the delay to a minimum, which is needed to allow reaction to be quick enough.
I really must wander in there next week :).
I don't care that it isn't autonomous (and probably needs a number of cameras to ensure the entire 3D movement space is monitored), it's still very impressive. I also like how the copter controls stop them from turning over after what are actually rather wild movements.
Hmm. Couple of cameras, fridge with cold beer, serving platter.. Must investigate :)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
