Hi Vinko, Yes, Kumar's quadrocopters *do* use Vicon cameras for tracking, and they're clearly visible throughout that video whenever they're interacting with the environment or acting cooperatively. (The bits where they're just demonstrating manouevrability and flying them in the auditorium are under tele-remote operation.)
I dislike the idea of flying robots in my home for many many reasons, but as just one motivating example - have you ever tried flying a quadrocopter near a miniature foxy terrier? ;-) As for other reasons - the payload capacity is minimal, power drain is huge so flight time is tiny, stability is nigh on impossible so image projection tasks are out, you have to do collision avoidance in 3D rather than 2D so processing requirements are huge, which means you need a real computer, which given all the previous problems means you almost certainly have to do the computation offboard, which means you need now to have a consistent low-latency RF comms channel to a moving target in a cluttered house with walls and RF noise out the wazoo - and you're not allowed to use any frequencies other than those available to consumer products and you must comply with maximum transmission power regulations. Frankly, good luck with that. Regarding your suggestions, with the utmost respect, they form a fairly typical response that I get when I ask a random person about robots. Essentially, you want a household slave. That's fine, and fair enough - I'd love one too - except that the technology required to do any of these manipulation tasks as well as a human can will cost you many millions of dollars and almost certainly require fixed infrastructure in your house. For a lot less than a million dollars, you can hire a maid to do this, and for at least the next twenty years I guarantee they'll do a better job than a robot. It's not so much that it's too hard, it's that it's too hard and expensive and improbable that it can be made at a consumer level of robustness. What I'm trying to do is fine a combination of technologies that *is* affordable, and which in combination can be made rock solid so that; It. Just. Works. The downside obviously is that it necessarily has less features and abilities than the million dollar solution. The upside is that I can actually make this happen. Cheers, Tom On 6 April 2012 16:14, Vinko Grgic <vinko.gr...@gmail.com> wrote: > Hi Tom, > > Firstly, good on your for dreaming big and deciding to go for it. > We've all seen robots in film for a looooooong, long time...but > nothing really noteworthy to show for it in real life (for me a guitar > strumming/beer gathering robot just doesn't really hit the spot if you > know what I mean). > > Regarding your comment on quad-copters, I am going to assume that you > don't mean Vijay Kumar's flying robots? > http://www.ted.com/talks/vijay_kumar_robots_that_fly_and_cooperate.html > ...cos they seem to handle themselves quite well without room cameras. > > I personally like the idea of flying robots much more than humanoid/ > wheel robots. Sam nailed it with his comment about being allergic to > stairs...indeed they are. Flying robots leapfrog the entire problem, > along with increased velocity/agility/energy efficiency/power-to- > weight ratio. I would much more prefer a flying quad-copter bot to get > my beer from the fridge upon the command "buzz, fetch me a beer!". > Flying robots can whizz through the house and let you know where > people are too...if you really want that. I personally find robots > completely uninteresting when it comes to doing "cool" stuff that > isn't at all useful. > > So to answer your question, and agree with all "reminder" vocal type > comments made by Sam, here is my list (I don't want to hear "it's too > hard"...until then I just simply wont buy one for ~$1500-5000): > > - do the dishes > - take out the rubbish/recycling > - fetch said beer from fridge > - clean/mop floors > - reach hard to get places (top of cupboards/under bed/ceiling...etc, > other very high places) > - clean out the gutter > - mow the lawn/garden snipping > - buy groceries from vocalized list (better still go to shop for me, > record live video and let me choose through it) > - operate printer (not a biggie...but easy to do in a "connected" > household) > - control pets or let them in/out of house...feed them? > - never talk to me uninvited...unless it's a reminder/warning alert > - through out old food from fridge (keep track of this > too...autonomous) > - and, of course the granddaddy of them all...I can see what it sees > (don't care how this happens) because the possibilities are endless - > especially if it flies! > > ...yeah, that's all I got for now. I was really aiming for the > "household help" thing. But generally I want to also have the option > to never drive, a robot can do it better/safer...I can read/watch > video/converse etc. > > All the best with it, > Vinko > @vinkogrgic > > On Apr 5, 9:19 pm, Tom Allen <t...@jugglethis.net> wrote: > > Thanks for the responses guys. Some very interesting points and ideas to > > consider. > > > > I'm not convinced by quadrocopters. Almost all the cool videos you see > out > > of ETH Zurich and UPenn's Grasp Lab involve rooms with numerous Vicon > > cameras ($100k+ each) tracking each robot. There's no way to do this in > the > > home. Even if the prices came down, I don't want a solution that > > *relies*on fixed infrastructure, even if part of what I'm suggesting > > involves home > > automation and implies infrastructure. The setup and calibration of a > > system like that might be suitable for B2B with a service contract, but > > doubtful for consumers. I would like the robot to be *capable* of > operating > > just fine on its own, but to be capable of greater things if part of an > > interconnected home. > > > > In terms of TVs/computers/smartphones being able to do the things I'm > > suggesting, this is true in part, but definitely not as a whole. Part of > > the reason for using a mobile robot is that one set of sensors can be > moved > > to perceive different areas of the home - fixed TVs and computers can't > do > > this unless you have lots of them, and then the costs blow out. > Smartphones > > don't currently have an RGBD camera or any other great means of building > 3D > > maps, and even if we tried to do this with camera data only, they don't > > have the processing power onboard to achieve it. Sure, we could offload > > that computation to the cloud, but if you're going to all that trouble > you > > should probably just buy my robot instead... :-) > > > > Incidentally, some of the people in my research group have done work in > > tracking people in known environments when they continually go out of the > > field of view of a robot (or group of robots). It is certainly possible, > > although obviously the longer they're out of view, the less trust you can > > place in the estimated solution. > > > > Someone I bumped into today suggested that even if I'm not able to > include > > a manipulator, I should at least sell a cup-holder attachment. > > > > Cheers, > > Tom > > -- > You received this message because you are subscribed to the Silicon Beach > Australia mailing list. Vist http://siliconbeachaustralia.org for more > > Forum rules > 1) No lurkers! It is expected that you introduce yourself. > 2) No jobs postings. You can use http://siliconbeachaustralia.org/jobs > > > To post to this group, send email to > silicon-beach-australia@googlegroups.com > To unsubscribe from this group, send email to > silicon-beach-australia+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/silicon-beach-australia?hl=en?hl=en > -- You received this message because you are subscribed to the Silicon Beach Australia mailing list. Vist http://siliconbeachaustralia.org for more Forum rules 1) No lurkers! It is expected that you introduce yourself. 2) No jobs postings. You can use http://siliconbeachaustralia.org/jobs To post to this group, send email to silicon-beach-australia@googlegroups.com To unsubscribe from this group, send email to silicon-beach-australia+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/silicon-beach-australia?hl=en?hl=en
