Slashdot Mirror
Kinect Revolutionizing Robotics
HizookRobotics writes "The Bilibot Project, an open-source robot platform based on Microsoft's Kinect, was just announced by MIT researcher Garratt Gallagher on Hizook.com. Bilibot is just the first in what will likely be a torrent of robots (both hobbyist and professional) utilizing the Kinect. This sentiment was echoed in an essay by Fred Nikgohar, CEO of RoboDynamics, who believes we've reached a watershed moment in robotics enabled by cheap 3D sensing. While much of the attention for the Kinect has focused on video gaming, perhaps robotics will be its greatest beneficiary."
I've thought for a long time that companies have been missing out on the prospect of just giving people toys to play with. All these tablets, portable gaming systems, calculators, kinect, etc. would be so much more useful if they were opened up and just plain allowed to let people play and/or tinker with them but the companies involved have been oddly non-forthcoming in this regard. It's like how with basic legos (yes, legoS!) you can build whatever you want but with the trend towards highly specific, pre-determined builds it's just so limiting. One (ridiculously expensive) thing vs. almost unlimited possibilities, which would you rather have?
Kinect was actually developed by PrimeSense.
This is not entirely correct. One part of Kinect was licensed from PrimeSense, the range camera. Microsoft developed the software technology internally. People can argue, and do, what is the most important part, but I think it is the combination that is unique.
http://en.wikipedia.org/wiki/Kinect#Technology
It's revolutionary because prior to the launch of the Kinect, if you wanted both visual input and a depth map on a robot, you had to spend hundreds, possibly thousands of dollars on LIDAR sensors, which are fussy pieces of equipment at the best of times. Within its design range, Kinect is as accurate as any LIDAR sensor, much more reliable, and waaaaay cheaper. For this reason, a lot of robot designers don't bother with LIDAR, which means you have to estimate distance and range with GPS, direct image data, or a host of other not-quite-as-accurate means. Not to mention it handles skeletal tracking, gesture recognition and other unpleasant programming tasks itself, leaving the robot designer free to do other things.
Kinect really has kind of changed the game overnight. People are very excited about being able to equip accurate depth sensors on all kinds of robots that they wouldn't have bothered with before. Even if no "new" innovations were to ever appear from Kinect, the increase in accuracy of old standbys like manipulator arms will be tremendous now that they can have depth maps. I'm at Cornell at the moment, and many of the grad students in robotics already have Kinect-based projects well underway, and even in the undergrad robot learning classes (where you typically do one semester-long project) the professor is pushing Kinect as an option.
Dislike the Electoral College? Lobby your state to join the National Popular Vote Interstate Compact.
Yes, and the combustion engine automobile was invented in 1862, but wasn't available to the general public at a low price point until 1903. Do we remember Lenoir, or Carhart, or the Duryea brothers? No, we remember Ford, who built the assembly line process that standardized and cheapened the production of automobiles.
The robotics field needs this jump to standardization of components, APIs and functionality. Yes, academia is coming up with designs all of the time, but each one is custom hardware & software, akin to Professor Carhart's steam-powered automobile in 1871. After that, it needs to move into the consumer markets, where the masses can tinker, hack and tweak the designs to add functionality, and truly innovate.
I love the kinect; I've done some great stuff in my robotics research with it already. It's a great sensor for testing out algorithms because of the high definition of the data, but it's next to useless as a long term solution for mobile robotics due to the nature of structure light sensors; the dot pattern projected by the IR camera can be easily interfered with by other kinects.
While there has been one example of two cameras working orthogonally, I can't see it expanding much more beyond that. To use more than a couple, you'd have to time the sensors to work together, or something more ingenious. Regardless, right now they're great in the lab, but the state of mobile robotics is still such that good sensors cost >$10,000.
I am not doubting that the Kinect is changing the game. I just question whether it is revolutionary or not.
Kinect is changing the game but it's not revolutionary? What's your definition of a revolution then? Before the kinect, it cost me close to $10,000 for a good 3d point cloud data. If I had more room on my robot, I might put a Hokuyo LIDAR on a pivot but that still put me back 6 grand. Today I use industrial sensors from IFM, re-purposed for Robotics. They cost about $1500, and only provide 50x64 pixels of range data, as compared to the Kinect's 320x240.
So the cheapest feasible sensor I can buy costs $1500. So here comes Microsoft. They're selling a sensors 10 times cheaper with 24 time the resolution. Now any old schmuck can buy this and test their idea for a new image segmentation algorithm. This has NEVER been possible before.
So yeah, Kinect is changing the game. That's the definition of a revolution. Just because it was done in a lab before by Ph.D.s after 10s of thousands of dollars of time, effort, and equipment doesn't diminish it. If a company started selling robot cars to the public, that would be revolutionary too, even though we can do that in the lab (for $1,000,000+).
And Microsoft can't get all the credit; none of this would be possible without ROS and the amazing Point Cloud Library. This is a second component of the kinect revolution, which, in itself is revolutionary.