Dec 13 2011

Quadrotors Will Do Everything (Well, Almost)

""){ ?> By Valkyrie Ice


About a year ago I wrote an article for H+ magazine on the use of quadrotors for a variety of purposes, ranging from VR telepresence units to sensor platforms for dangerous environments to construction.

So, you can imagine my reaction on reading this article on Singularity Hub.  In short, it’s about a demonstration of robotic assembly, done by quadrotors under computer control, building a 20 foot tall tower out of lightweight foam blocks. Foam might not sound impressive, but it’s a public demonstration, so I’m sure foam was chosen not only because it’s light enough to not place a major strain on the copters, but because it’s soft enough to not cause injuries if the tower falls over. The materials are meaningless however, because it’s the control systems that are the real story. Fifty quadrotors will fly under complete computer control, having to navigate not just the static environment, but the variable obstacle course of all the other quadrotors, the changing environment of the tower being built and maybe even having to dodge the occasional overly curious onlooker. As you can probably imagine, I had to grin. Not even a full year later, and already we’re seeing stories about quadrotors being used as I described.

But I’m not the only one who’s seen how useful quadrotor could be. In a recent blog post, K. Eric Drexler asked “Where are the Parrots?” He looks at the robots used to explore the Fukushima reactor, a pair of Monirobo’s, a track based one armed robot that have a top speed of 2.4 kph, and weigh 600kg, and has to wonder why such clumsy robots were being used when the Parrot AR drone makes a far superior platform for the job. He points out that  many “Very Serious People” are dismissive of “toys”

So I decided to do a review and take a look at what sort of developments have been happening with quadrotors over the last year. First up, I have a video from January of 2011, just a few months after my original article.

As you can see, this features construction with modular materials… in this case, magnetically connectable girders. It provides an illustration of the most basic concept of the quadrotor construction battalion.

However, to really appreciate the potential here I have another video for you

That’s a video of China’s Broad Group building a modular hotel in less than one week. Now replace every human worker in the video with a quadrotor and you can probably guess what the upcoming demo is going to look like.

Precision swarming has also made advances since that first video, as this one from September of 2011 illustrates.

These videos are from the ETH labs in Zurich, and are part of a great series of quadrotor developments they have made, but autonomous flight is not the only kind of developments they are working on.

I found a very interesting video in which they are demonstrating a “control interface” that is entirely virtual, powered by a Kinect.

While I think full “mind control” of quadrotors via an emotive epoch style headset is what will eventually become the primary control interface of an RTU drone, the Kinect demo shows how intuitive we can make the control systems for everyday use of quadroters. This ease of use is one of the primary advantages of using quadrotors as camera and sensor platforms for dangerous environment navigation, like the Fukushima reactors.

There’s lots more interesting videos out there covering the many capabilities of quadrotors, from DIY projects to various university reports, and they all continue to say the same thing I first thought a year ago. Quadrotors are neither a toy, nor a curiosity. They are the first primitive stages of a variety of useful tools that will reshape how we do many things. I’m looking forward to seeing videos of the construction demo, because I love seeing the future be developed right in front of me.

And of course, getting to say “told you so!” : )

  • By star0, December 14, 2011 @ 6:42 am

    I’m not convinced that quadcopters will be the builders of the future. Buildings still require drilling and sliding objects into place, even if they are prefab; and copters just don’t seem cut out for that job by themselves (maybe with some human assistance they might be useful in building construction, though).

    I think a much better task for quadcopters is litter cleanup. Imagine outfitting a small fleet of copters with a camera sensor and some AI to identify litter (which should be a fairly easy recognition task, relatively speaking). All the copter would have to do is to fly over the litter, lower a grasper to pick it up, and then fly the litter to a trash heap. One could clean up an entire city in no time at all this way!

  • By Valkyrie Ice, December 19, 2011 @ 6:38 pm

    That’s a a far more complex task than I think you realize star0. It’s far easier to program a complex task in bounded space than in unbounded space.

    Additionally, I have the impression you think the Quads will all be the same size, whereas I am confident that a range of sizes will be used, from heavy lift units able to move tons of material, to small units that have multiple graspers to enable the use of some for gripping a stable object like a girder while performing a secondary action such as sliding or drilling.

  • By David P Williams, May 8, 2012 @ 11:02 am

    The upside has to take into account human need for meaningful work when all “Labor” has been robotized. I am struch, when observing my grandchildren, with the extent to which ART is central to the ways they are living their lives, whether their “profession” is lawering, bricklaying, school teaching, “mobility mamagement” etc.
    When here is no need for humans to “make or manage” THINGS, we must build art into the central WORK of humans.

Other Links to this Post