A fascinating story about open source programmers deploying Microsoft’s Kinect hardware in amazing off-label applications.
For 25 years, the field of robotics has been bedeviled by a fundamental problem: If a robot is to move through the world, it needs to be able to create a map of its environment and understand its place within it. Roboticists have developed tools to accomplish this task, known as simultaneous localization and mapping, or SLAM. But the sensors required to build that map have traditionally been either expensive and bulky or cheap and inaccurate. Laser arrays cost a few thousand dollars and weigh several pounds, and the images they capture are only two-dimensional. Stereo cameras are less expensive, lighter, and can construct 3-D maps, but they require a massive amount of computing power. Until a reasonably priced, easier method could be designed, autonomous robots were trapped in the lab.
On November 4, a solution was discovered—in a videogame. That’s the day Microsoft released the Kinect for Xbox 360, a $150 add-on that allows players to direct the action in a game simply by moving their bodies. Most of the world focused on the controller-free interface, but roboticists saw something else entirely: an affordable, lightweight camera that could capture 3-D images in real time.
Within weeks of the device’s release, YouTube was filled with videos of Kinect-enabled robots. A group from UC Berkeley strapped a Kinect to a quadrotor—a small helicopter with four propellers—enabling it to fly autonomously around a room. A couple of students at the University of Bundeswehr Munich attached a Kinect to a robotic car and sent it through an obstacle course. And a team from the University of Warwick in the UK built a robot that had the potential to navigate around post-earthquake rubble and search for trapped victims. “When something is that cheap, it opens up all sorts of possibilities,” says Ken Conley of Willow Garage, which sells a $500 open source robotics kit that incorporates the Kinect. (The previous non-Kinect version cost $280,000.) “Now it’s in the hands of just about anybody.”
Robot freaks weren’t the only people to explore the Kinect’s possibilities. Researchers, visual artists, and pornographers have all begun cobbling together home-brewed Kinect projects and posting the results online. Artist Robert Hodgin built a makeshift motion-capture animation program that allows users to manipulate video of themselves on the fly, turning their bodies into bulbous cartoon characters or reflective mercury-like blobs. Two students at Germany’s University of Konstanz bolted a Kinect to a helmet, creating a bare-bones navigational system for the blind. And a company called ThriXXX built a rudimentary sex game that allows players to rub women’s body parts with a creepy disembodied hand.
None of these projects were sanctioned by Microsoft (especially that last one). Indeed, for the past few months, if you wanted to use the Kinect on anything other than an Xbox, you had to install homemade drivers cobbled together by a dedicated group of hackers. Yet the company’s official response to all this activity has gone from hostility to acceptance to vigorous support. In June, Microsoft expects to release a software development kit that makes it easier for any academic or hobbyist to build Windows applications using the Kinect’s camera and microphones. The company is also granting access to the high-powered algorithms that help the machine recognize individual bodies and track motion, unleashing the kind of power that was previously available to only a small group of PhDs. (Microsoft is also working on a commercial version of its software development kit, which will allow entire new businesses to be built using the Kinect’s technology.)
This is all blue sky stuff, right now, and if you watch the video you’ll see that some of the applications have that open-sourcey slapped-together look and feel. But the simple idea of using any surface as a virtual touch screen is revolutionary.
Looking for a real estate angle? Someone get Obeo on the horn. Virtual tours just got a whole lot more interesting!
Al Lorenz says:
Kinect allows not only navigation and visualization for robots, but measurement in real time. It is going to be fun to see all that comes from this technology.
July 14, 2011 — 4:05 pm
Greg Swann says:
How about a Roomba-like security guard that patrols constantly and delivers every bit of data it collects in real time to the web. Kill the power? It’s a Roomba. Kill the lights? The Kinect has IR video.
Incidentally, if all security cameras were broadcasting in real time to the web, petty crimes of all sorts would all but vanish. The bad guy spray-painted over the lens? What about all the other lenses that saw him coming and going?
That’s a web business idea, incidentally, the visual recreation of a route traveled at a secific time in the past. Totally doable if all the source video is available.
July 14, 2011 — 4:18 pm