Andy Rubin has twice changed the mobile phone landscape—first as co-founder of Danger, Inc., the company that created the T-Mobile Sidekick in 2002, and later as the founder of Android, the world's most widely used smartphone OS. Earlier this year, when Rubin suddenly stepped down from his position as leader of Google's Android division, all that was said about his future was that he would "start a new chapter at Google." Now, thanks to an interview with the The New York Times, we know what Rubin's next project is: Google Robots.
The interview is light on details, but it says the "the scale of the investment... indicates that this is no cute science project." The paper also reports that for now, Google's robotics initiative will be aimed at businesses rather than consumers. (Examples include electronics assembly and warehouse robots.) Despite Rubin calling the project a "moonshot" and saying that Google needs a "10 -year vision," The Times claims Rubin's new division is not part of Google X and is meant to get a product to market sooner than later.
As usual for a new Google initiative, the company has gone on a shopping spree to jump-start the project. Google acquired seven robotics-related companies in the last six months, covering humanoid robots, computer vision, robot arms, and wheel design.
The "Google I/O 2011: Cloud Robotics" talk.
Robotics has long been a hobby of Google's. Google I/O, the company's annual developer conference, is usually crawling with bots of various shapes and sizes. At I/O 2011, Google publically acknowledged the existence of a "Google Cloud Robotics Team," and members of the team gave a joint talk with Willow Garage, the robotics company responsible for the PR2 robot. In the talk, the Googlers made the case for a cloud-powered robot, saying that the majority of the PR2's power consumption is from the computing horsepower needed for the robot to understand the world around it. By harnessing Google's extremely powerful cloud computing, however, they could offload a lot of that horsepower to the cloud. The power cost to the robot would then be very minimal.
Google is uniquely positioned to build a cloud-powered robot, as many of its current projects translate very well to robotics. Google Goggles, the smartphone app that can identify objects and text in a photo, is really an exercise in computer vision. The voice recognition present in Android and the iOS Google Search app would allow a robot to understand human speech. Google Maps could help a robot navigate the world. Google's text-to-speech engine could allow a robot to talk back to its masters.
The new robot initiative may even involve Rubin's old project, Android. Google has ported the most popular robotics software, ROS (Robot Operating System), to Java so that Android and ROS can communicate with each other. The CEO of iRobot even gave a talk at Google I/O about the idea of a tablet or smartphone as a "robot head." Since mobile devices have many components a robot needs—think cameras, microphones, speakers, screens, Wi-Fi, mobile data, and GPS—a cheap, easy way to build a robot is to just put wheels on a mobile device. iRobot fleshed out this concept in a line of headless robots called "