If you’re a science fiction fan, or married to one, you’ve probably watched Terminator, Battlestar Galactica, and I, Robot. Widely entertaining, these shows revolve around the premise that humans create machines, and somewhere along the line, the machines become self-aware and turn on their human masters. Luckily, these stories are relegated to the land of fiction. However, the rise of robotics is not. In fact, RoboEarth recently announced that it’s developed an open source cloud engine called Rapyuta, which will allow robots to share knowledge and learn from each other. While this news might be creating nightmarish visions of SkyNet, it also presents a potential investor gold mine, as robotics could be akin to the next PC or iPhone. Here’s what you need to know.
Photo Credit: Ѕolo via Compfight cc
They’re heeere…
Believe it or not, robots already play a part in human life. iRobot makes robots like the Roomba vacuum, and FirstLook, a robot the military can use for situational awareness. Companies like Hansen Medical and Intuitive Surgical make medical-assisting robots that aid in complex surgeries. And companies like ABB make manufacturing and industrial robots. Whereas these robots probably don’t conjure images of Cylons or T-1000s, they are robots nonetheless, and part of a progression in technology.
The next steps in this progression are robots that can adapt to their environment, react to changes, and alter their behavior, all without human intervention. One such robot is Baxter, a robot made by Rethink Robotics. This robot is mainly for manufacturing companies, and was designed so that instead of going overseas for cheap labor, companies could stay stateside and use robots in their manufacturing plants, with humans overseeing the robots.
Rodney Brooks, Rethink Robotics‘ CEO, said that he believes robots “will become as common place in our lives as turning to a search engine is today,” and that as technology progresses, robots will become cheaper and more adaptive to their environments, allowing the average person to train robots to do everyday tasks – like the dishes.
This is where cloud computing, Google‘s Goggles — an image recognition service for mobile devices, and cloud storage, and Microsoft‘s Kinetics, come into play. As Brooks stated, a goal in robotics is to create inexpensive robots that can do everyday tasks. But one of the problems has been object recognition. However, researchers at Berkeley have developed a custom version of Goggles that runs on Google’s image recognition system that will facilitate training and recognition. According to the researchers: “The training endpoint accepts 2D images of objects with labels identifying the object. The recognition endpoint accepts an image, and based on the set of features, either returns the object’s identifier along with a probability of correctness, or reports failure.”
In other words, cameras, along with Microsoft’s Kinetics, allow the robot to take pictures and create 3-D scans of objects. Then, images are uploaded via the cloud. Based on …read more
Source: FULL ARTICLE at DailyFinance
