A robot that can accurately identify human targets and know exactly what we're doing. Thanks, science!
The Kinect is more than just a way to convince your parents to get an Xbox 360. Despite being billed as a casual gaming toy, the console add-on actually boasts a camera with a powerful object recognition system. MIT PhD student Philipp Robbel decided that the Kinect's Eye could be put to better use, so Macgyvered a Kinect onto an iRobot Create, and thus was born the KinectBot.
Robbel's mad-science creation can perceive and map out its environment in full 3-D. Kinect's optics gather data, then sends it wirelessly to a host. The host then creates a graphical 3-D representation of the KinectRobot's surroundings.
It can also use the Kinect as it was originally intended: for gesture controls. Once the robot detects a human in its field of vision, it suppresses its natural urge to kill. It will instead wait for the human target, er, operator, to point out where to go, then head off in that general direction. Let's all thank Philipp for having the foresight to give his robot a Roomba's body. Imagine if he hooked it up to a more mobile, weaponized machine.
[source: Singularity Hub]