Researchers at Cornell's Personal Robotics Lab have succeeded in creating arguably the greatest advent in technological history: a robot that anticipates when you're in need of a beer, and then pours a fresh one for you. While the beer-pouring robot is far from the most graceful mechanical butler in existence, it's probably one of the few that can dispense a Budweiser for you without spilling it everywhere.
The robot's accuracy is the result of a clever algorithm that tracks human behavior and uses it to predict the probability of someone's actions. For example, in the video, a user approaches a refrigerator holding a pot with both hands. The robot realizes that he is attempting to put the pot in the refrigerator and opens the door.
"We extract the general principles of how people behave," said Ashutosh Saxena, Cornell professor of computer science and co-author of a new study tied to the research. "Drinking coffee is a big activity, but there are several parts to it."
The robot uses the 3-D camera built into the Microsoft Kinect to record a user's behavior and compiles a database of 3-D videos from which appropriate responses are extrapolated. At any given juncture, the robot looks at what a user is doing and predicts whether they are most likely to clean, drink or eat something next, among other possible behaviors. The robot then assists.
Saxena notes that while the robot's reactions are intuitive, it is far from a true artificial intelligence. Though it perpetually studies human behavior and adapts its behavior accordingly, it still only has a few, pre-programmed responses.
"Even though humans are predictable, they are only predictable part of the time," Saxena said. "The future would be to figure out how the robot plans its action. Right now we are almost hard-coding the responses, but there should be a way for the robot to learn how to respond."
It's unlikely that the beer-pouring robot will be available to anyone for purchase in the near future, but if you'd like to meet him in person, Saxena along with Cornell graduate student Hema S. Koppula will be presenting their research in Atlanta at the International Conference of Machine Learning from June 18-21.
