Added: 30th April 2015
PR2, or “Robobarista”, as the allegedly “cute” little robot has been named, has been developed by a team at Cornell University in conjunction with instructions provided by internet users in an imaginative crowd-sourced project. The team however didn’t say how many cups of coffee either they or the users providing the instructions had drunk in the course of educating “Robobarista”…!
“Robobarista” is part of a larger programme designed to create robots that can interact with a range of household items. The apparently simple everyday tasks performed by humans are in fact extremely complex to break down and present in forms that will make sense to logical machines – as any skilled barista will know intuitively. In order to understand the tasks and translate them successfully for robots, the team developed “a novel approach…based on the idea that many household objects share similarly-operated object parts.”
PR2, or “Robobarista” had to acquire information in a general way and then apply the knowledge acquired to identify similar items and use them correctly. In this way, the Cornell team claims, “Robobarista” was able to utilise information that it already had about handles and other machine parts. Using this knowledge, plus an English-language instruction manual, the robot identified the relevant part to press on a coffee maker in order to produce its first cup of coffee. The process is described as “a deep learning neural network” by graduate student Jaeyong Sung, who is working with Ashutosh Saxena, assistant professor of computer science, on the project.
Can “Robobarista” produce any of the special coffees loved by coffee drinkers? So far it has got to grips with the concept of pouring hot coffee over ice cream for an affogatto coffee as well as producing a latte, identified as a more complex process due to the use of steamed milk.
The little robot is already capturing the imagination of the world. The good news for coffee bars and restaurants is that it seems that “Robobarista” has the potential to work using existing equipment and professional espresso machines, since the whole idea of the project is to teach “Robobarista” and his fellow bots how to do what humans do all day and every day with their labour-saving devices. It’s the fact that the robot can do this apparently autonomously that makes it special.
So – is this the future? Will a little army of “Robobaristas” elbow their human counterparts out of the way in their bid to provide the best service? Perhaps what’s more likely is that humans and robots will learn to work together as a team to bring the customer the most delicious smelling, delicious tasting coffee. And as yet, of course, robots can’t swap information about scent or taste with the customer – that’s definitely still in the future.