We all know the struggle: You're vegging out on the couch, eyes on your favorite Netflix show, hand tight around a half-empty craft brew. The last thing you want to do is get up and get yourself another beer.
Now imagine you can get a robot to fetch you a beer while you stay comfortably seated on the couch. That's exactly the idea behind BeerBots, a team of robots developed by researchers at the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology.
Here's the big bartender bot, which stays behind the bar and waits for the little waiter robots to deliver people's drink orders.
For their experiments, the MIT researchers turned their laboratory into a testing ground where the two small robo-waiters, called "turtlebots," could roam freely. The turtlebots, which resembled coolers on wheels, traveled back and forth from the bar to rooms where people were sitting, and they'd take drink orders from the human participants. When someone requested a beer, a turtlebot traveled to a third machine, the large robot "bartender," which supplied the beverage. Then the turtlebot delivered the beer to the person.
The BeerBots used algorithms to plan, reason and fill the orders. In addition, each turtlebot had to determine which orders were requested in which room and when the other turtlebot had delivered drinks, said MIT communications coordinator Adam Conner-Simons in a statement about the project.
Here's the big robo-bartender placing a drink into the cooler-on-wheels turtlebot, which then delivers it to the person who ordered it.
The project was the brainchild of Chris Amato, a former CSAIL postdoctoral researcher who is now a professor at the University of New Hampshire. Amato worked with a team of MIT professors, postdocs and graduate students to develop the BeerBots, which were presented in July at the annual Robotics Science and Systems conference.
As you might guess, coordinating a team of robots to deliver beer is far from easy. In a recent paper, the MIT researchers detailed some of the issues they faced in the experiment, including navigation problems, unpredictable outcomes and communication between robots and humans.
“Each robot’s sensors get less-than-perfect information about the location and status of both themselves and the things around them,” Amato said, per MIT News. “As for outcomes, a robot may drop items when trying to pick them up or take longer than expected to navigate. And, on top of that, robots often are not able to communicate with one another, either because of communication noise or because they are out of range.”
Anticipating these difficulties, the MIT researchers programmed the robots with complex planning algorithms, which allowed the machines to engage in higher-level reasoning about their location, status and behavior -- similarly to they way humans perform tasks. In other words, the BeerBots were aware of their neighbors and could adjust to several possible situations.
“Almost all real-world problems have some form of uncertainty baked into them,” said Amato, according to MIT News. “As a result, there is a huge range of areas where these planning approaches could be of help."
Sure enough, these planning algorithms have applications beyond bartending. For instance, Amato and his collaborators are currently testing the algorithms with robots employed in simulated search-and-rescue scenarios.
H/T MIT News