Picture an industrial robot. You might imagine a hulking, highly specialized piece of machinery, cordoned off by a cage from human workers on a factory floor. While that's one type of manufacturing robot, a new generation of robots designed to work side-by-side with humans is on the rise.
Some call them "cobots," and UW-Madison researchers are playing an important part in making human-robot collaboration more natural and efficient.
Bilge Mutlu, an assistant professor of computer sciences, will work with counterparts at the Massachusetts Institute of Technology (MIT) on determining best practices for effectively integrating human-robot teams within manufacturing environments.
Mutlu
Their research is funded by a three-year grant, announced last year, from the National Science Foundation (NSF) as part of its National Robotics Initiative program.
Mutlu directs UW-Madison's Human-Computer Interaction Laboratory and will serve as principal investigator on the UW-Madison side of the collaborative project. At MIT, Julie A. Shah, an assistant professor of aeronautics and astronautics, will fill that role on her campus.
The new crop of manufacturing robots includes platforms such as Baxter, introduced in 2012 by the company Rethink Robotics. Approximately the height of a tall person, Baxter is a two-armed robot with a tablet-like panel for its "eyes." Baxter's eyes provide useful cues to human workers, helping them anticipate what he will do next.
Referring to Baxter and other new platforms like it, Mutlu says, "This new family of robotic technology will change how manufacturing is done. New research can ease the transition of these robots into manufacturing by making human-robot collaboration better and more natural as they work together."
The new robots are less expensive and easier to integrate into the manufacturing process than older types. They are also capable of being reprogrammed for new tasks, rather than being meant for a single, highly defined task.
New research done by Mutlu and students in his laboratory will build upon earlier work they have done on topics such as gaze aversion in humanoid robots, robot gestures, and what is known as "speech and repair" (if you misunderstand a robot's instructions, or carry them out incorrectly, how does the robot correct you, the human worker?).
At MIT, Shah places her emphasis on breaking down the components of human-robot teamwork and deciding which tasks are best done by each.
"People can sometimes have difficulty figuring out how best to work with or use a robot, especially if its capabilities are very different from people's," says Shah. "Automated planning techniques can help bridge the gap in our capabilities and allow us to work more effectively as a team."
Mutlu's approach complements Shah's by focusing on how the human and robot actually interact. "We design the necessary models and algorithms so that robots can collaborate with people in a natural way," he says.
Furniture maker Steelcase, a global company headquartered in Grand Rapids, Mich., is a partner in the project. The company owns four Baxter robots, each located at a different facility in the United States.
"Working with world-class research universities like UW is critical to our strategy to evolve our industrial systems and develop industry-leading capabilities," says Edward Vander Bilt of Steelcase's Innovation Management Office. "We are very excited to partner with UW. ... Our hope with this research is that we will learn how to extend human-robot collaboration more broadly across our operations."
Over the summer, computer sciences graduate student Allison Sauppé traveled to Steelcase headquarters to learn more about its efforts to incorporate Baxter into the production line. She found that perceptions of Baxter varied according to employees' roles.
While managers tended to see Baxter as part of the overall system of automation, front-line workers had more complex feelings. "Some workers saw Baxter as a social being or almost a co-worker, and they talked about Baxter as if it were another person. … They unconsciously attributed human-like characteristics," she says.
Sauppé, who has observed the Baxter platform in companies other than Steelcase, comments that it is currently best suited to simple "pick and place" tasks — repetitive motions like bagging plastic cups or picking up tubes that will be filled with some material.
In the workplaces of the future, it's not a question of robots replacing humans, but rather humans becoming managers of robots, she says. "We want robots to take over monotonous tasks in our lives, things that don't really engage our cognitive function," says Sauppé. "And there are all these things that humans can do really well that robots struggle to do. … How can we use humans to their full potential?"
The United States is still in the early days of new-school "cobots" in manufacturing environments. But according to Mutlu, increased human-robot cooperation in industrial settings isn't a question of "if" or "when." The transition in manufacturing has already begun. Experts in human-computer interaction can make it more successful for everyone involved, from assembly-line workers to CEOs focused on the bottom line.