Robots Get Real
The social robots created by Carnegie Mellon University researchers can gossip, get cranky and even talk back to students.
In science fiction, robots are sometimes indistinguishable from people. However, until recently, the robots used by industry, the military and law enforcement looked like machines and were designed to complete highly specific, mechanical tasks.
Enter researchers of Carnegie Mellon University in Pittsburgh, who are creating robots with personalities.
Valerie, Tank, Athina and Victor are among the world’s first storytelling robots. They are the result of an interdisciplinary project between the university’s School of Drama and its Robotics Institute in the School of Computer Science. Carnegie Mellon’s robotics program is considered among the best in the United States.
The project grew out of conversations about how technology affects people’s day-to-day lives. Intrigued by the possibilities of marrying art to technology, professors Anne Mundell and Reid Simmons came up with the idea of “social robots,” which would converse while performing tasks. Mundell works with drama students to create each robot’s persona and Simmons helps robotics students develop the software to make the robots interact with humans.
The robots provide information, like directions and weather reports, but their quirky personalities encourage people to spend longer periods of time interacting with them. That is a breakthrough. As robots assume a greater role in healthcare, assisting the elderly with their medications, for example, it will be important for them to be approachable.
The students’ first collaboration produced Valerie in 2004, a “roboceptionist,” who is now used for classroom demos. Initially installed in the computer science building, Valerie answered visitors’ and callers’ questions while gossiping and, at times, complaining about her mother’s attempts to run her love life.
Valerie’s successor, Tank, took over as roboceptionist in 2005, “freeing Valerie to pursue her singing vocation,” says Mundell.
Tank, conceived as a rugged military veteran, has enjoyed a longer tenure as department roboceptionist. He has his own unique personality. For instance, he gets cranky if visitors are rude to him. Tank will eventually be replaced by Miranda, who is in the works.
Victor has the persona of an adolescent prodigy, attending the university on a Scrabble scholarship. He has been designed to play against human opponents and talks like a moody teenager. Installed in a student lounge, he taunts students by citing things that robots can do better than humans.
Athina is a talkative and lively chatbot, part of a permanent collection called “Roboworld” at the Carnegie Science Center.
The students developed the serialized stories told by these robots. Visitors type questions using a keyboard and the robots respond with computer-generated speech. “As you interact with the robots, their stories evolve. It’s like a running soap opera,” says Mundell.
The robots have facial expressions, courtesy the software devised by Simmons, and they tilt their heads and move their eyes. They are equipped with motion sensors that detect where people are. The collaborators have learned from successive creations; newer robots interact more naturally. Victor has “an emotional response spectrum, so he can respond appropriately when things happen during the course of a Scrabble game,” says Mundell.
For her, social robots, despite their real-world applications, are about “storytelling in a new format.”
Who’s Afraid of Robots?
An interview with Matthew T. Mason, professor at the Robotics Institute, Carnegie Mellon University.
Science fiction makes us believe that once robots reach comparable intelligence to humans, they could rise up. Are they coming to get us?
Not in the near future; maybe 500 years from now. I wish we’d been so successful in researching artificial intelligence and developing robots that it was a realistic fear.
What can today’s robots do or not do?
All things that are easy for us are hard for robots and all things that are hard for us are easy for robots. The intellectual challenges related to the chess game were addressed quite successfully by artificial intelligence—computers won against human competitors. Yet, if you want a robot to move chess pieces on the board, it’s tough. Also, look at how humans and robots deal with uncertainty. Humans use their senses and also common-sense physics. When they encounter obstacles, they can do a quick analysis to solve the problem. Robots have a hard time dealing with such obstacles.
What are some of the challenges in making robots more versatile?
Generally, researchers are looking at different mechanisms, structures and materials that can advance robotics. They work on controls, perception, intelligence and machine learning. I work on manipulation, which involves motion planning and control, including how robots use their “hands.” Today, we have robots with simple grippers that are more like a pair of tongs. Some people are designing better “hands,” similar to humans’. But, even if they succeed, I don’t think it will solve the problem. A human being with a pair of chopsticks will still be vastly more capable than a robot dealing with a novel situation, at least in the near future.
Mobile phones are changing our lives. Similarly, will robots affect us more radically?
Robots have been changing our lives, but not our minds. Often, we see an automated device as too simple to call it a robot. I’m excited about applications of robots in education.
At Carnegie Mellon University, we have an automated tutor, which listens to children read, corrects their mistakes, and guides and encourages them. Space exploration is changing. Our motion-planning software was used in the NASA rovers, which have explored Mars. The driverless car, medical robots and technologies that can watch and pay intelligent attention to all the space around us and inside our bodies will cause amazing changes in our lives.
Do robots themselves select inductees to the Robot Hall of Fame at Carnegie Mellon University?
That would take away all the fun from being a judge...and would scare me a bit!
Lauren Monsen is a staff writer with the U.S. Department of State’s Bureau of International Information Programs.