Semester exams are looming, with an extended holiday break on their heels. But before Dr. Andrea Thomaz closes the Socially Intelligent Machines Lab at the Georgia Institute of Technology for the season, the lab hosted a few more visitors last week for the final experiment of the semester.
The lab welcomes guests to interact with Simon, a humanoid robot developed with seed funds from the Office of Naval Research (ONR). These interactions allow student-researchers to adjust software models for Simon’s learning and behavior generation. And it all starts once Thomaz and her team wake the resting robot.
“Simon, can you hear me?” Thomaz asks.
“Yes, can you hear me?” Simon responds in kind.
Standing nearly 5 feet tall, the robot is surrounded by an arc of desks where students design, refine and stabilize his software to demonstrate his range of skills. He’s alert and ready for interaction.
The partially mobile ’bot is backed by funding from several ONR programs—including cognitive science, machine learning and autonomy—and he’s teaching the Navy about human-to-robot dynamics.
This robotic effort could, ironically, shed light on human activities and help the Navy on its track to build a hybrid force.
“We’re looking at people, how the social dynamics between two people work and how we can embody some of that in robots,” Thomaz says. “In the future, when you have teams of people interacting with robots in stressful environments potentially, you want to make sure that interaction is as easy as possible.”
Thomaz, a recipient of ONR’s prestigious Young Investigator Program (YIP) award, was highlighted this fall in Popular Science magazine’s annual “Brilliant 10” list of researchers. She and four other aspiring researchers funded by ONR made the cut for their innovative projects, accomplishments and career potential. Those include:
With an ONR Multidisciplinary University Research Initiative grant, Dr. Deva Ramanan, of the University of California (U.C.) Irvine, is studying how to endow computers with automated visual intelligence across a variety of platforms.
Before graduating and moving to U.C. San Francisco, Shawn Douglas contributed to a published study that led to ONR’s funding of a DNA project at the Dana-Farber Cancer Institute.
Harvard’s Adam Cohen is studying how electrical signals work in the brain with YIP support from ONR’s nano-electronics program.
Dr. Aydogan Ozcan, of U.C. Los Angeles, is using computational algorithms to transform smart phones and laptops into affordable, lightweight diagnostic tools to strengthen medical capabilities for use in remote areas.
Like Thomaz, Cohen and Ozcan are former YIP recipients, too.
Ozcan also received a Presidential Early Career Science & Engineering Award this summer, which will provide funding through 2016 to continue his mobile technology research under ONR’s Warfighter Performance Department.
“Today’s cell phones are more powerful than the supercomputers of the 1980s and even early 1990s in the sense that they’re getting extremely powerful in terms of what we can do,” he says.
Picture a small disposable tray holding a sample of bodily fluid. It attaches to a smart phone which, using optical illumination and computational reconstruction and image analysis techniques, detects trace amounts of a harmful bacteria. Early intervention and treatment can now begin.
Ozcan says it’s possible.
“This is actually a platform which could have various applications,” he adds. “Of course, the Navy is very much interested in it and they’ve been funding it from very early on in my career. And this funding has been extremely important in my group. It’s allowed us to mature ideas into real implementations.”
Ramanan also sees ONR’s role in evolving ideas. His team could develop a deployable visual intelligence (VI) system in as little as two to three years, making it possible to count individuals in a space, identify specific individuals by movement and/or posture, and offer some indication of the events taking place (e.g., a meeting, a marketplace).
Initial goals might center on surveillance or flagging suspicious behaviors, but VI has a broad range of applications including robot perception, human health care and many others, making it doubly worth the Navy’s investment.
“If someone has suffered a debilitative injury to their motor system, such intelligent systems could perform long-term monitoring of progress and possibly suggest rehabilitative exercises,” he suggests of VI’s potential to enhance rehabilitation programs for injured warfighters.
But that’s still a ways off.
For now, Thomaz and her students may deliver the first measurable results of the new year as the lab welcomes additional equipment funded by ONR in January.
Yes, Simon is getting a cousin, an upgraded model with a greater range of motion.
“The main difference will be in its mobile base,” she says. “Right now, Simon is upper body and in the lab only. Now, he’s getting his legs.”