One of Oregon State University professor Julie Adams’ projects involves building the infrastructure that will allow swarms of about 250 autonomous vehicles, both on the ground and in the air, to assist military troops in the field.
Another involves researching bees to understand swarm behaviors and apply those lessons to robots. Yet another involves collecting data in ambulances through cameras and wearable devices that would automatically recognize what procedures emergency medical technicians use on patients and generate a report detailing that for doctors receiving the patient.
Since joining the faculty at OSU in January, Adams has been part of research projects funded to the tune of about $12.2 million, although the money is shared with partner organizations, such as Raytheon BBN, Smart Information Flow Technologies, and Adams’ former employer, Vanderbilt University.
Adams is working on six different projects with seven graduate students, both at OSU and Vanderbilt, and she plans to add a postdoctoral researcher and five or so undergraduate assistants to her team.
Using large groups of robots in coordinated ways is a focus in a couple of those projects, including the $7.1 million Defense Advanced Research Projects Agency undertaking aimed at assisting troops with hundreds of robots capable of moving collectively and following voice or gesture-based commands from soldiers in the field.
Adams’ team will contribute to testing the user interface for the system in simulations and researching how the ways the soldiers interact with the drones to give them direction should change in different circumstances: For example, some situations might have too much noise for soldiers to give voice commands.
Adams said the project will start by having field trials with swarms of about 50 robots, including quadcopters, fixed wing unmanned aerial vehicles and six-wheeled ground vehicles about the size of radio-controlled cars. Over its three-year run, Adams said, the collaborators plan to work their way up to 250 such vehicles moving in coordination.
Adams said using groups of simple robots to act collectively can have advantages: For example, if one robot is disabled, the others still can function.
She compared this to schools of fish, which split into two when predators come or form a ball of moving fish to confuse the predator.
“You have simple individuals, not very intelligent, but when they act collectively they look pretty intelligent.”
The swarms would be similar, in that each member would be less sophisticated than say, a self-driving car. Collectively, though, the robots could act together to have greater capabilities and more flexibility.
Some of the inspiration for the tactics the systems will use are drawn from nature, and Adams has been working with the OSU Honey Bee Lab on a separate project to examine bee behaviors to see if similar behaviors can be applied to robot swarms. One area of focus is on how sick or contaminated bees will isolate themselves from the group: Adams said the hope is similar methods can be develop so that a drone in a swarm infected with a virus would shut down and not spread it to other robots in the group.
Adams said she was drawn to OSU because it was preparing to launch its Collaborative Robotics and Intelligent Systems (CoRIS) Institute and because of the graduate robotics program the school has. At Vanderbilt, she was the only one of her colleagues working with ground and aerial robots, but coming to OSU allowed her to be part of a team of people doing the same.
She’s also now working with OSU oceanographic researchers on projects involving underwater robots. And she’s hoping to encourage more faculty members to get involved with using autonomous vehicles in research as the new principal investigator for OSU’s participation in the Federal Aviation Administration’s Center of Excellence for Unmanned Aerial System Research.
Adams said moving to Corvallis from Nashville, Tennessee, has had its challenges, but she is adjusting to small-town life.