URSI: Meet “Landro” and “Tadro,” Evolving Robots
“Landro” was built out of the shell of a Roomba vacuum cleaner. “Tadro’s” chassis and tail were crafted in a 3D printer. Together, they are the subjects of some intense research this summer in Vassar’s Interdisciplinary Robotics Research Laboratory in the basement of the Bridge for Laboratory Sciences.
Six students and four faculty members are working with Landro and Tadro under the auspices of Vassar’s Undergraduate Research Summer Institute. Their work is part of an ongoing robotics project funded through a $1 million grant from the National Science Foundation. Most of the students will continue their research during the upcoming academic year.
The robots are equipped with electrical “nervous systems,” tiny computers that enable them to make their own decisions as they use specially built sensors to seek out light or avoid other objects. After the robots achieve their tasks, their genetic coding is modified.
“The robots have optical sensors that detect light and proximity sensors that detect distance from an object,” biology prof. John Long explains. “The genetic codes determine the pattern of connections between those sensors and the motors that move the robot. The genetic codes of the next generation of robots are both similar to that of their parents and different because of selection and random mutation.”
The six URSI students say they’re enjoying the collaborative nature of the project. “We’re all working as a team,” says Jonathan Snyder ’18, a physics major from Ortung, WA. “It’s not just professors telling us what to do; we have the freedom to make out own decisions on how we do things, and we come from many different disciplines.”
Snyder came to the team with a background in electronics gained during an eight-year stint in the U.S. Navy. “I’m the hardware guy on this project this summer, and it’s really been rewarding,” he says. “I plan on going to graduate school in electrical engineering, and this experience is really valuable to me.”
Meghan Willcoxon ’18, who heads up the “Landro” team, says the group has encountered some challenges along the way. “We’re trying to teach it to seek light and avoid obstacles, but as the brain was learning these tasks, the software wasn’t communicating with the hardware. We had to assess the problem and figure it out,” says Willcoxon, a cognitive science major from Doylestown, PA.
“What’s different about URSI is you are spending 40 hours a week designing, programming and tasking a robot, and you’re collaborating with your professors instead of being taught by them,” she continues. “This enables us to come up with solutions to problems as a group instead of simply taking direction from a professor.”
There have even been times when the students have disagreed with their professors and found their own way to solve a problem. “We have a level of control in our own design,” says David Wallach ’17 who heads the “Tadro” team. Wallach, a computer science major from Poughkeepsie, NY, says the students determined their robot would be able to perform certain tasks more efficiently if it weren’t equipped with its memory component. “We disagreed with our professors about the need for memory,” he says. “We realized it wasn’t necessary for the tasks we wanted the robot to do, so we eliminated it.”
Other members of the URSI team are Xiaoqing Xu ’18, a cognitive science major from Guangzhou, China; Mackenzie Little ‘17, a cognitive science major from Mansfield Center, CT., and Ben Tidswell ‘18, a cognitive science major from Northampton, MA.
Other faculty overseeing the URSI project are prof. of cognitive science Ken Livingston, associate prof. of computer science Marc Smith, and associate biology prof. Jodi Schwarz. The other member of the team is Nick Livingston, assistant director of the robotics laboratory.
Long says he is impressed with the work all six students are doing. “As the summer goes on, they’re less like students and more like colleagues,” he says. “This is the kind of work that is being done in graduate schools at major universities, and we know we’re competing with them successfully because we’re winning these major government grants.”
Long says the reason for the success of Vassar’s robotics project is simple: It’s the collaborative nature of the research. “This is true ‘liberal arts’ science, being done in this lab” he says. “We are blurring the boundaries between disciplines.”
Photos: Buck Lewis
Posted Monday, August 8, 2016