Tic-tac-toe is often the first game children, and sometimes robots, learn to play.
“With robots, they have to know not only how the game is played but how to pick up the X’s and O’s. What comes easily to us as humans is actually a very complicated maneuver,” said Thomas Herring, a junior at Rice University with a double major in computer science (CS) and electrical and computer engineering.
Already a research associate in the Kavraki Robotics Lab, Herring is no robotics novice. In his native Toronto, Herring spent four years as a member of his high-school robotics team. In 2013, he and his teammates, along with more than 10,000 other students from 37 countries, competed in the FIRST (For Inspiration and Recognition of Science and Technology) World Championship in St. Louis.
“Thomas is a driven student with quite a bit of robotics experience already. He started in high school, built his own robotic arm in the Oshman Engineering Design Kitchen (OEDK) and eventually found us. He’s been working with several students and a postdoc in our lab to help create a software infrastructure to increase the autonomy of robots,” said Herring’s adviser, Mark Moll, a Rice senior research scientist in CS.
Herring came to Rice with the aid of a four-year, $25,000-per-year Distinguished Trustee Scholarship. Since his freshman year, he has worked as a lab assistant in the OEDK. In 2017, he won the Freshman Design Excellence Award at the OEDK Showcase. His project, Lightweight Robotic Scanning and Control Platform, is an educational platform for teaching students the operating system most often used with robots.
The tic-tac-toe project is focused on motion planning. He researches constrained motion-path generation methods. In others, how do you make a robot move exactly how you want it to move? In Moll’s words:
“The immediate goal is to create an abstract interface that allows users to specify simple high-level pick-and-place commands for any robot arm. A command could be to move object A to location X. His code would automatically determine possible grasp locations and compute a feasible path for the robot to move the object from its current location to location X.”
Herring’s research would abstract away the specifics of grasping a particular object and how a particular robot moves, and permit the user to think about tasks at a higher level. Tic-tac-toe itself is essentially simple but it combines significant aspects of robotics: perception, human-robot interaction, grasping and motion planning. If successful, the abstractions Herring and his colleagues devise should make it a fairly straightforward task to move from tic-tac-toe to other games and, ultimately, more practically useful tasks.
“This all started with childhood wonder from movies and videos. I was always interested in robots, so there’s a thrill to sitting down and writing an algorithm and having it work,” said Herring, who over the summer spends on average six hours a day working in the Kavraki Robotics Lab in Duncan Hall.
Herring expects to work in industry after earning his B.S. in 2020. He wants to get to work.
“I would be surprised if he didn’t continue to pursue robotics after graduation. With his experience in both mechanical engineering and computer science he will be well-positioned to pursue a robotics career,” Moll said.