My team from Johns Hopkins University in Baltimore had the honor of receiving KUKA's annual Innovation Award for our project CoSTAR, the Collaborative System for Task Automation and Recognition. Our system allows end users to teach robots complex and reusable skills.
Every year, manufacturing companies from all over the globe converge in Hannover, Germany to join the biggest industrial automation trade show in the world. The event – Hannover Messe – is a preview of the changing face of manufacturing, so it is no surprise that one of the biggest exhibitors is Augsburg-based KUKA Robotics.
For the last three years, KUKA has hosted an annual Innovation Award, and every year the response has grown. This year, 25 teams from all over the world proposed novel ways of using KUKA's LBR iiwa robot in flexible manufacturing. From these applicants, KUKA chose the six finalists they thought were the most promising.
Our goal was to make robots into flexible, adaptive assistants that small manufacturers can use to solve problems. Our system can recognize certain objects and simple events, as well as understand concepts like "left of" and "right of." We implemented a user interface that allows us to teach the robot and create graphical programs.
At the show, we impressed the judges by quickly creating a variety of different tasks. This is best exemplified by using an anecdote about the show's first day. During an award ceremony, a robot was needed to present the prize. In a hall filled with industrial robots, our system was the only one that could be reprogrammed quickly enough.
The Johns Hopkins University team was composed of myself, Andrew Hundt, Matt Sheckells, Chi Li, Felix Jonathan, and Kelleher Guerin. We were advised by Professors Marin Kobilarov and Gregory D. Hager at Johns Hopkins University. We all plan to continue our work and make robots into more intelligent, adaptive, and perceptive tools for ordinary people.