Johns Hopkins team's task robot takes top prize at industrial technology trade fair in Germany

A student team from Johns Hopkins University took top honors last month at the world's leading trade fair for industrial technology, the Hannover Messe.

Team members Andrew Hundt, Felix Jonathan, Chi Li, Christopher Paxton, Matthew Sheckells, and Kelleher Guerin came home from with the KUKA Innovation Award, which honors outstanding innovation in robotics.

This year, the expo challenged participants to invent robotic applications for the realm of flexible manufacturing. The JHU team's device is the Collaborative System for Task Automation and Recognition, or CoSTAR, which is made up of an "intelligent industrial work assist," a personal computer with a touchscreen, and sensors. It is designed to be used by companies to customize—and adapt—programs for use in manufacturing.

"CoSTAR makes it easy for manufacturers to retask a robot in a matter of minutes. As a result, users with no programming background can quickly and easily create complex task plans that readily adapt to new situations," said Paxton, the team leader. "The goal is to let ordinary people in a factory use capabilities offered by cutting-edge research to solve their own problems."

Team CoSTAR was advised by Gregory D. Hager from JHU's Department of Computer Science and Marin Kobilarov from the Department of Mechanical Engineering. Team members say that they also benefitted from the resources and talent available through the Whiting School of Engineering's Laboratory for Computational Sensing and Robotics.

Team CoSTAR presented their design to a jury who judged their submission, along with those of six other finalists, over the course of four days.

"The CoSTAR team had some early setbacks due to equipment failures and missing components," Hager said. "However, after some early glitches, the system performed beautifully and the demos and presentation just kept getting stronger. The ultimate 'secret weapon' was the interactive programming environment and general perception capabilities that allowed them to continually add and improve demos, sometimes at the suggestion of onlookers."