AI Agreement Will Spur Advances in Robotics, Sensing, and Beyond

A five-year agreement between the University of Maryland, College Park (UMD); the University of Maryland, Baltimore County (UMBC, as a partner); and the U.S. Army Research Lab (ARL) will help scientists and engineers from multiple disciplines develop technologies to reduce human workload and risk in complex environments, such as the battlefield and search-and-rescue operations. The agreement, worth up to $68 million, supports research pursuits in the areas of AI, autonomy, and modeling and simulation. It brings together University System of Maryland researchers in disciplines including engineering, robotics, computer science, operations research, and cybersecurity.

“This is a big partnership with an ambitious vision. We want to change the world by quickly getting lifesaving autonomy into the hands of the people who need it,” said Derek Paley, lead researcher, UMD professor, and director of the Maryland Robotics Center. “No matter how autonomous we think a system is, a human operator will interface with it at some level. The goal is to migrate the dangerous, dirty, and dull work to the autonomous platform.”

Technologies identified in a shared (UMD and ARL) grant announcement that the agreement is poised to support range from wearable devices to unmanned aircraft — technologies that work intelligently and in cooperation and with human actors across environments. Outcomes identified for the collaboration include a shared, standard infrastructure for developing and testing collaborative autonomy — including a “virtual proving ground” that models and simulates existing facilities, to enable different interactions between virtual and real-world actors; and shared access to ARL’s Robotics Research Collaborative Campus (R2C2). The campus features a reconfigurable, multiple-terrain outdoor testing laboratory for scalable AI, autonomy, and robotics research.

Jeffrey Hermann, co-lead researcher and a UMD professor, said the agreement will additionally develop a consistent modeling and simulation infrastructure. That infrastructure, he said, will include libraries of simulation models and the tools for running them.

Currently, Hermann said, teams build their own simulation testbeds, considering their own algorithms and systems. The process increases cost and delays further innovations. Technologies such as self-driving vehicles or mobile robots that combine computer vision, remote sensing, robotics planning, and control, as well as other advanced specialties, would be supported by such an infrastructure.

AI-based networking, sensing, and edge computing will be the focus areas of other solutions, the collaborators said. In particular, the research team will work to strengthen the decision-making abilities of Army AI technology to meet the demands of national defense.

The cooperative agreement includes 27 investigators from UMD and UMBC.