ANDRO awarded $1.1M Navy contract

Posted

ANDRO Computational Solutions has been awarded a $1.1 million research contract by the Navy to develop a new unmanned aerial system technology that can automatically adapt to dynamic mission objectives and rapidly changing environments.

Dr. Jithin Jagannath, director of ANDRO’s Marconi-Rosenblatt AI/ML Innovation Lab envisions an end-to-end, human-on-the-loop framework called D-MARVEL, which leverages deep reinforcement learning to provide autonomy, adaptability, and human-machine teaming for drone applications. He is assisted by Nicholas Polosky, who leads the D-MARVEL algorithm research and software development and Tyler Gwin, who supports the associated flight-test activities.

“D-MARVEL leverages human visual cues with artificial intelligence and data science to provide for next-generation machine learning-based drones that can autonomously perform a variety of tasks and conduct missions in diverse environments with little to no human intervention,” said ANDRO President Dr. Andrew Drozd.

Drozd envisions significant growth arising from this work in cooperation with ANDRO’s business partners to incorporate the solution into advanced drone platforms during the next phase.

“D-MARVEL is the next step in a strategic plan to expand ANDRO’s research portfolio and footprint in 2021, including the research activities of the Marconi-Rosenblatt AI/ML Innovation Lab,” Drozd added.

ANDRO, headquartered at One Beeches Place, 7980 Turin Road, provides research, engineering, and technical services to defense and commercial industries. The independently owned company, established in 1994, is dedicated to research, development, and the application of advanced computer software for electromagnetic environment impact studies, spectrum exploitation, secure wireless communications for cognitive radios, multi-sensor and multitarget tracking, advanced radar data fusion, and sensor resource management.

For more information, visit www.androcs.com.

Comments

No comments on this story | Please log in to comment by clicking here