Embedded Artificial Intelligence Technology

An artificial intelligence (AI) test environment can help researchers evaluate promising embedded AI technologies and their impact on systems developed for Department of Defense missions.
The image shows three use cases for embedded AI: physiological monitors, drones, and radios.
Integrating intelligent capabilities into embedded platforms such as, L to R, physiological monitors, autonomous aircraft, and tactical-edge radios, will enable future “smart” devices for DoD applications.*

Researchers in the Embedded and Open Systems Group are developing an artificial intelligence (AI) research environment that can be used to develop and evaluate promising embedded AI technologies and their impact on critical Department of Defense (DoD) missions. Recently, industry and academia have made substantial progress and investment in AI technology in many areas, including natural language processing, computer vision, and medical diagnosis. Because military applications have vastly different requirements and operating environments, replicating the success or evaluating these technologies in the context of DoD-relevant missions is challenging. To address this shortcoming, we are developing multi-layer, comprehensive test suites that take into account specific application context, such as mission requirements, datasets, and deployment platforms. Leveraging this test environment allows researchers to quickly evaluate the impact of promising AI technologies on real DoD missions and develop innovative DoD-centric solutions and technology improvements. The applications of interest include advanced multimodal navigation, synthetic aperture radar object detection, and other AI-enabled military capabilities.


*Image credits: Heart rate monitor, MIT Lincoln Laboratory; autonomous aircraft, USAF Master Sergeant Scott Reed; tactical radio, USMC Corporal Shane S. Keller.