Automatic Target Recognition of Personnel and Vehicles from an Unmanned Aerial System Using Learning Algorithms

Black River Systems, teamed with SRC Inc., proposes the “REcognition with Aerial Learning” (REAL) approach which will address the challenge of targeting through novel integration of adaptable learning approaches on UAS. This approach will support a wide problem (i.e. mission, environment, application) and solution (i.e. sensing modalities, learning/training, taxonomy representation, autonomy) space with a clear transition path. We will implement learning algorithms that provide operational flexibility by allowing the target set and DRCI taxonomy to be quickly adjusted to operate in different environments, all while exploring and ultimately integrating our advanced SWAP capabilities, multi-modal sensing, machine learning (ML) applications, and autonomous control on a class 1 and/or class 2 UAS. The team is strongly founded in learning approaches and applications, SWAP constrained system design and deployment, HMI development, and autonomous distributed control. Finally, we will perform Phase I demonstrations of adaptable learning applied to DRCI with SWAP constraints in the lab and field to provide a clear story for transition to Phase II and Phase III. We will leverage ongoing efforts with learning, autonomous control, low SWAP compute platforms, and sensor development to assist with transition and mission relevancy.