Decentralized Computer Vision for Counter-UAS (C-UAS)

Unmanned Aerial Systems (UAS) are an ever present and increasing threat to the Joint Force. Prescient Edge proposes the development of a federated learning computer vision UAS detection, identification, and tracking AI that can be deployable into existing visual sensors and integrate into JADC2 systems.  Relative to traditional computer vision, Prescient Edge’s unique federated learning approach creates an AI that is (1) more secure, (2) requires less computational power (3) requires less connectivity, and (4) provides more tailored solutions for localized operational conditions.  Prescient Edge will baseline development from their commercially available AI platform and develop a federated computer vision software that integrates into current sensors, “train” the AI to meet the accuracy, sensitivity, and responsiveness requirements, and develop a “tipping and cueing” function that alerts and vectors a response.