Autonomous Optical Sensors

IERUS and Kitware propose to design a modern autonomous optical sensor (MAOS ¢ pronounced ōmouseö) to implement event-based sensing and machine learning into the test range portfolio. ĀEvent-based sensing is a modern sensing architecture with recognized value in the defense arena. ĀFor example, the DARPA Fence program selected research teams at BAE, Raytheon, and Northrup Grumman to develop infrared event-based imager arrays. ĀHowever, mature visible band event-based cameras (EBCs) exist and are well suited to tracking various aerial threats. ĀA modern system is a learning system. ĀData processing with EBCs is a rapidly developing field and is well suited to AI/ML algorithms. ĀThe proposed design will learn over time and over acquired test data to optimize configuration, tracking, and data extraction for a diverse array of arial weaponry including various missiles, drones, and bombs to provide the most accurate, valuable data in response to customer needs. ĀIn Phase 1 we will provide analyses and a design based to the extent possible on commercial off the shelf (COTS) equipment with a strong track record of reliability and performance. ĀIERUS will serve as the primary hardware designer and integrator, while Kitware will lead the AI/ML architecture design and development. ĀAlso, as part of Phase 1, IERUS and Kitware will work with the government customer to identify the most valuable ancillary AI/ML functions (organic computing, edge processing) or hardware upgrades (e.g., data transmission) for implementation during Phase 2 as schedule and budget allow.