Project Starling: Common Framework for Networked, Collaborative Autonomy with Tactical AI

Emerging threat environments and adversary capabilities, coupled with the extreme cost of advance weapon systems, present a challenge to continued USAF dominance in potential future conflicts. Advances in autonomy and Artificial Intelligence, allowing the use of large teams of lower-cost weapon assets while managing the required operator workload, offer a potential solution to manage capability at scale within budget, schedule, and manpower constraints. EpiSci’s Starling program will demonstrate the ability to perform demanding autonomous missions, integrate with operators as part of a manned-unmanned team, and demonstrate modular portability for a variety of platforms and missions. While previous autonomy efforts have used bespoke simulation environments, Starling will demonstrate open architecture integration with common and validated USAF modeling and simulation tools such as the Weapons Open System Architecture (WOSA), allowing for independent verification of performance and rapid adoption into diverse vehicles and missions. Built on a Digital Engineering framework with agile DevSecOps practices, Starling will delivery low-cost and low-risk capabilities. Upon successful completion, Starling algorithms will provide portable, trustable, and high-performance autonomous behaviors which are ready for integration into programs and platforms of record.