Belleau Labs
Back to contracts

STRATFI SBIR Sequential Phase II

AFWERX · AFWERX STRATFI · AFWERX

AI-Readiness Score
22/25
Pathway Speed
5/5
Timeline Realism
3/5
Problem Framing
4/5
AI / ML Fit
5/5
Award + Transition
5/5

Award

$12,434,000
Award ceiling
EPISYS SCIENCE INC
Awardee
Posted December 13, 2024

Description

EpiSci will develop, mature, integrate, and test a software autonomy suite to enhance the USAF strategic capability of Complexity, Unpredictability, & Mass. By allowing for large-scale mission execution with greatly reduced operator workload, Tactical AI integrates on diverse hardware platforms to enable widespread deployment of affordable & overwhelming mass. Directly supporting the operational imperatives of Tactical Air Dominance, Moving Target Engagement, and Global Strike, Tactical AI offers a common, reusable, and federated approach to deliver software-defined, tactically informed, operationally relevant, and hardware-enabled autonomy at scale. EpiSci will develop and demonstrate collaborative autonomy on operationally-relevant platforms ranging from small, expendable drones to Group 5 Autonomous Collaborative Platforms, directly meeting the emerging transition requirements for multiple programs, customers, and warfighters. EpiSci’s Tactical AI comprises both a developmental paradigm and a robust mission autonomy framework. The design methodology and philosophy have delivered unmatched operator alignment and acceptance in multiple competitive evaluations. The framework delivers a modular, adaptable system for building trustable, high-performance, and portable multiagent & swarm collaborative autonomy solutions. Tactical AI implements a common autonomy stack with reusable collaborative applications such as autonomous control of heterogenous UAV swarms, off-board tactical sensing, and manned-unmanned teaming (MUm-T) “loyal wingman” applications. Tactical AI was first deployed in support of DARPA’s Air Combat Evolution program, demonstrating advanced control while guiding simulated full-scale F 16 and L-39 aircraft in 2vX within-visual-range combat maneuvering. Utilizing a novel hybrid approach, Tactical AI systems combine the exceptional performance and task adaption of deep reinforcement learning with the assured, bounded behavior and rapid integration of traditional deterministic control systems and collaboration heuristics. EpiSci will adapt and deploy Tactical AI on systems ranging from single-use air and ground-launched Airborne Adaptive Enterprise platforms (such as the Altius 600 and Phoenix Ghost), group 2 and 3 UAV platforms such as the Firestorm Tempest 50, legacy and emerging NCA-capable weapons, and group 5 ACP such as the XQ-58, BQM-177, and VENOM F-16 aircraft. Additionally, in each case EpiSci will develop and demonstrate the required training and simulation systems to deliver enduring operational capabilities. At completion, the products of this effort will be ready for immediate inclusion in Programs of Record through a modular and portable software approach with demonstrated multivehicle adaptability. Operationally relevant demonstrations will inform both warfighter feedback and deployed CONOPS to minimize fielding risks.

Score Rationale

This is an AFWERX STRATFI award at $12.4M ceiling with explicit Program of Record transition language, earning top marks on pathway speed and award/transition — the gold standard for AI-native defense startups. The AI/ML fit is genuinely strong: deep reinforcement learning, multiagent swarm autonomy, and manned-unmanned teaming are core AI problems with real operational data (DARPA ACE heritage), not grafted-on buzzwords. Problem framing is well-bounded with named platforms, named programs, and identified warfighter end-users, though the scope breadth across Group 2 through Group 5 platforms and multiple mission types introduces some execution ambiguity; timeline realism scores mid-range because the response deadline is unknown and the multi-platform demonstration ambition is aggressive for even a well-resourced team.

Source

View original posting
Back to all contracts