Belleau Labs
Back to contracts

Sentinel Swarms: Drone Defense for Austere Airfields

AFWERX · AFWERX STRATFI · AFWERX

AI-Readiness Score
18/25
Pathway Speed
5/5
Timeline Realism
2/5
Problem Framing
3/5
AI / ML Fit
5/5
Award + Transition
3/5

Award

$1,247,078
Award ceiling
Mazer Intel
Awardee
Posted May 19, 2025

Description

Mazer Intel proposes the development of Sentinel Swarms, an AI-driven, multi-agent small Unmanned Aircraft System (sUAS) swarm designed to provide autonomous perimeter defense and persistent Intelligence, Surveillance, and Reconnaissance (ISR) for forward-deployed airfields. This system directly addresses the Air Force’s Resilient Basing initiative by enhancing airfield security in contested, GPS- and comms-denied environments, where conventional ISR and security assets are vulnerable. The Sentinel Swarm leverages Graph-Based Multi-Agent Reinforcement Learning (MARL) and Graph Control Barrier Functions (GCBF+) to enable autonomous threat detection, optimized task allocation, and collision-free drone coordination. By integrating Airborne Optical Sectioning (AOS), the system achieves enhanced target tracking and occlusion penetration, providing continuous situational awareness with minimal operator input. This decentralized AI-driven swarm autonomously adjusts coverage, detects and tracks unauthorized incursions, and optimizes ISR capabilities in dynamic battle environments. Phase II will advance Sentinel Swarms from simulation to prototype field testing with the Air Mobility Command (AMC) and 621 Contingency Response Wing (CRW). The solution aligns with DoD’s Replicator Initiative and has clear transition pathways through TACFI and STRATFI funding. Beyond military applications, Sentinel Swarms have dual-use potential in critical infrastructure protection, disaster response, and industrial security, positioning Mazer Intel as a leader in autonomous perimeter defense solutions.

Score Rationale

AFWERX STRATFI is one of the fastest modern acquisition pathways available, earning the top score, and the core problem — multi-agent swarm coordination, autonomous threat detection, and GPS-denied perception — is genuinely and deeply AI-shaped (MARL, GCBF+, AOS), warranting a 5 for AI/ML fit. The award ceiling of ~$1.25M sits in the lower-middle band and transition mentions (TACFI, Replicator) are present but described aspirationally rather than contractually committed, capping award/transition at 3. Timeline realism scores low at 2 because the response deadline is unknown, which introduces serious planning uncertainty for a complex multi-agent hardware-software prototype that must move from simulation to field testing — an ambitious scope that a vague or compressed schedule could make infeasible. Problem framing earns a 3 because the end-user (AMC/621 CRW) and operational context are identified, but success criteria are not quantified and data availability for training in GPS/comms-denied environments is unaddressed.

Source

View original posting
Back to all contracts