Air Force STTR DAF27TZ01-DV001 AI-Enabled Computer Vision Air-to-Air Refueling Performance Assessment and Automated Pilot Debrief System

FOCISAUS is an Air Force Direct to Phase II STTR topic for cooperative information sharing and sensor fusion across teams of uncrewed air systems, the multi-agent autonomy problem that sits between single-platform autopilots and genuine collaborative autonomy. The work funded here covers distributed data fusion across heterogeneous airborne sensors, shared world-model construction when platforms see different parts of the same scene, consistency and conflict resolution between agents holding contradictory tracks, and operation under degraded, intermittent and contested communications where naive sharing schemes either saturate the link or propagate bad state. Because it is an STTR topic, a formal partnership with a U.S. nonprofit research institution is mandatory, with at least 40 percent of the work performed by the small business and at least 30 percent by a single research institution, which makes this a natural fit for a startup paired with a university robotics, estimation or networking group. The Direct to Phase II structure carries a 2,000,000 dollar ceiling and no feasibility phase, so applicants must evidence completed Phase I-equivalent work in multi-agent autonomy or distributed fusion. The topic is part of Release 1 of the Department of War fiscal year 2027 SBIR/STTR announcement, which opened 28 October 2026 and closes 25 November 2026 at 12:00 noon Eastern; confirm topic availability and the live deadline on DSIP before committing to a teaming agreement.

HUBBLE (High-throughput Unmasking of Biological Interactomes with Binding Landscape Emulators) is a DARPA Biological Technologies Office program to build fast computational emulators that predict protein-protein interactions across the human body far faster than atomistic simulation. Instead of full molecular dynamics, performers produce low-dimensional signatures of each interaction from the topology of its diffusive binding landscape. The goals are to extend coverage into the uncharacterized 'dark interactome' and to produce a hypercatalog of context-aware interactions with predicted probabilities, affinities and kinetics. Traumatic brain injury is the first use case, with causal, directional interaction networks validated experimentally. The program is a strong fit for teams combining machine learning on structural and interaction data, physics-informed models, computational biophysics and protein modeling; teams without a modeling core are a weak fit. Phase 1 runs 12 months with up to $2.0 million per team, and an 18-month Phase 2 is planned. Awards are Other Transactions for Research, and the model OT requires either a nontraditional defense contractor or nonprofit research partner with a significant role, or one-third cost share. Questions are due 3 November 2026, abstracts of up to six pages are due 5 November 2026 at 3:00 p.m. ET, and oral proposal packages are due 11 December 2026 at 3:00 p.m. ET, submitted through the DARPA BAA Tool.

This Air Force Direct to Phase II SBIR topic funds robotic systems that can perceive, grasp, route and install wire harnesses in unmanned aerial system airframes, one of the hardest remaining problems in aerospace manufacturing automation because cable harnesses are deformable, self-occluding and never present the same geometry twice. The technical content sits squarely in embodied artificial intelligence: learned visual perception of thin deformable objects under poor contrast, pose estimation and tracking through self-occlusion, contact-rich force-controlled manipulation, and policies that generalise across harness variants and airframe configurations rather than being hand-taught per part. The payoff for the Air Force is throughput and cost in UAS production and sustainment, where harness installation is still overwhelmingly manual and is a recognised bottleneck as drone production volumes rise. Because this is a Direct to Phase II topic, the award ceiling is 2,000,000 dollars and there is no feasibility phase, but applicants must demonstrate that they have already completed work equivalent to a Phase I feasibility study, which in practice means prior robotics or manipulation prototypes with documented results rather than a concept. The topic sits in Release 1 of the fiscal year 2027 Department of War SBIR/STTR announcement, which pre-released 7 October 2026, opened 28 October 2026 and closes 25 November 2026 at noon Eastern; applicants should verify the topic number and final close date on the Defense SBIR/STTR Innovation Portal.

Release 1 of the Department of War SBIR/STTR fiscal year 2027 Broad Agency Announcement carries 25 topics, a substantial share of them artificial intelligence, autonomy and robotics work, and it is the single highest-volume non-dilutive funding window open to U.S. deep-tech small businesses this autumn. The cycle pre-released on 7 October 2026, opened for submission on 28 October 2026 and closes 25 November 2026 at 12:00 noon Eastern, giving roughly four weeks of writing time against topics that are written narrowly enough that proposal-topic fit, not general AI capability, decides outcomes. AI and autonomy topics in this release include AI-enabled computer vision for air-to-air refuelling performance assessment and automated pilot debrief, agile wire harness manipulation and installation for unmanned aerial systems, autonomous technical data package to robotic task specification, and fusion of cooperative information sharing for uncrewed air systems. Award structure differs by component and track: Phase I is capped at approximately 300,000 dollars for Air Force and Office of Strategic Weapons topics and 315,000 dollars for Navy topics, while Direct to Phase II topics award up to 2,000,000 dollars and skip the feasibility stage entirely for companies that can already evidence Phase I-equivalent work. STTR topics require a formal research institution partnership with at least 40 percent of work performed by the small business and at least 30 percent by a single research institution. One Army xTech topic runs an earlier white paper deadline of 23 October 2026, so component-specific dates must be checked individually on DSIP.