Merlin’s USAF Tanker Autonomy Advances: Bolstering Air Force Resiliency for the Future

When we first announced our work on integrating the Merlin Pilot onto the KC-135 last year, it marked a major milestone in our mission to advance autonomous capabilities for the USAF. Since then, we’ve made significant progress—launching our flight test campaign, securing USAF acceptance of our airworthiness plan, and advancing AI-driven autonomy for aerial refueling. These advancements highlight our dedication to creating dual-use technology that will help bring about the next era of military aviation.

Here’s a more in-depth look at our latest milestones:

We successfully conducted a System Requirements Review to ensure the Merlin Pilot aligns with USAF operational needs and safety standards. Additionally, we delivered a comprehensive Concept of Operations (CONOPS) detailing how our Natural Language Processing (NLP) capabilities will function within aerial refueling missions. This work sets the foundation for scalable autonomous operations across a variety of Department of Defense (DOD) use cases.

To ensure seamless integration, we’ve established Pilot Working Groups with USAF pilots. Their insights are invaluable in refining our product requirements and ensuring the Merlin Pilots’ seamless integration into current and future operational environments. The working groups are important as Merlin is gaining critical insight into the pain points and high workload areas that USAF pilots experience, ensuring that our system can address workload inefficiencies and challenges. Additionally, these working groups are helping us align our product and technology roadmaps with the needs of the USAF. 

Merlin has completed numerous working group sessions and will be heading to MacDill Air Force Base in the coming weeks to continue these collaborative discussions. This approach enhances adoption success and operational effectiveness.

A deep understanding of in-flight communication is crucial for NLP-driven autonomy. At MacDill Air Force Base, we’ve been collecting real-world audio data from USAF refueling missions to train our AI algorithms. This approach – curation of training data with domain experts and operational users – works with our advanced run-time assurance and system design tools to ensure that AI capabilities are safe, effective, and explainable for real-world applications.

This process, akin to our automated checklists development, is critical to ensuring safe and efficient mission execution. By leveraging real-world data, we’re improving AI-driven decision-making and communication capabilities.

Another key milestone we successfully completed is our first contractual deliverable, which provides a CONOPS for real-time dynamic flight planning and NLP tailored for aerial refueling. This achievement signals our ability to meet project milestones and brings Merlin and the USAF one step closer to tailored AI for operationally relevant flight autonomy.