DARPA’s AIR Program: Advancing AI-Driven Tactical Autonomy for Beyond Visual Range (BVR) Air Combat

Rajesh Uppal 4 hours ago AI & IT, Air Force & Aviation Comments Off on DARPA’s AIR Program: Advancing AI-Driven Tactical Autonomy for Beyond Visual Range (BVR) Air Combat 315 Views

DARPA has been at the forefront of developing artificial intelligence (AI) for air combat, beginning with its Air Combat Evolution (ACE) program, which focused on assisting pilots in dogfighting by automating low-level maneuvering tasks. Building on that foundation, DARPA has now launched the Artificial Intelligence Reinforcements (AIR) program, a more advanced initiative aimed at developing dominant tactical autonomy for multi-ship, beyond visual range (BVR) air combat missions.

This initiative aims to develop dominant tactical autonomy for multi-ship, beyond visual range (BVR) air combat missions. The AIR program is a significant step forward in autonomous air warfare, building on the successes of DARPA’s previous Air Combat Evolution (ACE) program, which focused on within-visual-range (WVR) engagements.

From Dogfighting to Beyond Visual Range Combat

The ACE program made headlines in August 2020 when an AI-driven system defeated a human Lockheed F-16 pilot in a virtual dogfight. The competition featured AI teams from industry and academia, with Heron Systems emerging as the winner by scoring five kills against zero for the human pilot. The program demonstrated that AI could outmaneuver human pilots in close-range combat, paving the way for greater autonomy in future air combat scenarios.

Building on ACE’s success, DARPA is now shifting its focus to BVR air combat, where engagements occur at distances beyond human visual perception, relying on radars, electronic warfare (EW), and sensor fusion. The AIR program will push AI autonomy to the next level by enabling real-time, AI-driven decision-making in dynamic, contested airspace.

The Vision for the AIR Program

DARPA’s Tactical Technology Office released a Broad Agency Announcement (BAA) in November 2022, seeking industry expertise to develop tools for achieving “dominant tactical autonomy.” According to DARPA, tactical autonomy refers to AI-driven systems operating with delegated human authority to execute short-term, mission-critical actions within a larger strategic vision.

The primary objective of AIR is to develop an AI-driven air combat capability that seamlessly integrates with existing sensors, electronic warfare systems, and weapons. This integration will enable autonomous systems to function effectively in highly dynamic and realistic combat scenarios, executing advanced tactics and strategies with precision. By leveraging cutting-edge AI, AIR aims to enhance decision-making speed, adaptability, and operational effectiveness, ensuring that autonomous combat systems can respond to evolving threats in real-time.

The AIR program will initially develop and test AI-driven autonomy solutions on manned F-16 testbeds. Once these autonomous systems prove effective in BVR engagements, they will transition to unmanned combat aerial vehicles (UCAVs), ensuring that AI-driven combat capabilities extend to next-generation warfighting platforms. This phased approach allows for rigorous testing, validation, and optimization of AI combat algorithms in real-world conditions. The ultimate goal is to create an AI-powered air combat system that integrates seamlessly with current military assets, operates effectively in complex battle environments, and rapidly evolves through continuous software iterations, maintaining a decisive edge over emerging threats.

Building on DARPA’s ACE Program

The AIR program builds upon the success of the ACE program, which demonstrated AI-driven air combat capabilities in within-visual-range (WVR) engagements, such as dogfighting and close-quarters aerial combat. While ACE proved that AI could execute high-speed, reactive maneuvers against human pilots, AIR takes a more advanced approach by shifting the focus to beyond-visual-range (BVR) combat. In these engagements, pilots rely on long-range radar, sensor fusion, and predictive decision-making rather than direct visual confirmation. By enhancing AI’s ability to assess threats, manage complex battle scenarios, and make split-second tactical decisions, AIR represents a crucial step toward integrating artificial intelligence into modern air combat operations.

BVR combat introduces a new level of complexity, requiring AI to handle tasks such as long-range target detection and identification, electronic warfare countermeasures, coordinated multi-ship tactics, and advanced missile engagement strategies. By incorporating AI-driven decision-making into these domains, AIR aims to exceed human cognitive limitations, processing vast amounts of sensor data in real-time to optimize tactical responses. This capability will allow the U.S. military to gain a significant technological advantage, ensuring superior situational awareness, faster engagement times, and increased operational effectiveness in future air combat scenarios.

Developing AI-Driven Tactical Superiority

The success of the AIR program hinges on the integration of advanced modeling, high-fidelity simulation, and real-world flight testing to develop next-generation AI combat capabilities. The program will focus on creating AI agents capable of making live beyond-visual-range (BVR) combat decisions, ensuring that these systems can process vast amounts of sensor data and execute optimal strategies in real-time. To support this, AIR will establish a robust testing and validation framework that enables AI-driven air combat to be rigorously evaluated under realistic operational conditions. Additionally, scalable methodologies will be implemented to ensure continuous AI performance improvements, allowing for rapid evolution and adaptation in response to new challenges.

One of the defining strengths of AIR is its ability to accelerate AI development cycles through an iterative process of design, testing, and deployment. By streamlining these processes, AIR will enable rapid updates to AI combat algorithms, ensuring that autonomous systems remain at the cutting edge of aerial warfare. This adaptability is crucial in an era of rapidly evolving threats, where technological superiority is determined by the ability to integrate new advancements seamlessly. By fostering a dynamic and responsive AI ecosystem, AIR will enhance the U.S. military’s tactical edge, ensuring that future air combat capabilities remain ahead of adversaries.

DARPA’s Plan for AIR: Phases and Technical Focus Areas

The AIR program is structured as a two-phase, four-year initiative aimed at developing and refining AI-driven air combat capabilities. The program’s technical focus is divided into two key areas: AI model development and multi-agent AI training. By leveraging advanced computational modeling and autonomous decision-making frameworks, AIR seeks to create AI systems that can adapt to the complexities of modern air combat, enhancing both individual and team-based aerial engagements.

Phase 1: Model Development (18 Months)

The first phase will focus on developing fast, accurate AI models that can process and react to uncertain combat scenarios in real-time. These models will be designed to improve with continuous data inputs, refining their ability to predict and counter adversary tactics. A key objective will be simulating adaptive adversaries, allowing AI systems to anticipate and respond to evolving threats dynamically.

Phase 2: Multi-Agent AI Training (30 Months)

Building on the foundation of Phase 1, the second phase will concentrate on multi-agent AI training, enabling AI-driven algorithms to execute coordinated multi-ship air combat maneuvers. This phase will emphasize distributed autonomous decision-making in dynamic, uncertain environments, allowing AI agents to work collaboratively while adapting to shifting battle conditions. Additionally, deception and adversary modeling will be integrated to ensure AI can counter sophisticated enemy strategies effectively. Ultimately, this phase will refine AI combat capabilities for operational deployment, ensuring seamless integration into real-world military applications.

AIR’s Game-Changing AI Capabilities

DARPA’s AIR program is pushing the boundaries of AI-driven air combat by addressing previously unexplored dimensions of autonomy and decision-making. Unlike earlier AI initiatives that focused on controlled scenarios, AIR is designed to operate in highly dynamic, open-world combat environments where unpredictability is a constant factor. A key innovation is the development of fully integrated sensor systems that enable real-time threat detection and response, ensuring AI-driven aircraft can process vast amounts of data and react instantaneously. Additionally, AIR is designed for scalability, allowing AI-controlled units to function effectively in large-scale engagements, including coordinated multi-ship formations.

Beyond real-time responsiveness, AIR emphasizes adaptability and advanced predictive modeling to anticipate adversary behavior. By continuously learning from new data and evolving combat scenarios, AI agents will be capable of adjusting their tactics mid-mission, ensuring optimal performance in rapidly changing environments. The program will also integrate deceptive techniques to outmaneuver enemy AI systems, further enhancing the survivability and effectiveness of autonomous aircraft. By tackling these challenges, AIR aims to reduce pilot workload, enhance combat efficiency, and enable autonomous systems to execute high-speed, precision-driven operations with unprecedented autonomy.

Investment and Industry Participation

DARPA is investing $70 million in the AIR program, with $30 million allocated for Phase 1 and $40 million for Phase 2. The agency expects to award six contracts in Phase 1 and up to four additional contracts in Phase 2, based on performance and concept viability.

DARPA Awards First AIR Program Contract to EpiSci and PhysicsAI

The Defense Advanced Research Projects Agency (DARPA) has awarded its first contract under the Artificial Intelligence Reinforcements (AIR) program to EpiSci, in partnership with PhysicsAI, for the development of tactical artificial intelligence (AI) algorithms for missionized, team-based air combat. The 18-month, $6 million contract marks a significant milestone in DARPA’s ongoing efforts to integrate AI-driven autonomy into air combat operations.

Under this contract, EpiSci and PhysicsAI will develop multi-agent AI systems designed for contested environments and beyond visual range (BVR) air combat missions. These autonomous combat drones will undergo rigorous training and evaluation—mirroring the tests human pilots undertake—to build trust, enhance capabilities, and allow airmen to focus on higher-level strategic tasks.

Key objectives of this phase include:

Enhancing AI decision-making capabilities for coordinated air combat.
Testing AI-driven autonomy on human-on-the-loop F-16 testbeds before transitioning to uncrewed combat aerial vehicles (UCAVs).
Developing AI agents capable of distributed tactical execution in complex, uncertain environments.

EpiSci has already established itself as a leader in AI-driven air combat. The company played a crucial role in DARPA’s AlphaDogfight Trials and the Air Combat Evolution (ACE) program, which recently modified an F-16 Fighting Falcon (dubbed X-62A VISTA) with combat-coded machine learning and AI capabilities. In May 2024 , Air Force Secretary Frank Kendall personally flew aboard the AI-assisted X-62A VISTA, as it maneuvered through a series of warfighting scenarios against a manned F-16, demonstrating the rapid progress of AI in combat aviation.

In another significant development, EpiSci has agreed to merge with Merlin, a company specializing in AI flight technology for fixed-wing aircraft. The merger aims to accelerate the development of autonomous aviation solutions across multiple platforms, further enhancing the AI-driven combat capabilities of the AIR program.

“With the acquisition of EpiSci, we are uniquely positioned to lead the charge in autonomous aviation,” said Matt George, CEO of Merlin. “We look forward to combining our expertise to develop versatile, trusted autonomy solutions that unlock human potential and deliver unparalleled value to our customers and stakeholders.”

What’s Next?

With DARPA’s AIR program now entering full-scale development, this contract sets the stage for future breakthroughs in AI-powered autonomous warfare. As EpiSci and PhysicsAI push the boundaries of multi-agent AI air combat, the transition from AI-assisted manned fighter jets to fully autonomous UCAVs is closer than ever. The coming months will be crucial in shaping the future of AI-driven air combat, reinforcing the U.S. military’s technological superiority in contested airspace.

The AIR program is expected to shape the future of AI-driven autonomous combat. By transitioning from manned fighter jets to AI-powered UCAVs, the U.S. military is positioning itself for next-generation air warfare, where AI-controlled aircraft will play a central role in securing air superiority.

As AIR progresses, it will set the foundation for fully autonomous air combat systems, redefining how wars are fought in the age of AI-powered defense technology.

Shaping the Future of Air Combat

The transition from manned fighter jets to autonomous UCAVs represents a paradigm shift in air combat strategy. AI-driven autonomy will not only reduce pilot workload but also enable swarms of AI-controlled aircraft to coordinate attacks, defend assets, and conduct reconnaissance missions with unprecedented efficiency.

By focusing on BVR engagements, DARPA’s AIR program is setting the stage for the future of autonomous air combat, where AI-powered UCAVs can dominate the skies without human intervention. As the program evolves, it will redefine modern aerial warfare, ensuring that the U.S. military maintains air superiority in an increasingly contested battlespace.

Conclusion

DARPA’s Artificial Intelligence Reinforcements (AIR) program is a groundbreaking initiative that will transform BVR air combat. By leveraging AI-driven autonomy and advanced simulation techniques, AIR will enable the rapid development of dominant AI combat capabilities for both manned and unmanned aircraft. As the program progresses, it will pave the way for a future where autonomous aircraft play a central role in securing air dominance.