Lockheed Martin Skunk Works, in collaboration with the University of Iowa’s Operator Performance Laboratory (OPL), has achieved a significant breakthrough in artificial intelligence (AI)-commanded missions. The successful demonstration involved the use of two piloted L-29 aircraft, acting as surrogates for uncrewed systems, to perform jamming support in a simulated air-to-ground mission.
The project, known as Enhanced Collaborative High-Frequency Orientation System (ECHOS), showcased the potential of autonomous uncrewed aerial systems (UAS) working in coordination with crewed tactical platforms. This collaboration creates a unified team that can effectively address complex threats in future combat operations.
By successfully executing the electronic attack task, the AI agents demonstrated their ability to provide data for rapid decision-making, reducing pilot workload and increasing mission effectiveness. Throughout the demonstration, experienced OPL pilots flew the L-29s while following instructions from the AI agents. They received cues in the form of heading, altitude, and speed, maintaining the option to override the AI if necessary.
To achieve this level of autonomy, the AI agents were trained using advanced deep multi-agent reinforcement learning techniques. The emphasis was on incentivizing emergent collaborative tactics while minimizing reward shaping. The agents were able to learn these tactics with sparse rewards, ensuring optimal execution of the mission.
The successful demonstration focused on sim-to-real transfer, training an AI agent in a simulated environment and then flying it on real aircraft to test its behavior. This achievement provides confidence in the ability of AI agents to deliver high performance and reliable behaviors.
Moving forward, the Skunk Works and OPL team plans to test AI in an end-to-end Suppression and Destruction of Enemy Air Defenses (SEAD/DEAD) mission by the end of 2023. The lessons learned from these tests will inform the development of AI/autonomy for future programs, such as the Air Force’s Collaborative Combat Aircraft (CCA), as well as ongoing crewed-uncrewed teaming development.
This groundbreaking work aligns with Lockheed Martin’s 21st Century Security vision, aiming to provide integrated, mission-focused capabilities that keep our forces and allies ahead of ready.
FAQs
1. What is the purpose of the ECHOS project?
The ECHOS project aims to demonstrate the potential for autonomous uncrewed aerial systems (UAS) to operate in coordination with crewed tactical platforms in future combat operations. It showcases the power of artificial intelligence (AI) agents in providing data for rapid decision-making and increasing mission effectiveness.
2. How were the AI agents trained?
The AI agents were trained using advanced deep multi-agent reinforcement learning techniques. The focus was on incentivizing emergent collaborative tactics while minimizing reward shaping. Sparse rewards were used to ensure the agents learned the best possible tactics for the mission.
3. What are the future plans for AI/autonomy development?
Following the successful demonstration, the Skunk Works and OPL team plans to test AI in an end-to-end Suppression and Destruction of Enemy Air Defenses (SEAD/DEAD) mission by the end of 2023. The lessons learned from these tests will inform the development of AI/autonomy for future programs, including the Air Force’s Collaborative Combat Aircraft (CCA) and crewed-uncrewed teaming development.
4. How does this work contribute to Lockheed Martin’s vision?
This groundbreaking work aligns with Lockheed Martin’s 21st Century Security vision. It aims to deliver integrated, mission-focused capabilities with speed and efficiency to keep our forces and allies ahead of ready. By unlocking the power of artificial intelligence in combat operations, the project enhances the company’s ability to address complex threats effectively.
(Source: militaryleak.com)