Can machine learning dramatically improve the operations and maintenance of sophisticated naval aircraft? We used data from the E-2D’s Automated Logistics Environment (ALE) to explore the concept to good effect.
Using a series of AI techniques to fundamentally unbalance simulations/games and investigating methods for applying this technology to achieve superhuman planning and coordination in complex systems.
Applied deep reinforcement learning paired with novel algorithm training methods yielded an AI able to consistently defeat other industry AIs and human pilots in VR.
Complex tactical behaviors via decentralized resource and task allocation of heterogeneous manned-unmanned teaming operations.
Novel methods for enabling decentralized resource and task allocation for coordination of heterogeneous multi-robotic systems.
A hardware-in-the-loop simulation/stimulation software architecture powered by OpenDDS technology enabling high-fidelity testing between geographically disparate laboratories