AI-Reinforced Fully Adaptive Radar Exploitation (AIR-FARE)

Guided by advanced US systems, foreign military radars are rapidly evolving toward full adaptivity in antennas, waveforms, and processing that pose a significant threat to forward-based US assets. As a counter, Toyon Research proposes to demonstrate a swarm-enabled electronic warfare (EW) concept based on realistic behavior and performance limits. The concept will train the EW swarm via reinforcement learning (RL) to observe and probe the radar defenses; gain understanding of what the threat radar network is reacting to; then surreptitiously inserting signals into the radar to affect the radarĺs behavior, thus presenting a distorted radar picture to the enemy. This concept is suggested by previous work at Toyon in advanced EW hardware development and cooperative EW tactics controlled by artificial intelligence. The proposed RL framework assumes reciprocity between red/blue side behavioral adaptation creating a robust AI-Reinforced Fully Adaptive Radar Exploitation (AIR-FARE) system. In the proposed effort, Toyon will augment the existing RL framework with high-fidelity models of sophisticated, cooperative radar and cooperative EW tactics. We will use the framework to conduct a demonstration of this swarming EW concept and show how such a system can be scaled up to planning and training many cooperative EW assets for larger scenarios.