Heron Systems builds industry leading AI for sensor exploitation, robotic control, and strategic reasoning applications. Actively building these systems since 2012, we offer an end-to-end capability that considers data acquisition, training, performance analysis, and system integration holistically. 

 

Our proprietary multi-agent deep reinforcement learning framework achieves superhuman performance across a broad range of tasks such as Atari, Doom, StarCraft 2, and simulated within-visual-range (WVR) and beyond-visual-range (BVR) aerial combat. Our industry-leading experts are transitioning these algorithms from the simulator to real-world hardware.

 

We have also synthesized the state-of-the-art in reinforcement learning, deep learning, and traditional ML disciplines into our AdeptRL framework offering our customers leading edge capability built from a mature, proven technology base.