HAMERITT

STR, with key contributions from Microsoft Corporation, is developing a simulation-based autonomy software framework called Hybrid AI Mission Environment for RapId Training and Testing (HAMERITT) that generates realistic Unmanned Aerial Vehicles (UAV) scenarios for neurosymbolic algorithm developer teams under DARPA’s Assured Neuro-Symbolic Learning and Reasoning (ANSR) program. As UAVs become more prevalent in search and rescue, emergency management, and law enforcement, hybrid neural and symbolic AI approaches for complex reasoning have the potential to safely decrease the level of human intervention for these applications. To support hybrid AI model training and testing, HAMERITT provides a scenario authoring pipeline to generate a large volume of randomized scenes with symbolic mission context, a UAV simulation environment with realistic camera and sensor feeds, and an Algorithm Development Kit with APIs for perception and maneuver planning and control.

Check out this joint paper by STR and Microsoft on ArXiv today! https://lnkd.in/eiq74mPr

Congratulations to the team on their impressive technical developments and paper!