Katie Washabaugh

NVIDIA Research presented eight papers at ICRA demonstrating how simulation-to-real transfer is enabling robots to operate reliably in dynamic, unpredictable environments. The work spans multi-arm coordination, cross-embodiment navigation policies, adaptive grasping, and deformable object manipulation, all trained in simulation without real-world robot data. The research shows measurable improvements: 3x speedup in multi-arm planning, 4.5x better navigation success rates, and 75% grasping success on real robots versus 41% baseline.

NVIDIA has released new open models and frameworks designed to streamline the development of production robots by integrating simulation, robot learning, and embedded compute into unified cloud-to-robot workflows. The tools aim to reduce friction in moving AI systems from simulated environments to real-world robotic hardware. This represents a shift toward making robot development more accessible and faster by consolidating previously fragmented tooling and infrastructure layers.