Nvidia Research was named an Autonomous Grand Challenge winner at the Computer Vision and Pattern Recognition (CVPR) conference taking place this week in Seattle, Washington.
The company won in the end-to-end driving at scale category for its Hydra-MDP model, which incorporates artificial intelligence (AI) in building applications for physical AI deployments in autonomous vehicles. The technology could also be applied to industrial environments, healthcare engineering, robotics and more.
End-to-end driving is a process that streamlines everything that is needed to bring autonomous vehicles to market. This system not just accelerates real-world testing scenarios but also avoid potholes to development of these vehicles. This system operates with three distinct yet crucial parts operating simultaneously:
- AI training
- Simulation
- Autonomous driving
Each of these parts requires its own accelerated computing platform and the end-to-end system is purpose-built to enable continuous development cycles of these sectors, Nvidia said.
Nvidia accomplished this with a model first trained on an AI supercomputer like its Nvidia DGX and then validated in simulation.