NVIDIA has introduced a new cloud-based system that will enable testing of autonomous vehicles using photorealistic simulation.

Expected to provide a safe and scalable method to bring autonomous cars to the roads, the new platform, known as ‘NVIDIA DRIVE Constellation’, is based on two different servers.

The first server runs NVIDIA DRIVE Sim software to simulate the autonomous vehicle’s sensors such as cameras, radar and lidar, while the second server manages the NVIDIA DRIVE Pegasus AI car computer, which runs the complete autonomous vehicle software stack.

The NVIDIA DRIVE Pegasus AI also processes the simulated data as the sensors of the car will respond to actual driving.

NVIDIA Automotive vice-president and general manager Rob Csongor said: “Deploying production self-driving cars requires a solution for testing and validating on billions of driving miles to achieve the safety and reliability needed for customers.

“Deploying production self-driving cars requires a solution for testing and validating on billions of driving miles to achieve the safety and reliability needed for customers.”

“With DRIVE Constellation, we’ve accomplished that by combining our expertise in visual computing and datacentres.

“With virtual simulation, we can increase the robustness of our algorithms by testing on billions of miles of custom scenarios and rare corner cases, all in a fraction of the time and cost it would take to do so on physical roads.”

The solution is capable of generating photoreal data streams to create multiple testing environments, including changes in weather and terrain.

Potential dangerous situations can also be added to the simulation process to evaluate the reaction of the autonomous cars, but ensuring no harm is caused to anyone.

DRIVE Constellation is expected to be made available to early access partners by the third quarter of this year.