Major car manufacturers are already testing their autonomous vehicles, over millions of kilometers every month, in rFpro.
rFpro were the first to deliver a commercial platform for training, testing and validating autonomous vehicles, in simulation. rFpro employs a three stage process:
1. Model the real world
2. Physically model your vehicle's interaction with the real world
3. Scale your testing, massively
This video shows examples of rFpro's digital library of real-world models in North America, Asia and Europe.
It shows vehicles physically interacting with the world, through sensors (camera, LiDAR, radar) and through vehicle dynamics enabling reward / cost functions for ride comfort and head-toss.
Finally it shows an individual experiment scaling across multiple CPUs and GPUs and then multiple experiments scaling across a cluster of CPUs / GPUs enabling millions of kilometers of testing in realistic timescales.
Please see other videos in this series for a more detailed look at the technology in rFpro that delivers training and testing for autonomous videos.