Nvidia recently introduced its plan to use a cloud-based system for testing the autonomous vehicles at the GPU Technology Conference (GTC) in San Jose, California. The cloud-based system is known as DRIVE Constellation Simulation System. It is an integration of two servers, majorly. The first is Nvidia’s powerful AI car computer, Drive Pegasus, which runs the complete autonomous server stack and processes the simulated data as if it were coming from the sensors of a car driving on the road. The second would be virtual simulation software to simulate a self-driving vehicle’s sensors, such as cameras, lidar, and radar.
For moving the production any further, the autonomous vehicles required a method to test and validate to achieve the safety and reliability for customers.
The company already has been in partnership with various companies which are already working on autonomous vehicles such as, Volkswagen, AG, Tesla, China’s Baidu and various others.
The new testing system is made according to the various driving conditions which are supposed to improve the testing capability of the system. The system creates a photorealistic simulation which helps in generating billions of miles of AV testing and all this is done by using the two sensors installed in the system. The simulation server generates the simulated data and processes it using the DRIVE Pegasus, the DRIVE Pegasus then returns the driving commands to the simulator thus completing the digital feedback loop.
This loop is completed 30 times in a second thus ensuring that the simulated vehicle is functioning properly. The software is also capable of simulating various weather conditions, different times, and different types of road surfaces. With this VR-based system, it is believed that it is possible to create a scenario to blind the sun 24 hours a day which couldn’t be done before.
The DRIVE Constellation product was launched at the same time when Nvidia and Uber put their AV testing on hold after the Tempe accident that took place a week ago.
This comprehensive solution is expected to be made available in the third quarter of 2018. It will allow NVIDIA customers like Google and Uber to improve their autonomous vehicle systems before putting it to the real-world test.



Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.