Nvidia today announced that it would suspend its testing of autonomous cars on public roads until it knows what went wrong last week, when a self-driven Uber vehicle featuring Nvidia’s technology struck and killed a pedestrian in Arizona.
The idea of self-driving cars is gaining a lot of traction these days, but the incident in Arizona may put the idea of autonomous cars on hold for a while. On May 18, a Volvo XC90 SUV equipped with Uber’s autonomous vehicle system (opens in new tab) struck and killed a pedestrian who was crossing the road.
Following the incident, Uber suspended its self-driving testing program in all cities, and earlier today the Governor of Arizona suspended Uber’s right to test automated vehicles in the state pending the NTSB and NHTSA’s investigations.
Uber isn’t the only company affected by this incident; the fatality puts heavy scrutiny on other autonomous car initiatives. Nvidia is currently leading the charge in developing technologies for self-driving vehicles, and it’s not taking the news of Uber’s fatal crash lightly.
Today at GTX 2018, Nvidia CEO Jensen Huang announced that the company halted all public self-driving vehicle testing worldwide until further notice, though it's still operating manually-driven vehicles with its AI technologies so it can continue to capture real-world data and work to make these systems safer. In the meantime, until Nvidia can sign off on the safety of its self-driving car technology, the company isn’t willing to put lives at risk.
Nvidia is now running virtualized experiments to test its AI systems. Last year, Nvidia revealed that you could use the Holodeck platform to safely train artificial intelligence systems in a controlled, simulated environment. Today, the company introduced the Nvidia Drive Constellation, which is a virtual environment of the same nature, but with the express purpose of testing autonomous vehicle technology.
The Nvidia Drive Constellation platform features the VR Autonomous Vehicle Simulator, which enables testing of autonomous driving systems in adverse conditions that could impair the sensors systems, such as fog, rain, and bright sunlight. With it, you can queue up any weather condition and run simulated tests without putting people in danger, making it the safest place to test Nvidia’s Drive AI system. (opens in new tab)