Level 4 self-driving cars may come to Europe next year, says Nvidia executive
Euronews Next sat down with Nvidia's vice president of the automotive team and spoke about how AI is changing the gears of self-driving cars, what it will mean for Europe, what we would do when we commute, and how it may change infrastructure.
Partial autonomous driving may come to Europe this year, and more advanced self- driving cars could arrive on the continent as soon as next year, vice president of Nvidia's automotive team, Ali Kani, told Euronews Next in an interview.
The chip leader and artificial intelligence (AI) giant revealed last week that, instead of building its own car, it has developed the software that supplies the intelligence layer for autonomy that robotaxi companies can purchase and build on.
AI is becoming a key accelerator behind self-driving technology and could cut the costs of the technology. However, Europe in particular stands at a critical juncture in determining how and when self-driving cars will navigate its streets and there may need to be infrastructure changes.
“We need to go as fast as regulation allows us, and I think what we see is it's opening up,” Kani said.
“My guess is in Europe it might end up being the end of this year,” he said, referring to Level 2+ driving, which means that the driver is still responsible for monitoring the driving environment, but the vehicle can steer, brake, and accelerate
As for Level 4, which means the vehicle operates completely autonomously under certain conditions and humans do not need to be ready to get involved, it could come in 2027, Kani said.
Trials are already being announced in major European cities, including London, though the timeline for full regulatory clearance depends on how well the systems perform in real-world conditions.
Europe currently allows Level 2 systems universally and has already approved Level 3 for controlled conditions. But this is not without its challenges.
This week, it was reported that Mercedes-Benz was pausing its Drive Pilot, an “eyes off” conditionally automated driving feature that was available in Europe and the US, according to the German publication Handelsblatt.
Nvidia announced last week that it will use its new technology in a new robotaxi alliance between Lucid, Uber, and Nuro.
Meanwhile, Mercedes-Benz, which also works with Nvidia, said it will launch a new advanced driver-assistance system in the US this quarter that will let vehicles operate autonomously under driver supervision. Nvidia’s founder and CEO Jensen Huang said the Mercedes-Benz technology will come to Europe and Asia in Q2 of this year.
Safety first
One of the biggest hurdles for deploying autonomous vehicles in Europe is the different regulations across different countries
But Nvidia’s technology can cleverly swerve the various laws, as while the core end-to-end model is consistent globally, the rules-based safety stack is customised for each country’s driving requirements.
Kani said that the company also has a different philosophy from its competitors, which is key to safety.
“Some other players talk about how we want to make sure we drive better than a human.
“We don't architect our system like that. We actually think of it in terms of how do we design this system so it doesn't ever cause an accident,” he said.
Nvidia achieves this by first, by using diverse sensor sets, so that if one camera isn’t working, there are other sensor options.
Secondly, it runs two stacks simultaneously: an end-to-end AI model alongside a dedicated safety stack that acts as a type of guardian. The stacks refer to a linear data structure that stores and manages data.
“The foundation is that two things are running, and the safety stack will make sure you never make a mistake. That means the AI model on its own is not something we depend on. We have a Safety Guard built into the system,” Kani said.
Nvidia’s Drive AV software recently earned a five-star safety rating from the European New Car Assessment Programme (Euro NCAP) in the Mercedes-Benz CLA, which is a notable accomplishment for a first-time production autonomous vehicle stack.
The long road
In five to 10 years, the Nvidia executive said one of the main challenges will be “long tail scenarios”—unexpected situations that systems haven't encountered before.
Such anincident occurred in December, when the robotaxi service Waymo was suspended for hours as vehicles struggled to read malfunctioning stoplights amid a power outage in San Francisco. Those in the driverless vehicles then found themselves stuck at darkened traffic lights.
One possible incident in the future could occur on European country roads. While Kani admitted he didn’t know much about the continent's road infrastructure, he said that he doesn’t think there would need to be a major overhaul for Europe’s roads to accommodate autonomous vehicles, but changes may be needed in more rural areas.
Kani said that while Nvidia’s technology can adapt to existing conditions on roads, he suggested it could be harder for autonomous vehicles to pull over on small, narrow country roads, as there is no shoulder.
“If that happens to be the case that on these country roads, there really is nowhere for you to pull over, you really have to stop in the lane. That's not safe because then there's another car that comes,” Kani said.
But despite these potential scenarios, there are many reasons to be optimistic about autonomous driving.
Kani said that self-driving vehicles make roads safer as human error or fatigue while driving is eliminated. But the vehicles could also reshape urban planning and how we manage our time.
He said an autonomous car could drop you off, go back home, and then pick you up when needed, effectively getting rid of the need for parking lots and using the space for society’s other needs, such as housing.
“You can just redesign a city so that we have more space for people to live and put parking lots further away,” Kani said.
For him personally, he is most excited about what he would do while commuting to work. He said he has a second home an hour and 40 minutes away from the Nvidia office.
With an autonomous vehicle, he said he could work in the car, or on longer journeys, or even travel overnight and sleep on the journey and arrive at the destination in the morning.
“I just feel like there are so many things we could do with our time if we had that [autonomous driving], and I would love that," he said.
“So I'm looking forward to that when we get there.”
No comments:
Post a Comment