GTC 2018: Autonomous vehicle development should continue, urges Nvidia boss

• Trials must go on despite fears from autonomous vehicle testing fatality 
• Investments in AV should increase; testing halted only temporarily for assessment

THE autonomous vehicle industry should soldier on and not scale back on investments, especially in the area of safety, despite concerns that have arisen after an autonomous vehicle killed a pedestrian, said the head of Nvidia Corp. 

Its company co-founder and chief executive Jen-Hsun (Jensen) Huang (pic, above)noted that what happened on Mac 19 in Arizona, New Mexico, was “tragic and sad,” but this shouldn’t be a reason to stop all autonomous vehicle (AV) development.

“What happened last week is exactly a reminder why we’re doing this,” said Huang in a Q&A session following his keynote address at its ongoing GTC (GPU Technology Conference) here in San Jose, California. 

“We’re developing this technology because we believe it will save lives,” he added.

READ ALSO: GTC 2018: Nvidia burnishes its AI credentials with new line up

Last week, a Uber self-driving car hit and killed a woman crossing the street in Tempe, Arizona. Elaine Herzberg, 49, was pushing a bicycle across a road in mid-block, outside of any crosswalks, Tempe police said. It was understood that the Uber self-driving Volvo SUV was in self-driving mode, with a human backup driver behind the wheel.

The case marks the first fatality involving an autonomous vehicle colliding with a pedestrian, and is potentially a blow to the technology expected to transform transportation. The accident prompted Uber Technologies Inc to halt self-driving car testing in four markets –  San Francisco, Toronto, Pittsburgh and Phoenix.

This is however not the first incident involving a self-driving car. In Feb 2016, Google reported a non-fatal crash when its self-driving car slammed a bus. General Motors Co is facing a lawsuit over a non-fatal accident while Tesla Inc’s accident in July 2016 caused a fatality.

And just only this week in Mountain View, some 20km from San Jose, a motorist driving his Telsa crashed into a median barrier on Mac 26. He later succcumbed to his injuries and died. Authorities are however not clear as to whether the crash had anything to do with Tesla's autonomous driving system, and the accident is pending investigation.

Initial reports by the San Francisco Chronicle suggest that Uber may not be at fault. A follow up story by the same publication however clarified that Uber may not have done enough to avoid the collision.

Experts have also waded into the controversy surrounding the accident, with some suggesting Uber could have avoided the crash. As expected, Uber’s bitter rival, Waymo, said the same thing as did another competitor, Intel-owned Mobileye, who develops vision-based, driver assistance systems.  

Tech portal Recode however reported that not all AV players are stopping their respective testings. 

Nvidia’s view

Asked how the case in Arizona would potentially affect the development of the AV industry, Huang said he believes that as a result of what happened last week, companies should double down on investments instead of cutting back. 

The Nvidia boss argued that anybody who thought they could get away with AV development without supercomputers and simulators and to not invest in engineers dedicated into making sure that this product is as safe as possible isn’t realistic. 

“Investments into [AV safety] should go up,” he said. I think that the world is going to, as a result, be much more serious about investing in development systems, which is good,” he said.

Quizzed as to why Uber suspended its self-driving programme, Huang was cautious to comment about the incident directly and he appealed for patience and not to be judgmental over what happened.  

“We simply don’t know what happened,” Huang stressed. “Uber should be given a chance to understand what went wrong. I can tell you that engineers at Uber are intensely serious about what they do [in self-driving testing] and we’ve got to give them the chance to go and understand for themselves.”

In a related development, Reuters reported that Nvidia itself has halted its own AV testing a week after Uber’s accident in Tempe, Arizona. The news wire noted that the Santa Clara, California-based company is testing its technology globally in New Jersey, Santa Clara, Japan and Germany.

Quizzed why this was done, Huang conceded that this move was done for safety reasons and that Nvidia believes in using extreme caution and best practices for testing its self-driving vehicles since its own engineers are actually behind the wheels of these vehicles.

“It’s something that we take incredibly seriously,” he stressed. “The reason why we suspended is simple: Since the accident, there is a new data point to assess. We must wait and see so we can learn something from this experience.

“We don’t know if we would do anything differently but we should give ourselves the opportunity to learn from this. 

“So we’re going to wait – and it won’t take long – to see what we can learn from this, and after that we will self-assess and decide what to do next,” he declared, noting that in the meantime, engineers will continue their AV work using simulators. 

Earlier at the GTC conference, Huang unveiled a cloud-based system for testing autonomous vehicles using photo-realistic simulation called Nvidia Drive Constellation.

The system boasts of the ability to give users a way to test self-driving scenarios in the rain, snow, during day or night driving, and to simulate various different hazard scenarios and refine the algorithm over and over to make sure it’s safe before deployment.

Edwin Yapp reports from GTC 2018 in San Jose, at the invitation of Nvidia Corp. All editorials are independent. He is contributing editor to Digital News Asia and an executive consultant at Tech Research Asia, an advisory firm that translates technology into business outcomes for executives in Asia Pacific.

Related stories from GTC 2018: 

GTC 2018: Nvidia burnishes its AI credentials with new line up

GTC 2018: Nvidia aims to leverage on AI, machine learning

Related stories:

Nvidia to further develop future AI talents in Singapore