Skip to main content

AVs need to be ‘100 to 1,000 times better than humans’, says Intel

Autonomous vehicles (AV) need to have a robotic system which is better than a human driver, because society will not accept machines killing people, according to Intel. Speaking at the Consumer Electronics Show 2019 in Las Vegas, Intel senior vice president Amnon Shashua said AVs probably need to be 100 to 1,000 times better than the human experience - which presents the question of how to validate such a system. “When you do your calculation, the amount of data you need to collect to verify somethi
January 14, 2019 Read time: 2 mins
Autonomous vehicles (AV) need to have a robotic system which is better than a human driver, because society will not accept machines killing people, according to 4243 Intel.


Speaking at the Consumer Electronics Show 2019 in Las Vegas, Intel senior vice president Amnon Shashua said AVs probably need to be 100 to 1,000 times better than the human experience - which presents the question of how to validate such a system.

“When you do your calculation, the amount of data you need to collect to verify something at a certain probability is one over that probability. If you do the calculation correct, comparing it to humans, the probability is that it will be billions of miles you will need to collect,” he added.

Shashua emphasised that a balance is needed between over-regulation and under-regulation is needed and that the US suffers from the latter.

Chris Urmson, co-founder and CEO at Aurora, insisted that the 324 US Department of Transportation has been supportive of the technology.

“I think it’s more about a thoughtful approach to talking about the safety regulations. Intel and 4279 Mobileye have done a good job at putting forward a first thought about this with the RSS (Responsibility-Sensitive Safety) and I think we need to have more of that kind of conversation,” Urmson added.

Both companies are part of the newly formed PAVE coalition, a group of private sector companies, academic institutions and other stakeholders whose stated ambition is to educate the public and policy makers on AVs.

Deborah A.P. Hersman, president and CEO of the National Safety Council, said: “It's in the best interest of the industry to engage with mobility advocates and other organisations and have a diverse group of stakeholders at the table because those differences will help create a positive path forward.”

For more information on companies in this article

Related Content

  • FHWA collaborative framework on automated driving systems: an explainer
    September 26, 2023
    USDoT FHWA has put together a collaborative framework to help secure the roll-out of automated driving systems in the US. John Harding of FHWA explains the thinking…
  • Lidar more use in ADAS than AVs, says Cepton
    September 18, 2020
    Silicon Valley start-up says it is already deploying Lidar with automotive manufacturers
  • Car-sharing locations need to be revised to tackle car ownership, says Drivy
    February 8, 2019
    The only way to reduce car ownership is to stop putting car-sharing vehicles in places where people travel - and start putting them in places where they live. This was one of the main messages from the 'Future of Global Urban Mobility' event, hosted in London last night by global research agency Kadence International. Speaking at the London Transport Museum, Patrick Foster, chief business development officer at car-share marketplace Drivy, says: "OEMs want to give you a car for free if you start sha
  • Loop detection still has a part in traffic management
    March 2, 2012
    Bob Lees, co-founder of Diamond Consulting Services, on why the loop detector just refuses to go away. The more strident proponents of newer and emergent detection technologies are quick to highlight what they see as the disadvantages, and hence the imminent passing, of the humble inductive loop. The more prosaic will acknowledge that loops continue to have a part to play in traffic management, falling back on the assertion that it is all a question of application. And yet year after year the loop, despite