Cloudera named a market leader in 2023 GigaOm Radar Report for Data Lakes & Lakehouses Get the report

What’s taking driverless cars so long? Part 1: The human factor 

Grace Dobush Freelance journalist, Editor of the ADP ReThink Quarterly

JUNE 21, 2023
Overhead of car on street showing digital sensors while people cross the street

Part 1: The Human Factor 

The technology for autonomous vehicles is rapidly accelerating, but the industry needs human buy-in and trust to reach the next level.

Elon Musk in 2015 predicted fully autonomous vehicles would be able to drive anywhere by 2018. In 2019, he predicted self-driving taxis would be on the streets the next year.

In 2021, the Tesla founder conceded the challenge was more difficult than expected. “Didn’t expect it to be so hard, but the difficulty is obvious in retrospect,” Musk tweeted

Enabling real-world autonomous driving — where a vehicle can safely navigate dynamic street environments in any weather — is an incredibly complex endeavor. 

“There’s been a lot of evolution in the past two years as people’s expectations have changed,” says Brian Carlson, director of global product and solutions marketing at NXP Semiconductors. The systems have to be trained to be better than a human driver. “Safety is the first thing we do. It has to be the first thing we do.”

“There’s been a lot of evolution in the past two years as people’s expectations have changed,” says Brian Carlson, director of global product and solutions marketing at NXP Semiconductors. The systems have to be trained to be better than a human driver. “Safety is the first thing we do. It has to be the first thing we do.”

 

One of the main selling points of autonomous vehicles is making roads safer for drivers, passengers and pedestrians. A total of 38,824 died in traffic crashes in the U.S. in 2020, the highest number of fatalities in more than a decade, in spite of the decline in miles driven during the pandemic. 

Despite the recent major advances made in getting advanced driver assistance systems (ADAS) into market-available vehicles, truly autonomous vehicles still seem to be years away. 

The first installment of this investigation into why that is focuses on the wildcard human factor. The next installment will address big data and the technological challenges of autonomous driving.

Where we are

The Society of Automotive Engineers’ ADAS scale is the gold standard of measuring progress of autonomous vehicles.

Consumer Reports found that half of all U.S. 2021 models offered Level 2 features, which includes adaptive cruise control and steering assistance. The most advanced ADAS features currently available in market vehicles occasionally get to Level 3. 

The fully autonomous vehicles on the roads are so limited in their capabilities that they’re basically remote-controlled cars, says Kelly Funkhouser, manager for vehicle technology at Consumer Reports. 

“Most of these cars are geofenced to a very specific location,” she says. “They’ve extensively mapped the area using very expensive technologies and have every speck of dust mapped in Phoenix. But if you move it over one street or one lane, the car can’t do anything.”

“We’re really far away from ordering a self-driving taxi,” Funkhauser says. “That reality is not going to be here until 2050.” A self-driving vehicle’s capabilities need to exceed a human’s to earn our trust, and we’re not yet to that point. 

Trust will come naturally if the technology tells the person in the car how and why it’s making decisions, she says. 

One way a human driver feels increased trust is when the autonomous vehicle provides information about what its systems are doing and how they’re working, Funkhouser says. 

“A year ago, I drove a $150,000 Mercedes S class with an augmented reality heads up display that included information about the driver assistance systems, like boundary boxes to identify and label a person. Or if it was merging into another lane, you could see it was intending to speed up and pass another car,” she says. 

Last month, she found out Kia is doing a very similar thing for its adaptive cruise control. “A ‘cheap’ car is going to have Mercedes-level navigation. I think drivers would benefit in terms of trusting and understanding the systems, and that will move the market.”  

And governments have to trust the capabilities of self-driving cars allowed on their roads. Countries including Germany, France, Singapore, the United Kingdom, and Japan so far have made legal changes to allow Level 3 and in some cases Level 4 autonomous vehicles to operate on public roads.

The road ahead

Interest in autonomous vehicles hit a fever pitch in 2017 and 2018, the year Waymo introduced its autonomous taxi in Phoenix, Arizona. But then consumers pumped the brakes in 2019, Deloitte reports. Climate change concerns shifted a lot of the innovation focus to going electric.

Within the industry and among the public, expectations have been ratcheted back during the pandemic, says Michael Ger, a long-time managing director of manufacturing and automotive at Cloudera, who recently retired.

The willingness of consumers to pay more for advanced technologies is a hindrance. In 2022, Deloitte found that 61% of U.S.-based consumers surveyed were unwilling to pay more than $500 more for a vehicle with autonomous features; and 56% of people said they wouldn’t pay an extra $500 for advanced safety features. 

But KPMG’s annual surveys of global automotive executives actually show increased optimism regarding self-driving vehicles.
Auto executives’ views on when autonomous driving will break through:

It’s a very exciting time to be in the automotive industry, as companies compete with different solutions to the same problems. 

“Suppliers play a way bigger role than people realize,” Funkhauser says. “If a supplier develops a device or algorithm before anyone else, they will steer the direction of where the technology goes.” 

As the tech becomes more affordable, advanced driver assistance features will no longer be limited only to the high end of the market. Consortiums of hardware and software companies are tackling those big challenges — read more in the next installment.

 

“We’re really far away from ordering a self-driving taxi,” Funkhauser says. “That reality is not going to be here until 2050.” 

 

Article by

Photo of author Pedro Pereira

Grace Dobush

Grace Dobush is a freelance journalist based in Berlin. She has contributed to Fortune, Wired, and Quartz and is the editor of the ADP ReThink Quarterly.

More articles


AI-first Benefits: 5 Real-world outcomes

Learn more


Space-based AI shows the promise of big data

Learn more

Your form submission has failed.

This may have been caused by one of the following:

  • Your request timed out
  • A plugin/browser extension blocked the submission. If you have an ad blocking plugin please disable it and close this message to reload the page.