Self-driving cars need to be better mind readers

Illustration: Sarah Grillo/Axios

Self-driving cars can be programmed to stay in their lane and obey speed limits. What they lack is human intuition — the ability to know what's going on inside someone else's head.

Why it matters: Autonomous vehicles are getting closer to reality but if they're ever going to drive better than people they need to learn how to share the road with other cars, pedestrians and cyclists. Turning them into social creatures with human instincts is arguably the biggest roadblock to autonomy.

Human drivers react to social cues all the time — and not just a wave of the hand or a nod from a pedestrian signaling for them to go ahead. People give off hundreds of signals that others use to understand  their state of mind and respond accordingly.

  • A person standing in a crosswalk looking at their phone is probably distracted, for example, so the driver knows to yield.
  • A group of kids playing around on the sidewalk could tumble into the street, so caution is advised.
  • A motorcyclist at an intersection touches their feet to the ground, so it's probably safe to proceed.

Be smart: AVs need the same instincts so they can recognize, understand, and predict human behavior. But that ability, which comes so naturally to humans, is hard for machines to learn.

Driving the news: Perceptive Automata, a startup whose backers include Toyota and Hyundai, has a unique approach — it's applying behavioral science techniques to machine learning.

  • Instead of training AI models by labeling objects (this is a tree; this is not a tree, for example), Perceptive Automata characterizes the way people understand others' state of mind and then trains its algorithms to recognize and predict human behavior.

How it works:

  • Researchers collect sensor data from vehicles interacting with pedestrians, bicyclists, and other motorists and then chop up the images into smaller slices.
  • Then they blur or cover a portion of each slice and ask groups of people what the depicted person is about to do.
  • They repeat the process hundreds of thousands of times, with a variety of interactions, and use the results to train models that interpret the world the way people do.

Humans don't always agree, of course — predicting human behavior is not as easy as labeling a tree — but capturing that ambiguity is important to program how an AV should behave, CEO Sid Misra says.

Yes, but: That also means self-driving cars could be prone to mistakes, notes Elizabeth Walshe, a cognitive neuroscientist who studies driving behavior at the Children's Hospital of Philadelphia and the University of Pennsylvania.

  • "Humans are not perfect. Are we happy with machines that make mistakes and are not perfect?"

The bottom line: It will take time for people to learn to trust self-driving cars, but understanding what they might be thinking is a good place to start.

Go deeper: Cars of the future need to be able to heal themselves

Additional Stories

What we're driving: Volkswagen e-Golf

Photo: Volkswagen

I'm driving the Volkswagen e-Golf, and perhaps just for sentimental reasons, since it's about to become obsolete.

The big picture: The e-Golf will go away once a wave of new electric models from VW hits, starting next year. VW plans 70 electric vehicles across all its brands by 2028 — a total of 22 million EVs worldwide. It's the automaker's way of finally putting the past behind it after a devastating diesel emissions cheating scandal.

A wave of showy AV deals

Illustration: Sarah Grillo/Axios

Waymo has inked an exclusive partnership with the giant Renault-Nissan Alliance on autonomous vehicles, the latest in a dizzying array of headline-grabbing deals in the AV space.

The big picture: Carmakers and technology companies need to work together on self-driving cars because neither industry has the expertise to do it alone.

  • But some so-called partnerships lack substance or seem to be driven as much by ambition and the need for credibility and capital as they are by true commercialization efforts.

No one would argue Waymo lacks credibility — it is widely considered a leader in autonomous technology — but its partnership with Renault-Nissan to take its technology to "a global stage" seems a bit of a stretch.

  • The deal is exclusive only "for an initial period" and is limited at this point to researching commercial, legal and regulatory issues in France, Japan, and potentially, China.
  • For Renault-Nissan, it is being portrayed in Europe as a salve between two bickering alliance partners in the wake of a failed merger with Fiat Chrysler Automobiles and the downfall of former CEO, Carlos Ghosn.

Between the lines: In a field as competitive as self-driving cars, it's instructive to pay attention to the language of partnership announcements and to the motivations behind them. This is especially true among lidar technology companies, for example, that are fighting for attention and hungry for investment.

  • Amid a glut of lidar manufacturers, companies often tout their "partnerships" with major carmakers.
  • What they usually mean is the carmaker is testing a handful of their lidar units on its pilot vehicles.
  • Though far from revenue-generating production contracts, such announcements could be parlayed into big investments if the companies are perceived as hot.

Not all partnerships are equal. Some are just a MOU — memorandum of understanding — signaling an intention to move forward on something more substantive, but not a legal contract.

  • One recent example is Aurora Innovation's deal with Fiat Chrysler Automobiles to develop self-driving commercial vehicles.
  • But MOUs typically expire or fall apart, as in the case of a similar partnership Aurora had with Volkswagen for the past 18 months. (It's also working with China's Byton and Korea's Hyundai-Kia, which just pumped about $70 million into Aurora.)
  • In a blog post earlier this year, Aurora explained it preferred working with multiple carmakers rather than being locked into an exclusive deal so it could deliver its technology more broadly.
  • But it's hard to imagine Aurora's relationship with Volkswagen, the world's largest automaker, wasn't a factor in securing a $530 million investment led by Sequoia Capital in February, at a $2 billion valuation.

Be smart: Unless money and board seats are involved, partnerships can be fleeting. Where big money is involved — as in the billion-dollar-plus investments made in Cruise by GM, Argo by Ford and Uber by Toyota — it's serious.

Uber's one-stop plan for transportation

Illustration: Lazaro Gamio/Axios

In 10 short years, Uber's ride-hailing service has become so ubiquitous that people use "uber" as a verb. For its next act, Uber wants to manage everything about how you get around, whether on the roads or sidewalks, underground or in the air.

The big picture: Like Amazon, which started selling books online and now delivers almost everything right to your door, Uber aims to leverage its digital expertise from ride-hailing to become a one-stop shop for transportation.

Read more at Axios
© Copyright Axios 2019