To Trust the Machine
Swiss holographic augmented reality (AR) company WayRay was just crowned the winner of the 2017 Top Ten Automotive Startups™ Competition at AutoMobility LA, with an AR execution that represents a big step forward for this exciting mobility technology. Improvements were made in not only the size of the unit, but also in a larger field of view and longer projection distance, making for a more natural experience.
The system has the ability to enhance your infotainment experience, with things like points of interest and incoming content. That’s definitely exciting, and it’ll keep us entertained especially once we make the transition from human-driven vehicles to autonomous technology, which the industry is seeing as increasingly inevitable. But the application that may have a larger impact is the one that will actually help us get through this transition to self-driving cars without soiling ourselves in the process.
The industry is making rapid progress in its technical development of autonomous vehicles, but it’s still reckoning with the role of human acceptance of the driving task being snatched away from us. Studies have repeatedly concluded that a major part of the buying public is very resistant to autonomous tech. AR can serve as a companion technology that can bridge the gap and help our primate brains subsume their stifling self-preservation instinct, using much of the same tech and data that enables autonomous driving itself.
It’s remarkable how uncomfortable people are with the loss of control, to the point that being a passenger can literally make us carsick. The fact is, no matter how much safer autonomous vehicles promise to be, many people still insist that the computer will have to wrest the steering away from our cold, dead hands.
This seems a bit strange, given that we willingly give up responsibility for our safety all the time. We get in the car with our friends and relatives driving; we even hop in with a stranger in our Uber or cab, or the city bus. Along with the basic presumption that a driver's license comes with a certain level of competence—doubtful given the sorry state of driver ed in this country—we also feel like we have a general handle on how the human brain works.
In a recent incident recounted in Car and Driver, an AV carrying a University of Michigan researcher, Anuj Pradhan, reacted with sudden braking to a bird that flew in front of it.
"If he had not seen the bird, Pradhan says, he might have been tempted to yank the emergency shutoff... A driver might have explained the reason for the incident, but the machine simply resumed."
We don't trust the machine, because we can't query its mind...at least not in the moment. If our Lyft driver swerves, we can look up from our phone and ask, "What the heck was that?" But we don't have a social connection to the autonomous algorithm.
So how do we bridge this gap? How do we ease humans into a world that researchers generally agree will be much safer, while circumventing our natural wiring that yearns for the status quo?
Augmented reality offers a useful tool in this battle. The same concept that took the public by storm last year, as millions of players tried to catch Pokémon theoretically lurking in the bushes, has a big future in automotive.
Augmented reality has its roots in other technologies that have existed for a while, like head-up displays and night vision. The combination of these ideas with the myriad sensors involved in the autonomous task can present a vivid tableau of the inputs that the system is processing splashed across the windshield of the vehicle. That sparrow that darted across Pradhan's path would have been highlighted in bright red, offering insight and reassurance to less eagle-eyed passengers that the vehicle has things under control.
When it comes down to it, the order of the day for autonomous vehicles is transparency. With the technology increasingly relying on adaptive machine-learning algorithms, which analysts are calling "black boxes," that cannot be picked apart and readily understood, we should at least be focused on making sure that the inputs are fully apparent.
With several more years before we see fully self-reliant "Level 5" vehicles hitting the road, we have some time to inure drivers—those would-be passengers—to the idea that their car is getting the right information to make the right decision. If we want to get to a safer future as soon as possible, we need to focus on putting those inputs out there for all to see, and that means a big push for augmented reality.
