In poker, it’s called a tell—a subtle habit, tic, or expression indicating how an opponent will act.
The communication is non-verbal, but the silent signal could speak volumes to observant types, allowing them to anticipate what their adversary will do next, and why. Combined with the capability to calculate odds, one could remove much of the risk from the game of luck.
But a novice to poker could throw it all off. A beginner is a true wildcard, making uninformed—and unpredictable—moves, forcing others to play more haphazardly, and opening up careful calculations to the wiles of fortune.
Driving, in many ways, mirrors the card game. Motorists and pedestrians seek subtle signs from each other—eye contact; smiles; a wave—to get a better read on how things will proceed in a world ruled by chance, then move forward accordingly.
In this realm of nonverbal cues, self-driving cars are the wildcard players. Without a human behind the wheel, we have no idea what the machines may do next—and without the proper communication in place, any move a pedestrian makes around them is a gamble.
But with Washington betting big on autonomous automotive advancement—and clearing the way for the cars to hit the road, soon—we have to figure out how to change those signals from tells to tell-alls.
The world offers plenty of ways to communicate—including, at last count, at least 6,909 spoken languages. But that abundance of options may actually make it more difficult to find the right voice for our cars. How can one guarantee the right message is sent—and received?
One way to work around the language barrier is circumventing the need for words at all.
Mercedes took a swing at the issue with its concept car, the F 015, coming up with highly-specialized headlights that utilize LEDs and tiny mirrors to project a crosswalk onto a street, indicating to pedestrians that it’s safe to pass in front of the vehicle. (The German auto giant also imagined the beams could fill in missing lane markings in construction areas and project arrows on the roadway for navigation purposes.)
Similarly, Google has come up with some visual cues to help bridge the communication divide. The company recently patented an autonomous vehicle pedestrian notification system that uses stop signs and the “safe to cross” figures found at most intersections to convey the car’s intentions.
And one Swedish firm put a positive spin on the situation, programming their cars to “smile” at pedestrians when the coast is clear.
One Swedish firm is programming its cars to smile to let pedestrians know the coast is clear.
That the symbols are already widely recognizable is key: it could help curb the learning curve for humans to pick up the new language of the road. And at least one company—Stanford spinoff Drive.ai—has taken that concept to a new-age level, proposing that emojis could be used to speak, at least symbolically, on behalf of a car.
But our autos may need to communicate audibly, as well.
The self-driving phenomenon has brought engineers back to the drawing board on nearly every part of our cars—including their horns. And just as the tone of a human voice can be used to indicate emotion, so, too, can the tenor of autonomous honks be utilized to convey a message.
In fact, Google reported last year that its driverless cars were learning different ways to honk at humans, depending on the situation at hand. Drive.ai has also considered the concept, with CEO Carol Reiley telling The Verge that vehicles should have the situational awareness to give a “more socially appropriate” honk.
Still, despite our new vehicles’ best efforts to talk to us, will anyone be listening?
Walk This Way
To determine how driver-free cars should talk to humans, programmers must first figure out how humans react to the cars. And just like with savvy poker players picking up tells, observation is key to predicting behaviors.
To help collect that objective data set, scientists have hatched some creative—and amusing—plans, but the results are no laughing matter.
One recent study, conducted by the Center for Design Research at Stanford, found that even while participants were perplexed by the vehicles, they crossed right in front of them anyway, with hardly any hesitation.
The study focused on the candid reactions of local pedestrians—who were unaware they were part of the experiment—when a seemingly driverless car rolled up to an intersection. And while most of the walkers eventually realized there was no one behind the wheel, only 2 out of 67 found a way to circumvent the unpredictable car.
In that case, the vehicle wasn’t attempting to communicate its intentions to others, but another study conducted by Duke University showed that such outreach attempts might not matter, anyway—and present huge problems for automotive designers.
The research team utilized several methods to convey ideas to pedestrians, including “walk” and “don’t walk” signals, as well as a projection of the car’s current speed, to indicate to others that it’s slowing down.
But a vast majority of the unsuspecting participants flat-out ignored the vehicle’s messages, according to the report. Even if they were reading the signs, the size of the screen would soon become a problem.
To broadcast words visible from a distance of 100 feet, a car would need to showcase letters 6 inches tall and nearly 4 inches wide, according to the report. A simple communique like “Safe to cross,” in that case, would require a screen at least 47 inches wide to show without scrolling, the researchers said.
While they may still need to deduce how the missives should be sent, however, many of those working on the problem agree that we must first learn to accept the new-age technology in order for any messages to be received.
Learning to Trust
Self-driving cars could not only change the shape of our communities, they could give us a new universal language to unite behind.
But their words will fall on deaf ears as long as we remain wary of the vehicles. It’s one reason why developers from Detroit to Silicon Valley have said trust in the technology is paramount to its effectiveness—and have invested heavily into determining what, exactly, could make the coming change less scary for us.
Regardless if we get the message, however, it seems we have no say when it comes to the eventual takeover of autonomous autos. The best we can do is go along for the ride—and try to hear them out.