The future has been all too real in California, from waves of Waymo’s autonomous minivans wandering the streets of Mountain View to self-driving rideshare services in Irvine. Last year, Los Angeles applied for a $10 million grant to get driverless cars on the street, and more autonomy companies are popping up on the streets of Southern California.
Even if you havent’ seen a self-driving car on the streets, you’ve probably heard the hype surrounding the vehicles. But what is it like to ride one? Pony.ai, a self-driving startup based in Silicon Valley, gave me the answer in the form of a recent collaboration with Hyundai to offer a self-driving rideshare service, limited to Irvine residents and UC Irvine students. More of a consumer satisfaction test than a profitable service (it was free), electric Hyundai Konas outfitted with imagery tech, along with two safety drivers, recently picked up people who had registered beforehand with its proprietary app, BotRide.
The day I went, my trip began at an office complex across from John Wayne Airport. I stood for a couple minutes next to two people who were excitedly waiting for their autonomous taxi, but seemed disappointed when they realized the incoming Kona was for me, not them. Inside, two men sitting up front greeted me. Then and throughout the drive, the man in the driver’s seat didn’t touch the controls, but his hands hovered around 5 and 7 on the wheel, his eyes glued to the road like a hawk. The second man monitored an iPad mounted between the seats and the dash, displaying a map of the surrounding areas, with data fed to it by the sensor suite mounted above the car. Cameras, radar and lidar (Light Detection and Ranging) all contributed to the visual narrative of the world around us displayed on the iPad, in real time. There were also two iPads strapped to the backs of the front seats, which displayed route information for passengers, along with a real-time map of the car’s location.
Before we moved, the man in the front passenger seat let the safety driver know that the Kona was about to shift into gear and get going. One of his jobs was alerting the safety driver behind the wheel about maneuvers the Kona was about to make. My destination was the H Mart on Alton Parkway, which was about a 10-minute drive from my pick-up point.
We got off to a slow start–the Kona kept a snail’s pace as we drove over a series of white speed bumps. But once the road ahead was clear, the Kona smoothly accelerated. As we waited for traffic to diminish at Jamboree Boulevard, the Kona crept into the crosswalk, and when the road was clear, pulled into the right-most lane, exactly like you’d expect from a competent human driver.
Almost everything in the ride felt predictable. Far from the impulsivity of a human driver, the Kona rolled along thanks to a plethora of algorithms that let the car know exactly what to do in every situation. Its reactions correlated to whatever object or situation it encountered, down to the centimeter. Every response landed on a linear scale, balanced to–above all else–avoid a collision.
Throughout it all, I felt safe. Really safe, though at times, it felt like I was being chauffeured by a sharp 15-year-old who’d just spent the last couple days before his license test memorizing the DMV handbook. There were full stops at every stop sign and we never exceeded the posted speed limits. The Kona signaled before every lane change: first by safety driver in the passenger seat, who narrated the decision the Kona was about to make before it made it, and second by the actual blinker, which gave ample time for surrounding drivers to react.
As we pulled into the Diamond Jamboree Shopping Center, I expected the safety driver to put his hands on the wheel and guide us through the parking lot, but he did no such thing. Instead, the car stopped at every stop sign, then gently pulled into an empty area of the lot next to my destination.
After a quick bit of ice cream, I ordered another Kona, which turned out to be the same vehicle I had just ridden in. It was on the ride back that I noticed a quirk–the car would hug the curb whenever it steered into a right turn lane. While most normal lane changes were soft and smooth, the Kona would rapidly whip into the turn lane, following the curve of the road with inches to spare. On the other end of the spectrum, the Kona wouldn’t dare to drive through a turn lane that wasn’t on its route, choosing to swerve around the painted lines than to pass through them like a lazy human might.
When a black Cadillac XTS cut in front of us, the Kona reacted smartly. It braked to allot enough room in front of it and the Cadillac, while also maintaining its stopping distance to ensure that we didn’t stop short and make the guy behind us spill his coffee. Seemingly impressed by the well-executed display of multitasking and appropriate decision making, the passenger-side safety driver typed a comment into his laptop, stating that the car did a good job.
Before I knew it, my ride was over. But when I stepped out, I felt less safe than I did when I was sitting in my little robot car. I was impressed. A Pony.ai spokesperson later told me how the company prioritized safety over comfort. This surprised me because I was more comfortable being driven by the Kona than I’ve been in an Uber Black. The safety I felt was an unexpected bonus.
The idea of self-driving cars was introduced more than 75 years ago by General Motors at the 1939 World’s Fair in New York City, and the reality of the project has come a long way since. Naysayers like Apple’s Steve Wozniak may think that driverless vehicles won’t be able to independently traverse public streets within his lifetime, but companies like Pony.ai are showing that a driverless future may be closer than we think.