Is Your Car Smarter Than You?
Near the beginning of a conversation that would take us through his time as a teenage pro wakeboarder traveling the globe, the job interview he had with Elon Musk, and how he turned down a shot at the Olympics to accept a scholarship to the University of Florida, Jonathan Azevedo asks me if I’m familiar with the classic cartoon, The Jetsons.
“Remember when they would show the cars flying and they’d be a foot apart and traveling really fast with no interruptions? That’s essentially where we’re going,” he says.
And while he isn’t suggesting we’ll be flying to the supermarket anytime soon, Jonathan believes our lives will soon get a little bit easier, thanks to the advent of the self-driving car.
College in a machine shop
Jonathan, president and CEO of the product development firm Outer Reef Technologies, has loved to tear things apart and put them back together since he was a kid. At 12, he gained an intimate understanding of two-stroke engines by restoring a dirt bike somebody had thrown out. At 15 he bought his first car for $900 and restored it from the ground up.
When he got to college, Jonathan joined University of Florida at Gainesville’s Formula Club, where students build miniature Formula One race cars from scratch. Through working a lot of nights and well into many mornings inside the campus machine shop, Jonathan got to know students on the school’s Grand Challenge team, which was performing some of the earliest work on bringing self-driving cars to market.
Hoping to hasten the advent of technology that will remove soldiers from harm’s way by allowing military trucks to drive themselves through hostile terrain, the Defense Advanced Research Projects Agency (DARPA) created the Grand Challenge for autonomous vehicles in 2004. These events invited researchers across the country to compete in designing off-road-capable self-driving vehicles.
As an underclassman, Jonathan’s work on the Grand Challenge team was limited to fabrication tasks and occasionally assisting with test runs. But halfway through his senior year, DARPA announced the Urban Challenge, which asked teams to build cars that navigate courses resembling busy town centers. Because of his work on the Grand Challenge car, UF offered Jonathan a position on the Urban Challenge team. The gig also paid for his graduate school tuition.
The challenge
The first task he and the team faced was figuring out how to convert a stock Toyota Highlander with mechanical control inputs into a self-driving vehicle with digital or fly-by-wire inputs. “We just started tearing into the wire harness, and if you’ve ever seen a wire diagram for a vehicle these days, it’s crazy because there’s a million different wires going everywhere.”
“Once the conversion from mechanical to digital control was complete, the SUV could be driven with an Xbox controller, like you would a toy remote-control car,” Jonathan says. From there, the team loaded in sensors for obstacle avoidance: cameras, GPS navigation, and Light Detection and Ranging (LIDAR).
“After that, it was 100 percent software,” Jonathan says, referring to the artificial intelligence the team had to code into the vehicle so that it could make sense of its environment and decide how to interpret all the data its many sensors generated.
“Due to the tight integration of all the vehicle’s sensors and software, the team shared most responsibilities,” Jonathan says. However, toward the end of the challenge, Jonathan’s personal project was designing the software that would allow the car to keep track of its location through a combined use of LIDAR, accelerometers, and GPS.
In a process of simpler design challenges that led up to a final obstacle course competition at an abandoned military base in Victorville, California, Jonathan and his UF teammates were among the 20 to 30 Urban Challenge teams that made it to the finals from a field of 200 applicants.
Getting better with every mile
Though he now spends his time developing everything from high-precision surgical robotics to a crate that isolates anxiety-inducing noises for dogs, Jonathan’s confidence in the technology for which he helped lay a foundation, lies in what he sees as the same invaluable benefits that robots provide surgeons: precision and consistency.
“Humans, we have our off days. We’re sick or distracted, we’re very inconsistent when you compare us to a robot,” he explains. Thanks to their highly-accurate sensors, Jonathan says these vehicles will be able to talk and listen at the same time, communicating their intent while anticipating the moves of those around them.
Over time, as cars route and reroute, tracing paths across the country, they’ll become more efficient, thanks to machine learning. As that efficiency improves, roadway deaths, traffic jams, travel times, traffic tickets, and insurance premiums will go down, providing a boost to the economy.
“The first thing everybody thinks of with self-driving vehicles is convenience, being able to read a book as they’re driving to work,” he says. “But there are so many more positives to this technology.”
In other words, we might not be so far from living the Jetsons’ life after all.