DAVID BELO
Driverless Tech Could Mean Faster Auto Races
By Drake Bennett, April 11, 2016
Q: WILL AUTONOMOUS CARS BE DESIGNED MORE CREATIVELY?
A: Once a vehicle is on the street, I think federal laws are going to have a hard time waiving anything. Even if the autonomous vehicles have safer drivers, you’re going to have ’60s cars on the street for years to come, ’80s cars. They’re not going to just disappear. A lot of other factors are going to impact what these autonomous vehicles look like. Visibility from the inside out is not a priority, especially if people aren’t even facing the windshield; they could be facing each other.
Q: STILL, IT’S NICE TO SEE OUT OF THE CAR.
A: But it could be optional. You could toggle it off or on, or you could be more selective. … Also, if the cars are going to be electric, then the front of the car may not need so much air coming in. This big grill that a lot of vehicles use as an identity would be eliminated. … If the vehicle is driving itself and doesn’t need headlamps, all it needs is marker lamps so other people can see it.
A lot of autonomous vehicles will fall into the fleet category: a cab company or Uber, who will get rid of drivers. They don’t care what it looks like; they care what it costs. No one’s going to have any passionate attachment to the damn thing. It could be more phone-booth-like, just so it goes. And they’ll have advertisements on the side. I think the appearance is going to go south.
Q: HOW MIGHT DRIVERLESS CARS CHANGE THE DESIGN OF CITIES?
A: One of the obvious issues is, assuming the cars are programmed not to hit what they recognize as a human being, there is a danger that in busy areas cars would never get through at all—they’d constantly be stopping. Congestion problems could actually get worse if pedestrians were able to assert much more priority. Most city authorities realize that what you want is pedestrian priority in the center of cities, which are the most complex busy places. But to ensure that vehicles can travel smoothly, we will have to control pedestrian crossing movements in a way we don’t at the moment.
Q: WHAT ARE THE OTHER POTENTIAL UNINTENDED CONSEQUENCES?
A: Well, if I’m in my driverless vehicle and I can’t find a parking space, I could jump out and just get it to circulate until I’m ready for it. Which would add to congestion. Or I could drive to work in the morning, send the car home, and get it to come back for me in the evening, and then go back in it, so then you’ve got four journeys instead of two. There’s the potential to create more vehicle movements.
David Belo trained as a mechanical engineer and he’s worked on cars for most of his career. He’s not a typical gearhead, though. He’s also studied computer science and physics, and his specialty is in the field of simulation: figuring out how to virtually reproduce the real world, in all its complexity and unpredictability, with enough realism that drivers can practice split-second decisions and engineers can test precision parts under high stress. Belo worked at the General Motors proving grounds before being hired by McLaren, a pioneer in the use of racing simulators—and in figuring out how to apply simulation technology outside the world of auto racing.
Q: McLAREN’S HERITAGE IS IN RACING—IT’S HARD TO IMAGINE THAT WITHOUT DRIVERS.
A: Yeah, the driver is the big part of the show, and in many respects it’s what people go to see races for. On the other hand, it’s also the technical challenge—one of the biggest you could aspire to solve as an engineer. The spectacle is just as much about the fierceness of competition between the drivers as it is about the marvel of what teams are coming out with this year. I think there is a lot of excitement for autonomous driving within Formula One companies for the same reason. Something that has been traditionally such a hard problem—driving—is now within our reach with these interesting and complex algorithms. At the same time, we don’t design without the human in the loop.
Q: WHAT OTHER ADVANTAGES ARE THERE TO DOING THAT?
A: Whether it’s Tesla or GM or Porsche or a racing company, you have to take into account that you’re designing this machine for a human to exploit it. I think what makes us and some of the other companies working in this field interesting is we’re starting to use vehicle simulators to understand how the brain is interpreting a lot of the signals the driver needs to interpret in order to react to what’s happening in the car and change the control of the car. That work is just as important as developing the engine to produce three more horsepower. The work interacts to achieve a lap time that’s lower or to get a passenger car from point A to point B with a higher likelihood of no accidents.
Q: WHAT ELSE DOES McLAREN BRING TO THE PROBLEM OF DRIVERLESS CAR DESIGN?
A: When you’re not driving, you’re a passive element. And this is another elusive field in that we don’t actually know how the human body reacts to different sensory inputs and physical inputs. As you’re being shaken by the vehicle and as you’re visually perceiving what’s around you and audibly perceiving what’s around you, that might set off some sort of nausea or other lack of comfort that is very hard to understand as an engineer. There are not a lot of virtual testing tools and virtual algorithms to predict how comfortable or uncomfortable this is.
At McLaren we typically deal with this in racing with a vehicle simulator. We’re relying on human drivers to tell us that it felt comfortable or uncomfortable or nauseating or unsettling or confidence-inspiring. We’re not relying on algorithms.
Q: AND YOU CAN APPLY THAT TO DRIVERLESS CARS?
A: We can simulate all of the sensory inputs basically to the human driver and to the human occupants.
There are vehicle design applications as well. So if you think about going from a single car to vehicle platoons, a cluster of cars, whether it’s on the highway or in a city, and they’re all autonomous. [By being electronically coupled, driverless cars could in theory drive much closer together than human drivers.]
So specifically in F1, for instance, aerodynamics is obviously a big thing. So imagine driving down the highway and you’re able to determine that there is a cluster of autonomous cars that are all going to the same region, and you can cluster them together, and you can make use of the fact that only the leading car is actually punching a hole through the air, and the other cars can reap the benefits of linking behind them.
If you can change the body of a car in a very specific way to just be able to gain a fraction of a percent, you’re now able to scale that gain across all the cars. We’re now going into territory where all these little gains will actually have big impacts.