To analyse VW’s emissions in real-world conditions, they used sophisticated devices mounted inside the trunks of two Volkswagens and one BMW that collected and analysed exhaust fumes as the cars drove along routes mainly in California.
You could be forgiven for reacting to the Volkswagen scandal by yearning for the halcyon era of dumb cars. Remember when our rides weren’t controlled by secret, corrupt software—when your father’s Oldsmobile was solidly mechanical and so simple in its operation that even a government regulator could understand it?
But emotionally attractive as it might be, the analog automobile isn’t a realistic option (which is perhaps why even Luddites aren’t asking for it). The real lesson in VW’s scandal—in which the automaker installed “defeat devices” that showed the cars emitting lower emissions in lab tests than they actually did—is not that our cars are stuffed with too much technology. Instead, the lesson is that there isn’t enough tech in vehicles.
In fact, the faster we upgrade our roads and autos with better capabilities to detect and analyse what’s going on in the transportation system, the better we will be able to find hackers, cheaters, and others looking to create havoc on the highways.
Right now we are at an awkward in-between phase in the transformation of the automobile—somewhere in the uncanny valley between the mechanical horse of Henry Ford’s era and the intelligent, autonomous, emissions-free, crash-free, networked fleet that will begin chugging along our roads later this century. This transition period will mean short-term turmoil. Cars today are lousy with code that can’t be inspected, opening the way for scary hacks and cheats and also the unforeseen complications of interactions between robots and humans.
Some of these problems call for obvious fixes. As many have pointed out after the VW admission, the code in our cars (and other life-threatening machines) shouldn’t be secret, allowing for better inspection by authorities and independent experts. Another obvious fix is to replace the sort of lab-testing that VW was able to game with the kind of real-world analysis that uncovered its chicanery.
But to do that, we will need more technology, not less. We need more sensors in cars and on roads and a network of computers watching the data to figure out when vehicles are behaving in aberrant ways. In other words, the best way to prevent cheating isn’t to make our cars dumber but to make the entire transportation grid smarter.
Look at how the VW cheat was uncovered. The software in the company’s diesel cars seems to have been designed to detect when a driver was pushing it through the specific routines that the Environmental Protection Agency uses in its emissions labs. When it detected these routines, the car’s software figured, “Hey, someone’s testing me!” and then put itself into an innocent, low-emissions mode that it didn’t use on the open road.
The cheating was discovered only when researchers at the International Council on Clean Transportation, a non-profit automotive research group, working with scholars at West Virginia University, used smarter technology. To analyse VW’s emissions in real-world conditions, they used sophisticated devices mounted inside the trunks of two Volkswagens and one BMW that collected and analysed exhaust fumes as the cars drove along routes mainly in California.
The devices are called “portable emissions measurement systems,” or PEMs, and they are a relatively novel way of measuring how vehicles perform in real-world conditions. The devices found that nitrogen oxide emissions from the VW Jetta were 15-35 times the acceptable standards, while those from the VW Passat were five to 20 times over the limit; the emissions from the BMW X5 were generally at or below the standards.
“This is a recent development—you miniaturise the equipment you would have inside a lab so that it fits inside a trunk,” said John German, a senior fellow at the ICCT. These devices are bulky and expensive: Each PEM unit fills most of a car’s trunk, and operating the system and analysing the data requires trained professionals. The council paid West Virginia University about $70,000 to study just three cars, German said.
He said it would take “a very long time” before such systems would become small and cheap enough to be used routinely in all cars. Still, as in all things powered by software, it’s likely these devices will become more accessible over the next few years, permitting more such tests, if not everyday use.
“What happened at Volkswagen had to do with embedded software that’s buried deep in the car, and only the supplier knows what’s in it—and it’s a black box for everybody else,” said Stefan Heck, the founder of Nauto, a new startup that is introducing a windshield-mounted camera that monitors road conditions for commercial fleets and consumers. The camera uses artificial intelligence to track traffic conditions; over time, as more vehicles use it, it could provide users with traffic and safety information plus data about mileage and other automotive functions.
The end goal for intelligent-car systems, Heck said, is to create an on-road network with data that is constantly being analysed to get a sharper picture of what’s happening on the road.
Sure, companies might still be able to cheat. But with enough independent data sources coming from different places on the road, it would become much more difficult.
He said there really isn’t any going back—software in cars is responsible not just for driver comforts like in-dash navigation but also for critical safety and performance systems, many of which improve the car’s environmental footprint. “I think people are fundamentally comfortable with software running in the car,” said Heck, who is also a consulting professor at Stanford who has studied the effects of intelligent transportation. “The real issue here is that it was a black box.”
When the cars are part of an intelligent network, the black box will be much more transparent.