When it comes to the much-anticipated advent of autonomous vehicles (so-called driverless cars, though that term isn’t completely accurate), there is good news and bad news. The good news is that autonomous vehicles will soon be driving among us. The bad news—or we could call it the challenging news—is that we will likely be in for a generation-long transitional period, when autonomous cars share the road with traditionally driven ones. It would be one thing to prepare for a brave new world where human drivers have been magically removed overnight from the equation altogether, but the messier reality for businesses trying to capitalize on the technology, their regulators, and the rest of us, is that we all need to prepare for a hybrid transportation system that can accommodate both humans and machines going about their daily commute, interacting safely.
The signs that autonomous vehicles could be commonplace in 8–12 years are everywhere. For starters, ever since cruise control became standard, the degree of automation present in our existing cars has been ratcheting up. For the leap into full automation, more money is being invested, more partnerships created, and vehicles tested. Apple invested $1 billion into Uber’s Chinese competitor, Didi Chuxing Technology Co., while General Motors Co. bought Cruise Automation, a software technology company. Strategic alliances between firms that can provide complementary innovations to one another appear to be a key feature of the impending AV era. Automakers understand that the current disruption of rideshare apps have spurred a small segment of the population to forego car ownership. They can surmise that eventually, ridesharing app companies will be able to purchase fleets of AVs to operate continuously, which will cut costs for everyone. As a result, they are investing early on to carve out their continued role in the transportation ecosystem. For instance, GM has invested in the rideshare company Lyft; the two companies will test a fleet of autonomous electric taxis by 2017. Volkswagen has invested $300 million in Gett, a popular European rideshare app. Alphabet (Google’s parent company) and Fiat Chrysler will utilize Alphabet’s AV technology in trial runs in 100 Chrysler minivans.
The cooperation of federal agencies in providing guidance for the AV future is even more remarkable. The Obama administration pledged $3.9 billion over 10 years to encourage new policy and research for AVs that will remove roadblocks and red tape. In the U.K., Britain’s Queen Elizabeth II announced in a speech to Parliament the introduction of bill that will eventually enable drivers to legally insure AVs.
Furthermore, automakers, rideshare services, and tech companies have formed the Self-Driving Coalition for Safer Streets to bring AVs to American roads and highways. They will work with regulators, lawmakers, civic organizations, municipalities, and businesses to “realize the safety and societal benefits of self-driving vehicles.”
The benefits of the collaborations on vehicle safety development are starting to emerge. In January, 20 automakers—representing 99 percent of the U.S. auto market—pledged to add automatic emergency brakes to almost all new vehicles (autonomous or otherwise) by 2022. According to National Highway Traffic Safety Administration estimates, the pledge will make automatic emergency brakes standard on new cars three years sooner than would be achieved via the formal regulatory process. The Insurance Institute for Highway Safety estimates that in those three years, the commitment will prevent 28,000 crashes and 12,000 injuries.
These early partnerships will prove invaluable in preventing the technologies from stalling. Public agencies are working swiftly (swiftly for government) to respond thoughtfully to inquiries and problems. In February, the NHTSA—in response to a letter from Google requesting the NHTSA interpret several Federal Motor Vehicle Safety Standards as they applied to the development of AVs—announced that it considers Google’s artificially intelligent driver the same as a human driver.
However, it is not all good news for the AV market—it will still have to grapple with state-level regulations. NHSTA has determined that it will not impose blanket federal laws regarding AVs; instead, it will leave that up to states. As we’ve seen with new technologies like ride- and home-sharing apps or drones, leaving regulations up to states can lead to a patchwork of regulation that stifles the growth of technology. When traditional vehicles share the road with autonomous ones, having different rules govern the relationship between the two each time you cross state lines could be a recipe for disaster. It’s safe to assume that the private sector, invoking the Constitution’s interstate commerce clause, will ask the federal government to reconsider imposing a national standard. Part of the challenge for regulators is a definitional one, as most vehicles become partly automated. More and more vehicles on the road approximate the aircraft we fly in that they will be mostly going on autopilot—with the option to be overridden by a human driver in tricky circumstances. And part of the challenge is that there won’t be an overnight switchover. Cars are not either automated or unautomated; certain features of a self-driving car are already showing up, like automatic braking. Furthermore, we cannot just say that, one day, all vehicles must be self-driving. Instead, the fully automated will have to share the road with the old-school.
So how will regulations and the law assign responsibility and driving privileges during this transition? For one thing, the current model of drivers’ education and licensing will need to adapt to new realities. Operators of autonomous vehicles will still need a license of a sort, particularly during the hybrid period. Drivers/passengers (who will be one and the same, sometimes) will need better knowledge of a vehicle, the computation system onboard, its fault-correcting mechanisms, and other details. You will need to demonstrate that you understand, for instance, when you are required to take control of the vehicle. While licensing is handled by the states, it would be wise to create nationwide standards so that a license to operate an automated vehicle in New York is also recognized in California.
And until we have 100 percent automation of transportation, we will have to figure out how to separate the automated cars from the old-fashioned—it would be too difficult for the two to speed along the same lanes. I believe highways will be the first to support the use of automated vehicles via dedicated lanes, similar to today’s high-occupancy vehicle lanes. Over time, the number of dedicated lanes for automated vehicles will increase as the share of these vehicles increase. In traditional American cities like Chicago, New York, or San Francisco, most side streets will remain the domain of driver-operated vehicles for the near future, while the major thoroughfares connecting neighborhoods (like Broadway in New York) will be redesigned to support both forms of vehicles through dedicated lanes.
Automated vehicles won’t rely only on their own intelligent systems to make decisions; they will also be linked with other vehicles (vehicle-to-vehicle networks) and environments (vehicle-to-environmental sensor) networks. During the transitionary period, even with separate lanes for self-driving cars, traditional vehicles will likely need sensors to communicate information about their speed, braking, etc., to self-driving cars. That’s because automated cars will need to know what’s happening at their sides as well as behind and in front of them—if a traditional vehicle in the left lane suddenly veers into the self-driving lane on the right, the self-driving car needs to know. But don’t get annoyed yet about how others’ embrace of new technology will affect your pocket: The price of installing these will be relatively low—perhaps a few thousand dollars—and most manufacturers will have an incentive to eat these costs to support the automated vehicle market. Adding these sensors will be as easy as going in for traditional vehicle maintenance where the auto manufacturer eats the costs of replacing a part that has been recalled/defected.
But the self-driving cars won’t be foolproof. Automated vehicles are being designed on the premise that they interact with mechanical sensors from the infrastructure and other vehicles on the road. What happens when these sensors fail? We have already heard of cases where blindly following a GPS has led to terrible outcomes. Now, imagine an automated vehicle enters terrain it thought would to provide data, but did not. Until confidence in such mapping increases, operators will have to pay close attention to the routes their vehicles plan to take—which brings us back again to the licensing question. Perhaps at the DMV you will have to demonstrate that you recognize when your car wants to take you down a Canadian logging road midwinter.
It doesn’t take much imagination to go on and on envisioning these tricky scenarios. And when you do, you quickly realize that, much like plenty of other technology debates, when we talk about the transition to a world of autonomous vehicles, we are really talking about a new era of human-machine relations.
This article is part of the self-driving cars installment of Futurography, a series in which Future Tense introduces readers to the technologies that will define tomorrow. Each month from January through June 2016, we’ll choose a new technology and break it down. Read more from Futurography on self-driving cars: