It’s long been the stuff of science fiction. You hail a cab, and when it pulls over there’s no one inside, nor is there a place for a driver, and no steering wheel or pedals. Slipping into a seat, you tap the screen, relay a destination, and take off in quiet comfort. When the journey is over, you push a button on your smartphone to pay the fare. No need for small talk.
That fantasy may soon become reality — faster than most folks might imagine, if tech giant Google has its way. Best known for its search engine and ubiquitous online ads, Google has become a leading player in the fast-emerging field of autonomous driving. It recently turned to Livonia auto supplier Roush Enterprises to produce 100 bubble-shaped “Google Cars” that look like they could have rolled off the set of the early Woody Allen sci-fi comedy, Sleeper. But the unusually-shaped cars have a very serious purpose. Over the next several years the vehicles will be part of a pilot program operating near Google’s Silicon Valley headquarters in California. The goal is to test whether self-driving cars can be operated safely and efficiently in day-to-day, real-world road and traffic conditions.
“This technology has the potential to be tremendously transformative,” said Chris Urmson, director of Google’s autonomous vehicle program, during a visit to Detroit in January. “Many people do love to drive. But a lot don’t. What they want to do is get from one place to another.”
Proponents suggest self-driving vehicles will not only take the drudgery out of the morning commute, but will shorten the daily drive by improving traffic flow without the need to add more highways. Other groups are working on autonomous air-based vehicles that can pick up a pod filled with cargo, medical supplies, or people and transport it to one of 20,000 airports in the United States or a remote location.
While it is considered at the forefront of the field, Google is just one of the many players in the nascent autonomous vehicle industry. Key rivals include General Motors Co. and BMW, the latter working on a system called Auto Park Valet that would allow passengers to disembark directly at their destination while sending their vehicle off to find a parking spot — returning to pick them up when summoned by a cellphone app.
In turn, Nissan Motor Co. has promised to put its first fully-autonomous vehicle into production by 2020, about the same time upstart Tesla Motors hopes to enter the autonomous market. A recent study forecast that by the middle of the next decade, the market for autonomous technology will top $40 billion annually. By 2035, another report projects virtually every vehicle sold worldwide will feature at least some self-driving capabilities.
Not everyone is convinced the autonomous revolution will happen that fast. “We think it will take a little longer than some other companies do,” says Steffen Linkenbach, director of the North American autonomous vehicle research program run by Continental Teves in Auburn Hills. The German mega-supplier is considered one of the leaders in developing the sensors and components needed to make vehicles like the Google car work, and Linkenbach thinks the technology will eventually go mainstream.
Even before it does, Michigan is vying to become one of the centers of that brave new world. Ongoing efforts in Lansing, as well as by groups such as the Detroit Regional Chamber, are targeting a potentially lucrative payoff — tens of thousands of jobs that could be created by the companies researching, testing, and producing autonomous vehicles and their components.
With its network of suppliers and automotive manufacturers, Michigan is already a powerful force in the new field. And one can get a sense of what this world will soon look like by paying a visit to Ann Arbor, where the University of Michigan Transportation Research Institute, or UMTRI, is setting up the 32-acre “M City.” Crisscrossed by four lane-miles of pavement, and scheduled to open in the spring, it is meant to simulate the various conditions that tomorrow’s smart cars will have to navigate.
“We were trying to design a facility with some particularly challenging scenarios for connected and automated vehicles,” says Jim Sayer, an UMTRI researcher and one of the leaders of the M City project. “We seldom think about just how complex the act of driving is, and how much information we have to take in as we proceed down a roadway.”
According to various studies, a human driver has to make at least 100 judgment calls every mile they drive — more when road conditions are poor or traffic is heavy. The satellite mapping system and cameras on an autonomous vehicle may know where the school zone is, but will they be capable of recognizing the time that school is letting out and where kids are dodging traffic outside the crosswalk?
Human drivers constantly have to adapt to obstacles and unexpected conditions on the fly, says Maarten Sierhuis, co-director of Nissan’s autonomous vehicle program. “We need to be able to drive any intersection anywhere and at any time to be fully autonomous,” says the former NASA scientist, ironically underscoring why many view the target date of Nissan’s project as being too ambitious.
Most experts agree that autonomous vehicles will ultimately reach showrooms, but they believe the rollout will come in a series of stages. We’re already entering the first, semi-autonomous phase with vehicles like the Mercedes-Benz S-Class. It brings together sonar, radar, and camera systems in what engineers call a “sensor fusion,” operating much like our own eyes and ears. The big sedan can spot a car — as well as a pedestrian or even a dog — that might wander into its path, sounding an alert. If the driver is distracted, it can bring the vehicle to a complete halt. On a freeway, it can start up when traffic begins to move again.
Cadillac plans to push things into the next phase with the 2017 launch of the Super Drive system on the CT6 sedan. It will allow hands-off motoring, but only on limited-access highways such as an interstate. “We’re not going to wait until we perfect a (fully) driverless system,” says Jon Lauckner, GM’s CTO.
About the same time, Cadillac plans to start equipping the smaller CTS sedan with a vehicle-to-vehicle and vehicle-to-infrastructure communications technology. Known among industry insiders as V2V and V2I, the system takes radio traffic reports to a new level. They will be able to more accurately track traffic flow, as well as weather conditions, on an ongoing basis, and at a granular level. You’ll instantly know if you’re heading into a tie-up, or about to run into a patch of black ice. Eventually, you might even get a warning if another vehicle is likely to run a red light, and your car will automatically slam on the brakes.
On a related path, the University of Michigan last September completed the Connected Vehicle Safety Pilot Model Deployment — which, at its peak, linked up 2,800 vehicles in the Ann Arbor area. Even as that effort was winding down, the U.S. Department of Transportation announced plans to set up a more advanced project covering 120 miles of metro Detroit freeways. The results of that program could lead to the creation of a national highway communications system. Meanwhile, the National Highway Traffic Safety Administration is considering the possibility that all future vehicles will be equipped with V2V and V2I capabilities.
There’s an ongoing debate among autonomous vehicle advocates over whether this will be necessary to make self-driving cars a reality. It clearly will help, but some experts stress that to be truly autonomous, a vehicle will have to be able to operate entirely on its own. After all, what would happen if hackers disrupted a V2I network? Would our roadways come to a grinding halt?
To get to the next step in creating a car that can truly operate on its own will require even more sophisticated sensing devices. From a hardware standpoint, the critical breakthrough is something called LIDAR (light and radar). This really is right out of rocket science, with the first use of the technology coming with NASA’s Apollo program, where astronauts used the laser-based system to map the surface of the moon. Each of the new Google cars features a distinguishing, if ungainly, dome on its roof housing a LIDAR unit, which paints the world with an invisible laser beam.
In turn, the system reads the returning image to create a vividly detailed “map” of what surrounds the vehicle. It can even spot obstacles that might be out of a human driver’s line of sight.
Getting rid of that awkward dome is one challenge, and by the time autonomous technology reaches production it’s quite likely manufacturers will settle for mounting separate LIDAR units in each of a vehicle’s four corners. That is, of course, assuming costs can come down to an affordable level. The good news, industry planners predict, is that the cost of self-driving technologies should mimic the rapid downward spiral seen in the consumer electronics world once production volumes reach critical mass.
Hardware is only half the equation, however. The data a LIDAR unit produces at blinding speed is useless if the brains of the system can’t figure out what it all means, and how to respond. That’s all the more impressive when you consider the most advanced vehicles on the road today already require as much as 100 times more lines of software code than an F-16 fighter jet, says Glenn Stevens, vice president of the MICHauto business attraction and retention program at the Detroit Regional Chamber. Yet, for now, even some of the simplest tasks a human has to deal with can overwhelm the most advanced autonomous prototypes.
“I’d give anything to be able to turn over control during (my) half-hour commute,” says Sam Abuelsamid, a Detroit-based senior analyst with Navigant Research and a former automotive engineer. He’s convinced it will take longer than many autonomous vehicle proponents would have you believe, especially in climates like Michigan where snow, ice, and salt can obscure lane markers and other visible cues the technology will have to rely on. In fact, the sensors now available can’t always distinguish between icy roads, gravel, and fresh snow. And considering “how you want to have a car respond on each surface is very different” for each of these scenarios, that’s a serious problem, Abuelsamid says.
Continental’s Linkenbach agrees that there are some major challenges ahead. Coming up with an autonomous vehicle that “won’t work because it’s snowing or the sun is shining at the wrong angle … will be unacceptable to customers.” That said, he also believes it’s just a matter of time before such challenges are overcome.
So, whether the first fully autonomous vehicles go on sale in 2020 or 2025, the general consensus is that they will ultimately reach showrooms. And when they do, they’ll begin to transform the driving experience. What that means is far from clear, and the transformation likely won’t happen overnight.
What’s all but certain is that things will change, as Ford Motor Co. CEO Mark Fields emphasized during his recent keynote speech to the Consumer Electronics Show in Las Vegas. “We see a world where vehicles ‘talk’ to one another, drivers and vehicles communicate with the city infrastructure to relieve congestion, and where people routinely share vehicles or multiple forms of transportation for their daily commute,” he told a packed audience.
No wonder Google may enter the fast-growing car-sharing field. Today, pioneers like Uber and Lyft are little more than high-tech taxi services, but that could change if a driver is no longer in the equation, Uber CEO Travis Kalanick suggested during a 2014 technology conference. Uber also wants to find a way to switch to autonomous vehicles because, Kalanick said, “When there’s no other dude in the car, the cost (of riding with Uber) gets cheaper than owning a vehicle.”
In some parts of the world, self-driving vehicles may become the only way to get around unless you have a bicycle or want to take mass transit. London set a global example when it recently began charging a hefty congestion fee on vehicles entering the central city. Several other cities, such as Hamburg, Germany, are considering outright vehicle bans, while others may limit access to autonomous vehicles powered by batteries or other zero-emissions technologies.
Nowhere might the benefits of autonomous vehicles be seen more than in China, which is today the world’s largest automotive market. The roads in key cities like Beijing and Shanghai are almost constantly in gridlock, and, as in much of urban America, there’s little room to add more roads. Now, add the fact that most of China’s mega-cities are choked with smog.
GM last year showed off a potential solution for China dubbed the EN-V 2.0. It looks like a modern-day rickshaw, but the little two-seater may point to the future of transportation, rather than the past. EN-V 2.0 uses a compact battery drive system rather than conventional internal combustion technology to create no harmful emissions. Relying on extremely lightweight materials would sharply improve the two-seater’s range, contend GM engineers.
In turn, the autonomous drive system is capable of “platooning,” riding all but on top of the vehicle ahead to pack far more vehicles on a given stretch of roadway, even while reducing collisions.
At the 2015 North American International Auto Show, show-goers got a look at a very different take on what an autonomous vehicle might look like — the Mercedes-Benz FO15 concept. Think of it as a Google Car on a grand scale, with all the luxury accouterments of a Mercedes S-Class. For example, the battery-electric vehicle can be operated simply by using hand gestures or voice commands. Its windows can turn into video displays at the touch of a button. And when operating in autonomous mode, the front seats can swivel around to turn the interior of the vehicle, dubbed “Luxury in Motion,” into a mobile lounge.
“Anyone who focuses solely on the technology has not yet grasped how autonomous driving will change our society,” says Dieter Zetsche, chairman of the board of management at Daimler AG, and head of Mercedes-Benz Cars. “The car is growing beyond its role as a mere means of transport and will ultimately become a mobile living space.”
In the near-term, don’t expect to be able to catch a snooze on your way to work. A human driver or, if you prefer, a co-pilot — will still have to remain on alert, ready to intervene in the event of a system failure, or some other problem, says UMTRI’s Sayer. It could be another decade or two before the technology proves reliable enough to allow you to plug in a destination and relax. And that raises a serious concern, he cautions. How do you keep a driver focused and at the ready if they’ve spent the last two hours sitting behind the steering wheel twiddling their thumbs?
Nonetheless, the general consensus is that autonomous technology should yield enormous improvements in vehicle safety, especially when one considers more than 90 percent of crashes are caused by human error, according to federal data.
Since reaching its 54,589 peak in 1972, U.S. highway deaths have plunged 40 percent to 32,719 fatalities in 2013. There are a variety of reasons, including safer road designs and a crackdown on drunk driving, but Adrian Lund, president of the Insurance Institute for Highway Safety, says perhaps the most significant factor has been improved vehicle design, especially the use of driver-assist features like electronic stability control. Early results show the first wave of semiautonomous technologies, such as forward-collision warning systems, already yield a measurable reduction in crashes.
To reduce further accidents, Volvo and Nissan have both set a goal of building ultra-safe cars, with Volvo targeting introduction for the 2020 model year. Indeed, the Volvo XC90 4WD was one of nine models that had no fatalities reported during a just-completed, three-year study by IHS Inc. For his part, Adrian Lund, president of IHS’ safety group, errs on the side of cautious optimism. “True zero, I don’t think we’ll ever see, but we’re going to get close to it,” he says. “I would be amazed if we don’t see a further decline in the death rate, perhaps a significant one.”
Considering all the potential benefits, the drive for autonomous vehicles has rapidly gained momentum and consumer support. Once skeptical, the latest polls show buyers increasingly ready to place orders. A study released in January by the Boston Consulting Group found two-thirds of those surveyed were interested in buying a semi- or a fully autonomous vehicle, and were willing to pay up to a $5,000 premium. There was some pushback: Some respondents were concerned about not being able to control the vehicle at all times, while others said they wouldn’t feel safe. “But overall, people clearly associate autonomous vehicles with safety,” says Xavier Mosquet, head of Boston Consulting Group’s North American automotive practice. “This is pretty significant, considering nobody has yet driven one of these vehicles.”
The study anticipated the first fully autonomous vehicles won’t reach showrooms until 2025, but even by that point, it forecast sales of more limited technologies would be worth $42 billion annually.
It’s no wonder that Michigan leaders are so anxious to stay at the forefront of the self-driving revolution, says the Detroit Regional Chamber’s Stevens. To do so will require a concerted and coordinated effort teaming government, education, and industry. More funding will also be needed to support projects like the M City autonomous testing village, and the state’s colleges and universities will need to place more emphasis on training the next generation of designers, programmers, and engineers.
There’s a solid base today, Stevens says, but, “We have to develop a pipeline of future talent.” The payoff could be significant, and put Michigan at the forefront of a new era of mobility, much as it took the lead putting the world on wheels a century ago. db