Communications of the ACM
Cars That Modify Driver Behavior
Credit: itechnews.net
"Having the car connected with the exterior world is a necessity," a BMW North America executive told reporters at the annual Consumer Electronics Show in Las Vegas in January. This was the dominant automotive-related message from CES, that increasingly sophisticated and immersive connective technology—the digital dashboard—is coming to the world's automobiles.
From a strictly consumerist point of view, this might be seen as a positive development, but for the car companies, it's critical, given that the most technologically advanced and useful thing we have in our cars today—at least in terms of the driver's experience—might be a smartphone. Pretty much everything the car companies unveiled—the ability to stream the Pandora Internet radio service, get real-time weather updates, check stocks, watch YouTube videos, find gas prices, and receive traffic updates—is already possible via an everyday smartphone.
The big automakers insist their emerging in-car systems will allow drivers to perform all these activities in a safer manner, because they will be integrated into intuitive, typically hands-free user controls. Doug VanDagens, Ford Motor Company's global director of connected services, said at CES that drivers will be able to manage their social lives, online and otherwise, "all while keeping their eyes on the road and their hands on the wheel." Granted, hands on the wheel and eyes on the road are preferable to the alternatives, but this conceit misses a central point raised by cognitive psychology: that looking toward something and actually seeing it are not necessarily the same thing.
Similarly, the manufacturers are being careful to include warnings with their technological suites; for example, Audi's "MMI" system, which enables drivers to read reviews of restaurants (in addition to giving directions), peruse Google Earth ("for the first time in a production vehicle," says Audi), or enter commands via an Apple Newton-like interface near the gearshift, warns upon startup: "Please only use the online services when traffic conditions allow you to do so safely." This self-evident declaration might assuage liability lawyers but overlooks a central issue in traffic safety: that traffic conditions change suddenly and unexpectedly and people are overoptimistic not only about their own driving ability but their innate human ability to react to such changes. An industry whose product leads to the death of some 40,000 people in the U.S. in an average year might be more sensitive to the idea of encouraging further distraction away from the responsibility at hand.
The curious point is that given how essential technology is in our everyday lives, it is perhaps inevitable that it is also the thing that promises the greatest benefit for improving the driving experience. Not in terms of accessing a Pandora stream but by making it safer and more efficient, helping traffic flow better, and perhaps someday, removing human drivers altogether from this activity from which they seem so eager to distract themselves.
Cars have long been dumb agents unable to communicate with other vehicles on the road or with the road itself. Likewise, they have been unable to communicate very well with their own drivers. The sole form of communication is the occasional "Check Engine" light that suddenly begins glowing and is often then disregarded by drivers willing to gamble that nothing important is truly wrong.
But consider the car's on-board diagnostics (OBD) port, a humble interface ignored by most drivers, through which mechanics access and assess a car's operating condition. Coursing through the port is a variety of very useful information, much of it related to the intricacies of engine performance; the OBD became standard on cars after California's strict emissions control legislation in the 1990s. Most of the time, the information is wasted. However, the recent emergence of other technologies, including GPS, wireless communication, and accelerometers, has made the OBD more interesting. Suddenly there was a method for someone—researcher, insurance company, fleet manager, parent—to discern not simply where but how a car was being driven, by tracking a host of variables, including excessive steering, hard braking, and idling. Companies like Green Road and Cartasite stepped in with solutions for clients with large fleets of vehicles, ranging from school buses to airport rental-car shuttles to delivery services, where a relatively minor improvement in fuel efficiency reduces costs significantly. Interestingly, it's not just fuel savings; studies by Green Road and others have shown a link between fuel-efficient driving and safe driving, in that the cars with the best mileage ratings also have the fewest injury and accident claims.
Here, technology provides driving-behavior feedback to the driver, something historically lacking. As feedback is crucial to the human learning process, these real-time systems are of special value to young drivers, and multiple trials involving in-car warnings and weekly progress reports show that keeping drivers "in the loop" improves their safety. However, feedback is not a neutral commodity, and care needs to go into its content and design; for example, Stanford University researcher Clifford Nass found that when male drivers were given real-time feedback by a female voice, their performance actually degraded.
These new data streams also have other potential applications. The engineers developing the technologies say they can, for instance, parse, from the data itself, whether a driver is talking on a cellphone. To take another notably vulnerable driving population—drivers above age 65—devices could be installed (perhaps as a condition of licensing) that would report patterns of erratic driving behavior, rather than waiting for a crash, which, for such older people, is more likely to result in injuries and death.
The combination of monitoring and predictive value (for example, more hard braking is associated with more crashes) is why insurance companies represent a primary, if still nascent, market for these technologies. Rather than grouping people into broad billing categories that may not reflect actual statistically informed risk differences, insurance companies are able to receive a more granular picture of individual drivers, not simply how much they drive ("usage-based insurance") but how they drive. This detailed information is useful for driving behavioral change. For example, the U.K. insurance company Norwich Union learned from a pilot study of instrumented vehicles (beginning in 2004) that young drivers are most likely to crash between the hours of 11 P.M. and 6 A.M., as well as on some road types more than others. So, under a novel pricing scheme it raised premiums during those hours; drivers received a monthly bill itemizing driving usage by time and road type. Following the new scheme, the number of claims made by drivers ages 18 to 23 dropped by one-third. As promising as this was, Norwich Union canceled the program in 2008 after failing to reach its desired customer base. Whether this was a marketing failure or consumers were uncomfortable with the technology or perceived privacy concerns is unclear. But it may not be the end of the issue, particularly, as The Daily Telegraph reported, approximately half of all U.K. insurers simply refuse to issue insurance for drivers under 21.
Knowing more about travel patterns is essential for a broader, and seemingly inevitable, transportation shift toward per-mile pricing, the so-called VMT tax (for vehicle miles traveled). As cars become more fuel efficient and shift entirely from the internal combustion engine toward hybrids, electrics, and other energy sources, a new funding mechanism will be required to pay for road construction and maintenance, both historically underfunded; for example, the U.S. federal gas tax has not been raised, not even to keep up with inflation, since 1993. The new funding system would tax distance traveled rather than fuel consumed, and, à la the Norwich Union insurance scheme, charge based on travel time. The Texas Department of Transportation, writing in a VMT primer published last December, said "Fuel taxes are largely a hidden cost to consumers and fail to send appropriate market signals to drivers, leading to overutilization of scarce roadway resources at peak periods of the day" (see http://utcm.tamu.edu/mbuf/).
The specter of these systems also raises privacy concerns. Americans are broadly of two minds about privacy on the road (a public space in which to even register one's vehicle is to yield some amount of personal privacy). License-plate tracking, transponders, and cameras are viewed as acceptable when used for collecting tolls or for supporting incident-management systems (to help manage traffic flow) but negatively when used as a safety- or traffic-enforcement tool. Even as some U.S. municipalities are removing nominally invasive technologies like speed cameras, the U.K. (despite its superior traffic-safety record) is pushing ahead with "average speed cameras" that, rather than encourage drivers to suddenly slow at a discrete tracking point, only to resume a higher speed, encourage adherence to the posted speed throughout the drivers' journey; U.K. studies predict that 70-mph cameras on motorways could reduce vehicle emissions by 1.4 million tons annually.
There is also an opening, and an irony, in the manufacturers' push to wire their cars so drivers are able to transmit via Twitter the most fleeting and intimate details of their lives. Most drivers are indeed already tracked, by virtue of the cellphones they carry. Technology being developed by engineers at the University of Minnesota uses a “Road User Charging” device (connected to the OBD port, linked via cellphone network) that would send a driver’s VMT information (possibly aggregated as a privacy measure) to a hypothetical billing facility. In theory, this would be no different from using a cellphone, which also provides more accurate locational information than GPS in dense city environments. Just swap miles for minutes. Drivers could get real-time updates on their usage behavior, bringing greater transparency and accuracy to the true cost of driving, in terms of both environmental impact and the externalities each individual driver imposes on all other drivers.
The connected car is long overdue—not as a vehicle for a tweet but as active agent in the social network that is traffic.
Tom Vanderbilt is author of Why We Drive the Way We Do (and What It Says About Us), Vintage Books, Vancouver, WA. He blogs at http://howwedrive.com and lives in Brooklyn, NY.
No entries found