Cars —

Death by GPS

Why do we follow digital maps into dodgy places?

The Jarbidge Mountains of Nevada, where the Chretiens got stranded.
Enlarge / The Jarbidge Mountains of Nevada, where the Chretiens got stranded.

One early morning in March 2011, Albert Chretien and his wife, Rita, loaded their Chevrolet Astro van and drove away from their home in Penticton, British Columbia. Their destination was Las Vegas, where Albert planned to attend a trade show. They crossed the border and, somewhere in northern Oregon, they picked up Interstate 84.

The straightest route would be to take I-84 to Twin Falls, Idaho, near the Nevada border, and then follow US Route 93 all the way to Vegas. Although US 93 would take them through Jackpot, Nevada, the town near the Idaho state line where they planned to spend the first night, they looked at a roadmap and decided to exit I-84 before that junction. They would choose a scenic road less traveled, Idaho State Highway 51, which heads due south away from the I-84 corridor, crossing the border several miles to the west. The Chretiens figured there had to be a turnoff from Idaho 51 that would lead them east to US 93.

Albert and Rita had known each other since high school. During their thirty-eight years of marriage, they had rarely been apart. They even worked together, managing their own small excavation business. A few days before the trip, Albert had purchased a Magellan GPS unit for the van. They had not yet asked it for directions, but their plan wasn’t panning out. As the day went on and the shadows grew longer, they were not finding an eastward passage. They decided it was time to consult the Magellan. Checking their roadmap, they determined the nearest town was Mountain City, Nevada, so they entered it as the destination into their GPS unit. The directions led them onto a small dirt road near an Idaho ghost town and eventually to a confusing three-way crossroads. They chose the one that seemed to point in the direction they wanted to go. And here their troubles began.

If Albert had been navigating the route in the daytime, he might have noticed that it was taking them through the high desert as it rose toward shimmering snowy peaks in the distance. In the dark, the changes were so subtle that they barely registered. And besides, he was on a road—“a pretty good road,” the Elko County sheriff would later say, that “slowly goes bad.” Through the night, it carried them higher into the Jarbidge Mountains, deeper into the backcountry. The road twisted, dipped, rose again, skirting canyons walled with sagebrush. It was the kind of terrain for which the Chretiens’ van was not designed.

Several days passed before their family and friends realized that Albert and Rita had never arrived at the trade show. The couple had not informed anyone of their detour, so nobody knew where to look for them. The manhunt involved police agencies in four states, scouring 3,000 miles of highway, with the most intense efforts in eastern Oregon, where they had used a credit card in a convenience store. On April 8, just shy of three weeks since Albert Chretien left Highway 51, authorities announced they were scaling back search and rescue efforts, a tacit admission that wherever the Chretiens had gone, it was too late to find them.

***

What happened to the Chretiens is so common in some places that it has a name. The park rangers at Death Valley National Park in California call it “death by GPS.” It describes what happens when your GPS fails you, not by being wrong, exactly, but often by being too right. It does such a good job of computing the most direct route from Point A to Point B that it takes you down roads which barely exist, or were used at one time and abandoned, or are not suitable for your car, or which require all kinds of local knowledge that would make you aware that making that turn is bad news.

This feature is a book excerpt from the forthcoming <em>Pinpoint</em> by Greg Milner.
Enlarge / This feature is a book excerpt from the forthcoming Pinpoint by Greg Milner.

Death Valley’s vast arid landscape and temperature extremes make it a particularly dangerous place to rely on GPS. In the summer of 2009, Alicia Sanchez, a twenty-eight-year-old nurse, was driving through the park with her six-year-old son, Carlos, when her GPS directed her onto a vaguely defined road that she followed for 20 miles, unaware that it had no outlet. A week later, a ranger discovered Sanchez’s Jeep, buried in sand up to its axles, with sos spelled out in medical tape on the windshield. “She came running toward me and collapsed in my arms,” the ranger wrote in a report. “Her lips were very dry and chapped with bleeding blisters and her tongue appeared to be swollen with very little saliva formation. I walked over to the Jeep and looked inside. I saw a boy slumped in the front seat with obvious signs of death.” Mother and son had wandered over ten miles of desert in search of water, and had resorted to drinking their urine. They had tried to share a Pop-Tart a few days earlier, but their mouths were too dry to swallow. As he lay dying, Carlos grew delirious, telling his mother he was “speaking to my grandfather in heaven.”

Most death-by-GPS incidents do not involve actual deaths—or even serious injuries. They are accidents or accidental journeys brought about by an uncritical acceptance of turn-by-turn commands: the Japanese tourists in Australia who drove their car into the ocean while attempting to reach North Stradbroke Island from the mainland; the man who drove his BMW down a narrow path in a village in Yorkshire, England, and nearly over a cliff; the woman in Bellevue, Washington, who drove her car into a lake that their GPS said was a road; the Swedish couple who asked GPS to guide them to the Mediterranean island of Capri, but instead arrived at the Italian industrial town of Carpi; the elderly woman in Belgium who tried to use GPS to guide her to her home, 90 miles away, but instead drove hundreds of miles to Zagreb, only realizing her mistake when she noticed the street signs were in Croatian.

These types of mishaps often elicit sheer bafflement. The local Italian tourist official noted that although “Capri is an island,” the unfortunate Swedes “did not even wonder why they didn’t cross any bridge or take any boat;” the first responders in Bellevue were amazed that the women “wouldn’t question driving into a puddle that doesn’t seem to end.” For their part, the victims often couch their experiences in language that attributes to GPS a peculiar sort of agency. GPS “told us we could drive down there,” one of the Japanese tourists explained. “It kept saying it would navigate us a road.” The BMW driver echoed these words, almost verbatim: “It kept insisting the path was a road.”

Something is happening to us. Anyone who has driven a car through an unfamiliar place can attest to how easy it is to let GPS do all the work. We have come to depend on GPS, a technology that, in theory, makes it impossible to get lost. Not only are we still getting lost, we may actually be losing a part of ourselves.

***

The modern era of GPS car navigation began after Operation Desert Storm in 1990, and no GPS startup exploited its potential better than a Kansas City-area company called Garmin. Gary Burrell and Min Kao were engineers at AlliedSignal helping to develop a GPS receiver. (A prototype was on board the Rutan 76 Voyager, the record-setting aircraft that in 1986 became the first to circumnavigate the world with no stops or refueling.) After the project was discontinued, Burrell and Kao formed a new firm in 1989, its name a portmanteau of their given names.

Garmin was able to design components that made the most of limited processing power, especially in the acquisition of satellite signals. Even at the end of the nineties, GPS devices designed for cars required the user to tediously download maps. Over the next few years, map data made directions more accurate, better processors made it easier and quicker for algorithms to compute turn-by-turn directions, and memory improvements finally put all the maps in the box. Garmin’s C550 receivers, which hit the market in 2006, achieved full functionality. “Once solid-state memory had enough density to hold the entire country out of the box, that’s when the market really took off,” says Jay Dee Krull, one of Garmin’s earliest hires.

By 2006, Garmin controlled 60 percent of the US market for navigation equipment. Americans bought five million Garmin GPS receivers that year, as the company posted $1.68 billion in sales, a 64 percent increase from 2005. Fully half of the company’s revenue came from car GPS units, with sales in that segment growing at an astounding 140 percent annually.

As Garmin enjoyed spectacular growth, the popularity of GPS navigation led to renewed interest in how these devices were affecting the behavior of users. But instead of just investigating navigation systems’ design and effectiveness, some experts confronted the question of whether these new navigation systems might be weakening our cognitive map. Some of the widely praised attributes of navigation devices, especially their ability to present information in smoothly filtered ways that removed us from the bother of map-reading, came under closer scrutiny.

A 2006 German study, conducted by a group of psychologists and artificial intelligence experts, tested the hypothesis that users of a navigation system will remember less about an environment than those who use a map. Participants in the experiment walked a predetermined route through the zoo in the town of Saarbrücken carrying handheld computers connected by Bluetooth to a computer carried by an experimenter who followed several meters behind. He transmitted directions to the next segment of the route, beginning with a photo of the subject’s current location. In addition to the directions, one group of subjects also received visual cues, such as a map and red line that appeared on the photo; others received only verbal instructions; and a third group received a combination of verbal and visual cues. A control group walked the route guided only by a crude map that showed the route with no landmarks, along with the same photos shown to the test subjects at the start of each segment.

Later, all participants were given a two-part test to gauge how well they remembered the route. They were asked to recall what directions they had taken—at what points they had turned left, right, and so on. The second part tested what the researchers called survey knowledge: “the spatial relationships between locations.” The participants were shown thumbnail pictures of the intersections, and instructed to place them on a road map of the zoo. The researchers were essentially testing participants’ ability to construct a cognitive strip map and a cognitive comprehensive map.

The data revealed a couple of interesting insights. First, while it seems obvious why map users would have better overall knowledge of the area they walked, why would they also have a better memory of the route itself? The researchers reasoned that the map users had to engage in more active learning, having to match the photos they were given of the route with the markings on the map. The researchers had predicted—wrongly, as it turned out—that the two test groups who received visual aids that provided spatial context would outscore the verbal-only group in the survey knowledge test. Instead, they concluded that because this information was not required for the act of wayfinding, the subjects were not forced to actively process it.

A similar study, conducted in Japan and published in 2008, involved actual GPS devices. Three groups of walkers were studied as they navigated routes in the city of Kashiwa. One group learned the route from direct experience, shown it by a guide who took them from start to finish, and then back via a circuitous route; the subjects were then instructed to walk from start to finish again, with no assistance. Another group was given GPS devices, with the complete route highlighted on the screen, while another was given a paper map with the beginning and end points marked, but no highlighted route. The results showed the GPS users exhibiting the weakest wayfinding acumen. They traveled at a slower speed and made more stops to reorient themselves than the walkers in the other two groups. They rated the overall task as more difficult than the group that learned the route by walking it. In post-walk tests, they had the lowest scores on memory of the configuration and topology of the route. The researchers concluded that the GPS system “was less effective than the maps and direct experience as support for smooth navigation.”

A Cornell University study published the same year looked at GPS’s effect on drivers, and reached similar conclusions regarding how GPS users “attend to objects in the paths they take toward their destination.” The study “found evidence for loss of environmental engagement . . . the process of interpreting the world, adding value to it, and turning space into place is reduced to a certain extent and drivers remain detached from the indifferent environments that surround them.” Their conclusion: “GPS eliminated much of the need to pay attention.”

Channel Ars Technica