The automobile defined 20th century America, and that momentum carried over into the 21st little abated. I live in rural western Sonoma County, and I’d be hard pressed to live there sans automobile. The nearest 7-11 is three miles away, the Safeway where we shop is four miles distant, and the feed store (Frizelle-Enos), which keeps our horses fed, is farther still. Sebastopol might as well rename itself “Prius Town,” but it still has cars, even if they’re environmentally friendly. (And, yes, it’s unclear whether burning gasoline or making lithium batteries is worse for the environment. It’s the automotive equivalent of cloth vs. disposable diapers.)
My new job, on the other hand, is located in the gray industrial wasteland of Fremont, along the eastern edge of San Francisco Bay. Nowhere is the dominion of the automobile more apparent. I’ve taken to walking to work from my nearby apartment, and it takes me alongside the 880 freeway, which is abuzz with cars nearly every hour of the day. The lack of sidewalks in much of the industrial park where AlterG makes its home emphasizes how important the car is relative to people and the healthy benefits of walking. People work inside blank industrial buildings, arriving and departing on asphalt paths built solely for the convenience of cars.
On Mondays and Fridays, I drive between Sebastopol and Fremont. It’s about an hour and a half when traffic is light and close to three hours on the Friday of a three-day weekend. Plenty of time to observe the problems that people create for themselves when driving: slowdowns with no apparent cause, drivers who don’t seem to understand their communal obligations, rubber-necking, young men weaving in and out of traffic in a effort to gain a few car lengths, only to arrive at the same traffic light that stops everyone.
Being of a technical bent, of course, I spend some of my driving time thinking about ways in which technology might improve the state of traffic. If you drive, as I do, making heavy use of cruise control, it’s easy to see that people have trouble maintaining a constant speed, even when traffic isn’t crowded. What would really be great is a system that optimizes traffic flow regardless of who’s sitting behind the wheel. Of course, I remember reading Popular Mechanics back in grade school, which promised that such things (like my elusive jet pack for commuting to work) would soon be reality. It’s harder than it looks.
Like other problems, you can try to build a system like this from the top down, with centralized control, or from the bottom up, where each automobile operates with relative autonomy. It’s not clear what Google’s intention is, but it’s made public some information about its efforts to develop a driverless car, most notably via a TED talk from Sebastian Thrun (if you’d like to watch it, Google “TED driverless car”). Thrun leads the development team at Google, which draws heavily from a Stanford team that won the Defense Advanced Research Projects Agency (DARPA) Grand Challenge in 2005. (The Grand Challenge is a Department of Defense competition to develop autonomous vehicles, something of which the DoD could obviously make good use in a variety of military scenarios.)
Driverless vehicles rely on a number of different technologies: GPS to establish where they are, radar and radar-like technologies such as LIDAR to locate surrounding vehicles and obstacles, and pattern recognition software to identify road markings, to name a few. In the Google fleet (including six Priuses, an Audi TT and a hybrid Lexus), there are humans in the driver and passenger seats. The driver can disengage the system by pressing on the brake, much like a cruise control. The passenger monitors the systems (obviously, they’re Google engineers).
Lots of interesting questions arise when a vehicle no longer obeys the assumption that a human is always in control. The obvious question: If an autonomous vehicle is speeding (perhaps because its database was out of date, since presumably such vehicles would be programmed to strictly obey traffic laws), who gets the ticket? But there are even more interesting questions. For example, what if a driverless vehicle, sensing the speed and trajectory of nearby vehicles, determined that a fatal collision was imminent and unavoidable? What if the car also “knew” it could save the lives of several others at the expense of the life of its “driver”? Surprisingly, this isn’t a new question for philosophers, who’ve examined the so-called “Trolley Problem” in depth (should a trolley driver do nothing and hit five men working on the track directly ahead, or turn the vehicle onto a siding where only one man is working, killing him?). Wikipedia has a fascinating article. The answer isn’t always clear.
For now, the Google driverless car is loose on the streets of California and Nevada (which passed special legislation to permit driverless vehicles to operate in their state). Like many Google announcements (such as Google “Project Glass” glasses, discussed last month in this column), the hype may exceed the reality. The car has logged many more miles where human intervention was required than miles where it was completely autonomous. Nevertheless, it’s an impressive accomplishment, particularly since no accidents have been recorded while the cars have been driving themselves. Ironically, one of the cars had a minor fender bender while under human control.
It seems pretty clear that in less than a decade, the technological challenges of driverless cars will have been solved. Of course, that’s only part of the problem, since each car needs information about what’s happening up ahead on the road. There’s also a well-known systems engineering rule, stating that optimizing the entire system generally requires some of its components to operate at less than full efficiency. In other words, to minimize the total travel time for everyone, each individual’s trip might take slightly longer.
Still, I’d be quite happy if my car could handle the drudgery of commuting. Would you? Let me know at mduffy@northbaybiz.com. And be careful out there!
Mike Duffy is chief software architect for AlterG, creators of the Anti-Gravity Treadmill. He’s been the technology columnist at NorthBay biz for more than a decade. You can reach him at mduffy@northbaybiz.com.
