Exploring Some Implications Of Driverless Cars

Driverless cars have been a popular and fertile topic for research and discussion for years. From the first high-profile Grand Challenge to the more recent work by Google, there has always been activity, though it’s rarely applicable to everyday life. But a few years from now we’ll really need to start pondering the potential effects of these robocars on cities, infrastructure, and markets.

ClockworkMod’s Koushik Dutta has written an interesting little post on Google+ about one potential major change that could come with the automation of vehicles. If a car can drive itself, that drastically increases its potential efficiency, and decreases the number of cars necessary per capita, especially in a city. You better believe the car companies don’t like the sound of that.

The comparison Dutta makes is to commercial fleets of planes, which spend a huge proportion of their “lives” in the air – i.e. in use. Private cars, on the other hand, sit idle 95% of the time. Naturally there are examples of the opposite in each case: private jets sitting in hangars and commercial car fleets being used constantly, but that isn’t really relevant to the comparison. With automation of cars, they could be put to use in much more ways and used collectively instead of independently.

There are already services that are exploring the potential of this model: Zipcar and Getaround, for instance, which attempt to maximize the utility of vehicles. But systematizing that and adding automatic navigation changes the game. What if you could drive to work and then send your car home to pick up your kids and bring them to school? Or if it wasn’t even your car, but one shared by five houses in your neighborhood, and after it dropped you off, it dropped off your neighbor, then took your spouse to the grocery store — and went to charge itself for half an hour while they shopped?

The implication is that if the average car is made even slightly more efficient, that results in a propotional decrease in the number of cars that need to be sold. Sure, there will be inefficiencies like empty cars (though taxis are in a way also empty much of the time) and individual vehicles will wear out faster owing to more constant usage. And, of course, many people will simply prefer to drive. But the amount of work a car can do per joule or hour or whatever could be increased, and if that utility ratio is improved enough, it starts having a serious effect on transportation.

And while at first the idea of driving cars around doesn’t seem like a Google thing to do, don’t forget that the most important part of this whole business, and the part that produces the most efficiency, is the logistics. Tracking the cars, locations, needs, routes, and so on — all information Google would love to sift. Google doesn’t care about the way the cars avoid obstacles – that’s an engineering challenge that researchers around the world are cracking. Google wants to power this network of nodes and be the unseen hand that points at this vehicle in this lot and tells it to go to this location by this route and pick up this person. Google’s forte is flattening deep data, and this would be a great application of it.

Not that I would trust Google to drive my car for me, exactly, at least not in Google Commute Beta, but I would certainly trust them to provide all the information my robocar needs to get where it’s going. And they’re jockeying for that position already, probably a decade before automated vehicles even start to be considered for road use.

That’s really the part of the equation that Google fits into. Where would Apple, Microsoft, IBM, Intel, Qualcomm, and everyone else figure? How will cities, and cars themselves, change? Like I said, it’s a fertile topic, and we’ll probably be talking a lot about these things in years to come.

Edit: Koushik does not work for Google, as I originally wrote (and still appears in the URL). Entirely my mistake, not entirely sure why I would even think that. He is an Android developer among other things and heads up ClockworkMod.

[via Hacker News]