Cars that Drive Themselves, Ctd - Momentum Builds

It comes as no surprise to me that after Google's quiet announcement of their work on driverless cars that the blogosphere would light up on the topic.
Doron Levin:

Cars that don’t need drivers also may not need private owners – since they could be summoned remotely and returned once their journey is complete. Why take on a lease if you can purchase a subscription to a car instead? Car owners who never want to spend a saturday under the hood or in the waiting room of a mechanic’s shop again might quickly adapt to a car subscription model.

 I have to say I side with Sean's dissent on this argument. Why would I want to climb into a car that someone else used (abused)? Taxis are gross enough, but at least some modicum of cleanliness is preserved by the driver or the dispatch. And the idea of not owning a car but instead relying on one to "arrive" via dispatch breaks down really quickly when you realize just as the car drives itself away at the end of your trip that you forgot eggs at the store and need to run back. It might be hours before you are queued up again in the list, and another car arrives.
That said, people in situations where they currently live a car-less life that involves either mass transit or taxis could be well served by driverless, subscription based cars. Also, there are probably a lot of taxicab drivers sweating about this; it seems like no big deal to set this up as a system where, for example, I could hop in a driverless car at the airport, punch in my hotel and it would take me there. Most taxis take debit and credit now, so no big deal, right?
Felix Salmon thinks electric vehicles will be boosted by this:

one of the big reasons why people are wary of electric cars is that every so often they want to take long car journeys which can’t be managed on a single charge. Up until now, the only solution to that problem is either to have a second, gasoline-based, car, or else to have a nationwide network of recharging stations which in any case are likely to take far too long to recharge the battery. Car subscriptions would be a much better solution. You use an electric car most of the time, and then when you need something with greater range, you just swap it out for one of those instead.

Yet many seem to think that the vehicle needs to retain some level of human control. Levin thinks that the vehicle will still need manual control for various activities. While I disagree with him on specifics (getting gas, getting service - these things could easily be done autonomously), I do agree that a steering wheel and brake pedal will need to remain. The reason for this is of course winter. Cars can have all the sensors in the world telling them where the road is, where other cars are, etc. But when there is 5 inches of snow on the road, it may prove difficult if not impossible to ever write software to tell a car how to navigate such things; humans are simply unsurpassable in their creativity and adaptability. Cue TAE:

It is reported in the BBC News that Spain is starting trials of a concept called Road Trains, where basically cars automatically draft each other until one car needs to pull off, at which point it drops out of the train. Cars can join up with the train at will, and should save up to 20% on gas by riding the slipstream of the car in front of them.

It seems to me that a likely future scenario is one in which inclement weather causes the car to require manual control on side streets and parking lots, but once reaching the cleared roads can take over for you.

Google, on the other hand, seems to be bending towards "safety" which is fine with TAE:

Wouldn't his tasks have been so much easier - and safer - if he hadn't been driving his car, but instead had been sitting in a car that drove itself? Talking on the phone while writing an email is multi-tasking most of us can handle easily. That is...easily if not driving a 2-ton machine at 60 mph.

That whole post is a diatribe about how much safer we'd all be if we didn't drive our cars. Automobile fatalities account for thousands upon thousands of Americans dead every year. And despite the fact that automobile fatalities are declining...going from 40,000 to 38,000 is still way too many dead each year. TAE has posted more than once that driverless cars are a great way to end drunk driving. If you are too drunk to drive...let the car do it instead. You can pass out in the seat and the car can wake you up when you get home. Or that idiot that simply MUST text while driving. How bout you text while not driving?
There is the other theory, though, that Google wants to free you up from driving so they can stick ads in your face or let you use wireless devices and the internet (where they rule the adsphere). TAE suggests this is a great idea, and not new:

I climb into my truck, and open up the console (where a steering wheel used to be). The console shows a map, my location, nearby landmarks, and includes options like satellite view, traffic updates (usually boring and uneventful), weather/road conditions, etc. I hit the "work" shortcut key and close the console. I sip some more coffee. My truck clicks into drive and pulls itself out of my parking space. I sit back and read my Kindle. I glance out the window at the sun coming up. "Radio" I say, then "AM 980", and the news comes on. I scan more headlines, as my truck smoothly navigates down roads towards work.
My reading is interrupted by a "ding ding ding..." "mute" I say to the radio, and I open the console. The truck informs me that a vehicle has broken down on my normal route, and it is automatically diverting me two blocks east. I clear the update and with an "unmute" go back to reading. The guy on the radio mentions the broken down vehicle, and informs commuters that their commute should be increased by an average of 8.3 seconds due to various diversions.

Perhaps the least discussed, but most advantageous reason to make driverless cars would be that congestion would be all but eliminated. While I have literally no idea where any of the other drivers on the road are going, a driverless network of cars could know where all the cars are going, and could streamline automobile flow by diverting traffic around broken down cars automatically, or by utilizing side streets during rush hour, or even converting two-way traffic roads into one way traffic roads. Why have a 6 lane highway be 3 lanes each way during rush hour? Many cities have commuter lanes that go in the rush direction...an automated network of driverless cars could easily just have a giant, superwide swath of concrete, and traffic flow could capture whatever percentage of that swath was necessary to create the most efficient traffic flow possible.

TAE thinks that driverless cars has become my pet issue, that is drowning out all others. I think that is probably because of all the technologies I dream of, driverless cars is the single most feasible. We literally have all the tech needed to implement such a thing, from GPS to in-car maps and navigation features, to proximity sensors, to robotic steering. Cars already can parallel park themselves with ease. in Europe they are experimenting with trains of cars that go driverless and draft a lead vehicle to save gas.

This is why Google is getting on board, now. The heavy lifting has been done. The technology is already out there. Google just needs to flex its muscles and integrate. Is it really so hard to believe that what Google is after is the last place free from visual advertising? Why let your local radio stations do a shoddy job of selling you stuff when Google could do it better? Does anyone really think Google doesn't see the incredible value of not only being able to stick ads in your face during your commute, but also be able to target those ads at you based on your GPS location? Americans spend on average 100 hours a year commuting. That's a lot of time to be looking at, and hopefully clicking on, ads.

Keep your eyes open for Android for cars.


_