at the mercy of the autobots
Tim's article about the social changes that self-driving cars might make possible is fascinating reading (so is his earlier installment of the series, in which Tim discusses the history and present state of robotic car technology). But I wonder about the conclusion that Ryan draws from it:
I think the net effect of autonomous vehicles on suburban areas would be to make them denser (I think the net effect of a lot of things on suburban areas will be to make them denser). As important as those shifts, I suspect, would be the boon such vehicles would be to explicitly urban areas.
Taxis and car-sharing services are very much complements to walkable urban life. Cheap and effective autonomous vehicles would likewise be complements to urban life, but in a much more significant manner. It’s hard to overstate the negative effect of parking on dense, urban areas. When fighting against dense, new development, NIMBYs cite parking and traffic concerns first, second, and third. Given the value of urban land, parking has massive opportunity costs. Street space in urban areas is very limited and is currently given over almost entirely to driving lanes and parking lanes. If self-driving vehicles freed up much of that space, it would make room for large increases in transit right-of-way, bike lanes, and sidewalk space.
I think the less parking/more density argument being put forth is a bit optimistic. Obviously self-driving cars would be phased into the fleet on the road slowly. So long as that's the case, it's hard to see how self-rearranging, super-dense parking lots can be made to work. We'll have to wait for the entire fleet to turn over — a process that, given the steadily increasing quality of modern cars, will take longer and longer as we wait for robot cars to arrive. I can see how a car that's easier to share — or even just one that can drop you off and then go its merry way — could help reduce the amount of parking we need and build. But it doesn't seem likely to make a significant difference during my lifetime.
The more immediate consequence is this, though: if automated carsharing drives down the cost of taking a trip in a private vehicle (you don't have to pay for the full upkeep or purchase of the vehicle, or pay for a driver's time); and if not having to spend time piloting the car drives down the personal cost even further, won't that encourage more trips in private cars? And won't that actually encourage suburban settlement patterns?
Don't get me wrong: I think self-piloting cars will be a boon to everyone. It's just hard for me to see how easier locomotion from arbitrary point to arbitrary point will push people toward the dense settlement patterns that Ryan favors. There's a widespread perception that city schools are bad and that having a yard is nice. I really like living in a city, but I'm not sure that either of those ideas is wrong. If you tell people they can have it all — and better, that it'll all be delivered BY ROBOTS — they'll take you up on the offer.
Comments
So long as that's the case, it's hard to see how self-rearranging, super-dense parking lots can be made to work.
This doesn't seem too hard to imagine. The easiest way is to just have parking lots with a human-driver section and a robocar section, with the robocar section having smaller, more densely-parked parking spaces. But even that shouldn't really be necessary. If self-driving cars have a way to signal their robot-ness to other cars, they can automatically cluster in higher density when they find themselves adjacent to one another.
But I actually think higher parking densities is a lot less important than the shift from ownership to taxi rental. This is a positive development for three reasons. Most obviously, taxis are shared so the total number of vehicles in circulation goes down. Second, taxis spend a lot more time on the road so they don't need parking very much. Third, taxis don't need to park anywhere in particular, so when they do have to park (which will be mostly in the middle of the night) they can go where the parking spaces are, which means much more efficient utilization of existing parking spaces.
Privately-owned cars require several parking spaces per car, because people expect a parking space at home, a parking space at work, a parking space at the mall, etc. Outside of the densest cities, today's parking spaces sit empty the vast majority of the time. In contrast, taxis require fewer than one parking space per taxi, because at any one time significant numbers of taxis are out and about.
Finally, something I didn't talk about too much in the article but I think is probably true is that I'm not sure that the concept of "parking" really makes sense in a self-driving world. Right now, most urban roads have one lane in each dierction (we don't think of them as lanes, but they are) for parking and 1-3 lanes in each direction for traffic. In a self-driving road, one can imagine a dynamic process where the mix of parking lanes and driving lanes gets adjusted gradually over the course of the day. On a six-lane road, all six lanes might be devoted to traffic during rush hour, while in the middle of the night 4 of the lanes might have cars "parked" in them (two rows of car double-parked in each direction). The total number of cars in circulation doesn't change day-to-day, so if most of the cars are on the roads during rush hour, it follows that there should be enough spare capacity that they can all "park" on the roads in the middle of the night.
This is obviously highly skeptical, but I think there's clearly more going on than the mere fact that self-driving cars would pack parking lots more densely.
You're right, of course -- Ryan was also kind enough to reply to this post, and in comments at his place I had to quickly admit that I was looking at the parking situation in too simplistic a manner. I still think that it's an open question as to whether automated cars will positively impact the overall question of density, but I'm now convinced that they'll at least let us spend less land on parking lots.
One more thought I'll add: it will be interesting to see how such a system will be implemented securely. Once cars begin to be able to impel each other to move around some pretty insidious exploits could arise. It should certainly be possible to engineer solutions to this sort of problem, but it's still a pretty interesting technical problem (as are the parking lot-reordering algorithms themselves).