Self-Driving Cars and Road Design

A fresh insight hit me as I was dozing off. I realized that there are really two ways self-driving cars will affect road design. This is because robocars will have very different needs and safety margins than human drivers; robocar roads with only occasional human drivers will therefore be designed around the robocars instead.

The first is that self-driving cars don’t need such wide roads. Geometrically, the effective space of an automobile at speed is governed by its driver’s reaction time. Human reaction time might be split-second, but it isn’t infinitesimal; that time translates into movement from the decision to hit the brakes being made and the pedal being pressed down. This stopping space varies based on speed being traveled, but it’s generally 2x to 4x the actual car length.

By contrast, a robocar’s reaction time is infinitesimal. What this means is that the minimal safe stopping space is instead governed by momentum. Robocars are also able to “platoon”, traversing the road with a high degree of physical precision — so much so that they need very little dedicated stopping space. The upshot of this is that the capacity of nearly every road balloons from double to four times its current capacity — far more capacity than can ever be justified.

The greatest impacts will be felt in urban and suburban environments, as most four- or five-lane roads find themselves with two lanes too many. Lanes that robocars don’t need are pointless to maintain; in many cases, this leads to narrower carriageways. Former carriageway and shoulder space can then be reclaimed for other purposes — busways, expansions of the bike network, light rail, and in some cases just plain old wider sidewalks.

Second is a proliferation of roundabouts. Robocars don’t have the hangups about them human drivers have; instead, roundabouts serve to increase vehicular throughput while saving space. That said, this is more likely to happen in the suburbs; urban environments often have development up to all four corners.

Third is designing to the speed limit. Robocars will have speed limit data programmed in and won’t ever exceed it (unless the vehicle is hacked, which violates its warranty and most likely revokes its insurance as well), which means that highway designers ca. 2050 will be designing their roads much more strictly to the speed limit, as there is no “need” to accommodate speeders.

And fourth, we have replacement of parallel-parking lanes with loading lanes. For a robocar, there is no need to be stored where its passenger is dropped off. Indeed, instead many will circulate in a fleet not unlike car or bike sharing; the excess that aren’t needed outside of peak periods will likely be sequestered in large, centralized garages, perhaps owned and maintained by the service providers.

A major upshot of this is that a lot of existing automobile storage space — especially in urban centers — is unnecessary, even moreso than today, and will become redeveloped for various other purposes. In a lot of cases, aboveground parking decks might get stripped to their structural frames and redeveloped as office or residential floors; in other cases, storage space is integrated into parking decks (particularly underground ones); in others, loading space might get moved; and so on. There will still need to be some private car storage space, of course, but with a robocar you can just store it at the cheapest place in town, rather than the closest. It’ll drive itself there … and come back to pick you up.

In the suburbs, private car ownership remains universal at the dawn of the age of the robocar, but as the excessive capacity becomes obvious, large suburban bike networks begin to develop — surprisingly rapidly, once a tipping point is passed. This in turn fuels a major reduction in car ownership as bikes replace them for local errands (particularly as fuel costs rise). About a generation later, it’s fairly likely robocars will be used for longer trips, and biking is the preferred mode for short ones.

The most difficult element to assess is the impact on city streets. Some, like the reduced need for capacity, the change to loading zones, and the strict speed limit, are positive; in other ways, robocars are the ultimate “road” invention. That said, it’s probable they can navigate a slow-speed city environment better than human drivers ever could.

Since less robocars are needed, and since, for the most part, they don’t need to be stored on city streets, there’ll be a stronger movement towards livable streets. Shared spaces become more common — first on an ad hoc basis, and then more formally. Blocks and neighborhoods endeavor to beautify their spaces, first incrementally with cute parklets and benches in the parking spots the block doesn’t need anymore and then with more and more parklike streets, often with permeable pavers and stormwater mitigation. American cities start to look beautiful again.

All of that said, it’s important to recognize that private automobiles are the least spatially-efficient form of transportation. Once you give up control (which the insurance premium will demand), their primary luxuries will be ones where you don’t share the climate-controlled cab with anyone, or need to wait on anybody else’s schedule. Robocars will alter the status quo — that much is a given. Some of the ways in which it alters it will be unpredictable. Others will be surprising. One that will be most surprising is that it leads to an expansion of mass transit.

On that, I will write more later.


2 thoughts on “Self-Driving Cars and Road Design

  1. Here’s an alternative outcome: human behaviour shifts in response to streets dominated by robo-cars. People on bikes will happily cruise down the middle of the lane at 15 kmh. People on foot will cross where they like, without bothering to look, safe in the knowledge that the robo-cars have seen them and planned for them even before they left the curb. Parents will let their children play near streets again. Bored teenagers will amuse themselves by pretending to dart into traffic just to watch the robo-cars all screetch to a stop.

    The end result will be that robo-cars will be slowed to approximately walking speed in order to provide a safe enough environment for the many people around them that aren’t protected by their own robo-car. For robo-cars to achieve anything they will need to be in places that explicitly banish people; freeways, and perhaps in the near-future suburban arterials. But urban streets will be their downfall because people.


    1. What you are describing is no different than the city street and main street environment that exists today in places with a lot of pedestrian traffic. Note the keywords here: “street” and “pedestrian traffic”.

      As an example, I feel safe crossing South Street in Philadelphia in the middle of the block, because nobody can go fast. Bikers also bike down the middle of such streets, though they tell me it’s a suboptimal solution for them. Cars are big things, after all, and dedicated infrastructure feels safer for cyclists.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s