Robocar impact on traffic congestion and capacity


Many people wonder whether robocars will just suffer the curse of regular cars, namely traffic congestion. They are concerned that while robocars might solve many problems of the automobile, in many cities there just isn't room for more roads. Can robocars address the problems of congestion and capacity? What about combined with ITS (Intelligent Transportation Systems) efforts to make roads smarter for human driven cars?

I think the answer is quite positive, for a number of different reasons. I have added a new Robocar essay:

Traffic Congestion and Capacity with Robocars

In short, a wide variety of factors (promotion of small, single passenger cars, ability to reverse streets during rush-hour, elimination of accidents and irrational congestion-fostering behaviour, shorter headways, metering of road usage and load balancing of roads and several others) could amount to a severalfold increase in the capacity of our roads, with minimal congestion. If you add the ability to do convoys, the increase can be 5 to 10 fold. (About 20-fold in theory.) The use of on-demand pooling into buses over congested sections allows a theoretical (though unlikely) 100-fold increase in highway capacity.

While these theoretical limits are unlikely, the important lesson is that once most of the cars on the roads are robotic, we have more than enough road capacity to handle our current needs and needs well into the future. In general, overcapacity causes building, so in time we'll start to use it up -- and have much larger cities, if we wish them -- but unlike today's roads which add capacity until they collapse from congestion, advanced metering can assure that no road accepts more vehicles than it can handle without major risk of congestion collapse.

Even before most cars are robotic, various smart-road efforts will work to improve capacity and traffic flow. The appearance of robotic safety systems in human driven cars will also reduce accidents and congestion along the way. Free market economist Robin Hanson believes the ability of cities to grow much larger will be one of the biggest consequences of robocar capacity improvements.


"Cities may double or triple in size, with far-reaching consequences." This point needs far more emphasis! This is the main reason to be really excited by all this.

One reason I don't predict a lot here is that I am not sure anybody has the urban planning chops to figure out where this goes. It's highly chaotic. It is a big change, but there are other pressures that robocars bring, including the ability to tolerate longer commutes (because they are productive work or leisure time) the ability to travel with a fleet of RVs that follows you around and more. These pressures might bend the city in other ways.

People like low density natural settings with safety and play space for their children. They also crave easy access to resources, interesting places and social interactions. These forces will push against one another. If you can live on a farm but can quickly (or not so quickly but comfortably) find yourself in a walkable, engaging social and commercial space, you might choose that, making the city vanish. Cities exist because of social desires and because they are cheaper due to economies of scale. Robots and telepresence and delivery-on-demand change many of these economies of scale in ways we are just learning to understand.

The Java link - - seems to be forbidden.

He moved those pages and has a number of new papers and videos up. I have modified the links

I agree. If computers were doing the routing, just think of how efficient we could make our traffic lights. Today, when the light turns green, you have to wait for all the cars in front of you to move, creating a wave, or accordion like traffic pattern. If computers are controlling intersections, all the waiting cars can be moved at the same time. The computer also knows how many cars are approaching an intersection, allowing cross traffic to move through openings.

There are going to be human driven cars on the road for many decades to come, and always some holdouts. Before the last holdout leaves the road, we may start seeing proposals that require human driven vehicles to at least have some computer networking equipment in them (even a mobile phone in a mount) to let them participate in virtual traffic signals, reversible streets and other ITS tools. Traffic lights are actually very expensive, even 4 way stops cost money.

Reversible streets already exist of course but today they are complex and expensive. It's mostly done on highways and bridges and involves remote controlled gates, or moving cones, or even movable dividers. On some bridges it's just frequent signal lights over each lane to tell if that lane is in your direction. While there are human drivers, each removable street segment would need to have a row of lights over each entrance, which is both costly and ugly, though worth the cost in the most congested areas. However, an app for the human driver that knows which streets reverse, gets positive confirmation and makes strong warnings if a driver appears to be about to headed down a street the wrong way could be cheap if it were mandated. Such a box need not be that expensive, and CBDs that are ready to do things like congestion charging could start insisting each driver have such a box -- which might also manage the tolls.

Seems to me that, instead of robocars, you could just institute a system where each car has a GPS-based speedometer and a forward-looking rangefinder. Both of these continuously report vehicle speed and separation from the preceding vehicle.

For every minute you spend less than 3 seconds' distance behind the vehicle ahead of you, you pay a dollar, starting after the first minute.

For every minute you spend driving less than 90% of the posted speed limit with no vehicle ahead of you, you pay a dollar, starting after the first two minutes.

Make "congestion pricing" a reality, instead of just a codeword for "gouge commuters".

This is a bit big brotherish. People don't want devices tracking their speed and location, though you could do some work to try to make this a bit more anonymous.

I don't think it makes sense to charge people for being within 3 seconds of the next car. Cars on city streets actually are meant to bunch due to traffic lights. 2 seconds is the official recommended headway, but most people drive with less.

Your two rules also conflict, in that if the vehicle ahead of you is going slow (for valid or invalid reasons) then you are charged if you get too close to it, but also charged if you hang off a distance behind at the slow speed. While I can see ways to fix that, I am not sure these are the real measures of congestion.

Congestion is mostly more cars on a road than it can handle, combined with triggers which collapse to stop and go, such as accidents, merges (even if volume is low enough for the resulting smaller road) and irrational slowdowns. When volume is well below capacity, irrational slowdowns don't cause too much trouble. Accidents still do as people slow to look -- and when the accident is fresh and still in the the road they slow down with good reason.

A key problem though is whatever system you design to keep the flow regulated in the congested zone, you may just move the congestion to a zone outside of it, as people wait to get in. Thus you need wait-at-home I think. Metering lights use the on-ramps for waiting, but sometimes that fails as the line spills out into other traffic. Congestion charging doesn't make people wait, but it's been known to bump congestion just outside the zone as people skirt around it to avoid the cost of going in it.

Based on trends with on-demand rideshare like Uber Pool, Lyft Line and Via, it seems that many riders would prefer sharing a ride if the cost is lower and the drive times are comparable. Therefore, autonomous vehicles may come in all sizes depending on what's the most efficient. It may be more efficient to run bigger vehicles on dedicated routes during peak times (perhaps a 40-bus, or a shuttle, or a larger capacity car), and mid-sized or ultra-small cars during off-peak times. The larger vehicles would work more like transit, especially in busy urban areas. For instance, Via routes are dynamically coordinated for efficiencies, so riders might be required to walk a block or two to save time on the overall trip.

"Robocar impact on traffic congestion and capacity"

The Marchetti constant (the average time spent by a person for commuting each day ; approx 1 hour) has proven to be difficult to beat due to the induced demand every improvement seems to create. Robocars promise many improvements to traffic flow, density and perhaps extra lanes. As a consequence and over time people will almost certainly start applying for jobs etc.. further from their homes until travel times again become too much and this frustrating equilibrium is re-established. Robocars could actually make it worse by reducing longer distance travel costs while allowing better quality time than simply driving.

On the assumption that mobility on demand services become the dominant I see Robocars and the associated pricing system as enablers for effective road pricing. It is effective use of road pricing. and there are many different options, that ultimately could have the greater impact on traffic congestion and capacity. While it is sure to be controversial, a pricing system within cities that subsidises shorter distance travel at the expense of long commutes could help. Something as simple as requiring all service companies to 50% discount say the first 10 miles of travel after flag fall fee could work. In a competitive market this discount would be recouped from longer distance commutes until the profit margin was restored. It may also be an advantage to charge a higher rate on empty vehicles to encourage greater efficiency.

There are many real costs that are spread over the general community (congestion, pollution, and infrastructure) which discourages greater efficiency. This has been due to the great difficulty of measuring and charging them back to the individual. Variable road pricing, as opposed to the crude area based systems used in London and Singapore, offers a way of locking in the efficiency dividends that Robocars offer. The alternative is that after a couple of decades the Marchetti constant will remain unchanged, as it has since Neolithic times.

Yes, the hour commute has been around as long as commuting, but that's not actually all that long, at least at volume.

Here are some factors which could change it:

  • The commuter robocar can be a comfortable workspace, with full communications, screen and desktop. You will be at work when you start. The only downside, the little bumps and motions. New suspensions might alter that.
  • It can also be a sleeping space, for those who find a nap as useful as regular sleep. Trips can take arbitrary times for those who sleep well in the cars and no human time cost, particularly overnight trips.
  • Travel can be quite cheap in smaller, single person vehicles. Cheaper for the individual, and cheaper in externalities as well.
  • Induced demand could be eliminated through road metering, but in truth, the extra congestion of induced demand is a lesser cost if the above factors apply.
  • In some cities, the middle class are already priced out of homes. Even today we see people taking 1 hour one way commutes.

On the other hand, time is still time, unless you can productively sleep. But today we see rich techies happy to take one hour commutes on things like the Google bus not to live somewhere cheap but to live somewhere fun and expensive. It provides just a taste of the future.

The factors you have listed look more likely to increase induced demand due to more relaxing journeys and the freedom to use the time for other activities. Also should not underestimate the possibility that longer distance travel could become relatively cheaper due to EVs lower running costs despite their initial higher capital cost.

Running parallel to the development of Robocars is the buildup of carbon in the atmosphere and what actions over the coming decades will be taken to mitigate it. My guess is that policies to reduce emissions will become much stronger and there are now proposed bans being placed on internal combustion engines in many European countries. The technology that allows companies to charge for shared mobility services could also enable efficient road pricing. This could also be used to encourage shorter commutes as another method for reducing emissions - and congestion. It would require government policies to drive it and again European countries rather than the US would be more likely to take such action. At present the case for Robocars seems to be based fairly heavily on improving safety but I wonder if environmental factors are a sleeper issue.

I've been talking mostly about that for 10 years.

Now, if the robocars are electrified, and the electric grid switches to mostly renewables, then the transport does not product many emissions. We are some distance from the latter but there are signs of both.

Add new comment