Why Autonomous Vehicles Won’t Solve Traffic

Many people who are excited about autonomous vehicles frequently cite a reduction of traffic as a bonus they expect to see when humans are no longer driving every car on the road. In many cases, though, autonomous vehicles may actually increase traffic instead of reducing it. Here’s why.

Different Kinds of Traffic

First, it’s important to distinguish between different types and causes of traffic on roads. While traffic science is surprisingly complex, most non-construction traffic that people encounter on a day-to-day basis can be divided into two categories.

The Slinky Effect

“Slinky Effect” Traffic is when there is a significant slowdown on a road for seemingly no reason that clears abruptly when you get to the front of the slowdown. This type of traffic is usually caused by someone rapidly braking or changing lanes, causing a chain reaction behind them that can slow traffic far away from the original cause.

Animated Slinky Effect Example from CGP Grey

This type of traffic can indeed be prevented and even solved to a large degree by autonomous vehicles, but only if human drivers actually cooperate. The best way to prevent slinky effect traffic is to leave a safe amount of following distance between you and the car in front of you, which autonomous vehicles generally attempt to do. However, many human drivers opt to take advantage of these safe distance gaps to change lanes, forcing autonomous vehicles to slow down to regain a safe distance, which can cause the very thing this strategy attempts to prevent. Once a slinky effect has already happened, the best way to fix it is to smooth traffic flow by slowing down all cars well in advance of the blockage, giving the cars ahead time to advance through the slowdown. But as any truck driver knows, this strategy is a one-way ticket to angry drivers cutting in front of you to seemingly get to stopped traffic as fast as they can.

Effectively, fixing the slinky effect can only occur when a majority of vehicles on the road are autonomous, which won’t happen for a long time. And more importantly, slinky effect traffic is not the most common type.

Gridlock Traffic

Gridlock Traffic is what millions of commuters experience every day at rush hour. It occurs when there are more cars attempting to use a road than the road was designed to handle. Once enough cars are on a road, trying to add more forces traffic everywhere to slow down, and once it does slow down it can’t speed up again until the number of cars on the road drops significantly. The problem with gridlock traffic is that the only way to fix it is to either add more road capacity, usually by constructing more roads or widening existing ones, or to decrease the number of cars on the road. Interestingly, adding more road capacity doesn’t always improve gridlock traffic since the number of people using the road will often increase to fill the newly available capacity (people switch away from using public transit or change commuting hours/housing location to take advantage of the new capacity).

Here’s a great simulator using realistic traffic modelling that you can use to recreate various traffic conditions and try your hand at solving them. And here’s a quick example of what happens when you add too many cars to the road:

Importantly, autonomous vehicles won’t solve gridlock traffic for one big reason: when autonomous vehicles are widely available, there will be more cars on the road, not less. To understand why, let’s look at two primary applications expected to promote autonomous vehicle sales:

1. Car sharing is expected to be the number one sales channel of autonomous vehicles for at least the first decade they are available. Uber, Lyft, and Tesla are all moving to develop or purchase large fleets of autonomous vehicles to offer autonomous rides to paying members, just like today’s ride sharing services, but without the overhead of a human driver. This is a fantastic use case for autonomous vehicles and will reduce the need for individual car ownership significantly. But ride sharing has already been shown to increase traffic, not reduce it because “idle” cars are driving around all over the place instead of sitting in designated parking areas away from active traffic flow. Autonomous ride sharing models will have the same effect, but possibly even worse due to the compounding effect of large numbers of people getting picked up and dropped off in locations without convenient pick-up/drop-off areas, blocking traffic in one lane while passengers get in/out.
2. Cost reduction for private vehicle owners is another big selling point for hypothetical autonomous vehicles. The reasoning goes that rather than paying for expensive parking, autonomous vehicle owners will instruct their cars to loiter, park far away, or simply return home and then come back later. Similarly, when not faced with the misery of having to remain attentive while sitting in traffic, many people will opt to move to cheaper housing farther from work to save money. In both of these instances, the amount of time cars are on the road will increase significantly from today, necessarily compounding traffic problems instead of alleviating them. Put simply, if your car drives itself between your home and work twice a day instead of once, it’s creating twice as much traffic.

What Are the Real Solutions?

The net result is a near-future in which autonomous vehicles may very well make traffic worse instead of better for everyone, with the only upside being that the additional time everyone spends in cars can be spent on things other than driving. But if this isn’t the rosy future many people expected, what can be done? The truth is that traffic solutions likely won’t change very much from what they are today:

1. More public transit investment with autonomous vehicle hubs. As many cities have learned, one way to reduce the number of cars on the road is to offer better public transportation options to commuters. Rather than seeing autonomous vehicles as a magic bullet that will solve all traffic problems and remove the need for infrastructure investment, cities need to be thinking of ways to develop more public transit options that complement autonomous vehicles via autonomous shuttle routes, driverless pick-up and drop-off zones, and partnerships with last-mile transportation services.
2. A wider range of commute times and locations. The number one cause of gridlock traffic is that the large majority of people all work roughly the same hours and thus commute to and from work at the same time. Changing this paradigm would go a long way toward reducing overall traffic on roads. The obvious ways to solve this are all challenging, but possible: build more housing near popular working locations so that people don’t have to commute by car, push for a wider range of “normal” working schedules for workers, and move away from the mentality that everyone has to work in the same building.
3. More Parking. One area the public and private sectors can work together is parking. Increasing available, cheap parking near popular work locations can go a long way toward eliminating the need for vehicle loitering and also provide centralized pick-up/drop-off locations for car sharing users. Since autonomous vehicles can park themselves without the need for human entry/exit, parking spaces and garages designated for autonomous vehicles can make use of much tighter parking spots to increase density. However, cities will still need to increase zoning for parking and loading areas.

Even though autonomous vehicles will be a great advancement in the areas of safety, cost, and time savings, they likely won’t help improve the worsening traffic conditions in many cities. It’s important to start thinking about alternative solutions instead of hoping for an easy fix.