Autonomous Congestion

Proponents of autonomous, self-driving, vehicles (AVs) argue that they will reduce roadway congestion and by doing so also reduce greenhouse gas production.  Much of the improvement that proponents anticipate is based on the assumption that autonomous vehicles will allow for higher vehicle densities on highways.  This is predicated on physically changing roadways by reducing lane widths and the assumption that computer and sensor-controlled vehicles can be safely packed on roadways in these higher densities.

…AVs could, using more precise control systems, follow one another at much closer distances. Similarly, lanes could be narrowed, accommodating perhaps six lanes where there are only five today. These promises were, and remain, the foundation upon which AV utopianism has been built: a greener, safer, faster, and more pleasant transportation future just around the corner.

On the other hand, studies of potential AV owners suggest a different scenario:

These studies, flawed as they were, found something very different from the rosy future AV companies wanted investors and the public to imagine. They found reason to believe AVs would drastically increase the number of vehicle miles travelled, commonly shortened to “VMT” in academic literature.

 And the more vehicles miles travelled, all else being equal, the more traffic and emissions we can expect, cancelling out many of the AV’s touted benefits.

The problem with those studies was their basis on surveys where drivers indicated their potential behavior, not actual on road activity for an obvious reason – there is no real-world experience with AVs.  Tough to predict how people in the real world will utilize AVs.  However, one researcher came up with a method to model that behavior. 

What if they hired chauffeurs to drive random people around? The chauffeur, Walker outlined, will do the driving for you. And, just like the most optimistic AV future of fully autonomous robot cars zooming around, you don’t even have to be in the car. “All these things the self-driving car can do for you in the future,” Harb summarized, “a chauffeur can do for you today.”

A small sample experiment was devised where a week of normal driving behavior was tracked, followed by a week of chauffeured vehicle availability.  The chauffeured vehicle was available in the same way an AV would be available including being dispatched independently to run errands or pick up children at school.   The results demonstrated a much higher number of vehicle miles travelled (VMT) for the pseudo-autonomous vehicles.

Harb thought they would see people sending their cars out more than if they were driving themselves, something like a 20 or 30 percent increase in VMT with the chauffeurs. Nothing to sneeze at, of course, but towards the middle of the wide range of the results the surveys had suggested. He was wrong. The subjects increased how many miles their cars covered by a collective 83 percent when they had the chauffeur versus the week prior.

As long as community planners believe that personal vehicles are to be the primary means of transportation our cities and towns will be dedicated to vehicles not people and everyone should be prepared for the congestion that will result.