One of the many promises of self-driving cars is that they will make traffic congestion a thing of the past, or at least reduce it. In urban scenarios, however, they’re likely to make congestion even worse.
That’s according to a new study from Adam Millard-Ball, a transportation planner and associate professor of environmental studies at the University of California, Santa Cruz, who argues that one factor that hasn’t been taken into account in autonomous car scenarios is parking costs.
The argument works like this: parking in city centers is both limited and (typically) expensive, which also serve as incentives to limit the number of cars coming into the city. Meanwhile, the usage of self-driving cars will initially be limited to services like ride-hailing and delivery, so what do they do between fares and delivery orders? Or supposing they arrive at the pick-up point early and the customer hasn’t yet arrived?
According to Millard-Ball, they will simply cruise around, which is cheaper than parking. And because cruising at slower speeds is cheaper than cruising at higher speeds, we’ll have empty self-driving vehicles cruising around city streets as very slow speeds. And it wouldn’t take that many of them to clog up the streets, Millard-Ball says [via UC Santa Cruz]:
Under the best-case scenario, the presence of as few as 2,000 self-driving cars in downtown San Francisco will slow traffic to less than 2 miles per hour, according to Millard-Ball, who uses game theory and a traffic micro-simulation model to generate his predictions.
“It just takes a minority to gum things up,” he said, recalling the congestion caused at airports by motorists cruising the “arrivals” area to avoid paying for parking: “Drivers would go as slowly as possibly so they wouldn’t have to drive around again.”
So essentially it’s an economic problem rather than a technical one. Consequently, Millard-Ball proposes an economic solution in the form of a congestion charge, which a few cities like London and Stockholm have already implemented. It could be a flat fee, or it could be usage based – so for example, self-driving cars could be billed based on metrics like location, speed, time of day, weekday vs weekend or which streets they use.
There is one snag to that solution, however – congestion charges tend to be politically unpopular with human drivers. Many cities around the world have introduced congestion pricing policy proposals, only to have them soundly defeated by political opposition. Here in Hong Kong where I live, for example, simply proposing toll price hikes on the cross-harbour tunnels (intended as congestion-reduction measures) is so politically toxic that it’s practically the one thing most political parties in LegCo can agree on. h
Millard-Ball admits as much, although he notes that it may be easier to pass one now that applies only to autonomous vehicles since they’re not commercially available yet:
“As a policy, congestion pricing is difficult to implement. The public never wants to pay for something they’ve historically gotten for free,” he said. “But no one owns an autonomous vehicle now, so there’s no constituency organized to oppose charging for the use of public streets.”
Except of course for the companies who make self-driving vehicles and the ride-hailing and delivery companies who will be their first customers. I don’t see either group going along with congestion charges that target their business.
It’s a thorny problem with no easy solution, but it’s a good illustration of how complex the transition to self-driving cars is going to be. All the benefits we hear about them will only really be realized a couple of decades from now when most humans are removed from the driving seat altogether and smart-city infrastructures are in place to support them.
That’s fine and worthwhile, but we have to remember always that how we get from here to there is going to be as convoluted, frustrating and messy as a rush-hour traffic jam.