How cities can prepare for a rise in AV commuters

Imagine you are trying out your own autonomous vehicle (AV) to get to your downtown office for the first time. You leave your home, the ride goes smoothly, but as you near the office, you begin asking yourself questions you’ve never had to consider before: Do I need to have the car stay in an expensive garage? Is there someplace close by where the car can just drop me off? Can I send it somewhere cheaper to park? 

Questions like these informed a recent study co-authored by Carnegie Mellon University faculty member Sean Qian. Alongside researchers at CMU’s Tepper School of Business and the University of Texas at Dallas, Qian helped model how urban planners can respond to the expected rise of commuters taking AVs to work. 

Qian is the H.J. Heinz III Professor of Civil and Environmental Engineering. He seeks to understand how new technology will change how people travel as well as how people change their behaviors in response to new transportation options. “Because I’m an engineer, not only do I want to understand the impact, but I also want to take some optimal action,” he said. For him, the question comes down to, “How can we improve the total transportation system performance and efficiency?”

Qian and his fellow researchers expect adoption of self-driving vehicles on the road to happen rapidly—and soon. That adoption could shift the dynamics of a decades-old headache for workers heading to a central business district: finding parking. With this in mind, the study recommends cities be proactive in anticipating how self-driving vehicles will change parking as well as traffic and related infrastructure needs. As the paper notes, “Now is a rare window of opportunity for urban planners to put in place policies that pave the way for the inevitable mass arrival of AVs.”

To that end, Qian and his co-writers proposed a model for how the morning rush hour could change as commuters begin to adopt AVs. They then examined how this would affect a city’s total system cost—factors like what it costs commuters to get to the office and their penalties for arriving early or late—and applied this methodology to commuting to downtown Pittsburgh. 

Qian expects parking considerations to be a major mindset shift for AV commuters. He described a game theory scenario where a commuter could pay $20 to park in the central business district for the day, or they could send the AV to park in a suburb for just a few dollars. Since it doesn’t cost the commuter additional time—no need to be in the car to park it—most people would choose the suburban parking option. 

At scale, this decision would have several ripple effects. First, cities would need far less space within their central business districts devoted to parking. Because it would become costly to maintain unused lots and garages, cities could look to repurpose these areas, increasing options for housing, shopping, or public spaces. 

It’s a trade-off, if you have too many parking spaces, you’re underusing your land value. But if you have too little, then people have to park outward, and you’re creating additional congestion.

Sean Qian , Professor, Civil and Environmental Engineering

“This is a very realistic real estate problem,” he added.

A related problem: how to manage an influx of vehicle drop-offs. As fewer commuters require urban parking, they will increasingly look for curbside drop-off spots near their offices. These spots are currently a limited resource in most central business districts, with existing options likely to become clogged as AVs gain prominence. Drop-off spots could also create downstream traffic as vehicles wait to pull into too few spaces. 

One recommendation the authors make is for cities to expand their curbsides. Some U.S. cities and airports have already begun increasing their designated drop-off areas in response to services like ride-shares. Qian said each city would need to find that optimal balance, similar to how cities have had to strike that balance for parking. 

Finally, the study considers the possible upheaval to city traffic, especially outbound traffic. If an AV goes to park in a suburb, that outbound route now becomes more jammed. A downtown worker could also send their AV on midday trips, like picking up other people or running errands. While these redeployed AVs could benefit local commerce, the added trips would create more traffic and exacerbate the related costs, from worn-down roads to more air pollution. 

Ultimately, the study is set up as a push for urban planners to consider these challenges before they manifest in real life. The paper’s authors describe their commuter model as serving “as a structured lens—an early warning system—for recognizing how seemingly small shifts in technology, costs, or incentives can lead to large changes in commuter behavior and system-wide efficiency.” In other words, if the cost of driving is lowered, more people are apt to drive, forcing cities to change how they shape and manage the larger transportation system. 

From a systems perspective, this means encouraging commuters to act in ways advantageous to everyone. One solution the authors believe cities could consider in the short term would be either charging AV commuters tolls or paying them subsidies. These tolls or subsidies would theoretically work to either discourage trips or encourage parking near offices at certain times of day, such as during the morning rush hour. Longer term, cities could build out and expand their infrastructure, such as drop-off spots. 

By optimizing around the shift to self-driving vehicles, the study found cities could lower their existing total system costs. Qian and his co-authors estimated that implementing these changes in downtown Pittsburgh would lower total system costs by more than 28%. A similar rate could equal hundreds of thousands, if not millions of dollars back in the budgets of large cities. 

The study acknowledges certain factors will complicate the end result for each city. For one, mass transit, such as commuter trains and buses, will still be more efficient on average at moving people than AVs. The authors’ model also assumes that everyone using an AV would own it—which, for now, isn’t really possible in the U.S. The federal government currently only permits AV licenses for testing and piloting purposes to private companies, such as Waymo or Tesla. Regulators will want to see automated vehicles are a safe option before allowing the average person to own one. Finally, questions of equity would need to factor into cities’ decisions: for instance, how AVs will change the quality of life for someone living close to a central business district compared to a suburban resident. 

So can AVs solve the commuter parking problem? Yes, Qian believes—but that will require purposeful, proactive coordination from urban planners, and cities may need to act sooner rather than later. With a timely approach, both commuters and cities can benefit from the solutions.

“We want to make sure it’s a win-win situation,” Qian said. “It’s a win for the city, that the entire system efficiency has improved, and also a win for every single person whose commuting cost has been reduced.”