High parking costs in central business districts and lengthy morning drive times can turn downtown commuting into an expensive slog. A new study from researchers at the University of Texas at Dallas and Carnegie Mellon University models exactly how autonomous vehicles might ease that burden.
Dr. Neda Mirzaeian, assistant professor of operations management in the Naveen Jindal School of Management at UT Dallas, led the work. She and co-authors Soo-Haeng Cho and Sean Qian published their findings in Management Science. They built a continuous-time, game-theoretic traffic model of Pittsburgh to track commuter choices on departure times and where to park—inside the central business district or outside it.
Autonomous vehicles change the equation. They drop passengers at the curb and then head off on their own. No driver waits. No need to circle for a spot near the office. The model shows that when all commuters switch to AVs, many choose cheaper parking on the perimeter. That shift raises total vehicle hours and vehicle miles traveled compared with today’s human-driven patterns. System costs climb until planners step in.
Here is where policy enters. Downtown parking fees and congestion tolls can steer behavior. In the Pittsburgh simulation, those tools cut total system cost—time in traffic plus travel time—by as much as 28.5 percent. The incentives push vehicles to park farther out. Core downtown blocks then free up for new offices, stores, and homes instead of garages and lots.
“Autonomous cars don’t necessarily need parking space,” Mirzaeian said. “They need a drop-off space—like a school drop-off in the morning.” The study treats that insight as a planning lever rather than a promise of instant relief.
Real-world data already hints at complications. A May 2026 analysis of Waymo robotaxi operations in California, published in Transport Findings by MIT Transit Lab researcher Awad Abdelhalim, examined 13.8 million trips and 86.3 million miles from August 2023 through December 2025. Only about 54 percent of those miles carried passengers. Deadheading—empty miles repositioning or awaiting the next rider—accounted for the rest. The share improved from 36 percent passenger miles early on but plateaued near 55-57 percent. Average empty time between trips fell to roughly 18 minutes by late 2025, down from 28.6 minutes overall. Still, the pattern mirrors conventional ride-hailing. Empty miles persist.
That persistence matters for curb space and congestion. In Westchester near Los Angeles, residents reported Waymo vehicles parking on residential blocks at all hours in June 2026, according to ABC7 Eyewitness News. The cars turned quiet streets into staging areas while drivers sought free spots near LAX. Waymo noted that parking between trips helps reduce congestion and conserve energy, yet neighbors called the practice frustrating.
Broader studies point to similar trade-offs. University of Texas at Arlington civil engineering professors Stephen Mattingly and Farah Naz reviewed travel behavior research and found that widespread AV adoption tends to raise vehicle miles traveled. Their meta-analysis showed an average 5.95 percent increase. Shared AVs produced a 5.3 percent rise; privately owned ones reached 6.9 percent. More empty trips and easier travel can encourage longer commutes or new trips that would not have happened otherwise. The researchers published their synthesis in Travel Behaviour and Society earlier this year.
Yet some modeling suggests net gains are possible with the right mix. A June 2026 study led by Southern Methodist University researchers, covered by Newswise and TechXplore, projected that full AV adoption in the Dallas-Fort Worth region could cut traffic delay by 33 percent by 2045. Faster flow and better routing delivered the improvement even without new highway lanes. The work appeared in the Journal of Urban Technology.
Land-use effects follow the traffic patterns. Less demand for downtown parking opens redevelopment. The UT Dallas–CMU model explicitly links lower central parking needs to more residential, retail, and office space in the core. Similar logic appears in earlier work. A 2019 paper by Adam Millard-Ball in Transport Policy showed that AVs seeking cheap or free spots could more than double vehicle travel in dense urban cores unless parking costs or regulations change. A 2023 Urbanism Next report funded by Waymo concluded that large parking reductions require shared, pooled vehicles deployed across entire metro areas and paired with strong transit—not private ownership alone.
Planners face timing questions. Mirzaeian’s team stresses that their model serves as an early-warning system. Small shifts in fees or infrastructure can trigger large swings in commuter behavior. Cities with morning-congestion problems stand to gain most from AV adoption if they adjust short-term pricing and long-term street design. Converting some parking spots to dedicated drop-off zones is one concrete step already under discussion.
National Parking Association materials note that autonomous parking systems themselves are advancing quickly. Vehicles that park themselves could pack tighter and reduce minor collisions, but those gains still depend on how fleets operate once empty. Cybersecurity, insurance, and liability remain open issues the association flags for operators.
Waymo’s own data shows gradual efficiency gains as fleets grow and service areas expand. Deadheading miles per trip dropped from about 5.1 to 2.8 over the study period. Geographic spread into suburban zones and freeway service helped push average passenger miles ahead of deadheading miles per trip in late 2025. Low sharing rates—around 1.4 passengers per trip—limit further reductions.
The picture is not uniform. Some corridors may see relief from smoother flow. Others may absorb more empty repositioning. Curb management becomes critical where pick-ups and drop-offs multiply. Cities already testing AVs in downtown zones, including Dallas, will watch these dynamics closely.
Mirzaeian and her co-authors avoid city-specific prescriptions. Their contribution lies in mapping the general trade-offs that emerge when AVs, commuters, and pricing interact. Urban planners now have clearer signals on where to set fees, where to build drop-off infrastructure, and which blocks might convert from asphalt to active uses. The model does not promise parking lots will vanish overnight. It shows how incentives and infrastructure choices can steer the outcome toward lower system costs and denser, more productive downtown cores.
Recent real-world friction in Los Angeles and efficiency plateaus in California data underscore that policy cannot wait for perfect technology. Fees, tolls, and curb rules will shape whether AVs ease parking pressure or simply relocate it. The UT Dallas–CMU framework gives decision-makers a tool to test those levers before full deployment arrives.
WebProNews is an iEntry Publication