Designing Cities Where Cars and Pedestrians Both Win

Urban planning rarely feels like a win-win. More lanes for cars usually means fewer paths for walkers. But researchers at the Singapore University of Technology and Design (SUTD) have flipped that assumption. In a study published in the Journal of Urban Design, they show how a generative urban model can produce layouts that support both road access and walkable neighborhoods. The team tested more than 3,000 designs on a 100-hectare site in Singapore, revealing that resource efficiency and pedestrian comfort do not always need to be at odds.

The Old Trade-Off

For decades, planners assumed that improving pedestrian infrastructure would cripple vehicular access. Wider sidewalks meant narrower streets. More parks meant fewer parking spaces. That zero-sum logic has shaped countless cities, leaving many residents with car-friendly but people-hostile streets.

The Generative Urban Model

The SUTD team took a different path. They built a computational tool that separately generates road and pedestrian networks, then evaluates them across environmental, social, and economic criteria. Think reachability, crosswalk delays, cul-de-sacs, and block shapes—all run through optimization algorithms that can simulate thousands of possible outcomes in minutes.

“We’re moving away from car-centric urban planning and placing greater emphasis on public well-being,” said Assistant Professor F. Peter Ortner.

The model was tested on Singapore, a dense and land-scarce city. Using the site as a proving ground, the researchers found that certain designs improved pedestrian reach and comfort without severely limiting road access. It was not, in Ortner’s words, a zero-sum game.

Two Competing Scenarios

The team compared two broad design approaches. One favored resource efficiency: overlapping footpaths with roads to cut costs and land use. The other emphasized pedestrian experience: creating direct walking routes, reducing crosswalks, and separating footpaths from traffic. The difference was striking. The pedestrian-oriented layouts produced safer, quieter, more inviting neighborhoods. The efficient layouts were cheaper to maintain but less pleasant to walk.

“Importantly, it was not a zero-sum game,” said co-author Assistant Professor Peng Song. “Our model shows that you can optimise for both – if you approach the problem with the right tools.”

Rebalancing the City

One surprising finding was that road expansion sometimes boosted pedestrian networks because city guidelines typically include pavements alongside new roads. At the same time, crosswalks—so vital for safety—often slowed walkers more than roads slowed cars. The nuance matters. Designing for people requires separating the two systems, not lumping them together.

The researchers stressed that their model doesn’t spit out a single “best” design. Instead, it creates a spectrum of possibilities, leaving the judgment calls to humans. “The goal is to augment human decision-making, not automate design,” Ortner explained.

Global Potential

Though tested in Singapore, the model is adaptable to other cities. Future versions could factor in thermal comfort, access for people with disabilities, or even noise pollution. The team envisions AI-assisted, human-centered planning becoming the new normal. Entire districts designed solely by AI? Not likely. But human-AI collaboration in shaping walkable, livable cities? Already here.

In the end, the study suggests that cities do not need to choose between cars and pedestrians. With the right computational tools, planners can design for both—balancing mobility and livability in ways that older models never allowed. The trade-off, it turns out, was not inevitable.

Journal of Urban Design. DOI: 10.1080/13574809.2025.2517652