A win-win approach to designing cities for cars and pedestrians

There is a general assumption that cities are often built on compromise. The more roads paved, the less space there is for walkways, parks, and people. What if this trade-off could be turned on its head?

In a new study, “Generative urban modelling for walkable neighbourhoods: design optimization for pedestrian-oriented street networks”, a team of researchers from the Singapore University of Technology and Design (SUTD) demonstrated how generative urban models—computational tools that rapidly produce and optimise design scenarios—can help planners identify urban layouts that cater to both pedestrians and vehicles.

The study introduced a method for simultaneously generating separate pedestrian and vehicular networks, then assessing their performances across multiple criteria, from street reachability to the number of crosswalks and road junctions. This method, which uses a multi-objective generative model, can evaluate environmental, social, and economic objectives and produce a range of optimal (or near-optimal) solutions much quicker than traditional urban planning approaches.

With a 100-hectare site in Singapore as a testbed, the team used a generative urban model and evaluated over 3,000 designs to examine whether resource-efficient and pedestrian-friendly cities could, in fact, be one and the same.

“We’re moving away from car-centric urban planning and placing greater emphasis on public well-being,” said Assistant Professor F. Peter Ortner, the first author of the study. “In a high-density, land-scarce city like Singapore, designing neighbourhoods that prioritise pedestrians is a matter of both liveability and sustainability.”

Urban planners have historically struggled with the complexity of designing for both pedestrians and cars. While many models conflate the two networks, the team’s study takes a more nuanced approach by generating and assessing them separately, factoring in elements like crosswalk delays, cul-de-sacs, block shapes, and walking distances to points of interest. Two design scenarios were tested: one prioritised resource efficiency, and the other favoured pedestrian experience.

The results challenged a prevailing assumption in urban design—that any improvement in pedestrian infrastructure must come at the cost of vehicular accessibility. Instead, the team found examples of well-balanced layouts that enhanced pedestrian reach without drastically compromising road access.

“Importantly, it was not a zero-sum game,” said Assistant Professor Song Peng, the co-author of the study. “In fact, pedestrian networks often benefit from the expansion of vehicular roads because these typically include pavements. Meanwhile, pedestrians are slowed down by crosswalks and intersections, especially if they have limited mobility. What our model shows is that you can optimise for both – if you approach the problem with the right tools.”

The pedestrian-oriented designs generated in the study feature more direct walking routes, reduced crosswalks, and greater separation between footpaths and traffic. These configurations offered improved comfort and safety, and potentially even lowered noise and air pollution, all of which are qualities that are hard to quantify but deeply felt in our lives.

“Residents in these neighbourhoods are likely to experience safer, quieter, and more pleasant walking environments,” said Asst Prof Song. “That in turn may motivate walking for short trips, reduce reliance on private vehicles, and support more active lifestyles.”

In contrast, the resource-efficient designs tended to overlap roads and footpaths to minimise construction and maintenance needs. While these layouts reduced land and material use, they did not always provide the most inviting walking experience, highlighting the trade-offs involved in planning urban spaces.

The model also allowed the team to map out how different objectives interact. For example, increasing the length of the pedestrian network often improved walkability but used more land. Reducing road length helped pedestrians but increased driving distances. Rather than prescribing a single best design, the tool helps planners explore a spectrum of possibilities, making the value judgments behind design choices more explicit.

“Every development has its own priorities,” explained Asst Prof Ortner. “The goal is to augment human decision-making, instead of automating design. Generative models like ours can simulate thousands of scenarios in minutes, providing planners with better references and evidence to guide their choices.”

Though the team’s work was grounded in the context of Singapore, the model is adaptable to other urban environments. They envision future iterations that will account for more real-world factors, including amenities, thermal comfort, and accessibility for people with disabilities. They also see Design AI—the collaboration between humans and AI systems—playing an increasingly bigger role in the process.

“We’re unlikely to see an entire neighbourhood designed by AI anytime soon, but AI-assisted, human-centric design is already a reality,” said Asst Prof Song. “What we’re building are methods that can support fast, informed, and inclusive urban design that is rooted in hard data but shaped by real people.”

Acknowledgment and Disclaimer

This research was supported by the Urban Redevelopment Authority (URA) of Singapore through the URA-SUTD Strategic Collaboration, under the project titled “Computational Modelling for Optimization of Planning and Urban Design Parameters”

Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not reflect the views of the URA.