Urban Planning in the Dawn of the Driverless Car

Mayor Curtatone and I recently returned from the Smart Cities Expo in Barcelona Spain, where we unveiled a new partnership between the City of Somerville and the car manufacturer Audi. We will be testing how autonomous vehicles work in an actual urban environment.

Driverless cars predominating city streets is in the realm of what Steven Johnson calls the “adjacent possible.” Uber just made headlines by purchasing a large chunk of Carnegie Melon’s robotics department. Several car manufacturers including Audi have indicated that autonomous vehicles are on the way, and Google has driven its fleet over 1.2 million miles.

Yet, despite all this, a recent study found that only 6% of large cities’ transportation plans consider the effects of driverless technology. Municipal governments are not preparing for what will likely be the most disruptive technology since the internet, at least in terms of its effects on cities.

Driving Blindfolded

Mayor Curtatone often compares running a city without data to driving a car blindfolded. While this analogy will be outmoded in a few years, in the meantime it also describes how transportation and land-use planners are preparing for changes in transportation policy. Somerville’s experiment should be one of thousands happening around the world; yet it seems most of the planning and testing is done by car companies without involvement from the cities they drive in.

People who are thinking about this are often considering it from a macro level: will the convenience of these cars induce demand for more vehicle travel and exurban living? Will access to fleets eclipse individual ownership of cars? Is it possible that a super-intelligent A.I. will weaponize smart vehicles, and then use them to transport the last living humans to a massive desert prison? (Ok, that last one is just something I wondered after reading this piece on the transhumanist philosopher, Nick Bostrom).

The most important questions, it seems to me, are more about the micro scale: should cities ditch parking minimums for new development? How can zoning incentivize fleet ownership? Should it? Assuming that machine learning algorithms will outperform human drivers, can cities plan to dedicate less space to travel lanes and parking spaces? Can we be more optimistic about Vision Zero goals - perhaps even calling for zero collisions? Will consolidated parking open up new public spaces and help to solve the affordability crisis?


Somerville will not be a static lab. One of the most exciting and intriguing aspects of our partnership with Audi is that we expect to test driverless cars in an area that will one day have unprecedented access to public transportation. The extension of the Green Line from Boston into Union Square will put a higher proportion of people within a 5-minute walk of the subway than almost anywhere in the country.

Somerville may help to solve one of the most pressing questions that driverless vehicles pose: how will they integrate with public transportation, cycling, walking, and other modes of transportation?

It’s not difficult to envision a future where you tap your wallet to hop on the green line, get off to an expectant Audi, and then have it drive you to the edge of Davis Square, where your smartphone tells you it will be faster to walk the final mile to your job at the hoverboard shop (at which point you will curse yourself for forgetting your hoverboard at work in the first place).

In any case, it is difficult to envision a future in which the practice of urban design is unchanged by driverless technology. Planners are usually very good at, well, planning. Why we are not planning for this is a mystery.

Written on November 25, 2015