Linda Poon is an assistant editor at CityLab covering science and urban technology, including smart cities and climate change. She previously covered global health and development for NPR’s Goats and Soda blog.
A radical proposal to solve congestion might be too beautiful for this world.
In the same week that I’ve written about several promising ideas for the future of transportation and one that’s doomed, yet another proposal has been catching headlines for its audacity. This one, called CarTube, envisions a network of underground tubes that would shepherd driverless cars to their destinations, leaving surface streets free for pedestrians, bikers, and probably any of those ancient cars still driven by measly humans.
It’s a bold idea, more remarkable for how desperately it wants to solve traffic congestion than for its chances of actually happening. But such a monumental problem demands dreaming big, even if it’s too big.
In this case, PLP Architecture in London did the dreaming, putting together this concept to target heavily congested cities like London, Beijing, and New York. Only fully electric driverless cars would operate in these tunnels, after having entered from existing roads above ground. It embraces smart technology, with an app to summon a car, which would be connected like cars of a train and driven to the passenger’s destination by artificial intelligence. They would automatically be spaced just six or seven feet apart for maximum efficiency. When passengers are dropped off at an underground station, they would be stored in “car stacks.”
“Once cars can begin to be intelligent in [their] own right, then you can have negotiation with other cars about how you get to your destination in the most efficient way possible,” says Lars Hesselgren, the principle designer behind CarTube. The idea is, in a sense, his response to the heavy push for multi-modal transport.
In simulations, it looks cool enough.
To urban planners, though, the feasibility of the idea and the car-centric approach may seem just as problematic as the doomed straddling bus in China.
“Technological fixes like this, however well meaning, buy into the traditional paradigm of focusing on movement, not people,” Robert Cervero, a recently retired professor of urban planning at University of California, Berkeley, tells CityLab via email. “It risks exacerbating the past half-century of ever-worsening problems attributed to automobile-dependent lifestyles, [like] urban sprawl, the engineering of walking and face-to-face contact of everyday life.”
Not to mention the challenge of finding the space for these tubes amid the existing underground pipelines and subway tunnels—and paying for them. The cost of all that, he says, would be better spent creating “world-class” public transit.
Still, it’s not an idea to cast aside entirely. The idea of “platooning” autonomous vehicles for efficiency’s sake has been around for nearly 20 years, and road tests have proven that this kind of computer-driven car travel in tight quarters is already possible on existing roadways. That helps explain why the firm is already looking for potential collaborators like Google, and it has piqued interest at a recent urban mobility conference co-hosted by Bloomberg.
Alexandre Bayen, the director of transportation studies at Berkeley and an autonomous vehicles researcher, thinks there is some value to the idea—strictly from a technology standpoint. “It essentially bypasses the most significant difficulty of self-driving vehicles,” he says. “How do you make manned vehicles interact with unmanned vehicles?” Yes, there is the advent of machine learning, but then there’s the risk of pedestrians adopting dangerous jaywalking behavior when they learn automated cars will always stop for them.
Assuming the infrastructure and technology are all there, Hesselgren proposes that driverless and manned cars simply be separated into individual networks. It’s not a totally groundless idea; similar technology was tested in 1997 when a platoon of automated vehicles drove in a separate lane alongside a highway in San Diego. And a proposal for a dedicated lane for self-driving cars along the highway between Vancouver and Seattle was only recently introduced. Bayen adds that if done right, the convenience of ride-sharing and the possibility of electric buses running in these tunnels as well could discourage private car-ownership.
But let’s be clear: The idea would have to get the green light first. Hesselgren intends for such a system to be gradually implemented over 30 or 40 years in small segments (rather than dig up an entire city at once). Yet with limited resources and a short timeline of just four years for cities to drastically curb carbon emissions to even get a shot at limiting global temperature rise, this will unlikely be high on mayors’ list of priorities.
Hesselgren acknowledges the obstacles, but says CarTube is a research proposal for the far future. He wants to know what parts could work and what rightfully should be scrutinized. “It’s a challenge for people to think,” he says, “not just react.”