The driverless car is coming. We’ve been promised the driverless car for a while, mind you, but this time it seems like it’s really here. Google is road testing one and everything. But cars aren’t independent entities. They are part of a larger infrastructure of transportation. As the technology of driving changes, it’s a good time to think about the design of the rest of the system.
Over at The Atlantic Cities, Emily Badger looks at what intersections might look like with driverless cars. Computer scientist Peter Stone argues that better sensors and software will allow for more efficient use of intersections. Gone will be the days of people stuck at lights, waiting for a red even though the opposite way is clear. Instead, Stone proposes a more steady flowing of traffic, with an animation of what that might look like.
Badger is intrigued but skeptical, calling the whole idea "far-fetched". The commenters, even more so, citing fears about high technology’s tendency to crash and the general unreliability of computers (which is so interesting to me, given the rampant unreliability of people). For my part, looking at the animation just made me think of traffic in Vietnam or India or wherever this is:
Watching these videos, it looks a lot like someone’s already implemented Stone’s algorithms.
What’s interesting about the skepticism towards automated driving is that it reveals how invisible the current systems of automation already are. Traffic control in most cities is already monitored and controlled by wired up intersections laced with a complex network of sensors and lights, designed to cause traffic to flow smoothly around town. In many ways, the fallible fleshy pilots of vehicles are the last remaining organic components of a city-wide cybernetic system.
It is instructive to trace the transformation of the traffic cop, from human agent fixing local traffic flow, to the robot servant of civil engineers. Here’s a passage from "Blocking All Lanes," an essay by Sean Dockray, Fiona Whitton, and Steven Rowell in The Infrastructural City which recounts that change.
Over time, the traffic cop was slowly transformed: his hands took on white gloves for visibility; his voice was replaced by a whistle; and eventually, he was elevated in a tower and communicated with the traffic via signs or coloured lights. The police officer slowly vanished, his body evolving into mechanical and electrical devices. His hands were replaced by standardized, colored signals. His eyes were replaced by sensing actuators, such as microphones, pressure sensors, electromagnets, or video cameras. All that was left was to replace his brain.
Back in 2010, when all of this was much more speculative, I saw a clear parallel between the replacement of traffic cops and the potential replacement of drivers.
Over time, the driver was slowly transformed: her hands took on a steering wheel for better maneuverability; her voice was replaced by a horn; and eventually, she was sealed in a cabin and communicated with the traffic via honks or coloured lights. The driver slowly vanished, her body evolving into mechanical and electrical devices. Her hands were replaced by high precision steering mechanisms, her feet by networked cruise control. Her eyes were replaced by sensing actuators, such as GPS chips, proximity sensors, local mesh networks, or video cameras. All that was left was to replace her brain.
Badger is probably right to be skeptical that people will be wary of such technology in the U.S. But what about in China, where driver norms aren’t quite set yet, and traffic fatalities are ridiculously high? Or in Japan, where the trust in robotic automation is an article of cultural faith?
[Via The Atlantic Cities; top image: The driverless Audi that Google is developing with Stanford.]