Mitch Turck


An Inconvenient Truth About Pedestrians

Question for you: when waiting to cross the path of a vehicle like the one at the intersection below, do you take Path A and walk in front, or Path B and walk behind?

As both a pedestrian and a driver, it infuriates me to see people cross in front of vehicles. At stop signs, at driveways, at parking lots, even at traffic light intersections when you know the light is about to change. Whenever feasible, I cross behind the lead car.

It’s not just because the lead car has a lot to focus on (me, signage, other vehicles, the path they want to take, etc), and because any car behind them only needs to focus on the front of their vehicle (where I am), and because the second vehicle’s space to accelerate is limited and predictable given where the lead vehicle is. It’s also because, all things being equal, there is less of the universe’s energy spent when I take a few extra steps around the car than when said car has to stop before their preferred point. In fact, I’d argue that all ped crosswalks should be behind the lead car. In theory.

Cars often need to nudge out, get better views, time their movements… if I make them stop for me well ahead of that optimal point of view, then I’m likely forcing their two-ton gas-burning vehicle to stop twice and wait longer. It’s unquestionably better for the environment if I use 3 extra calories and walk behind the vehicle.

And there is the inconvenient truth: because pedestrians are small, free-roaming, can turn on a dime, and use less energy, it’s just a lot easier to design streets without them in mind and have pedestrians simply deal with it.

That sounds okay in isolated situations. But when you take a broad-stroke approach to designing car-focused streets, you start optimizing only for throughput, which minimizes pedestrians’ freedom to choose their paths, which consequently decreases their engagement with the urban environment, which makes for a worse urban environment, which negates the value of that great throughput metric you nailed for cars.

We see some innovations to address that engagement problem, though many of them still function primarily to protect pedestrians from cars rather than free up more path choices.

You’ve got your unorthodox crosswalk:

Your idiot lights letting people know when they can walk:

And your “smart-whatever” streetlights to fancy up the crosswalk:

These are nice, but they don’t do much to improve pedestrian choice. The holy grail is to maximize both throughput and choice, which is a contradiction to conventional traffic planning. If everyone gets to choose where they want to go and when, throughput plummets, right? It sure would be great if all those street users could somehow be in coordination — that’s what’s stopping any optimizations from happening.

Oh shit, that’s right — autonomous vehicles create coordination. Well then, if we can communicate with all vehicle traffic at once, then maybe we can up the ante on these pedestrian communication tools in turn? While my knee-jerk reaction to “smart city infrastructure” is a dramatic teenaged eye roll, this might actually be a useful opportunity.

Here’s a static (albeit very well-designed) urban intersection, with an example of a traditional crosswalk. I added the roundabout feature as well.

What if, instead of forcing fixed 90-degree paths for our dynamos of direction, the pedestrians, we used tech to overlay temporary crosswalks based on pedestrian demand?

You step up to (insert smart curb/light/whatever here) at the middle of the block, because you’re trying to get to a point Southeast from where you are. You point, or stand, or tap, or set your GPS to that direction, and eventually a crosswalk opens up for you. If more folks join in, it opens up faster, because you’re creating more demand for that path. All the vehicles know this is happening, and are routed to avoid that area — possibly by stopping, but also by changing their routes or circling the roundabout/block to maintain flow. Whatever optimizes throughput based on the current priorities.

Note how this also educates the pedestrian to better understand efficient behavior, which is a potential benefit of AI that I am rabidly supporting. If you want to walk diagonally across an entire street, you’re gonna be sitting there for a long, long time waiting to be prioritized. The more perpendicular your crossing angle is, and the more you’re willing to follow where others are headed, the faster you’ll get prioritized.

Got a mass of people exiting a busy transit terminal like Penn Station? A smart grid could open up the Northern half the street to let peds pile into a safe staging area, which then halves their Southbound crossing distance once the traffic eases up and the whole block can temporarily become the crosswalk. Remember too that with an all-autonomous traffic grid, vehicles don’t have to stick to one side of the road. The option below could still let a few Westbound vehicles squeeze through before the crosswalk fully expands by shifting Eastbound vehicles into one lane (effectively turning a four-lane, two-way block into a two-lane, two-way block.)

Speaking of which, maybe the city wants to have a no-vehicle zone cover an entire block for lunch hour. Cool, that can easily be done.

How about an insanely large scramble that doesn’t actually have to exist as a permanent epilepsy-inducing paintjob like the ones we find today?

Also… sponsored crosswalks? I’d have to think long and hard about why we’d want a Mickey Mouse-shaped crosswalk, but sure.

Now I want to be fair to the folks who would actually have to engineer these algorithms that manage dynamic paths across an entire city and decide how to prioritize people vs. cars: I am aware this is insanely difficult. But it’s also an extension of where we’re going in general with a connected traffic grid. And as for feasibility, is it really so out of reach for the less-congested areas of the world? Suburban neighborhood streets? Perhaps roads like the ones they have in Tempe, Arizona?

Yes, it’s still probably extremely hard. Fine. Going back in my hole now.

More by Mitch Turck

Topics of interest

More Related Stories