Everything That Scares Us About Autonomous Cars Is In Windshield Wipers
Any given room full of self-driving car pundits will contain a peppering of sages and fools both arguing that autonomous tech is new; that it marks a clear separation from the cars of yesterday. For some industry audiences, I think that’s fine. For the general public, it’s unproductive bias confirmation.
That’s why, today, we’re going to talk about a meager little feature we’ve all had on our cars from the beginning; why it holds all the same mind-boggling questions as bleeding-edge AV tech, and why in fact, we could have reached the current state of self-driving innovation by working on absolutely nothing else but this feature.
It’s the damn windshield wiper.
Starting from the top: with rare and unnecessary exception, every car has a windshield. It serves two simultaneous purposes which, notably, are at odds:
- It protects occupants from the external environment, ensuring a safe and comfortable means of travel. This is critical for high speed transportation.
- It affords the driver an unobstructed view of the environment. This is a necessary compromise, as there are better materials (e.g., metal) with which to protect the occupant from the environment, but such materials would also obscure the driver’s view, negating the value of #1.
The need for #2 is why wipers exist. On both windshields and headlamps, light needs to pass through the protective surfaces to enable safe driving, and wiping down these surfaces repeatedly by hand while the vehicle is in motion is an unreasonable task.
I realize this all sounds laughably obvious, which is precisely why I’m writing the article: most of what goes assumed is highly dubious and poorly analyzed, which is why technological advancements often lead us to draw primitive lines in the sand and claim innovation is separate from convention, when in fact there is only constant flux. Whatever line we draw is merely one of an infinite number of arbitrary lines along a spectrum. The lines are neither objective nor fixed, but we like them there because our brains are lazy.
Hell, even I (and arguably, JR Hildebrand as well) was lazy in my passing comment about how windshield wipers are a metaphor for autonomy.
So without further ado, let’s get wet.
There was a time when windshield wipers, like many conveniences of the modern automobile, did not exist. Should any muss or fuss have its way with your windshield, you’d have to do the dishes yourself. Or, you didn’t have a windshield, and simply wiped your driving googles instead. All that changed when Mary Anderson invented the windshield wiper.
It would not surprise me to find you rolling your eyes at the banality of this history lesson. Well then, poindexter, riddle me this: what were the ethical repercussions of such a well-intentioned invention that took some degree of driver responsibility out of the picture? What was the lasting impact of a technology which improved the usefulness of an inherently contradictory design? Do those sound like familiar questions?
Think that over as we continue.
New approaches entered the fray as the decades zoomed by. Here’s a three-wiper set:
Seems reasonable for those of us who work at razor companies, right? More blades = better. But if better in this case means “safer”, then the notion that more blades are better implies that we had a less than optimal safety level at the time. So… who decided that the previous two-blade system was an “acceptable” safety level?
The clear view screen was quickly invented as an alternative to the distracting motion of wiper blades, though it was almost immediately relegated to duties on ships and trains. It worked by spinning a sealed portion of the windshield at high speeds so that, as centrifugal force flicked away rain, your view was never obstructed.
This too, challenged the idea that what we’d all been using was “safe enough”, but interestingly, it did so by calling out the blatant limitation of the wipers themselves, which ironically had to obstruct your view to clear obstructions. The obvious follow-up echoes our previous conundrum: what makes for an “acceptable” obstruction, and who makes that decision?
Is it too far a reach to tie Uber’s fatal software performance error to windshield wipers? Let’s see about that.
Window defrosters soon came into fashion after the advent of wipers. And a few decades later, headlamp wipers were introduced.
Lastly, we have an invention never commercialized (at least, not yet): an ultrasonic windshield:
What you’re seeing here is the deconstruction of the wiper as a concept, which is why I took great pains at the beginning of this article to explain what wipers were originally believed to be.
The defroster solves the same problem as windshield wipers, but does so with a different design due to the sub-problem of frost being an edge case. Headlamp wipers retain the design, but apply its purpose to a different surface. The ultrasonic windshield replaces wiping action with an entirely different physical force, and introduces electronics as part of the solution. In all cases, it seems difficult to make an argument against these innovations — agreed? The windshield wiper itself isn’t the thing we want; the thing we want is good visibility. Right?
No. The thing we want is safe, effective transportation. While good visibility would seem necessary to that outcome, it is not the outcome itself, just as windshield wipers are not the outcome. Both wipers and clear vision are tools to achieve said outcome.
With that in mind, we can identify a clear limitation of these tools we’re using to achieve our outcome: they only improve the vehicle’s surfaces. They do nothing to improve the driver. Would improving the driver also improve visibility, thus resulting in safer transportation? Certainly.
Here’s the most common pro-visibility driver improvement on the road today: automatic wipers.
Note that the innovation isn’t in the wiper at all, nor is it in the windshield as previous innovations attempted. It’s an electronic sensor that decides when it’s raining, and how hard. Consequently, the sensor relays to the wipers that they should wipe at a certain speed.
So, to recap: nothing about the wiper function has changed. What has changed is that a computer now automatically compensates for driver negligence or confusion by activating the wipers at the right speed. And that means, unless the technology fails, humans no longer have to pay attention to windshield wipers.
Unless the technology fails, humans no longer have to pay attention.
But the technology does fail, which means we either need to decide that we will hold drivers personally responsible for windshield wipers regardless of how well their automated features work, or that we will acknowledge the high performance of the technology as “safe enough”, and chalk up any catastrophic failures to insignificant edge cases that are a matter for the courts.
Of course, we have laws around drivers’ obligations as they relate to windshield wipers, so until those are changed, we can’t justifiably hand off wiping duties to a machine.
So we have a real lack of clarity around human vs. machine accountabilities and a realization that, at some unknown point, we progressed far beyond the original intent of windshield wipers. And yet, nothing has changed: wipers were always just a tool to achieve the outcome of safe transportation. Thus, we march on in creating solutions to improve visibility by enhancing or overriding the driver.
We said at the beginning that a windshield is made of glass, despite it being a poor medium for protection, because drivers need light to pass through the surface if they are going to successfully operate the vehicle. And wipers exist on windshields to maximize that visibility. Agreed?
Then it goes without saying that nighttime qualifies as an obstacle to visibility, and as such, night vision would help to mitigate that obstacle. In this sense, it’s really no different from the headlamp wipers, defrosters, or ultrasonic windshields that solve the visibility problems windshield wipers can’t.
Okay, this is interesting: until now, we wanted to keep the windshield as clear as possible; hence, wiper innovations. But since the human eye fails to see many obstacles at night, a clear windshield is of little value. Ergo, night vision actually assists the driver with a highly contradictory solution: a simulation of the road ahead and its obstacles, totally detached from the windshield itself and whatever natural light it lets in.
Can we hone this technology? If night vision can detect obstructions that the human eye cannot, then perhaps there are additional detection methods we can deploy in our cars to improve vision… or should we just call it perception?
And if these detection methods generate data, perhaps we can use that data to train software so that it recognizes obstacles on its own, which would in turn enable the vehicle to warn its driver of such obstacles.
To take it further: if we have both advanced detection methods and a system capable of analyzing the obstacle as a risk, then we could also give that system the access it needs to perform a safe driving task if the driver fails to do so.
And finally, if we are training that system to handle challenging driving tasks more responsibly than a human driver, the obvious conclusion is that we should eventually remove the humans altogether.
So when did we stop talking about windshield wipers here? Did we ever?
If windshield wipers are literally the thing Mary Anderson designed, then anything beyond that — even three wipers rather than two, or wipers on headlamps — would be an entirely different and unrelated innovation. That doesn’t sound like the line in the sand we were looking for.
Are automatic wipers the real line in the sand? We call them wipers, they look like wipers, and they perform the same task as the original invention… they just do it without a human’s interaction to ensure optimal safety. It would seem, then, that human involvement isn’t a vital part of what makes a windshield wiper — in fact, Mary Anderson’s original invention intentionally reduced the level of human involvement, so taking the driver out of the loop is par for the windshield wiper course.
But once we agree to that, we’re also agreeing that the human eye is not necessarily the standard we should be supporting from the standpoint of satisfactory visibility. And that’s how night vision, radar, LIDAR, and even V2V connectivity enter the discussion one by one over time.
In fact, you’re in for a long haul of a debate if you want to argue otherwise, because there has never been a standard “level” of visibility. Vehicle windshields have different sizes, wipers have different levels of quality, humans have different levels of vision, and most appropriately, humans have different levels of performance even when presented with the exact same level of visibility. Most people lack the mental clarity, discipline, and driving talent to get the most out of the visibility available… which is why we can’t confidently draw a line in the sand between the autonomous software which overrides human failure and the automated wiper which does the same.
Technology converges — we too often forget that. In fact, it’s taken a century of convention-defying innovation for us to realize an autonomous car is just a very fancy windshield wiper, which speaks volumes about the limitations of the vehicle laws we’ve made and the cultural norms we’ve cemented to date.
But perhaps our greatest epiphany as windshield wipers have progressed is that the biggest obstacle to safe driving isn’t rain, snow, or pigeon shit on the windshield. It’s the human behind the wheel.
More where this came from
This story is published in Noteworthy, where thousands come every day to learn about the people & ideas shaping the products we love.
Follow our publication to see more stories featured by the Journal team.
