I’m not sure if you’re aware of this, but, despite some disappointment from members of the Canoe Elite, I am now a shitty canoe owner. I haul my canoe around, like so many others, on the roof rack of my car. People haul all kinds of things on the exterior of their cars — cargo carriers on roofs, bikes at the rear, lumber hanging out of the back — every single day. I realised as I was driving around with my car’s canoe hat that this common practice would introduce a slew of issues for autonomous cars, and I’m not certain anyone has addressed this yet.
Autonomous vehicles are still very much in the developmental stage; incredible progress has been made, but they’re by no means finished, and we’re still, I’d think, at least a solid decade away from anything approaching Level 5 (as in pretty much full) autonomy, despite what Elon Musk says.
One thing we do know for certain is that AVs will require a lot of sensors. These will be cameras, radar emitters, LiDAR domes, ultrasonic sensors, and multiple instances of each.
We know these sensors will need to have unimpeded views of the world around them, which is why companies are already trying to figure out how to keep them clean, even in bad weather or if they’re targeted by a bird that just had chilli.
But what no one seems to be addressing is the fact that most of the ways we carry sporting or camping equipment or large tools like ladders or materials like lumber, and so on on our cars will inherently conflict with an autonomous vehicle’s sensors and ability to drive.
I’m not surprised this particular issue hasn’t been addressed: there’s still too much work to be done just to get these things actually working, but that doesn’t mean it’s too early to consider what this means.
If autonomous vehicles eventually become commonplace, I think this will be a bigger deal than perhaps we realise. People strap shit to their cars all the time; while it’s not what people do most of the time in their cars, it’s hardly rare, and, in certain market segments and areas of the country, it’s actually common.
How many times have you seen, say, a Subaru wagon with a bike rack? Plenty. What about Cherokees with kayaks on the roof? Or hatchbacks with the tailgate open and a couch sticking out, or a mattress strapped to the roof? I feel like pretty much everyone I know has pressed their cars into some kind of truck-like duty at least a few times, and I think people appreciate being able to do this.
Human drivers just need to keep the windows generally clear and to get a sense of how much the external space the car occupies with the strapped-on whatevers has changed, and we’re good to go.
Bu for an AV, this is much more complex; sensors cannot be blocked, and roof-mounted LiDAR domes would make almost any roof rack options useless. Corner-mount LiDAR or cameras could still be obscured by cargo tarps or rope or straps, and somehow the car would need to know that, say, it has three feet of canoe stern hanging off the back.
Some common types of racks designed for objects of known sizes, like ski racks, would be relatively easy. More general-use racks will prove a greater challenge.
Does this mean that AV owners or users just don’t get to carry things on their roofs or on tailgate racks? If we truly expect that autonomy will one day replace everyone’s automotive transport needs, can we realistically accept that this entire facet of automobile use — hauling bulky items and sporting equipment — is no longer an option?
I don’t think we need to accept that. There will need to be some kinds of solutions available; for example, autonomous vehicle roof racks will need to be “intelligent” and inform the car’s OS of its height and weight limits, as well as providing redundant/replacements sensors for any elements covered.
This could be expensive, though; a better solution would be to establish a universal sensor interface, like a USB for AVs, and allow sensors to be relocated from, say, a car roof to a pylon on a roof rack, with the roof rack plugging into the car’s network electronically.
It may also be possible, if you’re hauling something very oversized, to add Bluetooth or other wireless-protocol cameras or sensors to locations on the objects you’re carrying; this may also be needed if autonomous vehicles are to tow things like boats or campers as well.
These sensors and cameras could provide more information to the car about its new necessary perimeter, help it to see in areas that would be blinded by racks or objects or trailers, and, in the case of trailers, provide information about the trailer’s angles and location in relation to the main vehicle.
I don’t think this issue is anyone’s priority just yet in the development of autonomous vehicles, but at some point it will need to be. Unless we’re all willing to either give up a key ability of our personal cars or accept a new, possibly more resource-intensive way of doing things (arrange to get a truck whenever you want to take out your kayak or whatever), then I think these are issues worth looking into.
I’ve tried to find organisations or companies looking into the sticky problem of how to haul crap on a robot car, but so far, I haven’t found much. If anyone reading this is involved in such research and development, I’d love to hear from you.
I mean, I just bought a shitty canoe; I don’t want to have to get rid of it when I get an autonomous vehicle sometime in the 2200s or whenever I get around to it.