While watching some of that "B roll" video of Teslas being loaded onto ships, it occurred to me to wonder if the cars can be made to load and unload themselves onto/off cargo ships? I can't tell from the brief video whether human drivers are used for the loading process, but if they are, it seems like an interesting (although admittedly niche) application for autonomous driving. It would be cool to see an entire parking lot's worth of new cars efficiently loading themselves onto the ships!
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Practicality of having Teslas load/unload themselves for shipping
- Thread starter Levenkay
- Start date
The short answer is no...that would be more work than having a person do it.
Last edited:
E90alex
Active Member
Cars are very tightly packed for transport. Smart summon doesn’t even work well right now and when it tries to do it’s thing it’s slower than molasses.
"More work than having a person do it" A person? Looked like there were a couple hundred cars on the dock. And however many drivers are used, they have to disembark the boat and jog back to pick up another car.
"Tightly packed for transport" Yeah; wouldn't that make it even more cumbersome for human drivers to exit and wriggle out of the densely packed mass, and therefore more of an advantage for the cars to pack themselves?
The point that it would be a lot of work, and that smart summon doesn't work well is the best one. It's more of an application for "hive-mind-like" behavior than individual driving.
"Tightly packed for transport" Yeah; wouldn't that make it even more cumbersome for human drivers to exit and wriggle out of the densely packed mass, and therefore more of an advantage for the cars to pack themselves?
The point that it would be a lot of work, and that smart summon doesn't work well is the best one. It's more of an application for "hive-mind-like" behavior than individual driving.
Of course not. They do this for a living and have a system. They will park with the driver's side open in a line. Not randomly parked. Also they are FAST (again they do this for living) and could park several cars while the Tesla system was backing up, pulling forward, realigning and slowly trying to park itself. Hell I DON'T do it for a living and can park MANY times faster that my Tesla can UNDER all conditions.
EDIT: Just to add Tesla's still can't Full Self Drive through Boring Tunnel paths without a driver.
Last edited:
Just watch this video and you will see how preposterous it is thinking that Tesla's could even come close to navigating all the complexities, nuance maneuvers, overall mass coordination and on the fly decision making needed to load cars on a ship.
Maybe in a decade or so with specialized programing it could be accomplished.
Maybe in a decade or so with specialized programing it could be accomplished.
not a chance. Every transport ship is slightly different, every dock is different.
FSD can't even drive itself now, so why would they waste the time to automate something that is likely the most complex thing the car does once its life.
Rephrase - "some" of the cars. Only export cars travel by ship.
FSD can't even drive itself now, so why would they waste the time to automate something that is likely the most complex thing the car does once its life.
Rephrase - "some" of the cars. Only export cars travel by ship.
With a bunch more going by rail, requiring the same packing and unpacking process.
I like the idea. Instead of training FSD for those unique situations, implement a general system where special markers can be put down that the car can guide on. In theory, an entire network could be laid out, and the car could be given a destination in that virtual network just as it does with physical roadways. I'm sure that people would come up with innovative uses for such a precise tracking system.
But going back to the ship and rail use, I suspect that @Yelobird's concern about unions would be quite valid.
KArnold
Active Member
Similarly, as I understand Aurora's self-driving 18-wheelers possibly going live on one route this year; they arrive to a designated parking area. A human will still need to hop in to find the correct loading dock and back into it.
At least, that's all they've announced in forward speculation mode that I've seen.
At least, that's all they've announced in forward speculation mode that I've seen.
"Can't" in this case means "not allowed to". It doesn't mean "incapable of".
Las Vegas has a long history of artificial limits on people-moving systems. So far there are still no trams or even well-run bus service between the hotels and the airport. There are a few trams that run mostly between same-company properties, and a somewhat helpful monorail to the convention center. But even those have inconveniently placed stations requiring long walks through the casinos. Helping people to leave the property is not in the interest of the owners, and facilitating cheaper airport transfer is not in the interest of the taxi companies.
(Having said that, Vegas is far from incompetent regarding tr.ansportation or anything else. Things do work and there's a reason for everything, but saving customers money, or letting them go somewhere else to spend it, is certainly not the prime consideration.)
The Boring tunnel project is clearly constrained against becoming a true model of automation and efficiency. It's not a technology limitation. It's very clear that FSD, or a specifically-trained version of it, could easily replace the human drivers.
I haven't really thought before this thread, about the idea of Teslas doing their own parking at various holding lots and trucking/shipping assistance, but it makes a lot of sense to me.
All the demonstration runs in the first tunnels were always manual as well. I'm not aware of any FSD based runs even in Tesla test tunnels, so it sounds much more like "incapable of".
FSD is not autonomous and is only driver assistance - so "incapable of" really fits the bill here.
It absolutely is a technical limitation.
The Trailers that Aurora's Trucks are towing look like they are specifically setup for the Trucks. I don't see any airline gladhand hookups. Are Aurora's Semis restricted to only those Box Trailers or can they tow any Trailer?
aurora.tech
Aurora Horizon | Bettering business with self-driving trucks
We’re building Aurora Horizon to deliver the benefits of self-driving technology to freight logistics, offering extraordinary potential for growth, impact, and scalability.
I hear you but I respectfully disagree. The Boring tunnel is very simple use case. I and others have ridden in these and conversed with the drivers, as well as reading about it when it first opened. I stand by what I wrote; the restriction is regulatory and possibly Tesla-priority driven, more than the technical chaloenge.
I believe that the whole project was envisioned to expand well beyond the very limited demo service being offered today. But like some of the monorail and other people mover projects of the past, service to the airport, possibly downtown etc. has not been approved or funded.
I'm not saying they could simply turn on full autonomy tomorrow in the present Boring tunnel; I'm sure there would be specific ML training and a program of testing validation initially with safety drivers and all that. But given the limited scope and potential of what's been built so far, I would make the same decision to delay applying dedicated Tesla resources while they have bigger fish to fry.
Tiger
Active Member
You would need a Tesla Bot to accomplish this /s
Correct answer: it is doable, but not priority and not enough cost savings to do it only for a Tesla fleet as the volume would not justify cost of engineering and there is not enough training data.
Correct answer: it is doable, but not priority and not enough cost savings to do it only for a Tesla fleet as the volume would not justify cost of engineering and there is not enough training data.
KArnold
Active Member
I've only read about full size 18-wheelers. But they did buy the Uber self-driving team. You don't hear any news about them but they do continue to work as their own unit separately from the truck team.
My guess is they have long term plans for more but they have been very tight lipped about what it might look like. If they have self driving tech working on 18-wheelers, it would make sense to consider other applications as they mature.
