Auto-park and auto-load

[Crissa]

No, but Tesla could sell trailers mapped within the CT software. All dynamic information known and mapped. It would be added cost, but this isn't a short term investment. This is life from here on. peace

Oh, that's a good idea! Or have a lookup table of trailers.

-Crissa

Advertisement

 

[T3slaDad]

Depending on the dock, this would be really easy in retrospect. Our family has had a boat my entire life and we go out all the time! The docks where I live are very streamlined and well organized, whereas we've been to others that are discombobulated and cluttered.

So yeah, I think it's totally doable in a variety of situations, but maybe not all right away. Nothing software releases can't fix, though!!

 

[ajdelange]

I think it's wonderful that people still believe in magic.

 

[Crissa]

I think it's wonderful that people still believe in magic.

Why do you think that a system that can detect curbs with cameras won't be able to detect water?

-Crissa

 

[Sirfun]

Why do you think that a system that can detect curbs with cameras won't be able to detect water?

-Crissa

For that matter, you could back it down the ramp yourself, Put it in park, back the boat off the trailer then tell it to park itself. I would be very surprised if this isn't possible by sometime in 2022.

 

[ajdelange]

I suggest you look at the current Tesla vehicles' ability to find a spot and park without a trailer and especially look at the rate of improvement in that "feature" over the last couple of years before suggesting that it will, with two more years development, be able find a parking spot on its own, navigate there and park itself while towing a trailer. I also suggest that you think about how the autopilot is going to see the rear end of the trailer and learn the trailer's "control laws". Prepare yourself for a big surprise. But don't be surprised that other people aren't surprised.

I'm curious as to your experience with the Tesla automated driving suite and with boat trailers.

 

[FutureBoy]

I suggest you look at the current Tesla vehicles' ability to find a spot and park without a trailer and especially look at the rate of improvement over the last couple of years before suggesting that it will be able find a parking spot on its own, navigate there and park itself while towing a trailer. I also suggest that you think about how the autopilot is going to see the rear end of the trailer and learn the trailer's "control laws". Prepare yourself for a big surprise.

Given that until Tesla FSD came on my radar, my expectation of a truck was that I had to do EVERYTHING myself (short of turning the engine), I believe your characterization of "a big surprise" is entirely positive. Tesla FSD may not do everything my imagination could foresee as desirable functionality, but given the big 0 of other trucks ability till now, Tesla is already WAY better.

 

[Crissa]

Why would finding a parking space be impossible to reach, when today it can already navigate cities? Accuracy of inches vs feet? Sure, but... That's just dialing it in.

-Crissa

 

[ajdelange]

Don't get me wrong. I find the features of the automation package that do work to be very helpful in the circumstances under which they do work to the point where one feels safe in trusting them. I arrive at the end of a long trip much more rested than I used to in older ICE vehicles with "single axis" (could maintain constant speed) "autopilots". But what we have in today's Teslas is a far cry from FSD to the point where Germany is calling the use of that term false advertising. If, at the right speed and distance, you cruise along a row of parked cars, your car is supposed to be able to identify an empty spot and automatically park in it. Most of the time it cannot correctly identify a valid spot and often identifies arbitrary places along the road as valid spots. I have two Teslas and if anything the newer one is worse. The software was tagged "beta" in the old car. It's tagged "beta" in the newer one.

I won't even go into the technical problems of doing this with a trailer even if the basic parking algorithm worked. I said somewhere else that FSD is the stuff of SciFi movies and that, as such, it can do anything one can imagine. But that's not reality.

 

[Crissa]

But what we have in today's Teslas is a far cry from FSD...

Oh. So you believe that demonstrated technology will never be implemented or advance. Denying what it does today in the beta only makes you seem foolish.

It doesn't park now, but other much dumber software has handled that in the past. Saying that you don't believe it will advance because you don't believe it does what it does today is...

Unduly argumentative and misleading.

-Crissa

 

[Sirfun]

Don't get me wrong. I find the features of the automation package that do work to be very helpful in the circumstances under which they do work to the point where one feels safe in trusting them. I arrive at the end of a long trip much more rested than I used to in older ICE vehicles with "single axis" (could maintain constant speed) "autopilots". But what we have in today's Teslas is a far cry from FSD to the point where Germany is calling the use of that term false advertising. If, at the right speed and distance, you cruise along a row of parked cars, your car is supposed to be able to identify an empty spot and automatically park in it. Most of the time it cannot correctly identify a valid spot and often identifies arbitrary places along the road as valid spots. I have two Teslas and if anything the newer one is worse. The software was tagged "beta" in the old car. It's tagged "beta" in the newer one.

I won't even go into the technical problems of doing this with a trailer even if the basic parking algorithm worked. I said somewhere else that FSD is the stuff of SciFi movies and that, as such, it can do anything one can imagine. But that's not reality.

I see stuff like this and think, why shouldn't a CT towing an empty boat trailer, be able to drive at low speed through a boat ramp parking lot and pull straight in, between the parking lines, into an empty spot.

 

[FutureBoy]

I see stuff like this and think, why shouldn't a CT towing an empty boat trailer, be able to drive at low speed through a boat ramp parking lot and pull straight in, between the parking lines, into an empty spot.

I think the actual movement of the CT and the trailer will get automated relatively quickly. The difficult part will be curtesy and policy. Things like lining up for the boat ramp and letting others in line use the ramp first. Or knowing which parking spots are legal for which days/hours/vehicles sizes. There are lots of people including me who have a hard time with those parts. But it is pretty inevitable that AI will get to be more skilled than humans on defined tasks so I'm assuming the CT will eventually be able to do these things to.

 

[ajdelange]

I see stuff like this and think, why shouldn't a CT towing an empty boat trailer, be able to drive at low speed through a boat ramp parking lot and pull straight in, between the parking lines, into an empty spot.

What one sees depends on one's perspective. My perspective is that of an engineer with experience in AI, radar,signal processing, computing and navigation who has driven vehicles with "FSD" for about 2 years now. Engineers see things that lay people don't. There are plenty of jokes about this. (A priest, a rabbi and and engineer are about to be guilotined. The executioner pulls the rope and nothing happens. "It is the will of god" and they let the priest go. The same with the rabbi. They lead the engineer up and he says "I think I see the problem".) Anyway, did you notice the long sensor arm sticking out to the side? That covers my earlier question about how the tractor is going to see the back of the trailer. Did you see the LIDAR (I hope so as they specifically mentioned it)? Etc.

Anyway putting together my observations on GPS, assisted driving in Teslas, the history of AI and computing etc. it is my prediction that the probability of seeing, in my lifetime, a CT that can be commanded to go find a parking space and park whilst pulling a trailer is vanishingly small. Note the use of the word "probability". That word has special meaning to engineers and others who use science. In using that word I do not declare this to be impossible but rather that betting on this would be a very bad idea.

You all are, of course, free to think what ever you want and that's OK. What you want won't come in 2 years or even 10 years or perhaps ever but it is a goal to strive for. Were it not for visionaries like Elon Musk we would never move forward. The AI computer in the Tesla is remarkable and can do incredible things relative to what could be done 10 years ago. But it has its limitations.

 

[Jhodgesatmb]

What one sees depends on one's perspective. My perspective is that of an engineer with experience in AI, radar,signal processing, computing and navigation who has driven vehicles with "FSD" for about 2 years now. Engineers see things that lay people don't. There are plenty of jokes about this. (A priest, a rabbi and and engineer are about to be guilotined. The executioner pulls the rope and nothing happens. "It is the will of god" and they let the priest go. The same with the rabbi. They lead the engineer up and he says "I think I see the problem".) Anyway, did you notice the long sensor arm sticking out to the side? That covers my earlier question about how the tractor is going to see the back of the trailer. Did you see the LIDAR (I hope so as they specifically mentioned it)? Etc.

Anyway putting together my observations on GPS, assisted driving in Teslas, the history of AI and computing etc. it is my prediction that the probability of seeing, in my lifetime, a CT that can be commanded to go find a parking space and park whilst pulling a trailer is vanishingly small. Note the use of the word "probability". That word has special meaning to engineers and others who use science. In using that word I do not declare this to be impossible but rather that betting on this would be a very bad idea.

You all are, of course, free to think what ever you want and that's OK. What you want won't come in 2 years or even 10 years or perhaps ever but it is a goal to strive for. Were it not for visionaries like Elon Musk we would never move forward. The AI computer in the Tesla is remarkable and can do incredible things relative to what could be done 10 years ago. But it has its limitations.

When I see what has been accomplished with ML in the past 2 years alone (and I am by no means a champion of ML), it is not beyond my imagination to see Tesla tackle parking with great success, and perhaps try trailer mechanations. Your reference to engineers as seeing things that others don’t is narrow. Some engineers do, yes. Engineers, like everything else, are variable in their understanding of physics and in their vision. I meet people trained in engineering every day that don’t know the first thing about how things work. It may well be that a truck driving a trailer and parking it might exceed your lifetime and, given your experience with FSD, that very experience may preclude seeing the technology with as open a mind as you suggest engineers in general have. Sometimes when I pull my trailer back to a parking space it is a very simple task. Pull straight forward from the ramp to an already empty space. FSD could do that now no problem. So moving a trailer will start in some minor way, have fits along the way and, eventually, become viable. It is our way, as engineers, to see the world as a continuum, not in all or none ways. Well, you’ve laid down the gauntlet and now we can only wait and see what happens and then evaluate your vision and skills of prediction.

 

[MEDICALJMP]

Wondering if trailer towing will need and adaptor kit for FSD use. Stick on camera and sensor modules that link to CT for safety.