Autoparking without Autopilot Hardware?

[ChooseAnother]

Simple: Why we need Autopilot HW for a simple autoparking feature?

For example, BMW i3's parking sensors aren't better than Model S's (probably worse). But it still can "auto-park".

Why Model S Non-AP can't?

 

[gabeincal]

Simple: Why we need Autopilot HW for a simple autoparking feature?

For example, BMW i3's parking sensors aren't better than Model S's (probably worse). But it still can "auto-park".

Why Model S Non-AP can't?

Because to develop a system you choose certain requirements. Since the AP camera was available to Tesla, they chose to develop their IP using the cameras and in the process, they made AP even more vital to be purchased. I agree with you, it could be done using sensors just like other manufacturers did. Tesla chose not to. Same with automatic headlights. Same with rain-sensing wipers.

 

[ChooseAnother]

Because to develop a system you choose certain requirements. Since the AP camera was available to Tesla, they chose to develop their IP using the cameras and in the process, they made AP even more vital to be purchased. I agree with you, it could be done using sensors just like other manufacturers did. Tesla chose not to. Same with automatic headlights. Same with rain-sensing wipers.

Shameful.

 

[gabeincal]

Shameful.

You could say that... I probably used auto-park a handul of times.

 

[ChooseAnother]

You could say that... I probably used auto-park a handul of times.

Yes i know, probably i would be like this too but the thing is: Tesla's approach.

 

[MarcusMaximus]

Simple: Why we need Autopilot HW for a simple autoparking feature?

For example, BMW i3's parking sensors aren't better than Model S's (probably worse). But it still can "auto-park".

Why Model S Non-AP can't?

Because implementing and maintaining two separate implementations of that feature using two completely different sets of sensors and data would be a huge waste of resources.

 

[SabrToothSqrl]

Auto Park would be hard for us fossil cars with zero sensors.

Also, my understanding is that auto park only works in the very conditions of which I would never park... near enough to another car that would make me worry about other people damaging it with car doors...

 

[whitex]

Because implementing and maintaining two separate implementations of that feature using two completely different sets of sensors and data would be a huge waste of resources.

Unless of course not implementing separate system result is a Swiss army knife solution, does everything but nothing well, all-in-one solution. For example, Tesla's blind spot detection is about as functional and practical as a Swiss army knife saw blade is. AP2 auto-headlights just turn the lights on most of the time, even perfectly sunny weather, but they can put a check-mark next to the feature - "yes, we got it". The list goes on...

 

[MP3Mike]

Simple: Why we need Autopilot HW for a simple autoparking feature?

For example, BMW i3's parking sensors aren't better than Model S's (probably worse). But it still can "auto-park".

Why Model S Non-AP can't?

The cars without AP hardware don't have electronic brakes, so there is no way for it to stop the car.

They added the iBooster brakes when they added the AP hardware.

There could also be the issue of the required processing power wasn't available in the car, but really no electronic brakes=no auto-park.

 

[MarcusMaximus]

Unless of course not implementing separate system result is a Swiss army knife solution, does everything but nothing well, all-in-one solution. For example, Tesla's blind spot detection is about as functional and practical as a Swiss army knife saw blade is. AP2 auto-headlights just turn the lights on most of the time, even perfectly sunny weather, but they can put a check-mark next to the feature - "yes, we got it". The list goes on...

That's not really what's being discussed here. It's not about implementing a single separate system to handle auto-park in all cars, but rather implementing and maintaining two different versions of auto-park: one that uses the AP hardware and another that uses only other sensors(presumably ultrasonics) for non-AP cars.

I'll refrain from comment on your complaints so as not to hijack the thread.

 

[whitex]

That's not really what's being discussed here. It's not about implementing a single separate system to handle auto-park in all cars, but rather implementing and maintaining two different versions of auto-park: one that uses the AP hardware and another that uses only other sensors(presumably ultrasonics) for non-AP cars.

If you have a separate system that performs a function well, there is no need to create or maintain an AP hardware based equivalent.

I'll refrain from comment on your complaints so as not to hijack the thread.

Agreed, I was just using those as well known examples, not meaning to hash them out here. However, if you disagree, feel free to comment on relevant threads like "Blind Spot Detection is disappointing" which includes a poll on how many people are happy with they blind spot warning functionality.

 

[MarcusMaximus]

If you have a separate system that performs a function well, there is no need to create or maintain an AP hardware based equivalent.


Agreed, I was just using those as well known examples, not meaning to hash them out here. However, if you disagree, feel free to comment on relevant threads like "Blind Spot Detection is disappointing" which includes a poll on how many people are happy with they blind spot warning functionality.

That first part is possibly true, but it also requires both that you expect it to be of equal quality and requires less engineering effort. Right now implementing that other system would be superfluous and possibly/probably lower quality.

And, without looking through it, I think I've already responded to that thread. At least I've talked about it elsewhere and gone back and forth quite a bit about it with oktane.

 

[tuanyes84]

I use the auto park feature quite often, it has been very useful parallel parking in the city!