Auto-park fail - manually stopped just in time

[Nathanael]

So yesterday was my second time trying out a parallel auto park.
The first time worked perfectly, so I knew what to expect.

This was your standard parallel park on a quiet residential street. Lots of space between parked cars, quiet road, parking spot on the right side of the model 3.

As the car was backing up into the space and doing its final "counter clockwise" turn to tuck into the spot, I noticed the front right bumper was getting wayyyyy to close to the truck infront of me, and it was not stopping. I manually used the brakes to cancel auto park at the last minute. The front right corner of the bumper was ~20 - 30cm away from bumping a black truck. I got out to look, and there was no way the auto park would of successfully made it, unless it would of stopped and corrected itself, which it honestly felt like it was not going to happen.

I googled to see if anyone else had a problem, and I found a video of pretty much the exact same thing that happened to me, it was also a truck (only difference was it was on the right side of the car)

Has this happened to anyone else? Kind of disappointed that my auto park is only at a 50% success rate.

 

[PoitNarf]

Yikes. I've only used auto park a handful of times. I never seem to be able to get it to notice there is a parking spot. Either that or I just don't have the patience for it to come up on the screen and I just park myself anyways. Definitely going to be paying closer attention the next time I do engage auto park!

 

[Nathanael]

im curious if it had anything to do with the truck? Is the rear bumper / back end on a pick up truck too high for the M3 sensors? It seems like tesla would have thought about this though...at least I hope so

 

[chronopc]

So yesterday was my second time trying out a parallel auto park.
The first time worked perfectly, so I knew what to expect.

This was your standard parallel park on a quiet residential street. Lots of space between parked cars, quiet road, parking spot on the right side of the model 3.

As the car was backing up into the space and doing its final "counter clockwise" turn to tuck into the spot, I noticed the front right bumper was getting wayyyyy to close to the truck infront of me, and it was not stopping. I manually used the brakes to cancel auto park at the last minute. The front right corner of the bumper was ~20 - 30cm away from bumping a black truck. I got out to look, and there was no way the auto park would of successfully made it, unless it would of stopped and corrected itself, which it honestly felt like it was not going to happen.

I googled to see if anyone else had a problem, and I found a video of pretty much the exact same thing that happened to me, it was also a truck (only difference was it was on the right side of the car)

Has this happened to anyone else? Kind of disappointed that my auto park is only at a 50% success rate.

I've used auto park 3x so far and it's been working well for me. It does tend to get super close but at least from my experience it will stop in time.

 

[Zaphod]

im curious if it had anything to do with the truck? Is the rear bumper / back end on a pick up truck too high for the M3 sensors? It seems like tesla would have thought about this though...at least I hope so

Yes, exactly this reason. The field of view for the ultrasonic sensors tend to be fairly narrow and taller bumpers on trucks and SUVs are not always detected.

 

[WileyTheMan]

I only tried the autopark thing one time. The car did a good job until the end, when it decided my rim needed a good scrape on the curb. I'm good with my own parking skills for the time being.

 

[Pkmmte]

After reading this thread, I am now happy that my AutoPark rarely ever detects a real parking space.

 

[SOULPEDL]

Lucky for me, I can't seem to get it to find a spot in the first place. Only once so far, but I hate holding up traffic behind me while I sneak forward praying for a P. I'd rather just whip it in myself.

Could be a classic case of Tesla hiring smart, young designers with no real-world experience, who knows. FYI, trucks may also have something hanging out the back with a flag on it. And don't forget giant hitches too, but that's probably in sonar sight.

 

[chronopc]

I have tried using it for parallel parking yet (maybe I won't after reading this thread), but it works great backing in to a space between 2 cars.

 

[Wolverinegeoff]

I only tried the autopark thing one time. The car did a good job until the end, when it decided my rim needed a good scrape on the curb. I'm good with my own parking skills for the time being.

Had the same thing happen to mine. Autopark is dead to me.

 

[JeffK]

In a perfect world, it'd use the cameras to create a virtual 3D environment which it could map and "memorize" before the parking job is initiated and then use the ultrasonics as more of a redundant guide in case of unexpected changes. This is however not the case currently and who knows if that's a planned thing.

It seems right now the ultrasonics are the main decision makers and they are fallible, especially with raised truck bumper heights. As a human it's your job to eyeball this ahead of time before an attempt (and during of course).

In the future, it'd be nice if they either used a vision mapping technique and/or even included downward facing 180° cameras in the mirrors like some other manufacturers. It would help eliminate the limitations of just using ultrasonics.

 

[RDoc]

In a perfect world, it'd use the cameras to create a virtual 3D environment which it could map and "memorize" before the parking job is initiated and then use the ultrasonics as more of a redundant guide in case of unexpected changes. This is however not the case currently and who knows if that's a planned thing.

It seems right now the ultrasonics are the main decision makers and they are fallible, especially with raised truck bumper heights. As a human it's your job to eyeball this ahead of time before an attempt (and during of course).

In the future, it'd be nice if they either used a vision mapping technique and/or even included downward facing 180° cameras in the mirrors like some other manufacturers. It would help eliminate the limitations of just using ultrasonics.

Agreed. Until the car really starts making a model of what's around it, including things like low curbs that vanish to the sensors when the car is too close, this is all pretty fraught.

 

[SomeJoe7777]

Ya, truck bumper height was the culprit here. Ultrasonics didn't detect the bumper corner because it was too high.

I know Tesla wants to do all of this without Lidar, but we're now starting to see just how many shortcomings that approach has.

 

[Moderatefan]

He elaborared in comments:

The way the system works is not as you think : while you drive below a certain speed (35kph), the system constantly scans on each side to detect what appears to be a parking spot. It uses the ultrasonic sensors to do so. It detects an object at 1m... that's the car behind the spot.. then it's 3m (the concrete on the side of the parking) then it's 1m again (the truck in front). It uses the number of wheel rotations to calculate the spot's length and builds a model of the parking spot. It then shows you a "Parking available" logo in the instrument cluster. You put it in reverse and it asks, in a pop up showing the car, in real time, relative to the spot it detected "Autopark?". You hit go and it starts the maneuver. AT that point the system is purely working on what it "saw" when you drove past the parking spot. It does use its sensors to detect pedestrians / animals / another car stealing your spot but since it's purely reactive, what it can do is very limited. The car does have 360degres of coverage on the ultrasonic but it's 180 on the front bumper and 180 on the rear bumper. Nothing where the doors are. Not sure having more sensors would have helped since it seems I was driving to close the the car parked to my left and the rear part of the truck's bed was angled toward the ground, probably messing with the sensors. So.. too close and truck too high with angled bed = the car though the truck ended at its wheel and started its maneuver too early.

And then Tesla response:

they won't cover the damages, as I was expecting.

I personally don't feel good about this.
First, autopark is not Beta. So, you buy EAP, but can't count on it to do what you're paying for?

Second, it's been 2 years since he posted that video. Nothing fixed? Could they at least disable parking next to trucks or something?
Why is the blame entirely on the driver here, I don't quite understand.

 

[adaptabl]

It can't auto park safely but will soon be ready for autonomous driving. Fat chance that's going to happen.

 

[JeffK]

He elaborared in comments:
I personally don't feel good about this.
First, autopark is not Beta. So, you buy EAP, but can't count on it to do what you're paying for?

It does say in the manual that you are responsible for it.

Second, it's been 2 years since he posted that video. Nothing fixed? Could they at least disable parking next to trucks or something?
Why is the blame entirely on the driver here, I don't quite understand.

The driver has been warned in the manual, and the placement of the ultrasonics is fairly limited, there's not a whole lot that can be done. It's also an assistance feature meant for human supervision, it's not a fully autonomous feature.

 

[insaneoctane]

Reading this thread leads me to conclude that auto park is a useless feature that I should avoid. That makes it a "gimmick" in marketing terms. I honestly feel a bit duped by this. The time that AP would be most useful to me is when it is tight and difficult for me. The fact that it may hit the front car if it's bumper is too high or hit the curb which will most certainly result in damaging the wheel on my de-aeroed 18"....um, no thanks. I appreciate this thread because while I would have watched AP performance while using it, I will not bother using it now. Quite a shame IMO

 

[Moderatefan]

It does say in the manual that you are responsible for it.

The driver has been warned in the manual, and the placement of the ultrasonics is fairly limited, there's not a whole lot that can be done. It's also an assistance feature meant for human supervision, it's not a fully autonomous feature.

Ok, thank you for that reference, did not have a chance to read the manual, since don't have the car yet.
However, I agree with @insaneoctane that if you can't count on a feature, its best not to tempt yourself that it might work most of time, but avoid altogether.

Well, I guess we wait until August when Elon said exponential improvements to EAP should start coming out.

Currently, I am very conflicted on whether I want to spend $5k on the EAP in its current state.

 

[TT97]

Ya, truck bumper height was the culprit here. Ultrasonics didn't detect the bumper corner because it was too high.

I know Tesla wants to do all of this without Lidar, but we're now starting to see just how many shortcomings that approach has.

LIDAR is terrible at proximity detection. While I understand people wanting LIDAR to replace the cameras, it would be a downgrade to the ultrasonic sensors for close range.

 

[JeffK]

Ok, thank you for that reference, did not have a chance to read the manual, since don't have the car yet.

I don't either, but I'm too excited so I got bored.
https://www.tesla.com/content/dam/tesla/Ownership/Own/Model 3 Owners Manual.pdf