Model 3 crashed into a pole

[Daniel in SD]

It should require as much force to take over the steering in an autonomous car as it would to physically subdue an Uber driver.

Meaning that the car can exert enough force on the steering wheel that it would be impossible for many people to override it? Anyway, hopefully they have enough failsafes to make this failure mode impossible.
My point is that most people wouldn't think of trying to grab the steering wheel of a car being driven by human. I'd wager that many more people would try to yank the steering wheel of an autonomous vehicle just to see what would happen.

 

[sreams]

Meaning that the car can exert enough force on the steering wheel that it would be impossible for many people to override it? Anyway, hopefully they have enough failsafes to make this failure mode impossible.
My point is that most people wouldn't think of trying to grab the steering wheel of a car being driven by human. I'd wager that many more people would try to yank the steering wheel of an autonomous vehicle just to see what would happen.

Probably the same kind of people who would open the door and jump out just to see what would happen.

 

[tpatana]

And if the front wheels are losing traction, turning more (like you say you did) even makes it worse and more likely to push forward.

 

[Kant.Ing]

Sorry to see OP's loss.

TACC should have been disengaged by OP's brake at the freeway exit.

I guess OP was going too fast (taking the green light) on a sharp left turn. The intersection doesn't seem to have any pavement markings for guiding possibly 2 turning lanes so OP might have understeered. Jm2c.

 

[Daniel in SD]

Probably the same kind of people who would open the door and jump out just to see what would happen.

Apparently you haven’t read stories about what people do in Ubers when there is a driver in the car

 

[argon2018]

Hi everyone,

Last Thursday, I was driving from my work to home on the usual route. Autopilot was engaged on highway and as I was approaching the exit to my home, I applied brakes to slow down and took control of the vehicle.

I passed the 1st traffic signal and had to make a left at the 2nd signal. The signal was green so I just went and tried to turn the steering wheel to the left without luck, tried turning it fully but the next thing I realize is the airbag being deployed and I hit a pole. I noticed that regenerative braking did not kick in, and I always have it ON.
The car was not responsive to the turn which resulted in the car crashing into the pole. I also observed that the left front wheel had a flat tire, the accident was only to the right side of the car and no impact on the left side. There was absolutely no indication of a flat tire or low tire pressure indication before or after the crash.

This clearly indicates that there was some software/hardware malfunction in the car during the time of the incident. I am still in shock and scared to sit in a Tesla now.

Questions to be answered:

1. Why did the Tesla NOT detect the pole and brake automatically?
2. Why did regenerative braking NOT kick-in when I was making the turn?
3. Why was there NO alert on the screen from TPMS about low tire pressure before or after the crash?

I'm still waiting for the Tesla team to respond to my questions above, hope they investigate and provide a detailed/unbiased report. Meanwhile, I wanted to check if anyone faced something similar? Any inputs would be helpful.

Thanks

View attachment 397963 View attachment 397964 View attachment 397959

"The car was not responsive to the turn which resulted in the car crashing into the pole"

The sooner you accept that you caused the accident, the sooner you can move on. You were prob going too fast and lost control. The car didnt drive itself into the pole.

You hit a pole and you look at tire pressure and whether reg breaking kicked in? stop justifying and rationalizing why this accident occured. Just learn from your mistakes and drive more carefully. Be grateful that you didn't hit a pedestrian or another car and the fact that you didn't have any injuries. Dont waste your time blaming the car.

 

[animorph]

If the left tire quickly went flat before the left turn I think that would explain everything. The driver didn't need to stop before the left turn, so the speed was probably higher than usual. At that speed with a flat tire the car couldn't make the turn. I'm pretty sure regen would have a big problem with a flat tire as well. Traction control would have tried to keep the left front from "skidding" and reduced braking on the three good tires, so no real deceleration. ABS did that to me after a spin on a track day. All pretty reasonable. Although a TPMS warning would have been nice, it may not have been noticed or the flat may have occurred just before or during the turn.

 

[alsetym]

Sorry OP for your loss, but contrary to your hypothetical, the steering wheel is mechanically connected to the wheels. It is NOT I repeat NOT steer by wire. Unless your steering column snaps in half, you turn the steering wheel, the wheels turn with it. The computer can try to fight you, but it isn't very strong and that is not what you describe. A responsible person would correct their FUD in the first post, you'll scare the new guys lol.

 

[ggnykk]

Hi everyone,

Last Thursday, I was driving from my work to home on the usual route. Autopilot was engaged on highway and as I was approaching the exit to my home, I applied brakes to slow down and took control of the vehicle.

I passed the 1st traffic signal and had to make a left at the 2nd signal. The signal was green so I just went and tried to turn the steering wheel to the left without luck, tried turning it fully but the next thing I realize is the airbag being deployed and I hit a pole. I noticed that regenerative braking did not kick in, and I always have it ON.
The car was not responsive to the turn which resulted in the car crashing into the pole. I also observed that the left front wheel had a flat tire, the accident was only to the right side of the car and no impact on the left side. There was absolutely no indication of a flat tire or low tire pressure indication before or after the crash.

This clearly indicates that there was some software/hardware malfunction in the car during the time of the incident. I am still in shock and scared to sit in a Tesla now.

Questions to be answered:

1. Why did the Tesla NOT detect the pole and brake automatically?
2. Why did regenerative braking NOT kick-in when I was making the turn?
3. Why was there NO alert on the screen from TPMS about low tire pressure before or after the crash?

I'm still waiting for the Tesla team to respond to my questions above, hope they investigate and provide a detailed/unbiased report. Meanwhile, I wanted to check if anyone faced something similar? Any inputs would be helpful.

Thanks

View attachment 397963 View attachment 397964 View attachment 397959

I have yet to see the automatic braking function work when Tesla car is not in autopilot. In other words, you should expect the auto braking function only work in autopilot mode or dynamic cruise control mode. In other situation such as manual driving, the activation of auto braking is much less clear.

 

[Daniel in SD]

Sorry OP for your loss, but contrary to your hypothetical, the steering wheel is mechanically connected to the wheels. It is NOT I repeat NOT steer by wire. Unless your steering column snaps in half, you turn the steering wheel, the wheels turn with it. The computer can try to fight you, but it isn't very strong and that is not what you describe. A responsible person would correct their FUD in the first post, you'll scare the new guys lol.

Have you ever tried to steer a 4000lb car with 2 turns lock to lock steering without power? The steering motor is extremely strong, I would wager stronger than many Model 3 drivers and especially if you’re not expecting it. I don’t think that’s what happened but to say it’s not physically possible is silly.

 

[Leafdriver333]

Surely, if the OP sues Tesla, Tesla will do full analytics from the computer and explain what really happened.

 

[tpatana]

Have you ever tried to steer a 4000lb car with 2 turns lock to lock steering without power? The steering motor is extremely strong, I would wager stronger than many Model 3 drivers and especially if you’re not expecting it. I don’t think that’s what happened but to say it’s not physically possible is silly.

That's power steering. Is the Tesla autonomous steering done with same mechanics? I thought they were separate systems. One to do the power steering, another to do the autonomous steering. But I haven't studied that, so I could be wrong.

And yes, I had Audi 100 (pretty heavy, don't recall actual weight) without power steering. You should have seen my face the first time I drove car with power steering.

The more I think about the OP case, the more I'm certain he just cornered too fast, tires started slipping which causes the reg braking to disable. When front is pushing, turning even more just makes it more certain to push forward. Baam.

Like most accidents, the problem is the loose nut behind the wheel.

 

[Daniel in SD]

That's power steering. Is the Tesla autonomous steering done with same mechanics? I thought they were separate systems. One to do the power steering, another to do the autonomous steering. But I haven't studied that, so I could be wrong.

Yes, from the pictures there's just one motor and it's connected directly to the rack.I would assume that any electric power steering system would have some sort of failsafe to prevent the motor from overpowering a manual input.
I've got a 1600lb car without power steering and it's not easy to turn the wheel when the car is not moving.

 

[insaneoctane]

There are lots of possible theories offered in this thread.
Without video and/or Tesla's log data, it's very hard to understand what really happened.
I hope Tesla can help provide you more details and that you share that with us here in this thread.

 

[DrDabbles]

I’m thinking about your steering — could be you had a wire come loose from the steering column to have it not be turntable like that. But having this at the same time as the lack of regen doesn’t make sense. More likely a common failure (or feature) caused both...

A wire coming loose would make zero difference. There is still a mechanical linkage between the steering column and the steering rack. If you lose power to or control of the electric steering motor, you can still turn the wheel. Just like in a hydraulic assisted steering system we've been using for decades now.

OT and playing word games here but are headlines like “737 Max 8 crashes killing all passengers” or “Oil refinery explodes killing 8 workers” similarly problematic. Do they give agency to the airplane and oil refinery?

Generally speaking, yes. The plane did crash, but the phrasing seems off a bit. The plant wasn't flying itself, though with the MAX series that line is obviously blurred in the headlines.

Regen shutting down has happened to me a couple times as well. Battery was below 85% both times.

But the steering disengaging is what I don't understand. Even in AP, Driver steering input is always honored.
If it randomly disengages, that would be the worst car safety issue.
Or if a hacker can remotely disable it...

Regen power decreases when the wheel speed sensors detect wheel slip. So hitting a wet expansion joint on a bridge, or an icy patch on the road, or anything slippery for that matter can cause regen to significantly decrease momentarily.

Driver input is always "honored" because there's no alternative. There's a gear at the end of the steering column that drives a rack on the steering mechanism. The motor assists the driver under normal circumstances. But there are multiple safety measures that allow the driver to always override any outside electronic control of the steering system. This technology is decades old at this point, and very thoroughly tested and widely used. This isn't some Tesla invention that they decided to experiment with, these are off the shelf components produced for a wide array of vehicles.

Regardless, nothing can "disable" the driver from physically turning the wheel while the vehicle is in motion, and that having an effect on the car

Ok so this is super concern trolly but here it goes...
It seems like for the car to operate autonomously on the "Tesla Network" there must be a way to prevent someone riding in the front seat from steering the car.

Nope. Just the camera watching them, and the fact that the car would probably be programmed to stop driving.

The motor connected to the steering rack is quite large and likely powerful enough to override driver input.

Also no. There are safety systems in place, and regulations around this equipment. This is widely used hardware and present on countless vehicles at this point.

Again though it seems like the most likely scenario is that OP was following a car and TACC applied the brakes.

You give startled people too much credit. It's much more likely that OP doesn't remember what happened exactly, never hit the brakes, and just crashed into a pole. How many posts like this have you and I read here over the years? How many people blamed their Model S for just driving itself through a building? We know what happened here. We don't need to try to create elaborate scenarios where an otherwise ace level professional driver mistook the subtle feedback of the braking system, and the steering assist mechanism wrenched the wheel out of their hands.

 

[Kilotango74]

I'm going with going to fast to make the corner. Been there done that.

 

[GolanB]

I have yet to see the automatic braking function work when Tesla car is not in autopilot. In other words, you should expect the auto braking function only work in autopilot mode or dynamic cruise control mode. In other situation such as manual driving, the activation of auto braking is much less clear.

My car stopped and prevented me from hitting a car in front of me while AP was not engaged. His car came to a sudden stop (a school bus was making a wide left turn in front of him) for an upcoming intersection. My model 3 stopped so hard that I though I had hit him, but we were still a foot or so apart - all was good. At the next light, the driver of the car (an Audi A8) noticed me and has a ton of questions about the Tesla. He had no idea what had happened a moment before and I didn’t have time to discuss it before the light changed.

 

[KenC]

Turning left, with a left front tire going down shouldn't be so bad as to cause an accident, as the car gets light on the front left, and heavy on the front right. Also, it looks like when the car crashed into the pole, it may have distorted the whole axle, so that the left front tire got damaged by the wheel well, so a post-accident effect.

 

[SR22pilot]

Yes, from the pictures there's just one motor and it's connected directly to the rack.I would assume that any electric power steering system would have some sort of failsafe to prevent the motor from overpowering a manual input.
I've got a 1600lb car without power steering and it's not easy to turn the wheel when the car is not moving.
View attachment 398443

The key wording is "not moving." I bet it steers just fine when moving. I have driven many cars with manual steering. If they are moving, steering isn't a problem. It takes more effort and the use of both hands but it isn't a problem.

 

[Candleflame]

The key wording is "not moving." I bet it steers just fine when moving. I have driven many cars with manual steering. If they are moving, steering isn't a problem. It takes more effort and the use of both hands but it isn't a problem.

exactly. Even superheavy cars turn ok w/o power steering as long as you got at least a few km/h speed.

The OT either had a seizure, severe head injury with amnesia or he is just mad that he tried to floor the car in the city and subsequentially lost control...