Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Autopilot: Crashed at 40mph

This site may earn commission on affiliate links.
It assumes that an intelligent person is taking control and navigating the complex situation.
Maybe that is not a good assumption?

But I think it can calculate the amount of brake required no less accurately than an intelligent driver. The one that applies the brakes harder (driver or the car) should simply take effect. Is there any harm in that?

I don't have AP, so may be I am imagining it wrong. But the purpose of automatic emergency braking, from the ads I have seen on TV for other brands, seems to be that the intelligent driver was too slow to respond to an impending crash, and the system took action in time.

It is unnatural to expect the driver to not apply brakes at all if he is late. No matter how late, the driver will always spontaneously slam the brakes when he sees the collision coming.
 
Last edited:
  • Like
Reactions: calisnow and Ivo-G
But I think it can calculate the amount of brake required no less accurately than an intelligent driver. The one that applies the brakes harder (driver or the car) should simply take effect. Is there any harm in that?

I don't have AP, so may be I am imagining it wrong. But the purpose of automatic emergency braking, from the ads I have seen on TV for other brands, seems to be that the intelligent driver was too slow to respond to an impending crash, and the system took action in time.

It is unnatural to expect the driver to not apply brakes at all if he is late. No matter how late, the driver will always spontaneously slam the brakes when he sees the collision coming.
Emergency situations can get very complex (from a programming standpoint). While AP might be better able to calculate braking force than a human, what if the best course of action is not to brake but to take evasive action?
I believe that may driver safety people now say that it may be best to try to avoid a collision by taking evasive action rather than braking. If the car is braking hard while the driver is trying to take evasive action, it could lead to a sideways skid.
 
  • Like
Reactions: MP3Mike
Emergency situations can get very complex (from a programming standpoint). While AP might be better able to calculate braking force than a human, what if the best course of action is not to brake but to take evasive action?
I believe that may driver safety people now say that it may be best to try to avoid a collision by taking evasive action rather than braking. If the car is braking hard while the driver is trying to take evasive action, it could lead to a sideways skid.

Ah, I see what you mean. It is indeed a complicated design choice.

Although I still can't help imagining the possible scenario when the best course of action is to brake, and the driver realizing it a little too late. Let's say the car calculates it needs to apply 100 units of brake to minimize the impact, and begins applying it. A second later, the driver responds, and slams the brake. But it takes a fraction of a sec for the manual brake force to reach 100 from 0, while the automatic force of 100 falls to 0 as soon as the manual force is 1. This means there is a period of time when the brake force was not 100, while the computer calculated continuous force of 100 was necessary to minimize impact.

In most emergency situations, I imagine the first action of the driver would be to apply brakes - even if later than when the emergency braking began engaging. In other words, I see emergency braking being disengaged every single time.

That said, again, I do see your point.
 
I believe others did mention brake supplementation systems that are able to brake the car to a stop. However, I believe they only work at relatively low speeds (up to around 30mph), perhaps to avoid the risk of skidding as mentioned above.
 
The real problem isn't the AEB design. Or the TACC. Or the radar. Or the AP. Or even the driver.

The real problem is the fact that this system is good, very good, at making anyone using it think it has all bases covered. But there are a few situations that just aren't. And by not highlighting and educating owners about those situations, at least until they're fixed, Tesla is leaving its most valuable assets, their customers, hanging out to dry.

I now know from personal experience where this system can still fail. It was a very expensive and painful lesson to have to learn, and thankfully no one was injured learning this lesson, despite two pregnant ladies, 4 kids, and 5 cars being involved. All I'm now trying to do is create awareness and educate people to prevent this kind of mishap.

But what baffles me is how many would rather ignore the warning, place blame on the driver, think they would fare much better, than simply listen and learn from other's mishaps.

I've said all I can say about this subject, whoever decides to heed my warning will benefit from it, those who don't I wish Godspeed and Good luck.
 
I agree Ivo-G. Tesla does not emphasize enough that this is LANE KEEPING and DISTANCE CONTROL only.

In addition, Tesla seems to think that the car turning off AP automatically when its confidence is low is good enough. It isn't. The car doesn't turn off in many situations where it should. And sometimes when it does turn off, it leaves you literally a fraction of a second to react to avoid an accident. And finally, sometimes it turns off in a way that the driver is unaware that it turned off resulting in the car being driven by no one (happened to me twice so far).
 
  • Informative
Reactions: Bladerunner
I agree Ivo-G. Tesla does not emphasize enough that this is LANE KEEPING and DISTANCE CONTROL only.

In addition, Tesla seems to think that the car turning off AP automatically when its confidence is low is good enough. It isn't. The car doesn't turn off in many situations where it should. And sometimes when it does turn off, it leaves you literally a fraction of a second to react to avoid an accident. And finally, sometimes it turns off in a way that the driver is unaware that it turned off resulting in the car being driven by no one (happened to me twice so far).
It has always given me alert tones and message on the screen when it turns off... have you had a different experience?
 
  • Like
Reactions: JenniferQ
It assumes that an intelligent person is taking control and navigating the complex situation.
Maybe that is not a good assumption?

No, it's not a good assumption. Maybe you think you personally are perfect enough to make the right decision all the time, but I know Mercedes doesn't think most people are that perfect in their driving skills and their emergency decision making instincts - which is why their automatic emergency braking system adds braking force if you touch the brakes, but fail to press them hard enough to avoid hitting an object.
 
I don't have AP, so may be I am imagining it wrong. But the purpose of automatic emergency braking, from the ads I have seen on TV for other brands, seems to be that the intelligent driver was too slow to respond to an impending crash, and the system took action in time.
Tesla has been very consistent and specific in describing their AEB (automatic emergency braking) as a mechanism that attempts to minimize the damage of an imminent collision -- not as a mechanism to avoid a collision.

Some other manufacturers advertise their products as collision avoidance. I have no idea how accurate their depictions are.
 
Emergency situations can get very complex (from a programming standpoint). While AP might be better able to calculate braking force than a human, what if the best course of action is not to brake but to take evasive action?
I believe that may driver safety people now say that it may be best to try to avoid a collision by taking evasive action rather than braking. If the car is braking hard while the driver is trying to take evasive action, it could lead to a sideways skid.

Hard braking while turning could not lead to a sideways skid if the car has decent electronic stability control - the computer would prevent steering and braking input which would result in oversteer (what you are calling a sideways skid). Second, you can ask all kinds of "what if's" but the fact is the Tesla AEB system is inferior to what Daimler has in their cars right now. It might be better to steer away from the object - it also would be better to pay attention in the first place - but these systems are designed to compensate for error prone humans. The typical reaction of a person who fails to brake in time is to do at least *some* braking - and Mercedes recognizes this by compensating for people who do not brake hard enough to avoid a collision and by not turning off the AEB when a panicking driver adds some brake force of their own, but too late.

Tesla's AEB is a case of inferior programming and possibly inferior hardware, which I hope will be improved in the future. Tesla is a young company.

I believe others did mention brake supplementation systems that are able to brake the car to a stop. However, I believe they only work at relatively low speeds (up to around 30mph), perhaps to avoid the risk of skidding as mentioned above.

Mercedes new automatic emergency braking system in the 2017 E-Class will avoid collisions with stopped cars from as high as 90 kmph (55 mph). Tesla is simply behind.
 
  • Informative
Reactions: Ivo-G
Tesla has been very consistent and specific in describing their AEB (automatic emergency braking) as a mechanism that attempts to minimize the damage of an imminent collision -- not as a mechanism to avoid a collision.

True but this is simply a limitation of Tesla - Daimler claims its AEB in the 2017 E-class will completely avoid collisions with stopped traffic ahead at speeds of up to approximately 55 mph.
 
@iffatall
@calisnow

...(B)(A)

Some folks might remember another relatively recent thread with a Tesla driver moving from the B lane to the A lane when a car in front of him was "braking suddenly". The evasive maneuver chosen ended up causing additional drama with a nearby motorcyclist. But ignore the motorcyclist for a moment. Imagine the A lane had a car approaching (from behind). If AEB slammed on the brakes because of the obstruction car in lane B, then the maneuver to lane A would have included heavy braking which means the approaching car in lane A would be more likely to collide.
 
True but this is simply a limitation of Tesla - Daimler claims its AEB in the 2017 E-class will completely avoid collisions with stopped traffic ahead at speeds of up to approximately 55 mph.
You say limitation, I say design and marketing choice.


E-Class Sedan | Mercedes-Benz
When you click on the "(Disclaimer)" button, you are presented with the following:
COLLISION PREVENTION ASSIST PLUS may not be sufficient to avoid an accident. It does not react to certain stationary objects, nor recognize or predict the curvature and/or lane layout of the road or every movement of vehicles ahead. It is the driver's responsibility at all times to be attentive to traffic and road conditions, and to provide the steering, braking and other driving inputs necessary to retain control of the vehicle. Drivers are cautioned not to wait for the system's alerts before braking, as that may not afford sufficient time and distance to brake safely.

Granted it's probably lawyers just being careful, but that's a lot of disclaimer.


I prefer Tesla's approach here: underpromise.


More details on MB offerings here:
- NonPlus
Mercedes-Benz TechCenter: COLLISION PREVENTION ASSIST
- Plus
Mercedes-Benz TechCenter: COLLISION PREVENTION ASSIST PLUS

Regarding the latter, the phrasing here is "may be sufficient" rather than the "may not be sufficient" in the disclaimer. Interesting because the "may be" language to me comes across as even less committal than the disclaimer.


It then goes on to say...
a collision is no longer avoidable, the braking process will considerably reduce the effects of the accident.
... which is further indication that, despite the name, it's not promising collision avoidance. In fact the phrasing in the second half of the sentence sounds remarkably similar to Tesla's phrasing.

Model S Owner's Manual:
The forward looking camera and the radar sensor are designed to determine the distance from any object (vehicle, motorcycle, bicycle, or pedestrian) traveling in front of Model S. When a frontal collision is considered unavoidable, Automatic Emergency Braking is designed to automatically apply the brakes to reduce the severity of the impact.
 
Last edited:
Quick follow-up, some more info on when AEB will not apply from the manual:

Automatic Emergency Braking does not apply the brakes, or stops applying the brakes, in situations where you are taking action to avoid a potential collision. For example:
• You turn the steering wheel sharply.
• You press the accelerator pedal.
• You press and release the brake pedal.
• A vehicle, motorcycle, bicycle, or pedestrian, is no longer detected ahead.

Note the 3rd point. If you press the brakes, that doesn't cancel AEB. Press and release of the brakes does.
 
More details on MB offerings here...the phrasing here is "may be sufficient" rather than the "may not be sufficient" in the disclaimer... which is further indication that, despite the name, it's not promising collision avoidance.

The 2017 E class is completely new generation tech with Benz claiming capabilities far above what is currently on sale in current Benzes - the 2017 goes on sale in a month or two - if you do some research you can find the descriptions of its capabilities. I'll see if I can hunt some down and post them here. With the exception of accurate lane keeping, Tesla is about to be completely outclassed in active safety once again by Daimler when the 2017 E class goes on sale. I'm sure Tesla is aware of this and is working hard behind the scenes on its own next generation active safety systems.

The AEB from 55 mph collision avoidance is not on sale in any current Benz nor described on the current Benz website. I'll try to find the links.

However even the current Benz ADAS systems are ahead of Tesla's offerings (with the exception of lanekeeping - in which Tesla does seem to have the world lead in terms of accuracy and ability to hold lanes in difficult conditions with poor markings).

I have a 2016 70D and love it - but the level of knee jerk defense of Tesla around here gets to be kind of silly.

Mercedes has a long, long lead and a very deep research bench in the development of active safety technologies - it is no surprise that they would be ahead of Tesla in many ways despite Tesla's remarkable success with bringing world beating, best-in-class-accuracy lane keeping to market.
 
  • Informative
Reactions: Bladerunner