-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Sudden Unexpected Acceleration today
- Thread starter AZM3
- Start date
pilotSteve
Active Member
EXCELLENT if quite nerdy/technical article, thanks. I have real time, embedded sw development experience including avionics development under DO-178 safety critical methodology, and this changed my view from “no way could decent dev practices allow this sort of errors” ... to WHOA look at all the possible failure points and even more scary... all the corners that easily can be cut that stack up to a disaster waiting for the wrong circumstances to lime up.
Opened my eyes!!!
Im really glad you found it informative! It’s one of the deepest PowerPoint presos I’ve ever seen. (Which says what I think about PPTs!). I keep it at the ready to share anytime we review code quality and best practices.
It gives you some good insight into the auto industry as a whole ... and when Elon says things like “car companies are just not good at software” ... this is EXACTLY what he means. And a big part of why I think Tesla is so far ahead. They’re not perfect, but you read stuff like this and it’s really like the Flintstones and the Jetsons.
Spacep0d
Active Member
I'm new here, but I'm deeply skeptical.
1. First, the OP is reporting second-hand information and has an interest in protecting or vouching for his wife and her story (even if she believes it). Wife claims it was the car but I don't believe that. I think she thought she was hitting the brake but was really hitting the gas, and panic combined with a fight or flight response was no help. This type of thing happens with all makes of cars and humans predictably make errors. Women sometimes wear shoes which don't give them proper feel or can actively interfere with pedal-use, though this may have not been an issue in this case. Was she wearing heels or wedges or some other footwear not ideal for driving?
2. Humans will sometimes lie rather than admit mistakes, or they may blame someone else for mistakes they've made. Some females may make a false allegation of rape to cover up an unintended pregnancy, an affair, or a consensual relationship which has soured in some way, so naturally some people will risk ruining the reputation of a company to cover a simple mistake. It's the Cartman effect. Rather than be nice and get invited to La Casa Bonita, Cartman the sociopath opts to simply convince Butters that the Earth has been hit by a meteor where most people have been killed. He then proceeds to keep Butters in a fallout shelter whilst he takes Butters' place for La Casa Bonita....until he's found out.
Isn't this why some people run from the scene of an accident, or obviously cause accidents due to their own carelessness or error and then blame the victim? Some people genuinely believe they didn't cause an accident they clearly caused to the outside observer. Humans are not always reliable with memory, evidence, truthfulness, or interpretation of facts (which is why some people are Flat Earthers or believe other obviously false ideas). I would be suspicious of anyone claiming a car such as a Tesla simply accelerated on its own without thoroughly investigating human error. The unlikely scenario (car accelerates on its own) should not be accepted blindly over the likely scenario (human error). I am even more suspicious of a second-hand retelling of the claim by someone with a vested interest in supporting for the story (with penalties for not believing it).
3. Sometimes, humans sincerely believe in erroneous ideas, so they're not lying but they're sincerely incorrect. What a driver self-reports in a case like this matters less than what the sensor data reveals. Was the accelerator applied? Does the sensor data corroborate the story? I will believe that data over a claim for someone who has every cause to be embarrassed (and covering up) or sincerely wrong about what they did.
1. First, the OP is reporting second-hand information and has an interest in protecting or vouching for his wife and her story (even if she believes it). Wife claims it was the car but I don't believe that. I think she thought she was hitting the brake but was really hitting the gas, and panic combined with a fight or flight response was no help. This type of thing happens with all makes of cars and humans predictably make errors. Women sometimes wear shoes which don't give them proper feel or can actively interfere with pedal-use, though this may have not been an issue in this case. Was she wearing heels or wedges or some other footwear not ideal for driving?
2. Humans will sometimes lie rather than admit mistakes, or they may blame someone else for mistakes they've made. Some females may make a false allegation of rape to cover up an unintended pregnancy, an affair, or a consensual relationship which has soured in some way, so naturally some people will risk ruining the reputation of a company to cover a simple mistake. It's the Cartman effect. Rather than be nice and get invited to La Casa Bonita, Cartman the sociopath opts to simply convince Butters that the Earth has been hit by a meteor where most people have been killed. He then proceeds to keep Butters in a fallout shelter whilst he takes Butters' place for La Casa Bonita....until he's found out.
Isn't this why some people run from the scene of an accident, or obviously cause accidents due to their own carelessness or error and then blame the victim? Some people genuinely believe they didn't cause an accident they clearly caused to the outside observer. Humans are not always reliable with memory, evidence, truthfulness, or interpretation of facts (which is why some people are Flat Earthers or believe other obviously false ideas). I would be suspicious of anyone claiming a car such as a Tesla simply accelerated on its own without thoroughly investigating human error. The unlikely scenario (car accelerates on its own) should not be accepted blindly over the likely scenario (human error). I am even more suspicious of a second-hand retelling of the claim by someone with a vested interest in supporting for the story (with penalties for not believing it).
3. Sometimes, humans sincerely believe in erroneous ideas, so they're not lying but they're sincerely incorrect. What a driver self-reports in a case like this matters less than what the sensor data reveals. Was the accelerator applied? Does the sensor data corroborate the story? I will believe that data over a claim for someone who has every cause to be embarrassed (and covering up) or sincerely wrong about what they did.
I read the whole thing and it’s basically a red herring. Toyota’s code was sloppy but there’s no evidence it actually caused unintended acceleration. Fixing what the author thought needed improvements would not have prevented accidents.
Never said it was a smoking gun or any kind of definitive cause.
What it is, however, is a great insight into the software development lifecycle process Toyota uses (or, frankly, doesn’t.)
Even after decades of computerized vehicle controls, the manufacturers have lots to learn.
Standard coding for PCMs have a non-maskable interrupt for critical errors such as throttle position sensor not matching engine status. They override all bus commands, and jump right to the subroutine. Standard TPS's have a push-me-pull-you pair or triple set of sensors, like 0v = 0%/5v = 100% throttle paired with 0v = 100%/5v = 0% throttle. That way neither an open circuit nor a short can cause SUA.
Fly by wire throttles are very old now. 1997 was my first one, so 23 years or more. They are far safer than mechanical linkages.
Fly by wire throttles are very old now. 1997 was my first one, so 23 years or more. They are far safer than mechanical linkages.
MagnusMako
Member
I was sure I would never hit both pedals at the same time as I have never done this, they are very close on an S. Guess what, I did it many times getting the warning beep. Over the years I have seen so many people hit the gas and be 100% certain they were hitting the brake. The brakes on a Tesla will stop the car if you are hitting them even if the throttle was engaged. Some reasons we are seeing more of this is the power of an EV and the lack of engine noise coupled with the media fear mongering. The lady that drove into the gym, the guy that drove into this living room, the BOLT that took off on its own while parked, etc. I can see how people swear and think they are hitting the brakes but they are not. Tesla has much larger safety issues that are not completely surfacing like defective suspension parts and faulty ball joints that sheer off. Anyone purchasing cars this powerful should have to watch a 5 min video on the car screen to activate their keys because there are a bunch of really bad drivers and really stupid people driving high-power Teslas, Spend time at an SC and you can see them trying to park and it is very scary that they are actually driving these cars on a daily basis.
Hank42
Legacy Supporter
I am in the camp of user error to blame for all of this UIA non-sense. I experienced UIA several times, and I *know* it was user error on my part - in my case it was exactly the same each time:
I'd be using NOA and the max speed is set to say, 55, and the vehicle slows to like 30 but would fail to take an on ramp, so I would disengage NOA by jerking the wheel. That *always* disengages NOA as it should, but it does not disengage cruise control, which is still set to 55, and I find my self accelerating rapidly from 30 to 55. The expected behavior by me, and hence the user error on my part, would be that jerking the wheel would disengage NOA *AND* cruise control - returning all control to the driver.
Now I learn to just tap the brake and jerk the wheel. I fully accept the user error, and am working hard on re-training my expectations.
I know that's not exactly what this thread is about, but I wonder what the logs would show in my case - is cruise control being engaged considered "autopilot"; does it's acceleration show as a pedal press? (hopefully not)
I'd be using NOA and the max speed is set to say, 55, and the vehicle slows to like 30 but would fail to take an on ramp, so I would disengage NOA by jerking the wheel. That *always* disengages NOA as it should, but it does not disengage cruise control, which is still set to 55, and I find my self accelerating rapidly from 30 to 55. The expected behavior by me, and hence the user error on my part, would be that jerking the wheel would disengage NOA *AND* cruise control - returning all control to the driver.
Now I learn to just tap the brake and jerk the wheel. I fully accept the user error, and am working hard on re-training my expectations.
I know that's not exactly what this thread is about, but I wonder what the logs would show in my case - is cruise control being engaged considered "autopilot"; does it's acceleration show as a pedal press? (hopefully not)
If you have a 3 the proper way is to push up on the yoke it turns both off and does not send the log which is recorded when you grab the wheel as it you "reacted".
As I understand it, the pedal sensors are each logged separately from any resulting power (throttle) command.
Be sure to only tap up on the stalk a bit and don’t push it all the way up. A small partial push up disengages AP and CC easily and doesn’t risk putting the car into reverse (if you are at a stop or going below 5 mph) since shifting into reverse requires a full push up on the stalk.
roblab
Active Member
My recommendation to people who are just starting to drive a Model 3 is to put it into Chill mode and low Regen, since that most approximates what they are used to. After a couple weeks, when they've gotten used to the car, because you have to get used to the screen being to the right, and how to use the stalks, etc., then they can try out things like one-pedal driving. If you read thru the early pages of the thread, most of the people saying they've experienced UI, only just got the car.
Spacep0d
Active Member
Weird, the first time I rented a Tesla it wasn't in chill mode and regen was on. I had zero trouble with it. By the time I got to the end of the street I had it dialed. Then again I'm pretty tech-savvy and was paying close attention to the differences between this and my then FIAT 500 Abarth. I drove at night in the rain for the first time, but with no passengers, no music, no distractions. It was glorious.
I fell in love with the Model 3 very quickly, and decided in those two days that I had to have one. That was about two years ago. Either way, I thought the Tesla was extremely well-behaved and easy to manage, even in tight parking situations.
The LEAF I owned later had one-pedal driving and it worked amazingly well, but it's no Model 3.
I’m pretty sure if AP is on the first push only disables it without going into reverse, but now you have me doubting myself. I’m going to have to test it out now...
Hank42
Legacy Supporter
Thanks - good to know. It is a little hard to adjust to push on the yoke in a situation like that - I mean, hands are already on the wheel - finding small yoke is challenging when you're trying to regain control of a vehicle "careening out of control
" , but will add that to my re-learning - It's right up there with the turn signals - push the stalk in the same again direction to turn it off everyone should learn to drive a manual. if they can't even handle two pedals they shouldn't get a license.