I just read were some countries are requiring cars to notify drivers when they are exceeding the posted speed limit and also restrict the speed of the car. Thus far it seems the systems can be overridden.
This naturally got me thinking; could we see a situation, where like Boeing's 737 issue, where the car overrides the driver causing an accident?
We've previously mentioned the auto emergency braking coming on for no apparent reason. I could also envision someone pulling out to pass and the speed limiter kick in and this causing an accident.
Do note I am not a fan of a lot of this technology as software glitches all the time.
Thoughts?
Tesla cybertrucks are already doing just that across multiple people and often with dash cams. Ignoring the brae pedal.
It will get better over time, it takes regulation and proper testing to get there though.
Audi 5000 - claims of sudden unintended acceleration nearly put them out of business 40 years ago.
In reply to bludroptop :
Emphasis on "claims" with that one. 60 Minutes specifically modified a car to have the problem and didn't disclose it, then generated an incredible amount of buzz with their bombshell report. NHTSA eventually found user error to blame as Americans in the mid-80's weren't used to European pedal spacing.
During model years 1982–1987, Audi issued a series of recalls of Audi 5000 (the North American name of the Audi 100 at the time) models associated with reported incidents of sudden unintended acceleration linked to six deaths and 700 accidents.[38] At the time, NHTSA was investigating 50 car models from 20 manufacturers for sudden surges of power.[39]
60 Minutes aired a report titled "Out of Control" on November 23, 1986,[40] featuring interviews with six people who had sued Audi after reporting unintended acceleration, including footage of an Audi 5000 ostensibly displaying a surge of acceleration while the brake pedal was depressed.[41][42] Subsequent investigation revealed that 60 Minutes had not disclosed they had engineered the vehicle's behavior — fitting a canister of compressed air on the passenger-side floor, linked via a hose to a hole drilled into the transmission[40][41] — the arrangement executed by one of the experts who had testified on behalf of a plaintiff in a then pending lawsuit against Audi's parent company.[43]
Audi contended, prior to findings by outside investigators that the problems were caused by driver error, specifically pedal misapplication.[39] Subsequently, the NHTSA concluded that the majority of unintended acceleration cases, including all the ones that prompted the 60 Minutes report, were caused by driver error such as confusion of pedals.[44] CBS did not acknowledge the test results of involved government agencies, but did acknowledge the similar results of another study.[41]
With the series of recall campaigns, Audi made several modifications; the first adjusted the distance between the brake and accelerator pedal on automatic-transmission models. Later repairs, of 250,000 cars dating back to 1978, added a device requiring the driver to press the brake pedal before shifting out of park.[38] As a byproduct of sudden unintended acceleration, vehicles now include gear stick patterns and brake interlock mechanisms to prevent inadvertent gear selection.
Audi's U.S. sales, which had reached 74,061 in 1985, dropped to 12,283 in 1991 and remained level for three years.[38] — with resale values falling dramatically.[45] Audi subsequently offered increased warranty protection [45] and renamed the affected models — with the 5000 becoming the 100 and 200 in 1989.[39] The company only reached the same level of U.S. sales again by model year 2000.[38]
As of early 2010, a class-action lawsuit filed in 1987 by about 7,500 Audi 5000 owners remains unsettled and is currently contested in county court in Chicago after appeals at the Illinois state and U.S. federal levels.[38] The plaintiffs in this lawsuit charge that on account of the sudden acceleration controversy, Audis had lost resale value.[41]
Toyota had their own unintended acceleration issue as well, some were electronic faults, not just pedals stuck under mats.
Cybertrucks are steer-by-wire and it's been a particularly troublesome system, and its steering wheel is allowed to move much faster than the wheels can move, which can cause what appears to be a huge amount of lag in the steering system.
Some BMWs have a gearbox on the steering column connected to an electric motor that allows the steering angle to be electronically modified. It has a failsafe setup that should cause it to work like an ordinary solid shaft in case of a malfunction, but there could be more semi-steer-by-wire cars out there than we realize. Infiniti has one with redundant computers and a clutch in the steering column that disconnects in steer-by-wire mode.
ABS isn't perfect, it's clearly prevented orders of magnitude more accidents than it's caused, but it has caused a few through malfunctions.
This is more relevant to track driving but I'd often worried about what might happen with ABS-equipped cars in situations where you'd want to lock the wheels like in a spin or while sliding. Then I smashed the front of my Toyobaru in a crash that was at least significantly worsened because the ABS wouldn't let me brake much in a slide, while being able to lock the wheels would've given more deceleration and allowed me to change the car's curved path into a straight line. I have some ideas for controls that would let me quickly disable ABS to prevent any repeats.
wearymicrobe said:
Tesla cybertrucks are already doing just that across multiple people and often with dash cams. Ignoring the brae pedal.
It will get better over time, it takes regulation and proper testing to get there though.
That decline in commercial aviation is what I try to remind myself of every time I fly.
I know the chance of something happening is really small, but that possibility still scares me.
I think Japan has had speed limiters for a very long time.
Colin Wood said:
wearymicrobe said:
Tesla cybertrucks are already doing just that across multiple people and often with dash cams. Ignoring the brae pedal.
It will get better over time, it takes regulation and proper testing to get there though.
That decline in commercial aviation is what I try to remind myself of every time I fly.
I know the chance of something happening is really small, but that possibility still scares me.
More likely to get hit on the way to the airport for sure.
The hard part of this chart is that the government and regulation dictate service on Planes. When a Model 3 auto driver is 8 years old and maintained by Bubba in his backyard with aftermarket parts or modified with overvoltage software then this whole thing falls apart. Just the sensor alignment issues alone make this a very difficult project. As engineer adjacent as I am I want nothing to do with this tech and certainly not from the break stuff speeds that modern startups are working at.
The 737MAX problem(s) seem to be a symptom of the decade(s) long Boeing corporate culture.
IMO, that likely applies to the Tesla issues too.
It's hard to gauge where we're going with transportation, automation, always-accelerating technology...
It seems like trying to gauge which of two infinite things is bigger when we talk about trying to make autonomous vehicles (or driver aids) work and playing it against the alternative of actually making humans drive to a reasonable proportion of what they're capable of.
Not to tangent too badly, but the cost of a fender bender in today's world of per-model headlight modules and arrays of sensors really makes me wonder whether it's worthwhile. I mean, I have a gut reaction against it, but like many things in modern vehicles (e.g. auto stop/start) it might pencil out that the gas savings over millions of cars really is worth the expense or annoyance. I don't know, I'm just reasonably sure we don't tend to start these conversations at the actual beginning.
Keith Tanner said:
I think Japan has had speed limiters for a very long time.
I wonder if the bigger (safety) concern isn't the limiter, but how a "smart" system might decide to apply it. If you always know your car won't go over 75mph, you won't try a pass that needs 90mph. If you usually get to do what you want but something makes the system decide this is a time for auto-enforcement in the middle of such a pass, then you have (even bigger) problems.
I'm sure lots of folks here would also be against a "dumb" limiter, but it's a different concern.
This makes some sense to me as a valid concern when you look at systems that use either maps or cameras to note the current speed limit. It's one thing to just not let the car above some maximum speed, it's a different thing to limit it to the current speed limit. Our rental Citroen on a recent trip to Spain had a dashboard display of the current speed limit, and it was wrong a significant amount of the time.
kb58
UltraDork
7/8/24 7:29 p.m.
I can't help but think that Tesla is in for endless lawsuits. For example, say that someone steps out in front of a Tesla, there's another car coming the other direction, and a person is walking on the sidewalk. The Tesla software sees the person that just stepped out but it's too late to stop. It can't stop in time, so has to decide whether to hit the person crossing, or veer to the left and hit the oncoming car, or veer right and hit the person on the sidewalk. Regardless which way it veers, the Tesla may be seen as at-fault. Once Tesla claims that self-driving can be fully enabled, it seems like it'll open the flood gates to lawsuits, frivolous or real, rightly or wrongly. Sure, their system may work as well or better than a human driver, but with an important difference. When someone is hit by another driver driving Brand X, Brand X doesn't get sued, the driver does. With Tesla, they're about to confront the opposite case, where instead of the Tesla driver, Tesla themselves will be blamed. It'll be even worse since Tesla is a much more tempting target to sue, with their deeper pockets.
Keith Tanner said:
I think Japan has had speed limiters for a very long time.
It's also significantly higher than their highest speed limits. 180km/h limiter if I am not mistaken. A lot of their limited-access arteries have speed limits half that or less.
kb58 said:
I can't help but think that Tesla is in for endless lawsuits. For example, say that someone steps out in front of a Tesla, there's another car coming the other direction, and a person is walking on the sidewalk. The Tesla software sees the person that just stepped out but it's too late to stop. It can't stop in time, so has to decide whether to hit the person crossing, or veer to the left and hit the oncoming car, or veer right and hit the person on the sidewalk.
Ever watch I, Robot?
Regardless which way it veers, the Tesla may be seen as at-fault. Once Tesla claims that self-driving can be fully enabled, it seems like it'll open the flood gates to lawsuits, frivolous or real, rightly or wrongly. Sure, their system may work as well or better than a human driver, but with an important difference. When someone is hit by another driver driving Brand X, Brand X doesn't get sued, the driver does. With Tesla, they're about to confront the opposite case, where instead of the Tesla driver, Tesla themselves will be blamed. It'll be even worse since Tesla is a much more tempting target to sue, with their deeper pockets.
Oh, those laws are already on the books. The fault lies with the vehicle's driver, not the car or its software. As the driver it is your duty/expectation to take control if you are in a situation the software can't handle.
GameboyRMH said:
Then I smashed the front of my Toyobaru in a crash that was at least significantly worsened because the ABS wouldn't let me brake much in a slide, while being able to lock the wheels would've given more deceleration and allowed me to change the car's curved path into a straight line.
Maybe you were in too deep but your steering wheel can also turn a curved path into a straight path.
In reply to AngryCorvair (Forum Supporter) :
Hey, stay in your lane, Brake Guy (Joking! Couldn't resist...)
In reply to AngryCorvair (Forum Supporter) :
Yeah I was in pretty deep and quite worried that if I countersteered too much, I could transition into a slide in the other direction toward a larger number of much less avoidable obstacles.
I know it's not really what you mean, but the 737 Max issue was primarily the result of allowing only one sensor to be responsible for causing a pilot override. This of course is easily remedied by having redundant sensors (and disabling the system if more than 50% disagree). Obviously something to learn there and hopefully the car manufacturers have (?)
The real issue of course was creating a plane that was unstable in certain situations to save money by modifying a design instead of creating a new design. Car companies have certainly repurposed basic designs to save money, but I am not sure if there were any that created a safety issue because of it.
Tom1200 said:
This naturally got me thinking; could we see a situation, where like Boeing's 737 issue, where the car overrides the driver causing an accident?
I was near an accident that may have been caused by a situation described above. I was turning onto a side road with a newer F150 behind me and another truck behind him. F150 was following very closely and seemed to go into automatic emergency braking while I was getting off the main road. Truck came to a complete stop and the truck behind him could not stop in time.
I've also had the stability control get our old Five Hundred pretty sideways (while going straight) because we found some black ice. Car was fine, stability control light started flashing, and the car started doing crazy things.
Yes, that was a stability control system that was developed 20 years ago, but I agree that I question when this tech will stop preventing accidents and will begin to be a contributing factor.
I had two cars with shorted ABS senors. Both cars would hit the bump going into my garage, engage the ABS and then lengthen my stopping distance to the point that I puckered. One of then would also frequently engage in the rain, the worst part was that it only happened under 5 mph right when you expect to just stop.
One was melted in a previous repair the other wire was smashed, neither under my watch.
Moral of the story is, fix the bump going into your garage.
Not only am I against most automated nanny's in cars for control reasons as a card carrying genXer, have you seen an insurance claim for a repainted bumper lately? Yes it is but a scratch but we have to remove the bumper cover to paint so all twelve of your sensors need to be calibrated at a cost of $2500. For sure don't hit a sensor or two. No wonder car insurance is going up so much.
As mentioned above, the MAX being initially designed only requiring a single sensor input for MCAS was a massive engineering fail. Not fully explaning MCAS in the FCOM (Flight Crew Operating Manual) was a massive human factors/training fail. There's no real issue with the "need" for MCAS (which was more certification driven than safety of flight driven) but not making its inputs redundant boggles the mind given all the redundancy built into transport-category aircraft.
That said, both MAX accidents could have been prevented if each crew (may they RIP) correctly performed the Runaway Trim memory items - which would have turned the system off.
I don't see a similar issue presenting itself in automobiles, though I am a bit weary of the cybersecurity risk of highly networked new vehicles that get frequent over-air updates pushed to them.
20 years ago, I had a lot less worry about the code in cars. I'm not a fan of the auto code systems they use now.
That being said, part of the reason it take 5 years to design a car for production is code debug. The time frames for code to be released into production are held to pretty well- not perfect, as there are recalls to fix things. But for good companies, coding and debugging requirements are held to a really high standard. Somehow, all that stayed in place during the times where there were a lot of cutbacks and corner cutting.
But that's not to say it won't happen. The search for short term stock gains will not stop anytime soon. MD brought that to Boeing, and, thankfully, let Ford have Allan Mullaly.
aircooled said:
The real issue of course was creating a plane that was unstable in certain situations to save money by modifying a design instead of creating a new design. Car companies have certainly repurposed basic designs to save money, but I am not sure if there were any that created a safety issue because of it.
Note that the desire to save money was actually not on the part of Boeing here, it was on the part of the customers. The airlines didn't want to buy a new air frame, they wanted more efficient 737s. That meant they could share pilot training, mechanic training, spare parts, tools and yes, save money. Ultimately that leads to lower (or at least less raised) ticket prices for consumers.
The execution of MCAS obviously had problems, but the goal was worthy. The worst part about it was hiding the details from the pilots though.
akylekoz said:
Moral of the story is, fix the bump going into your garage.
Not "fix the ABS sensors"? Was it not displaying the warning light?
codrus (Forum Supporter) said:
aircooled said:
The real issue of course was creating a plane that was unstable in certain situations to save money by modifying a design instead of creating a new design. Car companies have certainly repurposed basic designs to save money, but I am not sure if there were any that created a safety issue because of it.
Note that the desire to save money was actually not on the part of Boeing here, it was on the part of the customers. The airlines didn't want to buy a new air frame, they wanted more efficient 737s. That meant they could share pilot training, mechanic training, spare parts, tools and yes, save money. Ultimately that leads to lower (or at least less raised) ticket prices for consumers.
The execution of MCAS obviously had problems, but the goal was worthy. The worst part about it was hiding the details from the pilots though.
Yes and no. To make the max competitive with the airbus, Boeing made the choices to modify the current 737 platform instead of a more major redesign, as it would save *them* money. The airlines would pay only current 737 and/or A320 prices for said new plane.
Yes, it was the customers who wanted the more efficient plane, but it was Boeing who chose to take the path they did to compete with airbus. And it was Boeing who offered the stripped down package and didn't fully tell the airlines/ pilots everything.
And the recent door plug blowouts follow the same background pattern.
WRT the auto industry, thanks to competition, it's far more cutthroat to not make as many mistakes. Thank you Toyota for making such and excellent car for so little money. They have saved many lives.