Software updates can have unintended consequences

[dimes4slim]

As reported in today's Wall Street Journal, "safety advocates have raised concerns about the growing use of software to repair faulty car parts. Software can be very complex, and automakers don’t have precise government standards to follow to ensure software fixes are validated, safety advocates say."

Peter Rothschild was driving home down a steep singlelane road in Northern California when suddenly his Volvo SUV started accelerating out of control. “I kept pushing on the brakes and pushing on the brakes,” said the 69-year-old retired radiologist. But for several seconds, nothing he could do would slow down the car. Rothschild was able to steer his gray Volvo SUV up a hilly roadside, bringing the car to a stop. The side air bags deployed, and the vehicle suffered some damage. “I don’t think I would’ve made the next curve and would’ve gone off the side,” he said.

As further quoted, "Volvo says it didn’t know that his 2025 XC90 plug-in hybrid SUV had a braking defect. It materialized after an April safety recall for 400,000 vehicles over rearview camera failures. That recall involved several software updates. Volvo later identified that about 11,500 plug-in hybrid and all-electric vehicles that received those updates could experience the same braking failure. Volvo said the software update to fix the camera issue was packaged with other software updates that introduced a separate braking issue. The carmaker also said that software updates allow it to resolve issues faster and that investments in software help make its vehicles better and safer."

We all know this could never happen with our cars with the elaborate software testing protocols before any updates are released. Anyway, it is something to think about after we receive an update. Several of the software updates I have received have caused some wonky unintentional results which likely got missed in the initial software testing.

Sponsored

 

[awp0]

I don't think that software is inherently more dangerous than hardware........unless the manufacturer doesn't know how to (and commit to) deliver it safely.

 

[dimes4slim]

I don't think that software is inherently more dangerous than hardware........unless the manufacturer doesn't know how to (and commit to) deliver it safely.

I respectfully disagree. If my car were at the dealer to have a camera replaced, it would be a long shot that somehow that service work would affect my brakes, unless they cut a brake line. A software update on the camera, which also has an unintended affect on the brakes can obviously easily happen with computer coding. Now that cars can basically drive by wire, it is disconcerting that something like that can actually happen. I am all for technology, but we are at the mercy of the manufacturers to exercise extreme diligence before releasing software updates that have the potential to have lethal consequences.

 

[awp0]

I respectfully disagree. If my car were at the dealer to have a camera replaced, it would be a long shot that somehow that service work would affect my brakes, unless they cut a brake line. A software update on the camera, which also has an unintended affect on the brakes can obviously easily happen with computer coding. Now that cars can basically drive by wire, it is disconcerting that something like that can actually happen. I am all for technology, but we are at the mercy of the manufacturers to exercise extreme diligence before releasing software updates that have the potential to have lethal consequences.

When implemented properly, software can have portable and near-instant automatic test suites to validate the safety of software and many hardware related changes (unlike, say, the lug nuts holding your wheels to your car). Are the manufacturers doing this today? Doubt it. They're all rushing as fast as they can. But I still maintain that software is not inherently more dangerous.

 

[hartmms]

A hardware replacement for a camera could involve new firmware on the camera module. The new firmware could introduce new communications on the vehicle bus that cause unintended consequences unless tested properly.

A second scenario: perhaps the camera module part was sitting on the shelf for 2 years. In that time, it's possible the camera module manufacturer has upgraded firmware of that module to fix bugs or interoperability with other modules. Old firmware camera module installed into current car network could also cause issues, unless installer ensures camera module is upgraded to latest firmware.

Per the article "That recall involved several software updates", so it's also possible more than the camera module was updated and that other item triggered the braking issue.

Let's hope vehicle manufacturers get better and better at testing out these complex systems so problems like this become more and more rare.

 

[Teslaeata]

I respectfully disagree. If my car were at the dealer to have a camera replaced, it would be a long shot that somehow that service work would affect my brakes, unless they cut a brake line. A software update on the camera, which also has an unintended affect on the brakes can obviously easily happen with computer coding. Now that cars can basically drive by wire, it is disconcerting that something like that can actually happen. I am all for technology, but we are at the mercy of the manufacturers to exercise extreme diligence before releasing software updates that have the potential to have lethal consequences.

Hmmmmm, I often investigated claims that somebody presses a pedal that doesn’t respond who often are advancing in years or otherwise not physically or mentally together but not always, and sometimes by somebody distracted when stationary doing e-mails etc forgetting the car is turned on in Ready mode when they say the car took of on its own etc etc.

The pre-crash data where available never once proved the car was the cause and it was not the loose nut behind the wheel with pedal confusion, for example.

The thing is, these anecdotes cannot be verified evidentially so I always look on them with suspicion.

 

[Jeff-NoVA]

Hmmmmm, I often investigated claims that somebody presses a pedal that doesn’t respond who often are advancing in years or otherwise not physically or mentally together but not always, and sometimes by somebody distracted when stationary doing e-mails etc forgetting the car is turned on in Ready mode when they say the car took of on its own etc etc.

The pre-crash data where available never once proved the car was the cause and it was not the loose nut behind the wheel with pedal confusion, for example.

The thing is, these anecdotes cannot be verified evidentially so I always look on them with suspicion.

First thought I had when I read about this nonagenarian's incident. It's an unfortunate occurrence, but is it connected to the recall? I'm old enough to recall Audi almost getting driven out of the USA by an unintended acceleration "problem" that was the usual PICNIC error sensationalized by an opportunistic news show.

In any event, bad software = bad results. Doubly concerning for vehicles, which is why OEMs need to get it right, every time. Sounds like Volvo didn't.

 

[awp0]

My opinion is that articles like this and the general publicity around software-related car accidents is overblown. It's like reading about shark attacks. If you didn't know any better, you'd think you should never set foot in the ocean again. The reality is that these technologies and software systems have provided much much more safety than they've compromised. But there's no money to be made with articles that sound like "ADAS got me to my destination safely".

 

[RSF]

I don't think that software is inherently more dangerous than hardware........unless the manufacturer doesn't know how to (and commit to) deliver it safely.

The human element is worst of all

 

[JohnFoxeSheets]

Hmmmmm, I often investigated claims that somebody presses a pedal that doesn’t respond who often are advancing in years or otherwise not physically or mentally together but not always, and sometimes by somebody distracted when stationary doing e-mails etc forgetting the car is turned on in Ready mode when they say the car took of on its own etc etc.

The pre-crash data where available never once proved the car was the cause and it was not the loose nut behind the wheel with pedal confusion, for example.

The thing is, these anecdotes cannot be verified evidentially so I always look on them with suspicion.

I had similar thoughts (though no forensics experience ), but the fact that Volvo admitted that there was a software issue that affected braking seems to bolster the guy’s story.

I experienced a wide range of bizarre and seemingly unrelated issues with my 2018 Volvo XC60 that ultimately were resolved by wait for it…. replacing the radio (and telematics, I presume) antenna. That Volvo ever figured it out is really rather amazing. (And I say Volvo because the issues were sent to them by my dealership, and they figured it out.)

 

[ChrisO]

What do you think the chances are that the government can somehow put in the regulations and oversight to actual improve the situation?

I can think of many ways that they can make things worse.

 

[rpr]

Cue up the old "if Microsoft made cars" joke. Google it if you don't know it, there are numerous versions out there...

 

[Clinton_B]

Not to worry. AI will fix it all.

 

[Teslaeata]

Not to worry. AI will fix it all.

For AI read AU - artificial unintelligence

 

[NorthlandPhil]

As per the article, the update included changes in several modules, not just the camera software. They did not list what other modules. I suppose because if they did they would have lost the sensationalism of "Camera Update Causes Brake Failure!"

I take offense at the idea that because he was in his 60's he probably didn't actually step on the brake. Volvo actually stated there were other incidences of this problem and they are recalling vehicles to fix it, as well as no longer downloading the update.

The camera update didn't cause the brake failure. Some other update did.

Still, when I think back to the 60's and 70's when highways were narrower, corners where sharper, roads in general were much less safe, and cars with no power brakes, steering that was incredibly loose and to top it off 65 MPH speed limits... Cars are much safer today even with occasional software glitches.

Sponsored