Hacker News Reader

Disabling a backup system like that strikes me as very negligent from an engineering perspective. They had reason to expect their system to fail because it's a prototype, but then they go and disable another system that is actively marketed and that could have prevented this death.

I'm so glad they're not testing their cars in my state. This makes me wonder if they even do basic safety engineering like FMEA or if they just wing it.

But it was never built as a backup system. It's a whole feature that was designed from the ground up assuming it's the only thing on the car doing anything like the job it does.

There's 2 chefs.

Actually, those systems are designed as backup. They are designed as backup, when the driver does not react. That is the reason why they warn first and initiate an emergency brake only, and only if an accident becomes unavoidable.

So that would, if the automated tech from Uber would have detected in time, it would have reacted like a human driver and the pre-installed system would may have not noticed it. But in case, it would have reacted as a backup.

On the two occasions my Volvo has decided to take over (both admittedly due to overhanging bushes in a parking space), it has taken over control immediately and entirely.

There is no warning.

No warning necessary because self driving cars are completely capable of being programmed to deal with varying ability to control the car. It can still have an intention whether or not the car is allowing it to proceed. It needs to make a best effort, allowing failure when a backup system takes control.

> But it was never built as a backup system.

The collision avoidance system? The fail-safe that's meant to ensure the car doesn't hit something when the human driver screws up? The one that works with an entirely separate autonomous intelligence (aka, the human) controlling the car? The one explicitly designed to trigger emergency brakeing even if the primary controller is trying to do something else? The kind of system installed on many other AI controlled cars to avoid this exact type of disaster?

That backup system?

Look, it's entirely possible that for some reason this specific safety system conflicted with the current version of Uber's system, but we have zero data to suggest that, and if so, it's something specific to those two systems in that particular vehicle (and suggests a severe issue with Uber's system to boot). There's no general issue here.

> There's 2 chefs.

Yes, and sometimes one of them takes over from the other one. That's how this works.

Have you actually driven with one of these systems?

There are plenty of intentional maneuvers required to integrate with real-world (ie. non perfect) traffic that trigger these systems.

They’re designed to take hints from human input (so for example, LKA responds less strongly to forceful unexpected inputs than a lack of expected input).

There is 0 chance the system wouldn’t interfere with a self driving.

Not to mention many of those safety systems include lockouts on the CAN bus which make removing their safety module a first step to even add your own inputs. I know that’s how it works on my Volt with a Mobileye sensor

The self driving should be capable of mitigating this risk by adapting to systems like these in general.

Problem is, that greatly complicates the problem-solviing process for SD vehicles---you're now trying to build atop a model of the world provided by the vendor, with its own faults and quirks (and you probably need an agreement with the vendor to see their code / dissect their model, and now your release schedule depends on their release schedule, etc.). It also makes Uber dependent upon Volvo's buy-in on the project, and that arrangement is fraught in a world where we don't yet know if SD cars are going to damage the business of traditional auto manufacturers 50 years out---Uber wants to be in a place where if a vendor relationship sours, they can pack up and change chassis without a complete overhaul of the SD logic.

If your SD system is supposed to be solving 100% of the same problem anyway, it's best to gut the software you don't control and work with the hardware as directly as you can. The notion of building atop a lower-level abstraction makes for great academia but doesn't work with industrial realities.

Thanks for the explanation and I agree that it won't be as easy to build SD for auto braking systems that change from chassis to chassis, but this is what I think people want. People want to have fail safes from different vendors.

Automatic emergency braking is quite definitely a backup system. If it ever trips in my car it is because the primary system, that is me, has failed to brake in time.

> we're currently at 1 pedestrian fatality and 0 fatalities for "people in cars killed due to overzealous/conflicting automated safety systems".

Not trying to take sides here, but the last part strikes me as "citation needed".

Sorry, I guess I left out the implicit "as far as I know".

Do you know of anyone who's died that way?

Yes: https://www.theguardian.com/technology/2016/jun/30/tesla-aut...

That's neither a "overzealous/conflicting" system nor a "safety system", so if anything, this would bring the score to 2:0 rather than 1:1.

I was already aware of that one, but sure, that's another example of the opposite of what I asked for. Rather than someone killed by conflicting or over-cautious safety systems, that's an example of someone killed by a single ineffective safety system, just like the pedestrian in Arizona.

This further reinforces my point, I think. :)

> Multiple independent systems with the ability to sound an alarm and slam on the brakes if they detect an imminent pedestrian collision...seem like they probably would be safer overall?

While yes - it would also force Uber to permanently partner with Volvo on technology they can't control. 'Uhuh, my Ubers a Chevy not a Ford - it could kill me because the Chevies don't built-in X is worse!'.

You do not want an arbitrary third party system controlling the car while the Uber system is in unknown ways; if the systems conflicted things could go out of control and be more dangerous.

> it would also force Uber to permanently partner with Volvo on technology they can't control.

That sounds like a win-win to me.

>You do not want an arbitrary third party system controlling the car while the Uber system is in unknown ways; if the systems conflicted things could go out of control and be more dangerous.

I find it interesting that, of these two systems, it is Volvo's widely-deployed system that is described as 'arbitrary'. If this argument had any merit, it would be unsafe for this system to be on the road, 'conflicting' with the car's human driver. On the contrary, the statistics clearly show that these systems reduce the frequency and severity of accidents.

I know this sounds bad, but the arbitrary is that the Ubers aren't designed to be as safe as possible currently - they're designed to be testbeds of the new fully autonomous technology uber is developing.

Maybe it makes sense to rely on arbitrary vehicle specific fallbacks, but long-term it doesn't as I'm sure They don't want to be dependent on a specific model. When you request a ride, do you want to care about whether your getting a Honda or a Volvo?