Uber Autonomous Car Kills Pedestrian.

I guess I take issue with every time this topic comes up it's a case of jump on the bashing bandwagon because your so afraid of technology and something stealing your job.

Technology doubles every year things are moving faster and faster. Servers don't crash like they used to this isn't the windows 95 days anymore. And redundancy is built into mission critical systems three levels deep.

Your argument about looking at people's windows and eyes are great you have two visual sensors how many can a computer have? Far more.

The technology is evolving it's not ready yet but that doesn't mean it won't be. And arguments like they said this 20 years ago but it never happened. Guys it's 2018 the technology is not even close to what you had back in the 90's.

Do you think they would be testing on roads if it wasn't close? You think a company could just send out an automated vehicle and think hey will fix what we find and to hell with everyone else on the road?

Look at cars all the new sensors smart cruise control lane departure warning, emergency braking these are all little pieces of the big picture and yes boys when it's ready machines will rule the roads and humans will be a fall back that as time goes on will be needed less and less.

Now go ahead and put your heads in the sand turn on your VCR's and keep searching for blockbuster.

I disagree with your point of "your (sic) afraid of technology". I don't think so regarding members here. It is something that must be embraced these days in the trucking industry by all.

Jan I believe the insulting in this thread was providing in your reply.

I intend to reply in detail when time permits but please enlighten my limited knowledge on real-time operational control systems...

Yes 2018 when an organized hack compromises many millions of people and a ransomeware attack on a major pharmaceutical firm shut down two production lines for weeks. Yup 2018...all is well...

And servers...they still fail but there is no interruption of service due to sophisticated failover and recovery that occurs almost instantly from a virtualized infrastructure. Yeah Jan...tell me how it is.

You are right about the head in the sand, only I think it might be yours.

Jan replied calling us "boys":

I guess I take issue with every time this topic comes up it's a case of jump on the bashing bandwagon because your so afraid of technology and something stealing your job.

Technology doubles every year things are moving faster and faster. Servers don't crash like they used to this isn't the windows 95 days anymore. And redundancy is built into mission critical systems three levels deep.

Your argument about looking at people's windows and eyes are great you have two visual sensors how many can a computer have? Far more.

The technology is evolving it's not ready yet but that doesn't mean it won't be. And arguments like they said this 20 years ago but it never happened. Guys it's 2018 the technology is not even close to what you had back in the 90's.

Do you think they would be testing on roads if it wasn't close? You think a company could just send out an automated vehicle and think hey will fix what we find and to hell with everyone else on the road?

Look at cars all the new sensors smart cruise control lane departure warning, emergency braking these are all little pieces of the big picture and yes boys when it's ready machines will rule the roads and humans will be a fall back that as time goes on will be needed less and less.

Now go ahead and put your heads in the sand turn on your VCR's and keep searching for blockbuster.

Afraid of technology? You must take me and Brett for idiots...

When you throw down like that and then offer superficial support to your points and verbal jabs, it's a rather weak defense of your opinion.

The systems on these autonomous trucks are way beyond "mission critical", they are "real-time operational control" with public-safety and the flow of commerce at stake if they fail. Minimally they require multiple layers of high-speed failover and 4-5 levels of redundancy. They are not as "ready" as you think because no serious stress testing or penetration testing has occurred.

The optical sensors you mentioned are failure prone and work effectively only in optimal conditions. Covering them with salt and grime renders them all but useless. Sensing a potential event is one element, the initial element, teaching the software to respond proactively and safely or not at all is entirely another, far more complex and variable element. Like I said here and in the other thread; stress testing.

I'll take my chances and continue to apply precautionary measures learned from years of experience avoiding trouble using an optical device "my 20/10 vision" that has yet to be effectively replicated with any level of current technology.

Please read all my replies on the other autonomous thread. My awareness and knowledge on technology is way beyond your assumptions and I do not have any fear of "it" replacing us. Like I said as it evolves, the job requirements will change with it but certainly not dramatically for the next 20 years.

Your right Jan it's 2018, and Microsoft still manages to chronically release bug-ridden software and OS releases so they can meet their financial targets promised to their share holders in spite of incomplete testing. Yeah...we've come a long way...profit always has and always will rule any technology business.

The software, firmware and hardware on autonomous trucks requires far more stringent and comprehensive testing, way more than what has been completed. I gaurantee heavy regulatory compliance from the Fed and likely an agency created to provide over site as adoption becomes more pervasive. Imagine a 60+ minute daily pretrip on our trucks that includes running computer diagnostics on the drive systems somewhat similar to what commercial pilots perform...that's likely to be the case but not in the near future.

Ready? Far from it. A disruptive technology like this has interrelated commercial, social, political, national security and public opinion issues yet to conquer...like I said, full scale adoption will not be swift. And unlikely to occur as it's currently assumed.

Have an intelligent exchange Jan, and chill on the insults if you want to be taken seriously.

I live in Arizona, not too far from Tempe where they’re testing these cars. It’s near the ASU campus and unfortunately there’s a bunch of students around there that like to “troll” the cars by jumping in and out of the street right in front of them just to see what they do. I’m not a huge fan of autonomous vehicles but if self-driving Uber vehicles can provide a huge benefit with minimal downside, I’ll keep an open mind but unfortunately I think ignorant actions like that will eventually put the kaibosh on their plans.

Dan chimes in:

I live in Arizona, not too far from Tempe where they’re testing these cars. It’s near the ASU campus and unfortunately there’s a bunch of students around there that like to “troll” the cars by jumping in and out of the street right in front of them just to see what they do. I’m not a huge fan of autonomous vehicles but if self-driving Uber vehicles can provide a huge benefit with minimal downside, I’ll keep an open mind but unfortunately I think ignorant actions like that will eventually put the kaibosh on their plans.

Somewhere in one of my replies I pointed out control automation of a vehicle operating in an unstructured, uncontrolled environment. It's a basic problem that will begin to surface as more and more of these cars become available.

The technology might be ready,...in a rudimentary fashion. But is our highway/road infrastructure, system of traffic control and current motor vehicle code ready for the technology? Add heavy trucks as part of that question and the answer should be obvious.

I've been keeping up with this story, and as it turns out, the police have been kind enough to share Uber's video of the incident with the public.

I don't know about you guys, but I'm very disappointed by what I see. This accident could have been prevented and this woman's life spared. Yes, technically the woman was jaywalking and she may or may not have been legally to blame, but I don't just run over pedestrians or other vehicles when they break the law. I legally could have rammed into the suv who cut me off with about 20 feet to spare in traffic yesterday, but I didn't because I anticipated it and reacted to prevent an accident. The pedestrian was in the road long before the vehicle approached, and the vehicle should have had no problem seeing it since lidar and radar work fine in the dark. Not only that, but the car never even slowed down or tried to avoid the acccident in any way whatsoever.

Here's an article which pretty much sums up how I feel about this. It is mostly well written except for the part where the author tries to play statistics analyst when he obviously has no understanding of statistics. Enjoy:

Video suggests huge problems with Uber’s driverless car program

This accident could have been prevented and this woman's life spared.

I agree completely. You also have to remember that cameras do not have the dynamic range that the human eye has. The human eye wouldn't have seen it the way that camera did. We would have seen her long before the camera shows she was visible.

Not only that, but as Pianoman mentions:

the vehicle should have had no problem seeing it since lidar and radar work fine in the dark

These cars are designed to see objects in front of them and this car never detected her at all. Never slowed or swerved or anything. This is why I always say that you don't judge a technology based on its successes or potential, you judge it based upon its failures. At this point they have no idea why the car didn't detect the object it was clearly designed to detect. This ability would have been one of the primary features of this software and it failed, the same way it failed to detect a tractor trailer across the highway and killed the driver of that car.

I embedded the video in Pianoman's comment above if you haven't seen it already.

I embedded the video in Pianoman's comment above if you haven't seen it already.

Thanks Brett.

The human eye wouldn't have seen it the way that camera did. We would have seen her long before the camera shows she was visible.

Absolutely! It might have helped if the human driver hadn't been texting for the 6 seconds right before the impact.

These cars are designed to see objects in front of them and this car never detected her at all.

This is the part that's a really big deal to me. If Lidar can "see" fine in the dark, that means that this car would have behaved the same way in broad daylight as well. That's a huge huge problem. If I were that lady's friend or relative I'd be suing Uber for every penny they have. I guess I've chosen my side lol.

There's some interesting information in this article concerning this subject.

Uber's Self Driving Cars Are Struggling