Opinions On Self-driving Trucks Please!

I just want my OnGuard to quit picking up the overpasses.

Gentlemen, 1. I cited trains as evidence that autonomous operation, even in a one-dimensional space, is proving to be beyond the scope of current technology (except the familiar driverless intra-airport "people mover" systems). I did not cite trains as evidence that driverless road vehicles are currently feasible. 2. No current technology is feasible to implement what G-Town calls "point technology"; to date, driverless vehicle development has focused on autonomous vehicle operation, not on infrastructure dependent vehicle operation, out of recognition that with currently available technology infrastructure build-out to support infrastructure dependent vehicle operation is not a realistic alternative. The progress made over the last 5 - 10 years on the development of "driverless" cars should be sufficient to convince most any tech-savvy observer that autonomous vehicles will ultimately depend not on infrastructure, but on strictly autonomous decision making capability, just as cars today depend on autonomous decision making by the human driver. The positive train control project within the railroad industry has abundantly demonstrated the pitfalls of trying to base vehicle operation on an infrastructure-dependent model. 3. The people who are working to develop autonomous vehicles are acutely aware of the "infinite amount of variables" in autonomous vehicle operation; that is the technological challenge that attracts some of the most talented engineers and physicists to work on autonomous vehicle design. And yes, there are mathematically sound and rigorous methods of testing systems against those "infinite amount of variables". No, software testing is not perfect, but software can be made fail-safe, as evidenced by the millions of lines of code that execute flawlessly in millions of modern passenger cars and aircraft. 4. I did not and am not arguing that "driverless trucks" are going to be a reality at any time in the future. (But I am not going to be foolish and say "never"; I suspect that when my great great grandparents took six months to travel from Boston to what is now Oklahoma, in the 1840's, they would have said that "never" would anyone make that trip in four hours - which, of course, is now done many times a day.) My argument is that there are many technological assists that are currently feasible now, and we need to take action to encourage development and implementation of those technologies. 5. With all due respects, may I point out that if everyone takes the position of "prove it is totally safe and reliable, and then I will use it", no progress will ever be made toward the adoption of technologies which would improve safety, efficiency, and driver working conditions. 6. May I remind you that essentially every manufactured thing in your life, from the truck you drive to the medicines that rescue you from certain death due to infection, exists because someone was chasing the dollar. If someone can develop a new technological advance that makes truck operation safer or more efficient, they certainly deserve to reap some monetary reward, just a you reap a monetary reward for putting your life on the line to deliver goods and materials to people who need them. Every new device that comes into any aspect of our lives exists because someone was motivated to "make bank". 7. It should be obvious that nearly every improvement of every mode of modern transportation has been invented, designed, and developed by people with no first-hand (practical?) knowledge of the tasks for which they were designing. Airplanes are not designed by pilots. Truck drivers do not design and manufacture automatic transmissions, pollution control systems, or hybrid power systems. New container ships are not developed by ship captains. The list is endless. 8. So, in summary, let's move the focus of the discussion away from the concept of "driverless" vehicles, and instead focus on adopting technology that has the potential to improve safety, reliability, efficiency, and ease of use. I think that we can all agree that right now, today, "driverless" vehicles are not going to be a reality within the next year, five years, or maybe many decades, but we should not point to that as a reason to drag our feet on many incremental improvements that can be realized within the next few years.

I plowed through as much as I could stand and finally couldn't take it anymore when I hit this:

No, software testing is not perfect, but software can be made fail-safe, as evidenced by the millions of lines of code that execute flawlessly in millions of modern passenger cars and aircraft.

Fail-safe software? Flawless execution?

Now this is a man who knows nothing about software.

I guess maybe you didn't see the news story about the autonomous car that ran down a woman who was walking alongside her bike without even touching the brakes?

Maybe you missed the time that an autonomous vehicle "didn't see" the tractor trailer completely blocking the roadway and ran directly into the side of it, killing the passenger?

Fail-safe software. Flawless execution.

we should not point to that as a reason to drag our feet on many incremental improvements that can be realized within the next few years

When I started driving in '93 they had "collision detection systems" on the truck that would beep when it detected a possible hazard in the roadway. That stupid thing beeped 1,000 times a day at every guardrail, bridge, and telephone poll you passed. Do you know what those systems do today? The exact same thing. They haven't changed a bit in 25 years.

Incremental improvement. pffffffff.

It's all hype. All of this talk about A.I. and autonomous cars and computers that will be more intelligent than humans. They're barely a millimeter closer to any of that happening than they were 50 years ago. It's all marketing. It's a bunch of research professors and tech startups trying to raise money to fund their little tech playgrounds.

They can't even make a robot that can vacuum my rug or clean up dog poop. Half the time I can't even get my bluetooth headset to connect properly.

Get serious.

Roy, when I started reading your reply I thought, "This guy is trying to sound like he's spectacularly intelligent when it comes to this stuff, which almost certainly means otherwise." I plowed through as much as I could stand and finally couldn't take it anymore when I hit this:

No, software testing is not perfect, but software can be made fail-safe, as evidenced by the millions of lines of code that execute flawlessly in millions of modern passenger cars and aircraft.

Fail-safe software? Flawless execution?

Now this is a man who knows nothing about software.

I guess maybe you didn't see the news story about the autonomous car that ran down a woman who was walking alongside her bike without even touching the brakes?

Maybe you missed the time that an autonomous vehicle "didn't see" the tractor trailer completely blocking the roadway and ran directly into the side of it, killing the passenger?

Fail-safe software. Flawless execution.

we should not point to that as a reason to drag our feet on many incremental improvements that can be realized within the next few years

When I started driving in '93 they had "collision detection systems" on the truck that would beep when it detected a possible hazard in the roadway. That stupid thing beeped 1,000 times a day at every guardrail, bridge, and telephone poll you passed. Do you know what those systems do today? The exact same thing. They haven't changed a bit in 25 years.

Incremental improvement. pffffffff.

It's all hype. All of this talk about A.I. and autonomous cars and computers that will be more intelligent than humans. They're barely a millimeter closer to any of that happening than they were 50 years ago. It's all marketing. It's a bunch of research professors and tech startups trying to raise money to fund their little tech playgrounds.

They can't even make a robot that can vacuum my rug or clean up dog poop. Half the time I can't even get my bluetooth headset to connect properly.

Get serious.

Or the bugs and security holes they find that have existed for years.

A recent Linux bug was found that has existed for 10 years or more (I don't remember exactly how long) and that is open source, meaning many many eyes have scrutinized it for all those years. Proprietary software is even worse.

Even if they perfect the safety, which I have my doubts about, all they need is one security hole to be exploited and suddenly some scumbag decides to have semis plowing through cars, buildings, people, etc. No thanks. It has already been done with cars. Not maliciously thankfully, but they have proven they can take control of a vehicle and force it off the road, etc.

Frankly, I don't think it should be allowed even in cars. And I am far from a Luddite, I've worked with computers and other tech for years. I just don't see the necessity for cars to drive themselves, nor the safety of it. If there is a person on board, then they can drive. We as a society are already lazy enough, now we can't be bothered to drive ourselves either?

We already have safe, self driven vehicles. They are called trains, and they are confined to tracks, which is fairly safe. I don't want an autonomous 40 ton vehicle barrelling down the street, alone, or in a caravan with a real driver in front.

The only way an autonomous vehicle makes any sense is if it's on a purpose-built infrastructure. If they were serious about autonomous transportation they would build the infrastructure for it.

I think drones have far more potential for changing the world than autonomous vehicles. Producing autonomous drones for package delivery, surveillance, and transportation is much more realistic than trying to get autonomous cars and trucks to function on today's infrastructure.

Brett wrote:

The only way an autonomous vehicle makes any sense is if it's on a purpose-built infrastructure. If they were serious about autonomous transportation they would build the infrastructure for it.

Thanks Brett. Exactly my point.

And please Roy don’t lecture me or anyone else in this forum on where our focus needs to be regarding safety. Our lives and livelihoods depend on it...

According to FMCSA , fatalities in crashes involving large trucks has increased from 3,380 in 2009 to 4,317 in 2016. Clearly, the current efforts to improve truck safety are leading to an increasing number of fatalities every year. So if you reject technology as a means to improve safety, what do you think should be done to improve large truck safety?

According to FMCSA , fatalities in crashes involving large trucks has increased from 3,380 in 2009 to 4,317 in 2016. Clearly, the current efforts to improve truck safety are leading to an increasing number of fatalities every year. So if you reject technology as a means to improve safety, what do you think should be done to improve large truck safety?

One reason for the increase, is that many drivers fail to put a very simple piece of equipment in their trucks. A CB... That and dealing with distracted drivers, texting and driving.

My bendyx system slammed on the breaks, because the way the light reflected off a street sign. If that had happened on icy roads, I may have been in a ditch. In some ways, our "state of the art safety systems" are anything but safe.

Roy...No one is rejecting technology that consistently functions as designed and reliably. I tried to have a sensible exchange of information with you Roy. Your intent seems to be showing us all how smart you are, and how myopic and simple we are.

I’ll give you a thought to ponder...

Don’t focus on the symptom, focus on the root problems. Once you actually get out here, you’ll quickly understand what they truly are.

It’s not an issue of rejecting technology, it’s rejecting technology that so far is far from perfected being used to propel 40 tons of vehicle among other traffic, in multiple instances.

An autonomous car crashing is bad, but nowhere near as bad as a 40 ton semi. And I don’t even want autonomous cars. There is already someone inside, why not just drive it?