A Vision Of A Driverless Future

[CHamilton]

Uh, they're called trains.

There may be big truck-like things that travel in convoys, but there will also be transportation vehicles that function essentially like interchangeable, road-based shipping containers, with no dedicated space for a person. Indeed, someone might even develop something like container “skates” that you just roll a container onto as soon as it comes off a cargo ship, or out of a factory, helping the thing zoom along roads to its programmed destination.

A Vision Of A Driverless Future

 

[caravanman]

Good luck with the "zooming along roads" bit... Hope they can find somewhere for cyclists and pedestrians in that "brave new world"

Ed.

 

[jis]

Good luck with the "zooming along roads" bit... Hope they can find somewhere for cyclists and pedestrians in that "brave new world"

If any cyclist of pedestrian chances onto today's 70mph interstates in many states in the US, they are as much toast as if they get on a railroad track in front of a train. They are really not supposed to be there.

 

[railiner]

Like it or not, driverless cars are the future....

With so many of today's driver's, driving while distracted by various distraction's, pedestrian's and cyclists will actually be safer, when eternally vigilant robots take the wheel...

 

[Anderson]

Like it or not, driverless cars are the future....

With so many of today's driver's, driving while distracted by various distraction's, pedestrian's and cyclists will actually be safer, when eternally vigilant robots take the wheel...

Until someone hacks a major system and 50,000 driverless cars have their guidance systems go down at one time. I agree things will likely work out in some respect here...I'm just seeing a major hole at the start that can go very, very wrong.

 

[jis]

That can only happen if it is a centralized system, which of course is a wrong architecture to start with.

 

[railiner]

And they will probably install a 'hack-proof fail-safe' system, that will stop vehicles that are compromised, some how.....

 

[jis]

And they will probably install a 'hack-proof fail-safe' system, that will stop vehicles that are compromised, some how.....

Some level of fail safe would be essential. The only hope of success would be using autonomous vehicles that can behave reasonably even with complete loss of communication.

 

[fairviewroad]

If any cyclist of pedestrian chances onto today's 70mph interstates in many states in the US, they are as much toast as if they get on a railroad track in front of a train. They are really not supposed to be there.

It's legal to ride your bicycle on an expressway shoulder in 15 states. (though most of those 15 states prohibit it in urban areas). Whether that's a good idea or not is a separate question, but as a rule interstate highways have consistently larger shoulders than 2-lane roads, FWIW.

 

[jis]

Shoulder of highway is not a highway lane. But of course who knows what wayward drivers do?

 

[railiner]

Whether it is legal or not, I would not feel comfortable riding a bike along a high speed interstate highway shoulder....driver's are more likely to be distracted while driving, as well as fall asleep, and veer off the lane onto the shoulder, at 60-80 mph.......

 

[Guest]

Driverless vehicles are an issue much as aviation was only a century ago. They will be made safer than the alternatives before they are rolled out for public use. Who knows how long that will be, and many security scenarios will play out.

Air traffic control has worked pretty well so far....

 

[Anderson]

Driverless vehicles are an issue much as aviation was only a century ago. They will be made safer than the alternatives before they are rolled out for public use. Who knows how long that will be, and many security scenarios will play out.

Air traffic control has worked pretty well so far....

Except when it hasn't. We've gotten a long way in the last century or so with ATC, but there have been some spectacular FUBARs over the years...and that has generally been without a central interface that could be crashed.

 

[jis]

Nothing is 100% FUBAR proof. It is a matter of keeping the FUBARs infrequent enough, and in that respect the ATC and associated stuff like TCAS and training of personnel etc. seems to have worked quite adequately. However completely eliminating human folly is impossible

Harking back to the original comment, as far as I know there has been no case of 50,000 aircraft in operation being brought to a standstill due to the failure of something.

 

[fairviewroad]

.driver's are more likely to be distracted while driving, as well as fall asleep, and veer off the lane onto the shoulder, at 60-80 mph.......

None of which would happen with a driverless vehicle, which doesn't fall asleep or look away from the road to send

text messages/change the radio station/pick up the Big-Mac etc. Additionally, the driverless vehicle doesn't stop for

"just a couple of beers" on the way home, hasn't sniffed glue, or gotten into an argument with its ex. If functioning

as designed, the driverless vehicle doesn't forget to use its turn signal or pass in a no-passing zone.

So yeah, it could be hacked. But all of the above examples happen all the time and cause countless crashes literally

every single day in this country. The more I read about this technology, the more I'm inclined to give it a serious

chance.

 

[Devil's Advocate]

As someone who works in IT I'm not sure how the Internet of Things is supposed to avoid becoming a massive hack fest. New routers and protocols can help but I wouldn't bet my life on it. It's doubtful a true disaster will happen right away, but as millions of connected devices pass beyond their support contracts they will eventually become targets of perpetual misuse and abuse. This may sound like a minor issue at first but these systems are expected to talk to each other and control everything from heavy machinery to streetlights to life support systems and even self driving cars.

Some of these devices will be extremely expensive but others will be trivially cheap. Some will receive security updates for the life of the devices but others will lose their support almost immediately. The latest embedded devices are being designed to last up to a decade or more without external power. A device that costs almost nothing is unlikely to receive much in the way of security updates. Pair an unsupported device with a lifespan measured in decades and you could end up with some pretty nifty tools for everyone from casual hackers and to state sponsored terrorists.

Many of the ATC systems in use around the world are remarkable for how basic they are. As ATC systems become more advanced the potential payoff for improved safety is likely to be squandered on even heavier workloads. Rather than some sort of panacea I see the potential for a series of major disruptions and catastrophes. The puny wrist slap punishments levied against institutions caught with poor technical security and safety controls is likely to seriously exacerbate the problem and substantially delay better solutions.

In my view the primary issue with building a real life Skynet isn't the sudden creation of a self aware supercomputer so much as the much more mundane potential for conventional hacking being used to bring sudden and serious harm to large numbers of people.

 

[Anderson]

As someone who works in IT I'm not sure how the Internet of Things is supposed to avoid becoming a massive hack fest. New routers and protocols can help but I wouldn't bet my life on it. It's doubtful a true disaster will happen right away, but as millions of connected devices pass beyond their support contracts they will eventually become targets of perpetual misuse and abuse. This may sound like a minor issue at first but these systems are expected to talk to each other and control everything from heavy machinery to streetlights to life support systems and even self driving cars.

Some of these devices will be extremely expensive but others will be trivially cheap. Some will receive security updates for the life of the devices but others will lose their support almost immediately. The latest embedded devices are being designed to last up to a decade or more without external power. A device that costs almost nothing is unlikely to receive much in the way of security updates. Pair an unsupported device with a lifespan measured in decades and you could end up with some pretty nifty tools for everyone from casual hackers and to state sponsored terrorists.

Many of the ATC systems in use around the world are remarkable for how basic they are. As ATC systems become more advanced the potential payoff for improved safety is likely to be squandered on even heavier workloads. Rather than some sort of panacea I see the potential for a series of major disruptions and catastrophes. The puny wrist slap punishments levied against institutions caught with poor technical security and safety controls is likely to seriously exacerbate the problem and substantially delay better solutions.

In my view the primary issue with building a real life Skynet isn't the sudden creation of a self aware supercomputer so much as the much more mundane potential for conventional hacking being used to bring sudden and serious harm to large numbers of people.

This is more or less what I've been thinking (and hence why when my 1987 Mercedes finally up and dies I want to replace it with a Lincoln from the 60s: As useful as Skynet OnStar and the like can be, I really don't like the potential for bad behavior with all of the fancy electronics.

With self-driving cars, all you really need to do in order to cause problems is to simply mess up some GPS calculations by a few yards...preferably at rush hour. The other problem I see is that if you can dispatch an unmanned car to pick someone up and you can dispatch an unmanned truck to deliver a crate of goods, you can dispatch some *ahem* not-so-benign cargo.