Driverless cars: Google in the front seat, hanky panky in the back

Canada Free Press ^ | 05/12/16 | Dr. Klaus Kaiser

[Sean_Anthony]

I now Understand...

What’s so enticing about the (just-around-the corner) self-driving cars?

In view of my slightly advanced age, I may be forgiven to be a bit slow in learning—but now I understand: It’s all about what happens in the back seat!

Special: Do You Like Entering Contests & Big Prize Sweepstakes? As the Globe and Mail reports, Kirk, of the Canadian Automated Vehicles Centre of Excellence, told the Canadian Press on Monday that “… once computers are doing the driving, there will be a lot more sex in cars.” Another (biased ?) pundit, Sergio Marchionne, is claimed to have stated that the new system “… will be fundamental to delivering automotive technology solutions that ultimately have far-reaching consumer benefits.”

[AnotherUnixGeek]

We don’t and the current systems don’t. Of course in this age of bluetooth they probably will BUT not through centralized control, they’ll actually talk TO each other.

We agree that cars will use some secure networking protocol to communicate and coordinate actions. This is de facto centralized control, or anyway, the ability for remote control - the ability of all cars to coordinate movement or come to a stop via input means the ability to receive communication over an encrypted channel which, with the appropriate credentials, will be able to bring your car to a stop, limit speed, etc. This is technically feasible, so it will happen.

There are even tangible benefits to such remote control - automatically slow or stop fast-moving traffic on a freeway at night if law enforcement knows that a severe accident a mile ahead will require shutting down the freeway, etc. And the growth of control over cars by the government will inevitably grow.

[discostu]

No it isn’t. There’s a massive difference between querying other cars on the road for what they’re doing in the next 2 minutes and centralized control. The biggest thing being the CONTROL. A bluetooth querying system gives ZERO control outside the vehicle, it’s just another source of data to be processed. It would be actually a REALLY BAD IDEA to allow that system to take control, because then you have a hackable system, dead people, and lawsuits the car maker will lose.

[Boogieman]

No, it’s just a random picture of a sweet tricked out van, dude.

[AnotherUnixGeek]

No it isn’t. There’s a massive difference between querying other cars on the road for what they’re doing in the next 2 minutes and centralized control. The biggest thing being the CONTROL.

The difference in terms of code is small - some new drivers and controller logic. The difference in terms of hardware is also fairly small - manufacturers will add or enhance on-board computing power to process input commands for automated systems. Some cars already have the ability to brake automatically under imminent-collision circumstances - an on-board computer reads radar input and hits the brakes. Not much code would be needed to also have the computer hit the brakes in response to a network command.

It would be actually a REALLY BAD IDEA to allow that system to take control, because then you have a hackable system, dead people, and lawsuits the car maker will lose.

The argument will inevitably be made that the number of deaths caused by hacking into a car's remote control ability will be far smaller than the number of deaths caused by simply letting people control their own cars. And it will probably be true.

[discostu]

No actually the difference in terms of code is HUGE. Designed correctly the communication system is completely separate from the control system and gives data to the control system upon request. A centralized control system couldn’t have that barrier. Automatic braking is based on any kind of communication system that’s in the control system.

Again, what you’re insisting WILL happen could have been happening for the LAST TWENTY YEARS thanks to OnStar and similar and IS NOT. The fact of the matter is it’s not a really good idea, involves way too much overhead, opens the door to way too much liability, and just isn’t going to happen.

[AZLiberty]

They’re going to sell a lot more self-driving vans than the EPA will stand for.

[AZLiberty]

And when a self-driving car drives itself into a bridge pylon, who will be blamed for the “suicide”?

[central_va]

Are they working on cyber road rage?

[AnotherUnixGeek]

No actually the difference in terms of code is HUGE. Designed correctly the communication system is completely separate from the control system and gives data to the control system upon request. A centralized control system couldn’t have that barrier. Automatic braking is based on any kind of communication system that’s in the control system.

The implementation of such a system is very doable - I'm not talking about central control of all driving, which would be an impossible computing effort today, but remote control on-demand. This would enable external control to stop a car on demand, limit a car's speed and then progress to limiting route access due to traffic conditions.

The ability to stop a self-driving car from the outside will be a law-enforcement and emergency medical necessity - or do you think the police will be OK with a criminal fleeing in a locked, self-driving car which attempts to avoid all collision and ramming attempts more effectively than a human being could and relentlessly proceeds toward a destination known only to the occupant of the car? Incrementalism will be the way forward in both self-driving technology and in governmental demands to regulate the operation of self-driving cars. The end-point over the next 50 years for self-driving cars will be as another form of mass transit, with no need to persuade drivers to give up their cars or to invest in massive new infrastructure projects.

[gigster]

Cue: AC/DC’s “Girls got Rhythm”. (Back-Seat Rhythm).

[discostu]

It is doable, it’s what they’re going to do. And doing that means central control can’t happen. Remote control on demand IS central control, even if only on demand, and nobody has any plans for that, and doing it would be a very bad idea.

No it wouldn’t. The only thing you’d have to do with self driving cars when it comes to dealing with emergency vehicles is program them to follow the law, oh look bubbles, pull over and stop. That’s all they need. That’s all they’re going to get.

All your arguments are refuted by OnStar. Police have technically had the ability to do those things to a significant chunk of the fleet for TWENTY YEARS and every time there’s a car chase on Fox it proves they DON’T. And these system will be specifically to not let that happen, allowing that to happen will get the car companies sued to hell and back. They’re already getting their butts kicked because they over-tied the entertainment systems to the main computer and people can hack in through there and disable the car. These lawsuits will teach them to do everything in their power to NEVER do what you think is inevitable.

[AnotherUnixGeek]

No it wouldn’t. The only thing you’d have to do with self driving cars when it comes to dealing with emergency vehicles is program them to follow the law, oh look bubbles, pull over and stop. That’s all they need. That’s all they’re going to get.

That's not realistic. The pattern of both government and law enforcement historically is that they seek more control and more power. The ability for a police officer to be able to safely stop a car which is driving unsafely, carrying a fugitive, etc. is going to be far too compelling to avoid, and that ability can't be based on the officer having a police car with lights flashing that another car will need to visually recognize as something separate from streetlights, Christmas lights, etc.

There's a RAND report which touches on some of this, which might be of interest. The report involved a Web Technologies workshop involving both technical people and law enforcement people. One of the needs identified by the workshop:

"Controlling automated vehicles

Need methods for control and manipulation of automated vehicles (when warranted). For example, law enforcement may want to direct a parked vehicle to move"

It's not listed as a top priority, but the fact that it's there at all is an indicator of things to come. As I've said, we're very early in the IoT age and that includes self-driving cars. Such demands will only grow.

[discostu]

It’s perfectly realistic. Again:
OnStar
Same predictions
All false
TWENTY YEARS

We already know you’re predictions WILL NOT come to pass. They’ve already failed for 2 decades, including a Clinton an Obama presidency. If they aren’t using OnStar, a system SPECIFICALLY DESIGNED to allow such a thing, they won’t with self driving a cars, which NEED to have a system specifically designed to AVOID such a thing.

There’s twenty years of proof that you are quite wrong. You can keep ignoring it if you want. But that’s on you.

[US_MilitaryRules]

Just think. The gov can execute people at will.

[AnotherUnixGeek]

There’s twenty years of proof that you are quite wrong.

Well, no. There aren't, or I wouldn't consider it an inevitability. OnStar has the ability to track drivers and to refuse them access to their cars. It does not have the ability to pilot cars. It's relevance to self-driving cars is limited.

You can keep ignoring it if you want. But that’s on you.

We're not going to agree on this, but events will tell us what will happen in the decades to come. I'd be perfectly happy if they don't, but there is no reason to believe that government and law enforcement will not seek exactly the achievable technical implementations I've described - in fact, projections against this take an astonishing amount of faith in the ability of government to restrain it's desire for ever-increasing control over people, a faith which defies all historical evidence and a faith which I don't possess.

[discostu]

Yes there are. OnStar has the ability to turn cars off remotely, which is one of your chief predictions, and they aren’t using it for that. It’s relevance is that for the last 20 years they’ve been able to do some of what you’re predicting but they are not.

It’s not about government and law enforcement. It’s about technical ability and legal liability. That’s why your predictions are NOT being used now with OnStar and will never happen with self driving cars. I haven’t said restraint of the government, that’s you making more bad assumptions. I’m talking exclusively about the massive technical problems for any level of centralized control, the even more massive legal liabilities that come with adding a hackable external interface, and the complete lack of gains for the companies for same. The government’s desires don’t enter into this.

[AnotherUnixGeek]

OnStar has the ability to turn cars off remotely, which is one of your chief predictions, and they aren’t using it for that.

OnStar has no ability to bring the car to a safe halt, manuevering out of traffic.It can disable the starter and control the locks.

It’s relevance is that for the last 20 years they’ve been able to do some of what you’re predicting but they are not.

There is no relevance. No one ever claimed the government would seek to stop drivers from starting their cars - this is a strawman. What I said, and what will almost certainly happen, is that government will seek to control how and where drivers use their cars.

[discostu]

In a high speed chase off is off. And yet they don’t do it.

Sorry but there is relevance. I know you don’t like it because it proves you wrong. But that’s a problem with your position, not the facts. The facts for the last 20 years clearly show the idea of the government taking control of self driving cars is 100% pure paranoid hokum. It’s no strawman, it’s empirical evidence that you’re barking at the wrong tree. What you said will not happen, because for it to happen car companies would have to make very bad mistakes that would get them sued out of existence.

[AnotherUnixGeek]

In a high speed chase off is off. And yet they don’t do it.

OnStar controls the starter - it has no ability to stop a running car, or stopping ignition in a running car. It can only disable the ignition starter. It's not hard to see why. Think about turning off the ignition of a car moving at high speed. Think of the damage this would do - you'd bring the car to a violent halt, also cutting off all power assist to the steering, most likely causing serious damage not only to the occupants but to surrounding traffic.

Think of instead bringing the car to a safe, controlled stop, and it's not hard to realize the attractiveness of such control to law enforcement. That's why they're already starting to propose ideas about this.

Sorry but there is relevance. I know you don’t like it because it proves you wrong.

It has nothing to do with what I like - it has to do with the facts. What I'd like is to be able to have the faith in government restraint that you have. Unfortunately, reality and technological feasibility argue against this.

[discostu]

Incorrect. OnStar can turn a car off http://www.zdnet.com/article/gms-onstar-remote-vehicle-slowdown-feature-stops-carjacking/ . Turning off a car that’s moving doesn’t bring it to a violent halt, it just makes it coast. Cutting off power to the steering makes it hard to avoid problems, but it’s slowing down so those problem are probably lessened.

It is attractive to law enforcement, but they don’t use it willy nilly. And hacked vehicles are very NOT attractive to law enforcement.

You finally said something correct. It does have to do with the facts. And the facts are that all your same predictions were made 20 years ago and none of them have come to pass. Thus proving you quite simply wrong. And once again I have expressed NO faith in government. Funny how people arguing your side always have to insist the other person is having faith in government even though that has not appeared once in my posts. And I’ve pointed that out to you before. So stop lying about what I’m saying. And pay attention to recent history, it proves you wrong.