How long until a group of high frat bros hijack a self-driving Pizza Hut car?

^That's the problem with sooo much of this tech stuff. The people making it can only visualize how a calm 25-40 year old gentrifide urban dweller or suburbanite would use it. No teens, no kids, no old people, nobody in rural areas, nobody in the 'hood, nobody who is in poverty etc. When put into practice techies have no idea how the Rest of Life's will use it... or not use it at all.

Or one runs over someone and they BLAH BLAH BLAH MAJOR SPOILERS -ryanlammi edit

 

*Black Mirror reference for those wondering*

^That's the problem with sooo much of this tech stuff. The people making it can only visualize how a calm 25-40 year old gentrifide urban dweller or suburbanite would use it. No teens, no kids, no old people, nobody in rural areas, nobody in the 'hood, nobody who is in poverty etc. When put into practice techies have no idea how the Rest of Life's will use it... or not use it at all.

 

Well unfortunately all of these driverless-everything vehicles will need tons of cameras inside and out to record vandalism.  But I think people will figure out how to make them "blind"...i.e. wait for one to pass under an overpass, then drop a tarp on it.  The thing will stop and then you can break into it. 

 

 

and steal the one pack of cigarettes that the owner sent the car to go get

^That's the problem with sooo much of this tech stuff. The people making it can only visualize how a calm 25-40 year old gentrifide urban dweller or suburbanite would use it. No teens, no kids, no old people, nobody in rural areas, nobody in the 'hood, nobody who is in poverty etc. When put into practice techies have no idea how the Rest of Life's will use it... or not use it at all.

 

Well unfortunately all of these driverless-everything vehicles will need tons of cameras inside and out to record vandalism.  But I think people will figure out how to make them "blind"...i.e. wait for one to pass under an overpass, then drop a tarp on it.  The thing will stop and then you can break into it.

 

Or they'll just figure out how to hack them.

What if driverless cars never happen?

https://jalopnik.com/what-if-autonomous-cars-just-never-happen-1820778692

 

This article echoes a few thoughts I've had over the past few years.  Mainly that we keep being told that it's about to happen, but never does. 

What if driverless cars never happen?

https://jalopnik.com/what-if-autonomous-cars-just-never-happen-1820778692

 

This article echoes a few thoughts I've had over the past few years.  Mainly that we keep being told that it's about to happen, but never does. 

 

There's a lot more industry pressure to remove drivers than most of the other "failed" techs that the article points out. Certainly the trucking industry would absolutely love to not have to pay the people who have to drive for them. Other delivery services like Amazon, Fedex, and UPS are also keen to ditch drivers asap. Heck, I bet Papa John's and Domino's would be on-board with the change as well. I think 2020 is ambitious, but I'd bet by 2030 autonomous vehicles will be a significant percentage of the total vehicles on the road...

This article echoes a few thoughts I've had over the past few years.  Mainly that we keep being told that it's about to happen, but never does. 

 

I think of driverless cars as today's Jetson's flying car. A little less pie-in-the sky, but likely to be just as impractical in reality, particularly given all of the things that can and do go wrong with digital technology. Not to mention maintenance and upkeep, which are likely to be expensive.

 

 

And network hackers. Need I say more?

What's amazing is that it's 2018 and no driverless car has driven cross country.  In 2015, it is claimed that Delphi drove a driverless (of course a driver was sitting in the thing the whole time an allegedly did drive 1% of the distance, so about 300 miles) car cross-country, but there is little footage of the drive. Nobody is willing to risk inviting the media to cover a cross-country drive of a car with nobody in it because they're so worried that the media will catch some embarrassing footage and run it over and over again. 

 

I don't think the average person is going to be riding in a driverless car anytime soon but I do believe that large companies in the trucking industry will use them as soon as possible, along with companies like Uber and Lyft. It'll take over commercial driving but that is going to be really obnoxious because we're going to be on the highway, having to trust software to operate 18 wheelers that could easily kill us.

 

Even if there's the occasional bugs/glitches in the software that cause wrecks, statistics will show that automated cars are going to be less likely than a human to wreck and that is enough reason to make them legal everywhere. My cousin refuses to wear a seatbelt because he was involved in a wreck where a car smashed into the driver's side and the force threw him over to the passenger side, preventing him from being crushed to death. If he was restrained by a seatbelt, he would have been dead. However, statistics are on the side of seatbelts and that's why it's illegal not to wear them.

 

The trucking industry (which employs more men than any other industry) will be able to cut $50k-$100k a year in labor on each truck while increasing productivity as software doesn't need to sleep 8 hours a day and keep logs. That's reason enough for them to invest in the technology right away. Owner-operators who are used to making a couple hundred thousand a year are going to see a big pay cut because their agents are going to be drastically outbid by a truck that doesn't even require a driver.

 

A lot of people are going to be hurting from this, including all those prostitutes that hang out at truck stops  ???

 

Can you imagine how automated cars might effect car insurance companies? If automated cars are actually a lot safer, those insurance companies are really going to be hurting. Another thing is that this will be entirely new for them. People are going to expect them to calculate the risk of driverless cars and give consumers a better premium if they're proven to be safer.

 

There's so many things to think about, when it comes to driverless cars. I think that's why people are so skeptical.

 

Oh. Another random thought: When driverless cars pull up to a gas station, the person working inside at the register is going to have to be alerted that it arrived and they'll be expected to run out there and pump gas. The problem is that gas stations don't make but a couple pennies on gas. They bite the bullet on that because they know that people who stop at gas stations are going to come inside and buy pop, candy bars or a couple other things that actually have a good profit margin. Driverless cars aren't going to impulse-buy.

 

How do driverless cars pay for gas, anyway?

 

I have so many questions about how this will all transpire.

This probably sounds weird, but my biggest fear with driverless cars is electronic warfare from a country like N. Korea, manifesting in the form of hacking all of the driverless cars through their manufacturer's systems and wrecking all of them. That would cause so much hysteria, emergency response, damage and death all at once if driverless cars were to scale.

What's amazing is that it's 2018 and no driverless car has driven cross country.  In 2015, it is claimed that Delphi drove a driverless (of course a driver was sitting in the thing the whole time an allegedly did drive 1% of the distance, so about 300 miles) car cross-country, but there is little footage of the drive. Nobody is willing to risk inviting the media to cover a cross-country drive of a car with nobody in it because they're so worried that the media will catch some embarrassing footage and run it over and over again. 

 

 

Yeah, it made headlines when one of Uber's self-driving cars got caught running a frickin redlight. Like that never happens.

 

Someone actually died in a Tesla that was on auto-pilot and they made that a huge deal, as well. However, if you look at the number of hours of Tesla's cars driven on auto-pilot (or whatever they call it) and its resulting deaths or injuries vs. all of the hours spent in a Tesla with a human operator, I wouldn't be surprised if the auto-pilot mode was proven to be safer because you have to factor in dumb@sses who hold a phone up to their ear while driving and even text and drive, plus drunk drivers.

 

It's a psychological issue, more than anything. People need to feel like they're in control of a situation. That's always been the joy of driving a car - being in control, and people really believe they can beat a computer at something.

 

In the future, the source of road rage will be decisions that computers are making. That's so weird. You can't pull up next to a car and call their software an @sshole.

 

 

 

What's amazing is that it's 2018 and no driverless car has driven cross country.  In 2015, it is claimed that Delphi drove a driverless (of course a driver was sitting in the thing the whole time an allegedly did drive 1% of the distance, so about 300 miles) car cross-country, but there is little footage of the drive. Nobody is willing to risk inviting the media to cover a cross-country drive of a car with nobody in it because they're so worried that the media will catch some embarrassing footage and run it over and over again. 

 

 

Yeah, it made headlines when one of Uber's self-driving cars got caught running a frickin redlight. Like that never happens.

 

Someone actually died in a Tesla that was on auto-pilot and they made that a huge deal, as well. However, if you look at the number of hours of Tesla's cars driven on auto-pilot (or whatever they call it) and its resulting deaths or injuries vs. all of the hours spent in a Tesla with a human operator, I wouldn't be surprised if the auto-pilot mode as proven to be safer. It's a psychological issue, more than anything. People need to feel like they're in control of a situation. That's always been the joy of driving a car - being in control, and people really believe they can beat a computer at something.

 

In the future, the source of road rage will be decisions that computers are making. That's so weird. You can't pull up next to a car and call their software an @sshole.

 

I just want to know if the driverless car I pull up next to, knows that I want to race when the light turns green.

Those things had better cut .500 lights every time or else they're not going to be safe at anything else either.

^Maybe this will make me sound like an idiot but what does your post mean? Cut .500 lights? Is that gear-head lingo?

Tesla on autopilot strikes parked fire engine:

 

https://www.mercurynews.com/2018/01/22/tesla-on-autopilot-slams-into-parked-fire-truck-on-freeway/

Ugh. That's tragic. I'm not at all a car guy but I still hate seeing a beautiful $80k-$100k car smashed and destroyed like that.

 

I'm really curious - does that wreck count against the driver for bad driving in terms of future insurance premiums? It technically wasn't the fault of an operator unless you count it against him/her that they chose to buy a self-driving car.

 

If you can afford to buy a Tesla, I guess the price of your insurance premiums aren't something you need to be very concerned about.

 

^What's you're seeing is that even if driverless cars are 100X safer than ordinary cars, every single accident will get much, much more attention.  Like how airplane and train wrecks get 100x more attention than car wrecks do now. 

Truthpaste.

^What's you're seeing is that even if driverless cars are 100X safer than ordinary cars, every single accident will get much, much more attention.  Like how airplane and train wrecks get 100x more attention than car wrecks do now. 

 

Partially because it will be some huge corporation's fault rather than just one anonymous person's fault.

^What's you're seeing is that even if driverless cars are 100X safer than ordinary cars, every single accident will get much, much more attention.  Like how airplane and train wrecks get 100x more attention than car wrecks do now.

 

Most new commercial planes can fly themselves (ironically the only issue with a complete auto-pilot is the taxi). Yet people will always be reluctant to let a machine take complete control and when something goes wrong the spotlight will be very bright. Hence, part of the reason there are still one or two people at the front of the aircraft. It's going to be very hard for humans to psychologically give up complete driving control without long-term safety performance.

 

I understand that 99.9% of commercial planes are flown practically entirely by a computer, but I still want a human able to take control if needed. Think about the crazy landing that person had to do at the Amsterdam Airport a week or so ago. I wouldn't trust a machine to be good enough right now to make that landing.

 

Or the plane that had to make an emergency landing in the Hudson River.

I understand that 99.9% of commercial planes are flown practically entirely by a computer, but I still want a human able to take control if needed. Think about the crazy landing that person had to do at the Amsterdam Airport a week or so ago. I wouldn't trust a machine to be good enough right now to make that landing.

 

Or the plane that had to make an emergency landing in the Hudson River.

 

I agree about Sully and the Hudson River landing. But common sense didn't prevail with that landing at Schiphol and I think a machine would have made a smarter decision if given the choice.  I understand everyone made it in safely but there's a reason why there are alternate airports in a flight plan.

^Maybe this will make me sound like an idiot but what does your post mean? Cut .500 lights? Is that gear-head lingo?

 

A "Perfect Light" in drag racing is .500 of a second between the yellow light (or lights) turning off and the green light turning on. People add electronics to their cars to assist with these launches in some classes.

Airplane autopilot isn't a good analogy because an airplane doesn't have much to do between the high drama of the takeoff and the landing.  With cars it's the opposite. 

 

The military has been building jets that require continuous computer assist for ordinary flying since the 1970s.  That's what enabled the first stealth fighter and then the stealth bomber.  People don't want to believe that a self-flying airplane is much easier to build than a self-driving car, but it is by many orders of magnitude. 

 

 

Airplane autopilot isn't a good analogy because an airplane doesn't have much to do between the high drama of the takeoff and the landing. 

 

I'm glad it's always sunny, light winds and 75 when you fly -- even at 35,000 feet. I never realized air travel was so simple.  :)

Airplane autopilot isn't a good analogy because an airplane doesn't have much to do between the high drama of the takeoff and the landing.  With cars it's the opposite. 

 

The military has been building jets that require continuous computer assist for ordinary flying since the 1970s.  That's what enabled the first stealth fighter and then the stealth bomber.  People don't want to believe that a self-flying airplane is much easier to build than a self-driving car, but it is by many orders of magnitude. 

 

There's some truth to this... it mostly lies with the fact that the auto infrastructure is built around "everyone" driving. This makes it difficult for autonomous cars to predict bad drivers doing stupid things. As non-autonomous traffic is phased out and infrastructure is geared more toward autonomous vehicles, making reliable autonomous travel would be much easier.

Airplane autopilot isn't a good analogy because an airplane doesn't have much to do between the high drama of the takeoff and the landing.  With cars it's the opposite. 

 

The military has been building jets that require continuous computer assist for ordinary flying since the 1970s.  That's what enabled the first stealth fighter and then the stealth bomber.  People don't want to believe that a self-flying airplane is much easier to build than a self-driving car, but it is by many orders of magnitude. 

 

There's some truth to this... it mostly lies with the fact that the auto infrastructure is built around "everyone" driving. This makes it difficult for autonomous cars to predict bad drivers doing stupid things. As non-autonomous traffic is phased out and infrastructure is geared more toward autonomous vehicles, making reliable autonomous travel would be much easier.

 

I'm not arguing your point it's just the idea that "there's not much to do between takeoff and landing" isn't entirely true either. Weather is constantly in flux and as long as almighty climatologists and ordinary meteorologists can't figure out problems like clear air turbulence there's a level of complexity to in-flight cruise autopilot systems before we can move to pilotless flying. Mother Nature is the "bad driver".

Another writer speculates that level 5 driverless cars will never exist:

https://medium.com/@parismarx/despite-the-hype-ces-showed-driverless-cars-still-need-work-58103ec0f122

 

Rain and snow are distorting or preventing the sensors from working.  So are we going to have driverless cars 90% of the time, but then suddenly during bad weather everyone has to learn how to drive again? 

 

The other day when I was in line at the car wash -- not only would you need to clean off the sensors when you start driving, but other cars would inevitably kick slush and dirt back up onto every surface of the car in short order.  So until they figure out dirt and snow-penetrating radar, this can never work.  And if it doesn't work all the way it's of no advantage whatsoever. 

 

You can't have a system built around 2x more efficiency on existing roads, then suddenly lose that efficiency during every storm. 

 

^ A few weeks ago I was at a cabin in Hocking Hills when a snowstorm hit. When the time to leave came, we had a few miles of ice/snow covered, extremely hilly backroads to navigate before we got to anything resembling a treated road. About halfway through flooring it in low gear as I went down one hill, hoping I'd have enough momentum to get up the next, I thought to myself that it'd be really interesting to watch a driverless car attempt this drive. There's a lot of nuance to driving in snow, in addition to the technical issues that article points out.

 

I think the only widespread usage of driverless vehicles we're going to see in our lifetime will be for transit and commercial use. They would be limited to only certain routes, and both the routes and sensors/equipment could be maintained on a regular, daily basis. Think of all the inspections and maintenance an airplane goes through between each flight - a driverless vehicle would probably need a similar gamut. And in very bad storm/snow conditions, these vehicles could simply stop running for a few hours. People won't be as upset about the buses not running as they would be about their personal vehicle telling them it can't drive them anywhere because of the weather.

 

From cnn.com:

 

But 5G is about much more than smartphones. Sensors, thermostats, cars, robots, and other new technology will all connect to 5G one day. Today's 4G networks don't have the bandwidth for the vast amounts of data all those devices will transmit.

 

5G networks will also reduce to virtually zero the lag time between devices and the servers they communicate with. For driverless cars, that means uninterrupted communication between a car and other vehicles, data centers and outside sensors.

 

To accomplish all that, 5G will need to travel over super-high-frequency airwaves. Higher frequencies bring faster speeds and more bandwidth. But they can't travel through walls, windows or rooftops, and they get considerably weaker over long distances.

 

That means wireless companies will need to install thousands -- perhaps millions -- of miniature cell towers on top of every lamp post, on the side of buildings, inside every home and potentially in every room.

 

So can driverless cars work at all until this huge network is installed?  If not, it doesn't matter what the car companies are doing.

 

 

Yes, that was how 5G was explained to me a year or two ago. 5G just won't happen in rural or even exurban areas due to lack of ROI. Maybe 6G or whatever it's called won't be like that, but a lot of areas are going to be stuck with 4G.

My understanding is that 5G is more like a hybrid system. It will use some frequencies that are good for longer distances but at slower speeds; and others that have shorter reach but much higher speed. So the rural 5G speed will be less than the urban 5G speed, but I would still expect that self-driving cars on any interstate highway or major US route should still be able to get an adequate signal. I think driverless cars will still be able to operate just fine if they lose temporarily lose their connection. They just won't be able to download up-to-the-minute information about traffic in order to choose the fastest route, and stuff like that.

Vehicle connectivity and autonomy are related, but not the same. The current crop of autonomous vehicles that are operating in a limited capacity have little to be connected to, yet are "self-driving." "Connected Autonomous Vehicles" (CAVs) is the industry jargon.

Apparently they already have self-driving cars & vans in England:

Omg... If there were self-driving U-Hauls, I'd be completely down for that. I hate driving big trucks. Especially U-Hauls because you can't see behind you at all; you entirely depend on mirrors. The worst is pulling into a gas station near where you're dropping the truck back off, to top it off and having to maneuver around and back up and readjust while the gas station is packed.

Yawn.  We now have a consistent pattern with these stories:

 

1. clickbait headline tells you that some driverless vehicle just did something (stunt was not announced ahead of time -- of course).

2. No video or photos -- certainly none from an independent journalist.

3. article tells you there was actually a driver aboard

4. article tells you the driverless vehicle had *almost* no problems

 

Also, a level 2 driverless truck?  What good is that except for free publicity?  These things are useless until they are 100% level 5 autonomous. 

 

https://www.yahoo.com/tech/embark-self-driving-semi-completes-203400076.html

 

I just saw some tweets from the City of Seattle's traffic engineer (great person to follow, BTW) talking about how driverless cars might misinterpret flashing red lights. At standard traffic lights, a flashing red means "come to a complete stop and then proceed through the intersection when it's safe." However at railroad crossings, the exact same signal means "stop, and do not proceed until the lights stop flashing." I'm sure there are many other situations that we haven't thought of yet, where it would be obvious to an experienced driver what to do, but much more difficult for a computer to decide what to do.

I would think a lot of those sorts of kinks could get worked out reasonably easily.  A driverless car would know it was at a railroad crossing, and not a 4-way stop, for example, because it has GPS and isn't just looking at the light.  It would just have to be programed to behave differently at those different intersections.  For something fixed, and rules based like that, driverless cars will be safer than human drivers very quickly.  It's the unforeseeable that I'm more worried about.

The problem is that the GPS will dig up some old interpretation of the crossing much like Google Maps gives roads some old name that hasn't been used since the 1930s. Or how it thinks abandoned roads are still functional.

Right, you can’t depend on using GPS and maps data because both of those might be inaccurate or outdated. GPS signals bounce off tall buildings, so in urban areas, your GPS location could be off by an entire block. Maps will not reflect temporary road closures, detours, etc. It will also take some time for new and realigned roads to be reflected on these maps. So the car pretty much needs to rely on its own onboard sensors.

Auto-braking feature on team car causes bicycle wreck during Tour of Dubai, seriously injuring Mark Cavendish:

http://www.cyclingnews.com/news/abu-dhabi-tour-organisers-blame-automatic-brake-sensor-for-cavendish-crash/

I would expect a lot more resources going into making sure GPS maps are accurate and up-to-date when CAVs start operating in significant numbers "in the wild."

To be clear, autonomous vehicles will absolutely use GPS technology and maps for navigation. What I'm saying is that they won't be able to use GPS and maps for driving.

With all the 5G boxes being installed (and likely 6G or 7G or higher before CAVs really take off), as well as supplemental networked infrastructure, sensors may become a small part of what the cars need for driving. Even with the rare weak signal, the cars will "know" where they've moved to in their internal map, as they record movements of the wheels. Networked infrastructure will be able to update maps on-the-fly to account for not just construction but also streets closed due to crashes or other emergencies. Changes in traffic lights anticipated beforehand as the signals will communicate their timing. Networked infrastructure is going to be a big part of all this, especially in urban areas.

To be clear, autonomous vehicles will absolutely use GPS technology and maps for navigation. What I'm saying is that they won't be able to use GPS and maps for driving.

 

What happens when a GPS satellite is knocked out?  If a driverless car has no steering wheel, the thing is useless.  So a pretty easy way to knock out the United States circa 2100 would be to simply destroy our satellites -- not as thought that wouldn't cause total havoc right now. 

What happens when a GPS satellite is knocked out?  If a driverless car has no steering wheel, the thing is useless.  So a pretty easy way to knock out the United States circa 2100 would be to simply destroy our satellites -- not as thought that wouldn't cause total havoc right now.

 

When everything is networked, cybersecurity and physical security come closer to being the same thing. If you think Russian bots are bad on Twitter, wait until they've hijacked your autonomous Uber. Beware the IoT.

 

Or is it the IoS?

https://twitter.com/internetofshit

Death by Click

Supposedly China is going to launch its own proprietary GPS system.  We forget that "GPS" was launched as a system for the U.S. military and that civilian use required a signature by former president Bill Clinton.  If they turned it on for civilian use, they can turn it off.  Bye, bye Uber, Google Maps, etc.  Frankly, it's pretty amazing that the U.S. government -- which spent tens of billions to develop and launch the system -- doesn't collect payments from civilian use.