Well as someone who uses the motorway to go to and fro from work everyday; I welcome these 'driverless' lorries, as the human lorry drivers really do not know what they are doing most of the time and I have to constantly watch out for them.
As a cyclist using a main road I feel a lot safer around lorries than around cars. Their drivers tend to know what they're doing and give you plenty of room. For one incident that was entirely my fault the lorry driver slammed the brakes, got out of the cab, checked I was ok and gave me a cheery "don't do that mate, you'll die". Contrast that with car drivers who will speed off even after hitting you.
Yeah it is different on the motorway though. It's much more boring driving and lorry drivers get tired, do other things, etc. I see them swerving a lot more than cars.
Is that really the case they don't know what they're doing, or more that they drive it all the time, so they know the limitations under normal conditions. (which of course can be fatal if anything unexpected happens, so an accident by definition) People do it all the time in all kind of jobs. They also do it with their own cars. (think how big the parking space had to be the first time you drove a car compared to space after a few years)
Either way, automated lorries will be better. Unless the designers + lawmakers say: hey they don't get into accidents see, now we can improve throughput by letting them go faster...
They could go slower. For long haul routes the truck could drive 24hrs straight, or longer, so driving slower could result in huge fuel savings while still getting to the destination on time.
A car cruising on a highway at 50 mph (80 km/h) may require only 10 horsepower (7.5 kW) to overcome air drag, but that same car at 100 mph (160 km/h) requires 80 hp (60 kW).[1]
What's wrong with increasing the speed? I'd rather the speed stayed constant and accidents went down than see speed going up and accidents staying constant.
It's important to consider whether increased speed does in fact cause accidents. Let's use data to determine an acceptable speed rather than arbitrarily decree "speed kills".
Are you implying that there is a scenario where higher speeds could result in fewer accidents? How could that work at current driving speeds? All the data we have right now about people driving shows that speed indeed kills. You have to consider both automated truck crashing into and being crashed by somebody - and that's usually going to be a human driver.
From my observations a major source of traffic clustering on 2 lane UK A and M roads is lorries overtaking each other. This seems to happen because some lorries drive at 50 mph and others at 60 mph. If all lorries were to drive at precisely the same speed they would rarely have cause to leave the slow lane and clustering of cars would be reduced.
Considering that car accidents are at least close to the #1 cause of preventable death in the western world...hell, maybe we should. There are lots of ways to handle reasonably high-speed, long-distance transit without giving every individual human or small handful of humans their own high-speed rolling projectile.
Driverless lorries might be more susceptible to hijacking IMO, because the risk for the attackers will be lower, both with regards to the act itself since they won't meet opposition from another human and also if they're caught because theft of an object is a crime less severe when the attackers didn't cause direct (physical or emotional) harm or threat to a human.
I'd suspect the opposite: It's pretty damn hard to take control of a vehicle that lacks human-accessible controls. Furthermore, I'd expect autonomous vehicles carrying any sort of valued load to also include tamper-detection; the truck would be calling for help within milliseconds of an unauthorised entry attempt.
Seems pretty straightforward to stop one of these with an impromptu roadblock. However they've also got an array of sensors with which to record the hijackers and alert the authorities with, precise location and all.
In a way, your point about authorities and highway patrol being alerted is actually a good thing. These trailers can go in lockdown and wait for authorities to arrive. Even if a human were on board, a silent alarm could be triggered without the human making any obvious moves for distress (perhaps a hidden panic button in case sensors don't trigger, assuming a human driver is required to be with the trailer). Also, trailers today aren't immune from hijacking/piracy if the driver isn't armed.
It's the UK, we probably have enough CCTV cameras everywhere to trace a masked hijacker all they way back to their house. Add to that that the trucks will doubtless be recording everything that happens around them and be able to be remotely disabled, I doubt we will see any of them stolen. Plus, is any truck driver really going to go above and beyond to stop someone stealing TVs off the back of current trucks?
One can only hope so. Or that they have some other major disadvantage. This is going to prevent a large amount of people from earning a living. Additionally it's going to kill a decent segment of SMBs who do truck driving and cargo transport on roads.
One more low-skilled job disappearing without replacement.
Long-term, jobs disappearing is not a bad thing. It really sucks for the people who are left without a means of supporting themselves, but do we, as a society, really want to use humans for jobs that machines can do equally well? I for one don't want to work just for works sake. Very few jobs are actually enjoyable and those jobs that we can currently replace with machines tend not to be of that kind.
Tech moves on and usually displaces labour but so far hasn't created mass unemployment. Instead it's made us richer. At any point in time people are concerned by the job destruction caused by progress, but after it has happened rarely want to rewind and go back to the old ways.
I'm curious - if you think this is a bad thing, are there technological innovations which have already happened which you would want to undo? How many jobs do you think it would create? GPS could provide be a good thought experiment.
There are clearly serious problems with wealth and income distribution. However, the size of the pie and the distribution of it are different problems (although not independent).
It doesn't follow that by prohibiting technological progress things would be more equal. I'm not sure what the relationship is between income distribution and progress; I think there are probably arguments both ways.
For example, would income inequality be rising even if technological progress was not happening? Probably yes, from my reading of Pikkety. Those with capital would probably continue to accumulate more. But it would affect returns to capital so perhaps it's not so clear cut.
Is income inequality increasing as a result of technological progress? Maybe, I guess? I'd be quite interested to see some real evidence either way because I don't think the answer is intuitive either way.
I know, it is sad we are eliminating all the Ice Carriers with Refrigeration. What ever will we do...
Will you not think of the Blacksmith when you replace your horse with one of these new motor vehicles....
Through out history we have had industries decimated by new technology, do you want to stop progress simply to protect jobs? At what point in history should we have done this? or why is today any different than at any point in our past? Should we time travel all the way back and kill that bastard that invented a way to control fire? Or the Guy that invented the wheel?
I like this underground autonomous cargo system better [1][2] (sorry in German, but you get the idea from the video images) It's autonomous, clean and inexpensive to build.
London used to have a similar system for distribution of mail throughout the city, officially known as the London Post Office Railway, or Mail Rail. I heard there were plans to open up some of the stations and tunnels to the public, not sure what they came of them though.
I used to work in Royal London House in Finsbury Square and the lowest subterranean level was sealed up but Royal Mail branded; given that it was a stones throw away from Mount Pleasant, I always wanted to gain access and see if it had access to Mail Rail or was even a station.
One station and a short section of tunnel are opening in 2017 as part of the Postal Museum. I forget all the detail, but the main limitation as to how far access is allowed is all down to the fact that evacuation was scarcely considered for the tunnels, given it wasn't really designed to carry passengers.
Here in Berlin they have been working to build 2.2 new kilometers of subway since 2010. They expect to finish around 2020, having spent 500 million Euros.
Building tunnels is really expensive. Barring radical longevity improvements, I don't think I'll live to see an underground transportation network.
At least from all the preliminary studies I've seen, electric power is not really practical for all but the smallest diameters. FYI I'm referring to the use of a rock-melting penetrator rather than a traditional cutting head. This is maybe just science fiction at this point, but it came to mind as the sort of technology that would make these sorts of projects economically feasible.
I think the term "self-driving" is preferable to "driverless" because it's not like there's no driver, the car is the driver. It's also less scary-sounding which could help speed up adoption. :)
Some jurisdictions have rules to restrict convoying, many trucks following each other as a group. I'm not sure I want to see road trains just to save a few bucks on that second/third driver.
Not having a driver in the cab supervising the robot is safer? I don't see any call to remove pilots from airplanes. And I doubt the extra 200lbs of driver makes much different given these are trucks.
Again, how does having or not having a driver in each cab affect this?
I would add that allowing multiple trucks to operate this way will require some changes to the law. And the redesign of many roads. Imagine trying to merge into traffic or onto a bridge when some vehicles are longer than merging lanes. It sounds like a great idea but isn't practical.
Truck five, if it has a human driver, needs to leave plenty of space so s/he can see what's going on with truck four and three. (Repeat this for all the lorries and they end up spread out) The computers don't need that visual space, because they're all linked with radio shuffling data back and forth.
> Imagine trying to merge into traffic or onto a bridge when some vehicles are longer than merging lanes
i) When the lorries pass approach a merging lane they either add some space between each truck, or they use a different lane of the motorway (because this article is talking about England) which allows people to move from the merging lane onto the motorway.
ii) Lorries will already drive in an informal ad-hoc convey. They'll keep a safe distance from each other. That safe distance isn't enough for a car to get into if the car is trying to merge from the entrance ramp. So if this isn't already a problem I'm not sure why it would be a problem with robot drivers. What's changed that suddenly makes it a problem?
trains
A tad flummoxed by this. I know rail freight has diminished, and there is the issue with transferring load from rail to train, but surely with clogged roads that is a better idea.
I cringe at the idea of self-driving trucks. If a car gets it wrong, it is probably a small bump in your door, if a truck gets it wrong, it can easily drive right over your car.
The flip side is that the cars will not be prone to exhaustion caused by unrealistic deadlines levied on human drivers -- speed will be regulated at the law (i.e., 55mph max).
Right now you have truckers driving for 18+ hours at 10-15mph over the legal limit. I'd much rather the road be trucked by software that won't speed and will defer to the side of safety always.
The fact that it is carrying freight, too, is a plus, as the truck driving software can always make the decision in favor of the other drivers on the road (whereas passenger self-driving cars will be focused both on the safety of the car itself and other drivers, a much more difficult decision tree).
e: my references to law are to US law, which isn't exactly applicable to this specific story, but the arguments stand (I'm sure there is a similar max-speed limit for trucks in the UK)
UK truck drivers are usually better drivers than most of the cars on the motorway. They don't like sudden changes in speed and their skill shows in traffic jams or roadworks, where the lanes of truckers will tend to keep moving. My major bugbear is trucks overtaking one another on 2 lane roads, when there'll often be a really small speed differential and so the overtaking manoeuvre will block the road for longer than needed.
That doesn't need robot drivers. That is and can be deployed today. Why it isn't ... that's an open question.
>>> "the truck driving software can always make the decision in favor of the other drivers on the road"
I've really wondered about this. Wouldn't software that acts like a jerk (ie not letting people in etc) be more valuable if it gets from A to B faster? Why think that the same pressures that force drivers to act like jerks won't also be applied to software?
> I've really wondered about this. Wouldn't software that acts like a jerk (ie not letting people in etc) be more valuable if it gets from A to B faster? Why think that the same pressures that force drivers to act like jerks won't also be applied to software?
This makes me think of VW.
> Winterkorn blamed "the terrible mistakes of a few people," whom he did not name, for the international scandal.
Now, to be clear I am not saying that Google's or Alphabet's CEO should go to jail. Rather the CEO of the company who owns the fleet should. Claiming that they didn't know what their employees did should not be allowed as a defense.
Yes, I am still mad at VW for throwing their software engineers under the bus here. I don't know what the solution is but it can't be to just say "well, the software engineers knew what they were doing and they should have said no when told to implement something illegal" (this is what I understand Uncle Bob suggested).
Wouldn't software that acts like a jerk (ie not letting
people in etc) be more valuable if it gets from A to B
faster? Why think that the same pressures that force drivers
to act like jerks won't also be applied to software?
Because with human drivers, a human is liable for failures. With software, the company is liable. If the software thinks "this is a dangerous situation", it will be written to defer to not causing injury, because there's really no way the contents of a truck are going to be worth more than the cost of paying out $x millions in the case of killing a human.
Check the fine print on your windows license. Then check the fine print on any aerospace autopilot. They disclaim everything. It's always the operators responsibility, never the manufacturers unless of a serious defect.
>> no way the contents of a truck are going to be worth more than the cost of paying out $x millions in the case of killing a human.
If that were true all sorts of other safety devices would already be on trucks, starting with gps-enabled speed regulators. There is always a cost-benefit analysis. That includes tradeoffs between utility and safety.
You are missing my meaning. By 'software', I mean the software that runs the self-driving trucks; both the companies building and using self-driving trucks could and would be liable, no matter what is 'in the fine print'. There is no operator, so blame has to go somewhere.
There are already safety devices on trucks, but since there is a human operator, much of the onus of using or installing those is on the human, not the truck manufacturer or trucking company; and again, you're missing my point that with no driver, the liability falls to the producers of the self-driving software.
>> with no driver, the liability falls to the producers of the self-driving software.
Not really. The owner of the truck is also the operator. Trucking companies are sued for damages caused by their drivers every day. If a truck company chooses to use particular software, as they choose a particular driver, the trucking company will be liable. They are the one that decided to put the vehicle on the road. That's why they must carry insurance and why that insurance attaches to the vehicle regardless of who is driving it on a particular day. Swapping out an employee-driver for a robot won't change the situation.
Even a truck that is stolen, a truck driven by someone the operator doesn't want behind the wheel, can result in liability should the operator not have taken reasonable steps to guard against theft.
> Why think that the same pressures that force drivers to act like jerks won't also be applied to software?
They will; but a human can improvise and lie, whereas software can (in theory) be audited and tested. The software will be built by manufacturers, not lorry owners; and they don't want to end up on the wrong side of car regulators -- see the recent VW affair.
You mean shrewdly duping regulators year after year so they can sell more cars, make more money, enrich shareholders, and become a market leader?
No matter what the cost today, I think VW shareholders would rather they have done what they did as opposed to giving up the market to competitors. (They all seem to have been playing tricks imho.)
But are they also automatically limited to 10/20/30 depending on the road? That tech has also been around for decades. Manufacturers haven't even attempted to market it.
That's actually a EU standard. Modern lorries are limited to 90 km/h (55 mph) today, to comply with EU standards.
In most European countries, lorries are limited to 80 km/h on motorways, with a few (including the UK) going fully to the limiter. Technically, the speed limit for lorries is 96 km/h on UK motorways, but newer lorries won't be able to drive that fast because of their limiter.
No, if a lorry is doing more than 96 km/h (older) or 90 km/h (newer), then they are messing with the limiter and basically putting it out of function.
A lot of lorry companies do this, unfortunately, to "comply" with the driving hours (or beat competition). So to ensure their drivers get further, they install magnets on their transmissions to prevent the Tachograph (and thus the limiter) from correctly reading the lorry's speed.
Indeed, this tactic is also used to trick the Tachograph into thinking the driver is taking a rest while he is actually driving. The driving hours are strictly enforced by police across Europe, and if the printout doesn't match the law, the driver and his employer stands to pay a huge fine. And both may also stand to lose their licence, depending on the scope of the crime.
(If they find magnets on the lorry, you can bet they are losing their livelihood.)
This already happens far too often. Just google "car crushed by truck". Given Google's results so far, driverless trucks are likely to reduce the incidence of this hugely.
A self-driving truck wouldn't have a blindspot, unlike a human-driven truck that you can't even safely pass on the right. It'll have a much easier time being able to see you in the first place.
I cringe at the breakdown rate. Loaded trucks cannot be moved nearly as easily as cars. But at least when they do there is a human driver to manage the situation. I've heard that the google cars are pretty slow at intersections. Lord help those stuck behind Optimus Prime when his gps fails.
Those concerns seem really mundane compared to the problem of a vehicle that can safely drive itself.
Just have humans be available to go deal with the trucks. Fine companies (a lot!) if they are too slow to deal with their trucks. This will encourage them to have arrangements in place.
As far as navigation, it's a technology product. They tend to get better over time, at least until some point where improvement becomes difficult is reached.
>> Those concerns seem really mundane compared to the problem of a vehicle that can safely drive itself.
Not specifically. Those concerns are to do with a driver-free truck, not necessarily one that is just driving itself. Autopilots still have pilots. I'd be all for an autopilot unit for a car/truck/plane alongside the driver. The difference is that these trucks seem to rely on the autopilot, at least to the extent that they are very distant from any human supervision. So the only real issue between autopilot+driver and autopilot solo is cost.
I don't trust people to pay attention, so I don't want them on the road if they require an attentive driver. I guess for testing requiring a driver is fine, but they should be taking video of the test drivers and evaluating their level of attention.
Planes don't really have so many things to collide with, so I don't think the choices made about autopilot there are instructive.
Regarding the issue of 10 lorries blocking an exit, our exits are very long, OK, not 10 lorries long but they wouldn't be block for more than a few seconds. If the convoy broke up into 3 or 4 while going past an exit then merged back together, or if they all pulled into the middle lane -- no worries.
Lorries drive up each others asses anyway, usually you just pull alongside and indicate and they'll make a little space for you to pull through. So pretty much business as usual.
30 yards a second at 60mph. Doesn't take more than a couple of lorries to cause a rolling roadblock, so 10 could easily start causing chaos. And accidents. You don't expect anyone to stop on the slip road, unless you see gridlock.
Now imagine you're in a continuous stream of rush-hour traffic and there are 10 contiguous lorries in the way. Actually even the existing lorries have to pull out into the middle lane there.
It must really suck to lose your livelihood like that, but its kinda inevitable. On the flipside, there will be jobs created for people to fix these machines, write the programs that run them, all the sensors and safety features that will be required before they can hit the road.
The UK is currently 50k short of drivers, same as much of Europe. There is a fear in logistics circles of this getting worse with fewer new entrants into driving - even with good pay.
Im not sure why George Osborne is going to talk about self-driving trucks in the Budget speech... Is he dropping phat-stacks of the budget on this programme to make the two related somehow?
