Level 5 driverless cars by 2021 can be done, say Brit industry folk Over the weekend, chancellor Philip Hammond boasted that “fully driverless cars” would be on Britain’s roads in four years’ time. Some in the driverless car industry think this is a dangerous fantasy, while more high-profile driverless car software companies are all in favour of it. The chancellor’s announcement was made as …
@ disgustedoftunbridgewells - "Wake me up when I can drive a car to the pub and legally have it run me home."
Bingo. That's what I want too. Autonomous cars save the country pub!
I'm also sure there is a market for cars that can drive people who are unlicensed or incapable of driving.
That would be a genuine game changer for some. Physically incapable of driving but can set a route on an accessible app/device which can be sent to the car and can drive own wheelchair into car.
More practically I can see convoys of autonomous trucks on designated motorway lanes running through the night. They'd stop at local depots and hand over to humans who will drive through towns filled with similarly unpredictable humans but in an easier to insure way.
But personally I favour the pub idea.
@ 0laf
Game changer for me if car able to take those physically incapable of driving - partner & I (whoever is available for that appointment) currently spend a lot of our "free" time essentially being a taxi service for ill relatives to doctors / hospital as public transport dismal and taxis extortionate.
...Being realistic, I doubt "sit back and let the AI do it all vehicles" will be avaialble when I'm too old to drive safely
Game changer or killer app for or me is, that I can go to a city center to take care of something (pick up package from post/shop/dentist...) and during that time car will drive around the block, as there is never parking place available where and when i need it. When I am done, I can summon the car to appear where I need it.
Short term parking place solution in other words.
The benefits are immense... and it's true 'drinking' is at the top of the list:
Can drive you home when you're too drunk to drive
Can come get you when you want a 'taxi'
Can drive people with mobility issues / sight issues
Can take control of heavy tonnage vehicles and drive at night time, freeing up daytime traffic for commuters
Can go pick up your kids for you or shuttle them to their sports / school / activities.
Will also pretty much dispense with the need for a personal vehicle since you just 'call' one on demand, so in theory, less traffic, more parking, less accidents.
I also think there'll be lots of negatives, mainly the car industry, since not everyone will need a car now, and noone needs to learn to drive, so that whole industry will tank. Include taxis, road freight and even today's uber drivers will disappear we'll have this massive pool of unskilled labor that we'll have to re-train to do 'something'...
Will also pretty much dispense with the need for a personal vehicle since you just 'call' one on demand, so in theory, less traffic, more parking, less accidents.
I am not sure I want to share a car with the remains of the previous occupant's kebab or worse, as it looks like the major use case is taking pissed people home. At least a taxi driver cares passionately about keeping his vehicle clean because he has to sit in it too.
"I am not sure I want to share a car with the remains of the previous occupant's kebab or worse, as it looks like the major use case is taking pissed people home."
It's hard to disagree with the imagery you just created, but on the other hand, just how many people really are that pissed or that inconsiderate on a night out? Is it really as bad as you portray or is that more of an impression caused by how people always remember the outliers?
Assuming most people end up using on-demand self-driving cars, then you will have a choice as to which company to go with. No doubt there will be a range of options. Some cars made all cheap and plastic and can be hosed out like a superloo by driving to a cleaning station, others may have people who do the cleaning. It's not beyond the whit of man to include sensors which can detect when people have vomited in a car. It tends to be smelly.
"I'm also sure there is a market for cars that can drive people who are unlicensed or incapable of driving."
And here in lies the elephant in the room when it comes to these driverless cars. It does, obviously, have the potential to bring millions of people back on the road. However, even today, roads are struggling to cope with the number of cars already on there. Plus London and (for some reason) Birmingham really don't want cars coming in to the city centres. They want you to use cattle trucks what pass for buses and trains to get there.
There's also the issue with charging the cars. Sure, plug them in to the grid. Well the grid is under pressure every winter, getting more so really, so how will that cope? Plus where do you park them? There are apartments being built in one part of Birmingham where there will be 200 apartments built but only room for 80 cars. So we can park them outside in the city or suburbs, but then we need charging points all around the place to cope with that. Then, say on a long journey and you need a bit of juice, where do you charge them? Can you imagine, with the increase in cars, the queue at the former-petrol-stations-turned-leccy-stations?
I think, if they do hit the road, then they will be the preserve of taxis and ambulances ferrying patients from home to the hospital for an appointment. I don't think, unfortunately, Joe Public will get them.
"It does, obviously, have the potential to bring millions of people back on the road. However, even today, roads are struggling to cope with the number of cars already on there."
It could get more people on the road but its not clear it would get more cars on the road. Currently most cars are privately owned and driven point to point with often the driver being the only occupant. With driverless cars you would either have multi-passenger vehicles or else driverless pods (essentially tiny 1-person cars used for commuters unwilling to share their space) that occupy much less space on the roads .Having a completely new technology such as driverless cars and continuiing to use them in the same way as human-driven cars does not make sense (althogh car ownership has a strong cultural status / significance among many people)
"Plus where do you park them?"
As above, more people does not necessarily mean more cars. Also, driverless allows them to be parked further away from the origin/destination as they can be called when needed, so there is plenty of scope for finding enough parking spaces.
"There's also the issue with charging the cars"
It's not a given that driverless cars will be electric, they can well be ICE or hybrid, In any case, yes, with increased use of electric cars, grid capacity will need to massively increase.
"Then, say on a long journey and you need a bit of juice, where do you charge them? Can you imagine, with the increase in cars, the queue at the former-petrol-stations-turned-leccy-stations?"
Motorway petrol stations currently have 8-12 pumps and hundreds of parking spots, the physical space is there. If it takes 5 minutes to fill a tank and 30 minutes to charge a car it would require 50-60 charging spots if all vehicles are electric. We're still a few years from charging a battery in 30 minutes, but then again we're years away from phasing out ICEs. Yes, it will require big investment in infrastructure over 20-30 years, but no more different to the investment in petrol stations over the years.
I'm not sure you can argue that "driverless pods" will take up less road space than a car, as most of the space taken up by any vehicle is actually the safe stopping distance in front and behind it. You can use this argument for trains of pods on motorways, but that mode of operation probably won't be useful in urban commuter environments where a lot of the congestion is.
We already have shared ownership schemes for conventional vehicles and it hasn't revolutionised motoring - there's really no reason to assume most people will suddenly take to it just because the cars involved are autonomous. Where there will be new takeup of shared ownership AVs it will be from people who cannot drive, but that's really only going to move passengers from public to private transport. This will increase traffic overall as cars take up more space per passenger than buses do.
If driverless cars take themselves further away to park, it will mean they are staying on the road longer. You can also bet that some driverless car users will tell the vehicles to keep driving around nearby to facilitate a quick pickup, rather than letting the machine park at a distant location and having to wait for it to come back. Both of these things would tend to increase traffic as driverless cars became more common, even if the numbers of passengers remained static (which they won't).
I'd also suggest that keeping traffic flowing is one thing that human drivers are actually very good at. They will often cede "right of way" to let oncoming traffic pass, especially in unusual situations (e.g. a stopped vehicle blocking one side of the road). The communication between drivers that permits the temporary bending of normal road rules can be quite subtle, so I suspect AVs might get stuck in situations that a human driver would negotiate quite easily - possibly creating a tailback nehind the AV as a result. The worst case scenario is two AVs getting mutually stuck in opposite directions, causing gridlock.
My gut feeling is that increasing use of AVs will massively increase congestion.
Ever tried driving in heavy congestion with adaptive cruise control turned on? Do it and I bet you'll turn it off again in very short order because the ride is just f'ing annoying. Constant cycles of acceleration followed followed by braking as it tries to maintain a fixed gap. Much smoother with a human at the controls (generally speaking). Granted the tech is probably rather crude compared to what is being developed for AV's however I personally suspect that most people who drive themselves today will find being driven around in an AV a frustrating and stress inducing experience.
> Ever tried driving in heavy congestion with adaptive cruise control turned on? Do it and I bet you'll turn it off again in very short order because the ride is just f'ing annoying
I suspect ACC is getting better and better. I just got a new car with it and it's perfectly ok in congestion and I suspect is at least trying to maintain a gap proportional to the speed as in slow traffic it closes the gap to a few feet. Equally it's wary of undertaking with a large gap in front of me, but less concerned if there is a small gap in front. It's pretty good in city traffic, too. The only annoying bit is in mild-to-heavy traffic on multi-lane as it starts to slow down quite early with a slower moving vehicle in front, meaning you have to pull out a lot earlier than you might normally do, but it does seem to take account of relative velocity when pulling out behind a faster moving vehicle, so it doesn't instantly brake.
"You can use this argument for trains of pods on motorways, but that mode of operation probably won't be useful in urban commuter environments where a lot of the congestion is."
I don't see why not. The linkage between vehicle is virtual. There's no reason a line of cars/pods can't be driving along with inches between them at slow city speeds in a snake-like formation getting most of it's advance data from the front vehicle. Most are going more or less the same way and when a car/pod needs to turn off, it and the ones behind only need to slow a little to give it room to manoeuvre. Even if we still use existing traffic light systems, a platooned line of cars can all start moving at the same time so more will get through the junction before the lights change back. With manual cars, you are relying on people paying attention and all pulling off one at a time, after the one in front has started.
Most of the improvements will only happen when there's a critical mass of self-driving vehicles and manuals are banned from city centres and/or other zoned areas.
"Where do you park them?" is, I think, the best use case of all for driverless cars. Because you can park them wherever you like.
Car takes you to work in the morning - you tell it to go away, as far away as necessary to find a free parking spot, then come back and pick you up at 5 p.m. Or you can send it home to park safely in your garage, until it's time to come and pick you up. It'll take a while, but within 20 years or so commercial car parks will be a thing of the past - no-one will pay anything to leave their car anywhere for more than a couple of hours, max.
Of course, the next logical step is to stop owning a car entirely, and rely solely on Uber/whoever's driverless fleet to ferry you about on demand. But that's a whole further step.
"Of course, the next logical step is to stop owning a car entirely, and rely solely on Uber/whoever's driverless fleet to ferry you about on demand. But that's a whole further step."
That is only a feasible step for some use cases.
In other cases, long term control of a vehicle is useful or necessary. Given charging schemes that may involve rental, leasing, or ownership, the economics of which must be analyzed on a case by case basis.
Consider someone who leaves home for the office (with a computer and other work related materials), then goes from the office to a scuba lesson (with 50 kg of scuba gear), then goes to a late dinner in entirely different clothing...
Or the person who goes out for a day that includes some photography with $15,000 worth of cameras and lenses, secured in the car when not in use... possible a work day with recreational photography, possibly a photographic assignment followed by non-photographic relaxation. Trust me, you do not want to constantly babysit even one $5,000 camera/lens combo when you don't have to.
Or the person who goes off for a weekend, or a week, in the woods, for some relaxed hunting, where the car serves to store food, camping gear, guns, and ammunition.
It is far more likely that future cars will be hybrid, internal combustion, or external combustion.
A recent Swedish study estimates that the carbon footprint for manufacturing a car battery pack is equivalent to 8 years of driving a conventional vehicle... and I believe I have seen that the life of such a pack is 5 to 7 years... all that before accounting for electricity generation.
Using a hybrid, ICE or ECE means refilling in minimal time, from existing infrastructure. That also gets around the significant range limitations, which are likely to be even more crippling when the temperature hits -20.
Win, win win.
"Yes, it will require big investment in infrastructure over 20-30 years, but no more different to the investment in petrol stations over the years."
On the contrary, the infrastructure will be much more expensive, complex, and failure prone.
And dangerous... indirectly.
Various problems - ice storms, electric grid failures, hurricanes, floods, forest fires - knock out the electric grid for days or weeks at a time, even in well prepared first world countries. When that happens, gas and diesel vehicles can still be refueled, whether by connecting a generator to a gas station pump, or by bringing in fuel trucks or containers of fuel... which helps a lot with coping with an actual or potential disaster. If most of your vehicles are electric, everything is going to be immobilized pretty quickly. And I shudder to think what could happen to an evacuation if electric vehicles start running out of power on the roads.
@ wolfetone: There's also the issue with charging the cars. Sure, plug them in to the grid. Well the grid is under pressure every winter, getting more so really, so how will that cope?
I know I risk being accused of pedantry but I'm not sure how true the statement is. Generating capacity is certainly under pressure, but whether the grid - i.e. the distribution network - is under pressure is less clear.
However, if electric vehicles do form a significant part of the total the grid certainly will come under severe pressure then. IIRC this house has an 80 Amp fuse on the incoming supply, which is about half of what it would need to be for a fast charger to be included. Then of course there is the bonkers government scheme to move away from gas for domestic heating, and that will require another 30 kW or so from the supply. OK let's have a 300 Amp fuse on the incomer, along with all the neighbours having the same.
It's a toss - up as to what would fail* first; the underground cable feeding us or the substation. The whole thing reeks of "politician think" rather than "engineer think" and as we all know from experience the engineering view will be ignored as being too "negative".
Thinking about it, with all the demand that could fall on the electricity industry I suspect that cables to estates would be of such a diameter that they couldn't be drummed ready for transport and installation; the coiling radius would simply be too big.
* spectacularly
Edit: @ jmch: I think you're a bit wide of the mark in your comment about the investment comparison with conventional petrol stations; petrol stations don't need a truly humungous electrical cable supplying them; their "electric only" counterparts will.
"However, if electric vehicles do form a significant part of the total the grid certainly will come under severe pressure then. IIRC this house has an 80 Amp fuse on the incoming supply"
This is the start of a good analysis.
Other things to consider... you are looking at charging at home, or work, I suspect.
Charging on trips is a bigger issue. Consider a fueling station on an expressway with about 20 pumps, capable of recharging a vehicle with enough gas for another 800 km of driving in less than five minutes. Most quotes for projected electric vehicles seem to be about 250 to 350 km for range... which means on a trip you will have to refuel three times as often... and it takes a lot longer. You would require hundreds of charging stations at each fueling station to handle the same expressway traffic...
Specifics -
A Tesla 'supercharger' provides about 450 km of range (IF you get the claimed range!) in 75 minutes, at 120 KW output (so the input will be more), or so says the maker... so you need 30 superchargers in two carefully situated refueling plazas to replace 1 gas pump. The nominal draw would be .120MW x 30 = 3.6 megawatts, plus overheads for losses... call it roughly 4 MW. A plaza with 20 pumps can be replaced by two plazas with 600 charging stations, each plaza drawing about 36 MW for chargers alone. I predict fun with wiring as transmission lines are strung out all over the countryside to connect these refueling stops with generation facilities.
Nasty reality:
You can do a 1,000 km trip in a day with a gas or diesel powered vehicle, but adding two or three 80 minute stops for charging makes it a much more difficult proposition. Given that sometimes I have done such trips in almost constant freezing rain and snow, running heaters, fans, wipers, etc, I rather doubt that I would get 450 km out of a battery, which would just mean more charging time and more charging stops... and more charging plazas.
With a little bit of investment, each charging station could have on-site generation using the same sort of fuel that is dispensed by petrol stations: burning the petrol, diesel or anything else that can be bulk transported to the site could probably be done cleaner than running an ICE in every vehicle, maybe with a gas-turbine generator...
"each charging station could have on-site generation using the same sort of fuel that is dispensed by petrol stations"
In general, small thermodynamic power generation units are less efficient and more costly per Kw than large ones. You may win on not laying cable, but lose on fuel, pollution, capital cost, maintenance, reliability, and thernal efficiency, compared to large central generators.
"And here in lies the elephant in the room when it comes to these driverless cars. It does, obviously, have the potential to bring millions of people back on the road."
That's the beauty of it.
Get rid of obsolete buses, streetcars, taxis, and replace them with vehicles that increase flexibility and options.
Convert transit-ways and streetcar tracks to 'autonomous control only' routes for express routing.
Use integrated local traffic control to improve traffic flow and reduce accidents, with automatic re-routing around problems.
Optimize delivery schedules and routes for material goods to minimize impact on peak human movement activity.
Provide facilities for sharing vehicles serially or concurrently (mode selected by users) to maximize effective throughput.
Use smaller vehicles to avoid shuffling empty seats around - a 20% full bus is a waste of space and energy. Avoid unnecessary vehicle re-positioning.
Optimize use of secondary routes and roads on a dynamic adaptive basis.
Minimize on street parking to provide more traffic lanes... use 'auto-valet' parking or sharing of vehicles to reduce need for parking.
Use data from vehicles, with appropriate modeling, to plan new roads.
To know that at least 40% of people dont live in big cities with infrastructure
To go shopping from my parents involves a 20 mile each way trip
As saturday's is the only day the bus runs local, instead of from the next village 5 miles away, how long would they have to wait for their AV?
hint, village has 50 houses, average 3 people per house, all will probably shop on Saturdays as that's market day in the town
Not so easy as you think, and add Surge Pricing as it's a good tax on the use of AV's
Doubt anyone outside London or Manchester size city's will want it
Newmarket Taxi's has a Tesla BTW, ask them how it fares?
@Olaf - Many who can not drive for medical reasons are not confined to a wheelchair. So are quite ambulatory. But still a game changer if it occurs in the next few years.
However I doubt it will occur in 4 years (40 might more likely) as there are numerous issues to be learned and resolved.
"@0laf - "I'm also sure there is a market for cars that can drive people who are unlicensed or incapable of driving."
I'm amused that you see those as two separate groups."
Perhaps because they are two different groups?
There are actually four groups:
licensed and capable of driving
licensed and incapable of driving
unlicensed and capable of driving
unlicensed and incapable of driving
The latter three can all use the services of a level 5 vehicle.
I should have thought that was fairly obvious.
@coppice
since a license has a pretty long expiry period, and a lot of things can happen to an able bodied person to render them incapable of driving, it's quite likely there is a group of people out there who are licensed but incapable of driving. Strictly speaking they can be separate groups. There is however major overlap between the 2 ofcourse.
"I'm also sure there is a market for cars that can drive people who are unlicensed or incapable of driving."
Level 5, the subject of the article, is defined as "steering wheel optional, full self-control in all conditions", so in theory you should not need a driving licence to be a passenger in one as "Johnny" takes you to your destination, especially if it's a model without the optional steering wheel.
"In the past - rural communities used a horse which knew its way home."
When a horse while eating is distracted by something that it needs to look at and listen to, it will stop chewing because it's brain is unable to process eating and paying attention to a potential threat etc, at the same time.
The horse is nevertheless still way smarter than your AV will be, the exhaust is good for your roses too.
That's pretty much my use-case for driverless cars - going to pretty countryside pubs, getting absolutely shatfaced, and getting home nice and safely afterwards. None of that crap city centre pub 'fun' on a Friday night followed by public transport 'fun' or wallet emptying taxi 'fun'.
Could do wonders for pubs that are a bit out of the way of the main routes.
"Wake me up when I can drive a car to the pub and legally have it run me home"
Nuts to that.
I want a car where I can say "Go thee hence and do not return until I can see that you have driven over Hammond (the politician, not the Top Gear chap). Make sure you reverse over him several times."
Wasn't me officer, It was the car wot dunnit..
Wake me up when someone tells us what these are going to cost and if there is any real reason to believe that enough people want one.
You can sell some of anything and there is a ready supply of suckers but this might need suckers with money and for the trendy ones it may lack appeal. Even footballers prefer to crash their own cars.
Driverless cars before driverless trains?
Am I the only one thinking that's the wrong way round? If you can do trains which are on tracks then some of what you learn from that can be transferred to cars. (think redundancy controls etc..)
In my mind the only way you are going to initially get self driving cars is on specified routes with sensors in the road and no meat sack drivers or laws that the meat sack is always in the wrong. (not good)
"They're much simpler roads than inner city or small town traffic."
Or rural routes.
Another aspect of rural routes is navigation. It's bad enough already with drivers following satnavs. At least the driver can (eventually) form an understanding that he's gone wrong and decide to call for the local HGV rescue to haul him out. I wonder how far an autonomous vehicle would carry on.
"At least the driver can (eventually) form an understanding that he's gone wrong and decide to call for the local HGV rescue to haul him out."
There are actually special satnavs for trucks that know all about restrictions. The problem is cheapskate operators and foreign truck drivers that won't pay for them.
Making it illegal to use a car satnav on a truck and checking them as part of our post-Brexit Customs fustercluck could go a long way to solving the issue.
"There are actually special satnavs for trucks that know all about restrictions."
I'm not sure how effective they are.
Most if not all the routing websites seem to know that the junction at the bottom of my road is too sharp a right turn for an HVG so will advise me to turn left and then turn back to avoid that. Because they don't ask what sort of vehicle I'm in I assume that they're taking restrictions into account when they know about the,
But the same websites have traffic cross over the route I'd normally drive out to to pick up the motorway and direct them down a twisty set of lanes from which the larger vehicles have had to be rescued. So I'd guess in that case that the navigational databases are sharing data which doesn't mark this as a bad route for HGVs.
"At least the driver can (eventually) form an understanding that he's gone wrong and decide to call for the local HGV rescue to haul him out. I wonder how far an autonomous vehicle would carry on."
If the stuck HGV driver had not cheaped out on a consumer grade SatNav aimed at car drivers, s/he'd probably not be there in the first place. Proper HGV oriented SatNavs take narrow lanes and low bridges into account (although not perfectly AFAIK)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
