Occasionally someone will argue that we don’t need to invest in public transport as new technology like driverless cars will come along and render our roads much safer and more efficient. In some cases they say the impact on the transport system will be similar to the one that private vehicles started having on transport. At the transport debate just over a week ago Transport Minister even spoke about a future of diverless vehicles after being given a ride in one of Google’s prototype autonomous vehicles.
We’re told this technology is just around the corner and it will soon be on the market. However in many ways it seems much like a carrot that’s constantly hanging from a stick in front of us as despite the progress being made by Google the technology is still a long way away from becoming a reality.
Many motorists dream of the day they can sit back and relax while their car drives itself.
And while Google and other companies are working hard to make autonomous vehicles a reality, it could take years to create a car that can negotiate complex situations on the road – including wet weather conditions.
Google’s self-driving cars can’t currently cope in heavy rain or snow – or find their way around 99 per cent of the US, an insider has admitted.
According to MIT Technology Review, the current prototype cars are very reliant on maps to navigate and can’t react like a human driver, dodging potholes and other hazards.
Google’s cars have driven themselves over 700,000 miles (1,126,540km) but they can’t cope in snowy conditions and cannot negotiate heavy rain.
Chris Urmson, director of the Google car team, said this is because the detection technology is not yet strong enough to separate certain objects from weather conditions.
While the cars’ cameras can spot a traffic light changing, they can be confused by strong sunlight.
They don’t distinguish between an empty plastic bag – which could be easily driven over – or a rock, so cars must drive around both. They also can’t detect uncovered manholes or potholes.
Mr Urmson told the publication: ‘I could construct a construction zone that could befuddle the car.’
The cars ‘see’ pedestrians as moving blocks of pixels and know to stop, but unlike a cautious human driver, they could not spot a traffic policeman at the side of the road, waving for traffic to stop – which could lead to trouble.
Those seem like some fairly serious issues that need to be addressed before the technology could even be considered for public use. For their part Google thinks the issues could be resolved in 5 years-time but I suspect that timeframe could turn out to be a pretty wild guess. Going further, even if Google manage to get everything working fantastically within that time frame it’s unlikely there’ll be more than a handful in the country for quite some time and it would take decades before they’re owned in any quantities. For the time being at least it seems like we still not going to see any change to the status quo.
Processing...
21 century Jetsons stuff but beloved by oddball types like libertarians and the people that run the AA.
With many people going between the same places they could connect these driverless pods together, then if they’re connected you could make it possible to move between them for social interaction. Voila, a Brilliant User Space connector
We could make them bigger. Run them in their own corridor free from other vehicles. Pick up points where locals can ride or walk to.
Better still, instead of having each vehicle with it’s own power plant why not put one big motor at the front to pull all the others. It’s much more energy efficient. And if they ran on metal rails instead of rubber tyres you wouldn’t waste so much energy with flexing the wheels…
Oh, right….
….and attach it to our mostly renewably generated electricity grid. Then run these car things separated from other traffic at such a frequency that people could just turn up, for a small fee, and hop on and off when ever and where ever, paying with some kind of card or maybe just their smartphones: Freedom!
It would take a bit of investment to set up but once done would last for decades if not centuries and the city would grow up around the stopping points….
Now that, Patrick, is my idea of freedom. Being able to make choices on how to get around this city, perhaps much of the country.
I’ve been lucky enough to see one in action in Mountain View (Google’s home town). Two guys in the front seats working on their laptops, while the car drove itself. Quite impressive to see it change lanes (on a 3 lane road), stop at the lights, keep pace behind the car in front, even if it was a perfectly sunny day in moderate traffic.
Nikki Kaye ranted on about how cities were focusing on new novel technologies for cities and not relying on old fashioned ones (presumably PT and walking and cycling) in the same breath that she claimed ‘her’ Vic Park tunnel had taken 10,000 cars off the road and how important it was to accelerate the AWHC. These driverless cars are just a red-herring to continue doing nothing except build motorways, the National Party knows it, but manages to convince most people that they’ll ‘solve’ mobility issues, much like it’ll solve climate change.
I struggle with this idea:
‘Many motorists dream of the day they can sit back and relax while their car drives itself.’
While no petrol head myself, I have known many, including members of my own family, and I can assure you that in as much as people ‘love their cars’ this is not the experience they are longing for.
If driveless road vehicles [of course driverless trains have long existed] are to have as big an impact as their boosters claim, then the first thing they are likely to disrupt is the Taxi industry, probably be an opportunity for rental players, but really what it completely spells the end of is massive road building as championed by the current Minister.
The claim is that the vehicle fleet will drop by 80%, if even half of that were to happen we will never need to build a road again, and will struggle to maintain those we currently have. It will also mean that we will only ever be a passenger, so the issue of whether we move in a single vehicle at higher cost or in bigger road vehicles like buses, or smoothly together in trains won’t make much difference. I do not see any threat from driveless cars to the critical space efficiency offered by urban Transit systems, in fact what a great way to expand the reach of stations without wasting land for carparking. Don’t Park n Ride; get the robot car to drop you off.
And the only way to retain your Freedom on the roads will be on a bicycle! Every m’way will be able to spare a lane each way just for cyclists…. bring it on Brave New World.
I’m a self confessed car nut and I can not see myself buying a self driving car. The better cars get, the more I want to get an old MX5. I enjoy driving. Transportation is different. I’ll use the method that makes the most sense.
I struggle too, what motorists really want is to drive with no traffic and that ain’t going to happen.
Also traffic will be worse with driverless vehicles, because computers can’t/don’t take risks, hence cars will go slower, keep more distance ect. Human beings have a very strong optimism bias while driving, that make them take risks they don’t even realize.
Yes it will be very very easy for pedestrians to take back city streets if risk averse robots are controlling vehicles: All we’ll need to to is step out onto the carriageway whenever we feel like it, instant universal pedestrian crossings everywhere.
I’ll certainly trust a botcar to not run me over more than the average distracted or entitled driver…. will be a little frustrating for the occupant though.
Shared Spaces surely won’t work at all for the bots, excellent instant full pedestrianisation.
Cars that will stop as soon as I step out in front of them? Fantastic. Can’t wait.
The traffic jams will be ginormous.
Wait for the inevitable law change the same as what motordom did in the 20’s. (Jaywalking in the US)
I agree that the better attentiveness for peds & cyclists will be a possible big boon of any driverless car future (if and when it arrives).
I don’t think shared spaces would be particularly problematic for even a semi-mature driverless car system. You just drive slow. As you are supposed to.
What I think is more scary is that at some stage we might get just too expectant of good driving “skills” from the cars around us – and then get hit by a careless person still driving, rather than letting the computer handle it. Of course by that time we (might) be so used to be safe in traffic that rather than the “it happens, you killed somebody” shrug traffic deaths get these days, the bloke will be pilloried for how he could act so recklessly as to actually hit somebody…
Completely agree that fully functional driver-less cars will doom the taxi industry first. Then taxis will become so cheap and ubiquitous that owning ones own vehicle will no longer be necessary for many, as you’ll be able to call a self driving car that will arrive in minutes (seconds?) and the per trip cost will be considerably cheaper than the per trip total cost of owning ones own vehicle. Sadly for those who like driving, if this technology really does work and and is as successful as boosters say it will be, there will come a time where society will have a decision about whether ordinary citizens are any longer allowed to drive cars on public roads, since relative to the self-driving cars our own driving is so dangerous and unnecessary that that the risks will significantly outweigh the benefits. We’ll see how this all develops, but I agree it’s not displacing public transport within any realistic investment timeframe.
Looking forward to confusing the hell out of these things on my bicycle!
The prospect of driverless cars making far more efficient use of roadspace (and parking space) as a reason not to invest in public transit infrastructure seems to me to be equal in strength (or weakness) to the argument that it’s also a reason no to invest in road infrastructure.
It is a solution looking for a problem. Driving is the fun bit of travelling, it is riding as a passenger that sucks.
+1
mfwic, that’s my point above, for those who like driving, and just a glance at car advertising shows that this is clearly a big group, having a quivering risk averse bot in control is a horrible idea.
I’d go further and say that it is the end of driving outside of motorsport arena; once this tech takes hold it’s going to be very hard to accept the carnage that human control routinely produces.
I find it funny that Moar Roadz advocates like the current minister seems to think this tech is any kind of pro-driving innovation. It’s the death of driving.
The only place we’ll be freely in charge of a vehicle on a public road in this future will be on a bike. Get peddling kids.
Driving *can* be the fun bit of travelling by car, but on a regular commute in traffic it’s dead boring.
Given the choice, I’ll drive for trips on holiday, out in the country or on sunny afternoons. For crawling in traffic I’d rather be a passenger so I can play with my phone or try and do something else.
It would be interesting to see how the technology accommodates wildlife. As a driver in Australia I need to be prepared for kangaroos along certain roads. Roos usually remain motionless beside the road, blinded by the headlights. Then they will either hop off into the bush, or onto the road. This is not just a problem for remote rural roads, because I have seen dead roos beside the main freeway into Adelaide, and my brother has crashed into a roo in urban Canberra. An algorithm would need to recognize the silhouette of a stationary roo, slow down, and be prepared to stop if it decides to go onto the road. Other algorithms would need to recognise moving animals and birds that have not yet become aware of the vehicle. Some questions
1. Who pays for damage to the car if it collides with the roo ?
2. How would the designers prove that their system is better than a manual driver ? This would involve animal ethics applications for each software design, each type of animal, and under a wide variety of weather conditions including heavy rain. Will it be worthwhile paying for all these development costs ?
3. Once the general public know about how such a system is triggered, all that a pedestrian would need to do is to hold the silhouette of a local animal beside the road, and all the driverless cars would come to a halt, allowing the pedestrian to cross. Bag manufacturers would emblazen such silhouettes on their products, so a user only needs to orient the silhouette to face the traffic and could then easily cross the busiest road, even when there is a red pedestrian light.
4. If the automated system slows the vehicle too much, there would be a temptation for the manual driver to take over.
5. If there are situations where the algorithms are not working reliably, there would always need to be a licensed driver in the vehicle who takes responsibility under difficult conditions. The system would then be little more than a next generation cruise control, which can guide the vehicle along busy motorways and main roads.
I’m sure better sensors and video processing will solve many of the issues eventually. 5 Years ago the cars were stumbling around in the Nevada desert, now they are driving with other people. Who knows where another 5 years will take us.
Who pays for the damage if you drive into a kangaroo now? If you drive in the outback and you warranty says you can’t, then it’s your problem.
Basically the car will avoid colliding with anything and can react much faster than a human driver could. The problem is that the car can’t understand that it can drive over a balloon, but not over a boulder. It can’t tell the difference and I’m not sure how they will ever get around it. You will have to have a driver ready to take control if the computer gets confused.
Not at all convinced by driverless technology. How exactly can we be sure that the onboard technology can keep track of and avoid ALL the other vehicles approaching busy intersections at all sorts of angles and speeds. Given that different manufacturers will use different technologies and protocols – are we prepared to stake our lives on them ALL being “state of the art”.
Transitioning to the promised utopian future is also problematic. Can ordinary and driverless cars safely share the same road space? I suspect not so do we force everyone to write off their existing car and buy new? Not feasible so we would have to limit driverless cars to a limited number of special purpose routes reserved just for their use – but then how would you get from home or work to one of these reserved roads? Perhaps the driverless function could be “part-time”, allowing the vehicle to be manually controlled on ordinary roads.
As I said years ago, driverless cars will never, ever be possible; the pattern-recognition problem is basically unsolvable. You want driverless cars? Put them on tracks, with switches, and a central dispatcher. We already have driverless trains (such as Vancouver SkyTrain or Docklands Light Rail).
Going through manicured, super-expensive roads in Mountain View, CA is nothing like driving on real roads in the countryside in bad weather — which is what cars are needed for. We don’t, frankly, need cars at all in the cities where Google has made the technology sort-of-work.