GTPNewsWire
Contributing Writer
- 21,489
- GTPHQ
This is the discussion thread for a recent post on GTPlanet:
This article was published by Joe Donaldson (@Joey D) on December 29th, 2017 in the Car Culture category.
Sounds like Uber, or Lyft.I think this is a failure of imagination. Fully autonomous cars will enable things that people aren't completely appreciating.
Imagine a large van pulling up to your house at the time that you've scheduled as being ready to leave for work. You get on, and there are 3 other people that work near you each in their own little booth on their laptops, already busy working on... whatever, personal affairs, business, doesn't matter. You get on, take your spot in your little booth, sip your coffee and check your email while the van drives you each to work. It takes an hour to get there, but you were getting things done the entire time. Maybe you even billed your employer for the time. Same thing happens on the way home. This isn't walking to the train station, it's not parking at the bus stop, it's ultra-convenient from farther-out cheaper housing.
You don't have a commuter car, because you don't need one. Your rented self-driving car takes you to and from work every day barely interrupting your routine. You don't park, you don't fill up the tank, and you share the expense with others. It's never late, it's never on vacation, it's regularly serviced, it's a perfect driver, it doesn't pick the music. And the car is designed from the ground-up for this, meaning no space is dedicated for a driver.
The expense of this car is low too because it keeps doing this all day long. After dropping you off, it picks up others. It ferries people around all day, and maybe a different one picks you up on your way home. Need to reschedule? Not a problem, a different one will be along shortly.
What does that leave you in your garage? A porsche? Which you can afford because it doesn't get worn down as a daily driver. When you want to drive for fun, it's there.
I'm still blown away by the option of renting a self-driving van that shows up with a bed in the back. You hope in, tuck in for the night, and go to sleep while it drives you 10 hours to another city. When you wake up in the morning you get changed, maybe have a bite to eat, maybe use the built-in bathroom, and hop out ready to go when you arrive. It could kill short-distance air travel.
Sounds like Uber, or Lyft.
You know that the number of deaths among all road users, and the KSI (killed or seriously injured) figures, are already the lowest they have ever been since recording started - despite the total number of road miles driven being the highest they have ever been - though, right?But in the near future it will become a must. There are 30.000 fatalities in Europe only.. every year. The drivers are getting more and more distracted during their trips and I'm sure that at some point it will be regulated by the law.
You know that the number of deaths among all road users, and the KSI (killed or seriously injured) figures, are already the lowest they have ever been since recording started - despite the total number of road miles driven being the highest they have ever been - though, right?
What concerns me about this technology is the fact that machines can make mistakes. Maybe they may not make idiotic human mistakes, but I'm concerned about the "buggy, rushed, faulty software" mistakes that a self-driving car could possibly make. Machines make mistakes too. Could you imagine the cluster that would befall a company that released a self-driving car that does something stupid like turn into oncoming traffic because of buggy software? I'm not sold on this. I think for me to be comfortable with this would, several manufacturers will have to produce a perfectly working example to prove that this isn't a technology that is getting rushed for the sake of "look what we did!"
The future is fully electric and fully autonomous. People may want to drive but because we are so poor at it in comparison to the soon to arrive autonomous cars govts will legislate against human driven cars. Perhaps there will be a car in the garage to take to the track and enjoy though.
I know. But I'm sure you would agree with me that the number is still high.You know that the number of deaths among all road users, and the KSI (killed or seriously injured) figures, are already the lowest they have ever been since recording started - despite the total number of road miles driven being the highest they have ever been - though, right?
Maybe so but road tolls are still staggeringly high, and if you factor in the non fatal injuries too then it's a wonder we are allowed to drive at all. One day we won't be allowed to.You know that the number of deaths among all road users, and the KSI (killed or seriously injured) figures, are already the lowest they have ever been since recording started - despite the total number of road miles driven being the highest they have ever been - though, right?
But not by much - same as everywhere else in Europe. It all seemed to bottom out in 2012-2015 and it's changing by literally thirty or so each year.Road kill numbers have been rising here since 2014.
Nope. 30,000 people annually across all of Europe is less than one in ten thousand. You're ten times more likely to die from seasonal flu.I know. But I'm sure you would agree with me that the number is still high.
Tolls? You mean... like... paying to use the roads? I don't see how that's really relevant. If you mean death tolls then see above. They're tiny.Maybe so but road tolls are still staggeringly high
That's what the KSI (killed and seriously injured) figures are for. In 2015, there were 23,869 road users (as above, all road users including pedestrians) killed or seriously injured in the UK. Again, that's still less than the extra deaths caused by a wonky flu vaccine in one winter, and represents one serious injury or death every 13.6 million miles. At 75 KSI per billion vehicle miles, that's less than half of just the deaths in 1949...and if you factor in the non fatal injuries too then it's a wonder we are allowed to drive at all.
This right here completely slams the notion that we need self-driving cars.But not by much - same as everywhere else in Europe. It all seemed to bottom out in 2012-2015 and it's changing by literally thirty or so each year.
Nope. 30,000 people annually across all of Europe is less than one in ten thousand. You're ten times more likely to die from seasonal flu.
But even one in ten thousand doesn't do just how rare road deaths actually are. Let's take the UK, because I have numbers for the UK to hand. In 2016 there were 1,792 deaths on the UK's roads - and that's among all road users, including pedestrians, cyclists, motorbikes, cars, buses, lorries... everything. In 1926 - the first year of recording - it was 4,886. The highest ever recorded was 9,169, in 1941, and just to put that into context, you were about 25% as likely to be killed on the roads during a WAR than you were by enemy bombs during THE BLITZ.
Anyway. In 2016 we had 65 million people, who drove 36 million road vehicles a cumulative 325 billion miles. This means that one death occurred for every... wait for it... 180 million miles. If you want to compare that to 1926, I can't as I don't have numbers, but there were 5,250 deaths from 32 billion miles, or one every 6 million miles. These figures are usually reported as "deaths per billion vehicle miles" and peaked at 165 in 1949, but hit 5.6 in 2013.
So not only have road deaths (and again, that's among all road users) fallen in raw number by about 80% from its all time high and 65% since records began, they've fallen in frequency by about 97%.
Road deaths in the UK are not a high number - and they're fairly representative of Europe.
Tolls? You mean... like... paying to use the roads? I don't see how that's really relevant. If you mean death tolls then see above. They're tiny.
I mean, if you want even more context, there's an expected base line of annual deaths in the UK that is roughly half a million people. That is to say that during any one calendar year, 500,000 people are expected to die. 1,792 road deaths represent 0.3% of that figure. By comparison, an ineffective flu vaccine in the winter of 2015 (wrong strain) was among the leading causes for an additional spike of 24,000 deaths that year...
That's what the KSI (killed and seriously injured) figures are for. In 2015, there were 23,869 road users (as above, all road users including pedestrians) killed or seriously injured in the UK. Again, that's still less than the extra deaths caused by a wonky flu vaccine in one winter, and represents one serious injury or death every 13.6 million miles. At 75 KSI per billion vehicle miles, that's less than half of just the deaths in 1949...
The number of all casualties, including all minor injuries ("slightly injured") like whiplash, were unsurprisingly higher, at 186,209. This is 40% lower than in 2000 - I mean... 2000! That's not even a blip in the 24 million visits people make to A&E each year
All of these numbers are at historic lows. Pedestrian fatalities/KSIs since 1950? Reduced by 80%. Cyclist fatalities/KSIs? Reduced by 90%. Motorbike fatalities/KSIs? Reduced by 75%. Children KSI? Reduced by 80%. Distance driven? Up 1000%...
The fact is that road deaths among all road users are now ridiculously unlikely. The greatest single reduction actually came from the three point safety belt, but other passive safety systems like airbags (officially known as a "supplementary restraint system" or "SRS") and straightforward crash protection structures, along with electronic aids like standard ABS and ESP have dragged deaths, injuries and boo-boos down to levels that mean they barely register any more.
I mean... as many children are injured by DOGS annually as by cars (around 1,700 in each case)...
Self driving cars don't have to be perfect at driving. They just have to make them better drivers than humans. Not perfect, just better.What concerns me about this technology is the fact that machines can make mistakes. Maybe they may not make idiotic human mistakes, but I'm concerned about the "buggy, rushed, faulty software" mistakes that a self-driving car could possibly make. Machines make mistakes too. Could you imagine the cluster that would befall a company that released a self-driving car that does something stupid like turn into oncoming traffic because of buggy software? I'm not sold on this. I think for me to be comfortable with this would, several manufacturers will have to produce a perfectly working example to prove that this isn't a technology that is getting rushed for the sake of "look what we did!"
What concerns me about this technology is the fact that machines can make mistakes. Maybe they may not make idiotic human mistakes, but I'm concerned about the "buggy, rushed, faulty software" mistakes that a self-driving car could possibly make.
There will be no autosport if self-driving car happens.
That's an interesting point. However, if machines demonstrably made fewer mistakes (and the early, limited data suggests that's the case) then isn't it better than having a larger number of human mistakes?
There's also an RAC-sponsored Ipsos Mori poll covering UK attitudes. The outcomes are broadly similar to the OP. Link.
Yes there will. There'll be self-drive and auto-drive. And both will be fantastic. Auto-drive cars as a TV sport will be astonishing, no pesky grandstands and huge mad tracks in the middle of nowhere.
That said...This right here completely slams the notion that we need self-driving cars.
The irony is that I grow up hoping to drive a car, only to find that they no longer need to be driven. Then again, motorway driving is a pain...That said...
The reason we're fairly static on road deaths is that we're pretty much at the limits of what we can do to prevent injury once a crash has happened. The biggest drop was with the three-point safety belt. Second was the supplementary restrain system (SRS... airbags to you and me). At around the same time we got US NCAP (via the NHTSA), Euro NCAP and such providing a cold, hard, scientific assessment of the effectiveness of crash safety structures. That's essentially driven us down to the point we're at today, along with standard safety aids like ABS and ESP - which are basically electronic driver assists that take a level of control away from our ham feet to prevent us from hurting ourselves. It's marginal gains at this point, with some manufacturers popping in whiplash protection, some looking to prevent rollovers, new crash tests that are harder for the cars to dissipate energy from.
But we can't really do much more during and after a crash any more. I spoke to Euro NCAP in March about this, and they're pretty much at that point themselves. The only way to drive the numbers down further now is preventing crashes from happening in the first place.
Step one is autonomous emergency braking (AEB). That's mandatory in Europe from next year and basically prevents you from driving into the back of the car in front (or, if you're going sodding fast, slows you down and reduces the impact), or pedestrians, cyclists and suchlike. That's likely to have a major effect on "slightly injured" stats. And if you think that sounds like something a self-driving car would do, you're right; it's a level 2 autonomy function. There are a number of other level 2 functions - adaptive cruise control, park assist, active lane keeping - that are getting pretty common but are arguably not quite as safety-oriented as AEB (they can prevent prangs, but their main goal is to carry out ordinary driving functions autonomously without increasing risk; AEB's main goal is to prevent prangs). Tesla Auto Pilot is basically a level 2 system.
If we want fewer road deaths, cars with some level of autonomy are now the only way to achieve it.
There are, of course, other benefits to (level 4/5) autonomous cars, as @Danoff and @Venari mention earlier. I used to do a ~5hr commute twice a week, from about 5.30pm-11pm one way and about 4pm-10pm the other. It was 320 miles of constant motorway (with only two places where I might have to slow for traffic lights) and 6 miles of marginal interest before and after. It took a giant chunk out of the productive time I had for the week - time I could have used for a bit of working, a bit of reading, or a bit of Gran Turismo, that I instead had to use for the totally thrilling task of concentrating enough to drive at 70mph in a straight line on a 3-4 lane motorway. An autonomous car would have been great.
And think about all the people who don't care about driving at all. They buy whatever generobox they want because they like the colour and they're all over the place because driving is the least important task they have (way behind texting, make up, eating a burger). They do 40mph in a 60 zone, a 40 zone, a 30 zone and a 20 zone - so you get stuck behind them on the quick road, then they nick off at warp six through town and you catch them back up on the quick road on the other side and get stuck behind them again. Get them into autonomous cars and they're happy... and you're happy.
Oh yea, that can be implemented in DAKAR RALLY.