Self-driving car kills pedestrian

I don’t see any way they can take over 100%.
What about emergency services or even using them for emergency services ?
The possible nightmare scenarios don’t even bare thinking about.
I’d much prefer a real genuine amblance wizzing thru the traffic to whisk me off to Hospital should the need arrive.

If it gets to the stage where even fifty percent are driverless, I predict total chaos, as many of the other half will no doubt delight in taking the piss by carving them up with little worry about come back or getting their heeds bashed in.
There's stuff that needs addressing, but I reckon they can make the cars look like any other cars and the biggest use will be on motorways where you can have a sleep. (really looking forward to that)
 


Unlikely that they will be stuck. That problem has been known in computing for a long time:
Dining philosophers problem - Wikipedia

If all the cars are independent of each other, then the way to program this is for each car to back off and wait to see what the other cars do. If they don't move then try to move forward, if they appear to move at the same time then break. If every car backs off for 2 seconds then they will remain stuck like that for ever. The most likely scenario is that they will back off for say 2 seconds plus a random amount of time based on a mathematical function of something unique, say the car registration or internal serial number (though BMW drivers will want that to be the value of the car). You then get cars backing for say 1 (the hash may come up with a negative number), 2, 4 and 5 seconds. Once that first car gets out, then the problem goes away.

To me, the real power of self drive cars is not to be independent and have some degree of communication/negotiation between them. One example is pulling out at a T junction, taking a right when you are on the minor road and there is high volumes of traffic in both directions on the major road. As a human we can often sit and wait for ages for a safe gap. A joined up system could work out that if two cars, one to the left and one to the right, slow down from 1000 yards away by just 1 mph, then that creates a gap at just the right time to get one or two cars out. The passengers on the minor road do not even notice the difference it is so subtle.

Another example is on a motorway. A car breaks down in the fast lane. In the current human system, you only often have a view of 5 cars ahead on a busy motorway. You sit in a jam for miles wondering why and only when you get close do you see the problem and have to slowly take action. In a joined up AI system, a single car could report as much as 15 miles in advance of a hazard to be avoided. Cars gradually merge in a safe gentle way until the hazard has passed. Cars in the middle lane allow those in the outer lane with only a slight drop in speed as there are no egos saying "You can't be in front of me" etc. Passengers are reading books or watching films and don't even notice the problem.

In your roundabout scenario, the 4 cars could negotiate directly between them and just decide which one goes first. Again this could be a mathematical hash of the serial number to denote the priority. That might mean certain serial numbers come in high demand! Another way is to give each car a priority counter, 1 to 10. In a congested situation, a 10 will get priority but then it's number reduces to 9. Anyone who is 5 or less and has lost 5 competitions goes back up to a 6. Too much detail there, but these things are workable and have the potential to be more efficient and safer than humans.

I agree that the way in which they could work together has huge, huge potential

One too many gone now ! RIP that person.

Human driven cars kill over a hundred in this country every month.

Deaths and serious injury are the cost of mass transportation. People lose their shit though because a human error killing someone is an accident but a machine error is justification for shutting the programme down? Doesn’t make sense when you think about it
 
Last edited:
There's stuff that needs addressing, but I reckon they can make the cars look like any other cars and the biggest use will be on motorways where you can have a sleep. (really looking forward to that)

I think thats where whey will come intot heir own initially. it may be motorways have points when you get on a junction that allows you to put it into autodrive.

My example earlier for instance when I'm going to france skiing. would mean I could get 4 hours kip through the nigth while it goes down the toll roads.
 
I think thats where whey will come intot heir own initially. it may be motorways have points when you get on a junction that allows you to put it into autodrive.

My example earlier for instance when I'm going to france skiing. would mean I could get 4 hours kip through the nigth while it goes down the toll roads.

Think that might be its limit for the next few decades at least. They haven't got a clue how to get them to work in dense urban environments, too many things going on, can't deal with people crossing roads, cyclists etc.

They're not going to solve issues of congestion or pollution either.

We're decades away from full autonomy, that's if we ever get there at all.
 
On the flip side these cars will overall be safer and kill far fewer people. So jobs lost but many lives saved.
The problem is that we accept humans are fallible, make mistakes and have accidents. When robots or in this case these cars make a mistake and people die I don't think we'll accept it quite so readily. Even though less people are dying it will be seen as unacceptable by the public.
I wonder if you'll lose your no claims bonus if your car goes a bit radgie and causes an accident?
 
There's stuff that needs addressing, but I reckon they can make the cars look like any other cars and the biggest use will be on motorways where you can have a sleep. (really looking forward to that)

There is also the potential for a law change or closer scrutiny of existing laws. For example if you are waiting at a junction and pull out at the last minute, at the moment you don't know if the human driving the car will stop, can stop or will chase you down and give you a kicking, but still some people do it. If the police don't see you, you get away with it. The fear is that if you see a self drive car coming, people can pull out with the safe knowledge that the oncoming car will stop and they will get away with it, even if this jolts all the passengers as the self-drive does an emergency stop.

However the self drive car will be equipped with a dozen different sensors and cameras and recording. You do something dangerous to make a game of getting one up on a self-drive and before you have even stopped laughing a report of the whole incident has been uploaded to the police. Someone then reviews the footage, presses the 'charge them' button and there is a dangerous driving charge on the way before you have even made it home.

It is one of the biggest things that I have not seen really talked about yet, surveillance. We already have a lot of cameras on roads, causing ANPR hits and the like, but self-drive cars by their very nature have the ability to assess and record the movement of other cars. You are doing a steady 70 on the motorway, the car knows the person overtaking is screaming past at 110. If this information can be accessed by our law enforcement, then the amount of electronic surveillance on the roads increases by a factor of thousands - and it is already pretty high. I really like the idea of self-drive cars but this is one area that does worry me, how could all this data being gathered be used?
 
There is also the potential for a law change or closer scrutiny of existing laws. For example if you are waiting at a junction and pull out at the last minute, at the moment you don't know if the human driving the car will stop, can stop or will chase you down and give you a kicking, but still some people do it. If the police don't see you, you get away with it. The fear is that if you see a self drive car coming, people can pull out with the safe knowledge that the oncoming car will stop and they will get away with it, even if this jolts all the passengers as the self-drive does an emergency stop.

However the self drive car will be equipped with a dozen different sensors and cameras and recording. You do something dangerous to make a game of getting one up on a self-drive and before you have even stopped laughing a report of the whole incident has been uploaded to the police. Someone then reviews the footage, presses the 'charge them' button and there is a dangerous driving charge on the way before you have even made it home.

It is one of the biggest things that I have not seen really talked about yet, surveillance. We already have a lot of cameras on roads, causing ANPR hits and the like, but self-drive cars by their very nature have the ability to assess and record the movement of other cars. You are doing a steady 70 on the motorway, the car knows the person overtaking is screaming past at 110. If this information can be accessed by our law enforcement, then the amount of electronic surveillance on the roads increases by a factor of thousands - and it is already pretty high. I really like the idea of self-drive cars but this is one area that does worry me, how could all this data being gathered be used?
Excellent question . I'm afraid I know what the answer in this country is. Hopefully companies will realise this and give us some control of the data. You could make an for having the last ten seconds before a crash that deploys air bags be saved to the cloud.
 
Will wealth be shared or will it be monopolised by those who own the technology?

The latter obviously but you know...

Automation will rob opportunities from people IMO

Who knows how it'll pan out but the very concept of wealth and work we hold now will be turned on its head. We'll have a shit ton of leisure time and access to a shit ton of low cost automated services.
 
Can a robot determin what a pedestrian might do on a pavement ? does it slow down when a mother of a little kid leaves go of their mothers hand a couple of seconds before the car passes ? safe driving takes years of experience, I'm not sure a robot can have that.
This 100%.
Did you see the Guy Martin TV show where they were testing a driverless racing car and it crashed at the first corner because it could't cope with cold tyres? How would they cope with the weather we had in the last month?
Driverless cars will work on closed roads where there they don't have to cope with humans in their environment, like pedestrians and other drivers; horses and other animals, or rapidly changing conditions not recognised by their programming. So not in the real world, then.
 
This 100%.
Did you see the Guy Martin TV show where they were testing a driverless racing car and it crashed at the first corner because it could't cope with cold tyres? How would they cope with the weather we had in the last month?
Driverless cars will work on closed roads where there they don't have to cope with humans in their environment, like pedestrians and other drivers; horses and other animals, or rapidly changing conditions not recognised by their programming. So not in the real world, then.

its still early days but you only have to see how far they have come in 5 years to guess where will be in 20 ...

its not an overnigth solution and also th reason I think Uber will be on a precipice at some point as driverless cars is their only way of making money at this moment in time. without reducing overheads they wither put prices up or continuee to lose 2 or 3 billion dollars a year
 
This 100%.
Did you see the Guy Martin TV show where they were testing a driverless racing car and it crashed at the first corner because it could't cope with cold tyres? How would they cope with the weather we had in the last month?
Driverless cars will work on closed roads where there they don't have to cope with humans in their environment, like pedestrians and other drivers; horses and other animals, or rapidly changing conditions not recognised by their programming. So not in the real world, then.

Did you see the early attempts at flight, where people had flappy wings attached to bicycles? How could we have ever thought about flying to the moon 60 years later? Thankfully people saw them as problems to be overcome rather than stopping at the first hurdle and quitting then saying it could never be done in the real world.

Already there have been a lot of real world trials of driverless cars in the US, which has mostly been quite successful. There is a lot to learn, but the rate of progress and the number of companies looking at getting in on this new era is impressive. You have a lot of the major car manufacturers but also newcomers to the car market like Google. I can't find the link now but one car interior was very high tech and full of screens. I'm sure it was Panasonic or Philips looking at partnering up. There is huge investment into the technology, which there would not be if it was a crazy pipe dream doomed to failure.
 
how will they grind to a halt ? and no need to speed...

Depends how carefully they programme them. If it’s about realism, will all lorries drive on the motorway at 50 miles an hour in the slow lane apart from the ones overtaking in the next lane at 50.1 miles an hour? Will 40% of all cars hog the middle lane at 58 miles an hour causing tailbacks, and will anyone attempting to get into the fast lane at a standard 70mph get smashed into by an ultra-realistic driverless Audi which never signals and does 110 up everyone’s arse?

And when they get onto the North Circular will the car AI automatically switch to Homicidal Fuckwit Mode?
 
Last edited:
I'm assuming the cars won't speed, so not only will there be a massive loss of revenue the countries roads will grind to a halt.

On the other hand, a 100% self drive motorway has to potential to go faster. If you calculate your average speed on our roads over a long journey, then it is pretty difficult to average over 70. Delays and problems are caused by a lot of human issues, like middle lane hoggers, people not letting you pull out or pull in, or my main annoyance - a massive traffic jam because people have slowed to look at a problem on the other side of the road.

There has been talk of 'road trains'. If all the cars can react quicker than a human or get the message from much further up the road about a problem, then cars can drive very close together in each others slip stream at much higher speeds while still being safe. If one car has to break then all break. I worked on some robot control projects in the late 90s when we had a team of small wheeled robots which were autonomous but could take collaborative actions like this. Technology has moved a lot since then.

My drive up to my parents is 330 miles, but typically takes 5.5 hours of driving - which is a average of 60 mph. When the road is clear I do get up to about 90. If you could cut out a lot of the needless human caused delays and get that up to say 75, you could shave that time to 4 hours 25. That is much further in the future as it will take from launch 5-10 years for the majority of new cars to become self drive, then usually it takes 10-14 years for cars to flush through the system to the scrap heap.

but they wont. its people accelerating and braking and being erratic that causes the stop start issues we experience on the roads.

traffic flow should be smoother as you havent got dicks with egos

You said it better in one line than my 3 paragraphs!
 
but they wont. its people accelerating and braking and being erratic that causes the stop start issues we experience on the roads.

traffic flow should be smoother as you havent got dicks with egos

This combined with higher levels of car sharing (initially anyway) would improve the efficiency of the road network. Still not the most efficient way to move people around dense urban areas mind, still polluting, take up tons of space etc.
 

Back
Top