cycling and danger - actual and//or perceived?

Tony EH

Wishbone Ash wrote: »

Indeed, but a computer doesn't get tired, get complacent, take chances, run late, use a smart phone, have a hangover, have a bad day at work, have marital problems, have financial troubles, have ill health or any range of other things that preoccupy the minds of many motorists when driving.

All true. But a human being still retains the ability to think outside the box as it were and is not simply limited to the programming that was put in - by a human being.

veganrun

I just started back cycling earlier this year after a 20 year break or so here in the north. I was very nervous starting back so I stuck to off road cycle lanes as much as I could. My commute is only about 15 minutes each way but the last bit is through some city streets with no cycle lanes. It hasn’t been too bad and I felt at times the biggest risk were pedestrians glued to their phones or not paying attention. I’ve had to slam the brakes on several times recently as they have walked into a cycle lane with no warning and not looking around.

I know one part of my journey some p**** in a lorry beeped at me as he was trying to turn left when there was a no left turn sign and I had waited until the lights were green before going. Another time I saw some driver in a car do a similar thing at the same junction and he seemed to half clip another guy on a bike. The guy didn’t come off the bike but rightfully seemed annoyed.

Something I realised when I lived in Dublin and drove was that I simply had to assume all other drivers had no brains at all. I knew if I expected a driver to act sensibly that usually they wouldn’t, so I just assumed they were all idiots and tried to make sure I allowed enough room for whatever c*** they did. I guess the same is true now I cycle.

Now I’m a bit more confident on the bike but still stick to cycle lanes as much as I can. I’ve been a bit scared going to roundabouts as I don’t like them (even in a car) so on the bike I feel very exposed.

I’ve found a decent 8 mile loop near where I live that’s on a fairly wide road with two lanes in places and a 30mph speed limit but a few times recently when I hear traffic coming up behind me I keep hoping they can see me (I wear a hi-viz jacket or vest as well as front and rear flashing lights).

Agreed that taxi drivers can be bad. I’ve had a few squeeze me when other drivers let me through a tight street.

Tony EH

veganrun wrote: »

I just started back cycling earlier this year after a 20 year break or so here in the north.

As a matter of interest, what are the cycle lanes like up there?

smacl

Tony EH wrote: »

All true. But a human being still retains the ability to think outside the box as it were and is not simply limited to the programming that was put in - by a human being.

Possibly so, but an autonomous vehicle will drive in a predictable and conservative manner so as not to cause an accident in the first place. In terms of cyclist and pedestrian safety this includes no blind spots, continuous 360 degree awareness, correct signalling, no risk taking, etc... Basically, you're removing nearly all accidents that relate to driver judgement error on the part of the vehicle causing the accident.

I think what most drivers object to is the loss of freedom, where driving a car has become much more than just getting from A to B quickly, comfortably and safely for many people. Unfortunately this includes the freedom to drive dangerously, and as safer options become available this is likely to become unacceptable on most public roads.

GreeBo

07Lapierre wrote: »

Maybe, but IMO driverless cars will always need Human input of some kind. there will always have to be an override facility, where a human can take control of the car. A person will always have to be responsible for the car's actions. Computers /technology cannot make moral decisions. At least I hope that the way it goes!

This is already not the case.
The driverless bus from the Las Vegas incident has no controls accessible to humans, just a display.

Tony EH wrote: »

It doesn't matter who was at fault. The computer was unable to deal with the situation, because computers can't think. It was a complete embarrassment.

Also, I could throw your words back at you and say that you cannot say that driverless cars will be flawless in 2025.

The essential point still stands however. A computer doesn't learn. It cannot make decisions outside of its programming or what a human has told it to do.

That, right there, is an extremely serious thing.

I think you are being a bit disingenuous tbh.

During the incident in downtown Las Vegas, the bus automatically stopped to try avoid an accident after its sensors detected the truck, the city said.
"Unfortunately, the delivery truck did not stop and grazed the front fender of the shuttle," it said. "Had the truck had the same sensing equipment that the shuttle has, the accident would have been avoided."

The problem in this incident was a human, not a "robot".

Computers do learn, just look at the Google cars in California, just look at any Machine Learning as already mentioned. Computers can actually learn far faster than humans as they can gather data from millions of scenarios and can share that with others in nanoseconds.

Tony EH

smacl wrote: »

Possibly so, but an autonomous vehicle will drive in a predictable and conservative manner so as not to cause an accident in the first place.

Says who?

In the Las Vegas example the computer on board the minibus couldn't even think of reversing to avoid a collision - albeit a minor one. It simply just froze. A human driver would have put the vehicle in reverse, instead of just coming to a halt, because it's algorithms didn't extend to that particular scenario and it just didn't know what to do.

One shudders to think what would have happened if an automated vehicle decided to simple stop on a busy highway, because it's programming didn't allow for a random situation.

The problem people keep glossing over is the limitations of a program that goes into a computer and I can foresee a lot of issues there, especially when not all vehicles will be automated.

Tony EH

GreeBo wrote: »

I think you are being a bit disingenuous tbh.

I'm being perfectly genuine.

GreeBo wrote: »

The problem in this incident was a human, not a "robot".

The salient point here is that the "robot" froze, because it couldn't think of reversing.

Am I really the only one seeing this point?

smacl

Tony EH wrote: »

The salient point here is that the "robot" froze, because it couldn't think of reversing

By that logic, if the truck had reversed into a wall, you'd blame the wall.

Tony EH

That analogy doesn't make the slightest bit of sense.

A wall is a stationary object. It can't move.

A functioning vehicle should be able to move, if the driver is able to think where to move to.

Lumen

Tony EH wrote: »

The problem people keep glossing over is the limitations of a program that goes into a computer and I can foresee a lot of issues there, especially when not all vehicles will be automated.

In principle robot cars don't need to be perfect, they just need to be better than humans.

The difference is that in our current effective road laws, humans drivers are allowed to perform incredibly poorly, far below known best practice*, without significant penalty, even when that poor performance kills people. The bar for death by dangerous driving is ridiculously high.

Whereas a robot cannot be inattentive so any fault is down to its programming and therefore the creators are liable.

The consequence of this is that robot cars are likely to be much safer and more cautious than human drivers once they hit our roads in numbers.

But it also means they will be bullied mercilessly by the remaining humans, including pedestrians and cyclists.

(* "drive at a speed such that you can stop in the space that you can see to be clear and is likely to remain so" is routinely ignored )

magicbastarder

you do realise that this was a *trial*? a trial by its very nature is there to identify problems so they can fix them in the next iteration. the position of 'well, here's a situation where the computer was not programmed to a respond' is an ever decreasing platform to stand on.
it's like criticising a child for not acting like an adult. the technology is not yet fully mature, and no-one is claiming it is.

hesker

Tony EH wrote: »

I'm being perfectly genuine.



The salient point here is that the "robot" froze, because it couldn't think of reversing.

Am I really the only one seeing this point?

You seem to be fixated on the fact that the computer is not perfect.

If you accept the fact that neither computer or human is perfect then it comes down to risk and probability.

In the long term I’d put my faith in the computer.

Also we humans get to reprogram them iteratively so problems will be reduced over time even if never ultimately completely eliminated. By then there will be no argument.

Insurance premiums will dictate the direction probably.

Ronaldinho

There's an article in the Times today by Sorcha Pollak decrying the Luas tracks as 'unavoidable'.

It doesn't affect me on my commute so would be interested to what people here think?

I'd have thought it was OK so long as you're paying attention?

GreeBo

Tony EH wrote: »

That analogy doesn't make the slightest bit of sense.

A wall is a stationary object. It can't move.

A functioning vehicle should be able to move, if the driver is able to think where to move to.

Oh you mean like the human truck driver who drove into the stationary bus?

magicbastarder

Lumen wrote: »

But it also means they will be bullied mercilessly by the remaining humans, including pedestrians and cyclists.

yeah, this is one of the more interesting quandaries - the notion of pedestrians walking out in front of cars because they know the robot car will stop where a human driver might not. the grauniad had an article about this, where someone posited 'all bets are off' roads where cars would be programmed not to stop.

GreeBo

Tony EH wrote: »

I'm being perfectly genuine.



The salient point here is that the "robot" froze, because it couldn't think of reversing.

Am I really the only one seeing this point?

It didnt "freeze", it detected an obstacle and stopped, like most people should do.
Swerving or taking some other unexpected response will likely lead to causing another incident.

Tony EH

GreeBo wrote: »

It didnt "freeze", it detected an obstacle and stopped, like most people should do.
Swerving or taking some other unexpected response will likely lead to causing another incident.

But you don't remained stopped if something something coming towards you.

Alanstrainor

Tony EH wrote: »

But you don't remained stopped if something something coming towards you.

1. It was a trial run of a slow moving (15mph) autonomous bus around Las Vegas.
2. It responded to the obstacle by stopping.
3. Despite stopping the infallible truck driver somehow managed to still hit it.

So we have a trial run of new system, which since they were trialling it would be safe to assume that mistakes were to be expected. But no mistakes were made, the system stopped in response to a obstacle. That human truck driver made an error and kept reversing, despite being able to "think", and hit into a stopped vehicle on the road.

So can we please move on from this. We're going around in circles here. Googles autonomous cars have been driving around San Fran for the last 8 or 9 years with 14 accidents over 2 million miles. 13 of which were not due to errors on the part of the autonomous system. Their system has no doubt changed and improved hugely as hardware and software improves. Still not perfect, but it's getting to the point where it'll better than most drivers.

Tony EH

Alanstrainor wrote: »

1. It was a trial run of a slow moving (15mph) autonomous bus around Las Vegas.
2. It responded to the obstacle by stopping.
3. Despite stopping the infallible truck driver somehow managed to still hit it.

I know all of this and is nothing with my point.

Alanstrainor wrote: »

So can we please move on from this.

Please.

smacl

magicbastarder wrote: »

yeah, this is one of the more interesting quandaries - the notion of pedestrians walking out in front of cars because they know the robot car will stop where a human driver might not. the grauniad had an article about this, where someone posited 'all bets are off' roads where cars would be programmed not to stop.

Fines for jaywalking would certainly become a bit stiffer. On the plus side, traffic lights start to become obsolete;

https://www.youtube.com/watch?v=kh7X-UKm9kw

Not sure where the cyclist fits in with the plan above. Seems like we're heading full circle to segregated infrastructure.

Lumen

smacl wrote: »

On the plus side, traffic lights start to become obsolete

Yeah, if there was only some way of building a sort of junction which people could drive through not directly but in a round about sort of a way without having to stop at a traffic light.

smacl

Lumen wrote: »

Yeah, if there was only some way of building a sort of junction which people could drive through not directly but in a round about sort of a way without having to stop at a traffic light.

If only. Means we could get rid of all those traffic lights at major roundabouts

Chips Lovell

Ireland, where our enthusiasm for building roundabouts is only matched by our inability to use them correctly.

GreeBo

smacl wrote: »

Fines for jaywalking would certainly become a bit stiffer. On the plus side, traffic lights start to become obsolete;



Not sure where the cyclist fits in with the plan above. Seems like we're heading full circle to segregated infrastructure.

These junctions have been in place for years in some countries...

https://www.youtube.com/watch?v=izEH_jUIfo4

McTigs

07Lapierre wrote: »

I also think there's a whole generation out there who have never cycled, who were driven to the creche, to school and probably drove their first car to college! As far as they're concerned, the roads are for motorised vehicle only and drive accordingly. they dont understand what its like to to have a car whizz past you at 100kph with only mm's to spare.

I cycled a ten mile round trip to secondary school from when I was 12 and have been cycle commuting for the 30+years since. In that time I've only had two accidents and one was entirely my own fault so while I don't feel unsafe that could be down to cycling competence acquired over many years.... I don't think there is any substitute for getting kids cycling as early as possible ( each of mine had the stablisers off before their 4th birthdays) so they will feel confident to cycle on the road and develop the 6th sense awareness that comes with experience

I recently passed my driving test and I can confirm that not 5 minutes of the 12 hour RSA approved compulsory driving lessons are given to cyclist awareness.... the only time it was mentioned was the instructor reminding me to check my blind spot on the left because "you wouldn't know when one of these cyclists would arrive up beside ya"

veganrun

Tony EH wrote: »

veganrun wrote: »

I just started back cycling earlier this year after a 20 year break or so here in the north.

As a matter of interest, what are the cycle lanes like up there?

I’ve only been cycling in Belfast so I can’t speak for other places but they seem ok. One of the streets has plastic bollards which stops cars driving into it unfortunately that’s only on one street that I’ve seen so far. People still park in other cycle lanes though, including outside a police station?! Although there’s usually someone in the car so that makes it ok right?

They have a few greenways here that are fully off road or the majority off road on tarmaced paths which is nice.

uck51js9zml2yt

I was driving down the north circular this evening with a squad car in front of me.
We passed 3 cyclists with no lights and AGS just kept going.....so much for enforcement.

hesker

Reina Thousands Moisture wrote: »

I was driving down the north circular this evening with a squad car in front of me.
We passed 3 cyclists with no lights and AGS just kept going.....so much for enforcement.

I see them all the time too when I'm out cycling. I just don't get it especially when you can pick up decent flashing lights in lidl for half nothing.

They don't frighten me though or threaten my safety through their behaviour, unlike the motorists I encounter daily.

Kaisr Sose

Reina Thousands Moisture wrote: »

I was driving down the north circular this evening with a squad car in front of me.
We passed 3 cyclists with no lights and AGS just kept going.....so much for enforcement.

Slightly off topic but on topic for danger/risk to all road users:

Tuesday/Wednesday morning, I seen a Garda motorcyclist taking details of a cyclist outbound on College Green/Dame Street (most likely for running a red). Good job, nice to see.
Meanwhile at George’s St and down to Trinity, vehicles (and cyclists) were going through red, with little risk of getting caught. When will AGS/Govt realise that the only way to deter red light runners is with technology? Until then, it’s going to be normal for many to do so.

ED E

Kaisr Sose wrote: »

Slightly off topic but on topic for danger/risk to all road users:

Tuesday/Wednesday morning, I seen a Garda motorcyclist taking details of a cyclist outbound on College Green/Dame Street (most likely for running a red). Good job, nice to see.
Meanwhile at George’s St and down to Trinity, vehicles (and cyclists) were going through red, with little risk of getting caught. When will AGS/Govt realise that the only way to deter red light runners is with technology? Until then, it’s going to be normal for many to do so.

Red light cameras = less jobs for the boys