The car **** up and loves to blow goats

Thank you. (happy smiley)

There’s nothing wrong with giving us all the safety enhancing aids that the research for these projects create though.

Absolutely.

There’s nothing wrong with giving us all the safety enhancing aids that the research for these projects create though.

yeah that’s true. The overprotective auto panic brake feature on my Golf isn’t AT ALL annoying when it gets spooked by stationary traffic in a different lane… 😂😡

Interesting but obvious insight to human attention span.. Personally I think the ‘safety drivers’ should be swapped out every hour or so to insure they are as fresh minded as possible and have a chance of intervening effectively before becoming bored and subconsciously switching off.

This is an experimental test on a live road after all.

Imagine going on a two hour taxi journey… Twice a day, 5 times a week, can’t really blame the human here.

The car in question clearly isn’t ready for such a test,  as demonstrated. It should have easily detected the pedestrian with the bike, but it didn’t.

Actually, I think the safety enhancing aids are going to make things less safe.  An automatic braking system means you can check that text message really quickly because the car will take care of it if anything goes wrong.

Total automation would make it more safe. The two hundred million Nigel mansells on the US roads kill 30000 a year-ish do they not? I reckon the machines would cut that dramatically despite this aberration.

One day, the stabiliser, not yet. It isn’t an aberration now because the cars’ systems aren’t good enough – yet. IMO they are on the roads prematurely. A bit like the code for this forum’s relaunch, it’s full of bugs but hopefully they’ll sort it out one day.

An automatic braking system means you can check that text message really quickly because the car will take care of it if anything goes wrong.

Sorry but that’s a guff statement. We already have that with ABS, but you don’t stomp on the brakes and let the car deal with it, to send a text, well you could but it would end in carnage.

yeez is it hard of thinking day in here or something:

1) I never said the car wasn’t at fault.

If you think i did, please POST THE QUOTE.  If you can’t post the quote where i said “the car wasn’t at fault” then shut up (btw the car was at fault. Quite clearly at fault)

2) I maintain that the AVERAGE human driver would have had a low statistically probable chance of avoiding this accident.

I’ve worked my whole life in the automotive arena, spent 6 years at Prodrive as a Senior Principal Engineer (not a “tech” btw!!) and have literally thousands of hours of seat time in a massive range of cars and with a massive range of drivers, from Damon Hill the F1 driver to Nancy Hill from 3 acacia avenue.  I’ve been involved with several objective studies on driver behavior, and i’ve spent a lot of time sat next to “average” drivers doing driver training, both on the road and under high dynamic conditions on a test facility. I’ve consulted to various advanced driver organisations and been published as an “expert” in books written by those organisations.

My experience, along with objective data i have personally gathered, suggests that in reality, with an average driver, paying the average amount of attention, with the average reaction time, and average attitude and skill set, they would not have prevented this accident.

This is backed up by the fact there WAS a human driver in the car at the time.  Who failed to even spot the victim (just like the car failed to spot that person). Despite, remember, the entire reason for them being in the car (and remember this was their job, they were paid to be there) to be a “failsafe” to prevent faults with the autonomy from resulting in accidents.

The big difference is what happens next.

Usually, in a typical ‘human only’ case, the driver would have said “i just didn’t see them” and that would have been that.  The driver may have been prosecuted if there was enough evidence, but in the vast majority of cases, often without even basic 3rd party witness statements, the death would be recorded as “accidental” the family would bury their relative with little closure or real explanation for what ocured and how.

This time though, we have hard, objective data and the ability to analyse, process and apply that data as a hard lesson.  That data can also be analysed in the cold light of day by external entities, be that the judge in any subsequent legal proceedings, or by Uber, or any other entity.

(i should also mention that “threshold braking” IS important, very, very important. Studies have shown that in many cases average human drivers simply fail to press the brakes hard enough (hence various automated assistance systems that have been developed).  Stop at 0.1g not 1g and you may as well not bother braking. And it matters most in the inital stages of the braking event, because your velocity (distance divided by time) is the highest. Ie you travel further per unit of time, and it’s the distance to the impact that matters and not the time to impact. (This is also why do just a few extra mph can mean a large difference in the speed you hit something)

The danger point of this is now onwards, until the day it is total autonomy…

A combination of fallible human drivers, less than perfected and proven software, new behaviours and reactions from humans and digital controllers.

The key thing is the ‘driver’ of the test vehicle wasn’t paying attention. That is both good and bad for the purposes of the test…

A) you wouldn’t expect a driver to be paying any attention in a fully automated vehicle.

B) the car failed, see point A.

What we do know now is uber, and possibly others are being negligent with thier testing on live roads in the race to get a product to market.

Negligence is for the courts to decide.

However, if they can demonstrate (to the courts satisfaction) that their system is, on average, at least as “safe” as the average for human drivers, i see no long term issue.  People will continue to be killed by cars for many years, and yet,t he sooner we start the sooner we can bring the numbers killed down.

Lets face it, we give driving licences to pretty much anyone. Never need a retest, never need to do more than pass some very basic theory questions and drive round the block without crashing.  Watch those “learner” or “OAP” driver progs on the telly and marvel at how many people are , frankly, completely unsuited to the act of driving, for example:

http://www.thisismoney.co.uk/money/cars/article-3623773/Liverpool-man-dubbed-worst-learner-getting-driving-licence-39th-attempt.html

39 times before he passed.  Do we think he’d nail a 0.5sec reaction time and then a perfect ABS stop, timed with an instant and yet mm correct swerve around an unexpected pedestrian in the road?  Fat Chance.

Every day i see poor driving, excessive speed, poor obeservation, anger, in-attention, lack of mechanical sympathy etc etc etc, not from one or two drivers but from LOTs of drivers. And that’s before we get to drink, drugs and other substance abuses.

If you are a road cyclist i’m pretty sure every time you go out on your bike you may well have some sort of close call, a pass that’s too close, the horrible sound of a scrunch of tyres into tarmac as someone suddenly spots you, too late, and realises they can’t actually fit past you and the car coming the other way, the SMIDSYs, and even the “you don’t pay any road tax” anger directed at cyclists by everyday drivers just for you being on a bike on the road.

Yes but there is something just a little bit.more organic about being mown down by a person in a car, the feel of metal on flesh just isn’t the same whe  a robot does it. Hence people would rather continue with bad drivers.

For all those talking about the driver of an autonomous vehicle ‘intervening’ – in my opinion that simply isn’t an option.

I don’t know how many of you have ever driven a semi autonomous vehicle but its a bit of an eye opener. I had a Tesla Model S on extended test drive (a week) as was thinking of getting one and it had AP2.(AA) This allows reasonably autonomous driving i.e adaptive cruise, lane control plus auto lane changing,traffic aware radar, 360 sensors etc etc.

1. Its scary as hell for the first day and you say to your wife – well we ain’t using this bag of tricks again.

2. Then curiosity gets the better of you and you start to use it

3. Within 3 days its like you have always had it.

4. in 5 days you have forgotten how to drive.

I have been driving for 30 years. I do a fair few miles and rarely do less than 20k PA. I have never crashed or even had a speeding ticket despite driving some reasonable cars and a few bikes. I am the sort of driver that concentrates hard – probably because I also ride a motorbike and commute by cycle where the stakes are higher if you mess up.

But the few days after I handed back the Tesla I nearly crashed up the back of other traffic (mainly entering roundabouts and in slow moving traffic) several times till I got used to a ‘dumb’ car again. I even checked my phone etc…..I NEVER do that when driving!

My point is it takes years to learn to drive but only days to hand that responsibility over to a computer and once you do you are no way going to intervene in a situation like this terrible accident. i don’t know what the uber driver/monitor brief is but I wouldn’t take that job as if uber expect the guy in the driving seat to be a ‘driver’ then that is an impossible task.

Personally, looking at that video I think most ‘good’ drivers (those that concentrate) would have been on the brakes at point of impact. However I suspect that many drivers would not have been concentrating any more than the rubbish uber software. But I certainly don’t blame the person in the driving seat.

What the video shows is that an imperfect driverless system is not made any safer by putting in a bored human monitor. You really might as well not have the human in there because there is a high likelihood that they will fail to monitor the car correctly and when an incident happens they won’t be able to react.  That is the lesson I think I would be learning form this incident.

 You really might as well not have the human in there because there is a high likelihood that they will fail to monitor the car correctly and when an incident happens they won’t be able to react.

Indeed. There’s plenty of data out there to show that there’s no way a human will be able to stay attentive while ‘monitoring’ this level of automation. Nothing to do = distracted car sitter.

Plenty of people saying the car failed to stop when it should have. Clearly the car should of detected the pedestrian before she steeped out on the road. Did it? We don’t know, all we know is that it didn’t stop. If the car behaved as some commentators would like then it’d never get anywhere as it would be stopping every time there was a potential hazard in another lane.

I doubt most human drivers would have stopped in time for the same reasons as MaxTorque has discussed. They may have been able to slow down enough to not kill her, but I doubt that as well tbh.

It’s most certainly a tragic accident, but it’s one that happens every single day on the roads of the world and no one bats an eye lid.

I seriously hope that self driver car technology doesn’t get binned due to this as the majority of human drivers are crap. The sooner we hand over cars to (well written and tested) software, the better.

I don’t know how the law is applied to these tests but in the UK I’d have expected that the person supervising the car would be held to be the driver legally i.e. the same laws and potential for prosecution would exist as if they’d actually been driving the car. In that case then under UK legislation I’d expect that person could be looking at jail time because there is a very good case to say that it’s at best death by careless driving, or quite possible even death by dangerous driving (given the obvious risks of taking eyes of the road as long as that person did). I’d also suspect that there is going to be a massive lawsuit heading Uber’s way.

It’s obvious to me from the video that a reasonably attentive driver would have seen that person and her bike in plenty of time for some action to be taken. To anyone with experience of comparing what a video camera sees at night to what a human eye can it’s obvious that the cyclist would have been visible well beyond the point where the video shows her – quite likely while she was still on the other side of the road. it’s also worth noting that an observant driver would have had several courses of action possible – none of which the car even attempted to take:

1) Hitting the brakes – and there is obviously time for that to have had at least some effect

2) Steering round her – there was no oncoming traffic and it wouldn’t haven’t taken much of a deviation for the person to have survived, even if the bike itself still got hit

3) Hitting the horn when she was still on the other side of the road to alert her to the danger.

Even if you believe that the cone of light shown in the video was the utmost limit of what a driver could have possibly seen (despite other videos clearly showing that not to be the case) then you’d have to accept that the car (which was speeding) was going too fast for the conditions (i.e. the capabilities of its own lighting).

Wherever you sit on this argument though you’d have to accept that it’s 100% proven that the technology used in the vehicle is not ready for road testing. It didn’t observe the speed limit and it didn’t take action when an object was in the road in front of it – two things you’d think should be advantages to a system like that. If a cyclist had been cycling along that road it’d have presented a smaller radar target than was the case here, so it’s highly likely the car would have just plowed into the back of them without slowing as well.

More pseudo tech from Maxtorque.

The last time you posted some nonsense about recovering a car from completely sideways on an autostrada thanks to fantastic rally-tuned reflexes (but you’ve never driveen a rally).

You slowly toned it down to a bit sideways and then spouted half a page about each tenth of seconds with yaw angle and driver inputat each stage – and all that without telemetry data to refer to.

If the incident ever happened the reality was – something broke, it went a bit sideways and you caught it. Or that’s how any properly competitive driver would explain it.

So having made claims that you caught something that would require a reaction time of less than half a second from breakage to steering imput you’re now claming it takes over 1.5 seconds from something happening to “threshold” braking.

Seiously, Maxtorque you need to make your mind up about reaction times and how long it takes to push a bit of hydraulic fluid through a pipe. Do your own test in a car without ABS. If you’re as good as you claim theer will be labout half a second from a big red light on the dash to smoking tyres. Even if you’re not anticipating the red light it’ll be a second.

Yup, anecdote and opinion always always always trumps actual hard data. No question about it.

Sorry but that’s a guff statement. We already have that with ABS, but you don’t stomp on the brakes and let the car deal with it, to send a text, well you could but it would end in carnage.

I don’t think you can dismiss the idea that driver aids make people pay less attention on the road.  This advert is a good example of that.

https://youtu.be/pjQt2lEZIXg?t=151

The idea that sometimes you get distracted and that’s perfectly normal is the attitude of the majority of drivers.  I can see driver aids making this even more ingrained.  Prior to driver aids you would never try to text and drive. Now you have automatic braking, lane keeping, etc so a surely quick text can’t do any harm?

I think the most dangerous point in this transition is going to be when driver aids become the norm but AVs haven’t taken over completely.

When will the driving gods of STW realise that not everyone driving on the road today is as good as they are?

When did you last practice / do an emergency stop?

I practiced a few when learning,  did one to pass my test in 1996 and one in 1998 (when I was still inexperienced and driving like I wouldn’t these days).  I’d like to think I’m capable of doing a good quick stop – am I really – how do I know?

We all like to think we have amazing reactions and would have braked in time (because it’s unpleasant to think the opposite – that we’re a bit rubbish and would have killed someone)  Not everyone’s reactions are that good particularly when you haven’t practiced the skill following the reaction.  Even doing the classic drop a ruler through the fingers to measure your reaction time at school is scary – yes most kids can do this successfully with a 30cm ruler, some can’t and you need a 60cm – 1m ruler and there are some who will require multiple practice attempts with a 1m ruler before they manage to catch the ruler.  In 5 years time these kids will be driving around on the road and we expect them to do an emergency stop!

Annecdotes are great.

Whether a human driver would have done better isn’t the point.

We were told these cars were safer, and if something went wrong the human driver would take over.

This has been shown to be 100% wrong. The technology failed in the most basic way possible, and the human didn’t do anything (not really surprising).

It wasn’t a difficult situation, with lots of traffic, it was one object that either was missed or ignored. I love tech, but this is many years away from being something I’d be happy to share a road with.

I don’t think you can dismiss the idea that driver aids make people pay less attention on the road.

You *can’t* dismiss it as it’s very very well documented.

Which is why most of the major players (or at least those who have a reasonable sense of self preservation) are striving for level 4 or 5 and missing out level 3 altogether.

Level 3 seems to be the worst of all worlds. At least with level 2 you get regular warnings and wake up calls and need to interact with the controls on a regular basis. Level 3 you get nothing until there is an issue. (or until after the issue has been and gone and is now a huge problem.)

Level 4 and 5 you get nothing at all. As the car is driving for you. Once it decides it’s not driving for you, it’ll pull up and park if you don’t respond to multiple requests to take over over a relatively long period, i.e. when you get to the end of your planned autonomous journey. (At least level 4 anyway, level 5 doesn’t have anything to take over……)

When did you last practice / do an emergency stop?

Practice bout half a dozen times in the last three months. Each time there was fresh snow on the road to assess how much grip there was and if the Winter tyres were adequate or whether I shoud fit chains.

Real emergency, a few years back to let in an on-coming driver who had misjudged an overtake.

The main question for this is:

a) Did a person die “despite” the driverless car

or

b) Did a person die “because” of the driverless car

Based on what I’ve seen I think it’s option b in this case. That’s a real disappointment because I for one am really looking forward to being able to stick my car into auto mode for the long drives back and forward between Edinburgh and London that I do several times a year. Based on that incident though I’d have said it’s necessary for the testing to take a step back as it doesn’t look safe for real world use yet,

Practice bout half a dozen times in the last three months. Each time there was fresh snow on the road to assess how much grip there was

Unless it was done as a complete surprise and directed by a 3rd party that isn’t practicing an emergency stop it’s just braking hard.

@epicsteve I’ll wait for the report

Level 3 seems to be the worst of all worlds. At least with level 2 you get regular warnings and wake up calls and need to interact with the controls on a regular basis. Level 3 you get nothing until there is an issue. (or until after the issue has been and gone and is now a huge problem.)

This times a lot. It’s a big human factors fail to expect someone with nothing to do to stay alert and focused. I’d imagine you’re all “eyes on stalks” at first, but after 1000 uneventful miles, you’re clearly on your phone.

From a consumer point of view, it also doesn’t deliver what I’d want from an autonomous car, which is transporting me while working/sleeping/drunk. If I’ve still got to sort of drive but not, what’s the point of me buying one?

When did you last practice / do an emergency stop?

Practice bout half a dozen times in the last three months. Each time there was fresh snow on the road to assess how much grip there was and if the Winter tyres were adequate or whether I shoud fit chains.

Real emergency, a few years back to let in an on-coming driver who had misjudged an overtake.

So we know that STWs resident driving God has done an emergency stop in the last few years and done some hard braking in the last few months.  Is everyone else on the road as conscientious and diligent at practicing their braking?

No they’re not – absolutely nowhere near.

ghostlymachine

You *can’t* dismiss it as it’s very very well documented.

Which is why most of the major players (or at least those who have a reasonable sense of self preservation) are striving for level 4 or 5 and missing out level 3 altogether.

Level 3 seems to be the worst of all worlds. At least with level 2 you get regular warnings and wake up calls and need to interact with the controls on a regular basis. Level 3 you get nothing until there is an issue. (or until after the issue has been and gone and is now a huge problem.)

Level 4 and 5 you get nothing at all. As the car is driving for you. Once it decides it’s not driving for you, it’ll pull up and park if you don’t respond to multiple requests to take over over a relatively long period, i.e. when you get to the end of your planned autonomous journey. (At least level 4 anyway, level 5 doesn’t have anything to take over……)

Good to see someone else pointing out that there are different levels (which is a point I think several people are missing). Ford are ignoring level 3 and aiming to release level 4 cars in 2021, or so they say, and I’d have to imagine that tallies with what you’re saying.

epicsteve

The main question for this is:

a) Did a person die “despite” the driverless car

or

b) Did a person die “because” of the driverless car

Based on what I’ve seen I think it’s option b in this case.

Again, just to flog a dead horse, surely that depends on the level of autonomy the car is purported to have and the legal status granted to cars of that level in that state. From what I can gather, if the driver had been in any other state in the US or anywhere in Europe he would have been at fault legally speaking.

Couple of things to note:

1. It’s very likely that there was more visibility to the human eye than the footage suggests. We cannot take this as being what the occupant could see without closely examining how it was recorded.

2. Doesn’t this technology work in the dark? (This is a genuine question, and there may well be limitations, but radar, lidar, etc… it doesn’t require light to function?) This is where any comparisons to humans fall down, because in this case, it may as well have been the middle of the day – the amount of light becomes irrelevant. (I’m happy to be corrected on that one)

3. My understanding is that it did not react at all prior to the collision despite a slow moving pedestrian being visible to the cameras for 2 seconds: one of the major selling points on this technology being that it has virtually instant reaction time.

I’m all for autonomy, but failing to see someone crossing the road isn’t really acceptable.

Sorry but that’s a guff statement. We already have that with ABS, but you don’t stomp on the brakes and let the car deal with it, to send a text, well you could but it would end in carnage.

ABS really isn’t comparable to automatic braking systems. One improves the performance of your brakes to assist your own driving. The other takes over your driving duties for you, if you fail to do so yourself. If you have technology in your car designed to prevent collisions without any human intervention, then it stands to reason drivers will become less attentive. This is something well documented and observed in real world scenarios. It’s one of the major factors in Google pushing for full autonomy.

I for one look forward to hearing about the reasons for the failure in this case. But it’s notable that the police response was to blame the cyclist 100% and not even consider that the car/driver might have been a teensy bit at fault.

  But it’s notable that the police response was to blame the cyclist 100% and not even consider that the car/driver might have been a teensy bit at fault.

That’s an American thing. Jaywalking is a crime. Therefore the driver/car is unlikely to get into  major trouble. Will see if I am proven to be wrong on this.

Doesn’t this technology work in the dark?

Some of the very basic/early systems need visible light to function at 100%. As they rely on cameras. So  at a very basic level, less light = less functionality. Anything driving around now, engineered by anyone who is paying attention,  will have at least 3 systems using different media. (Visible light,  lidar, radar, ultrasonic etc) and you’ll need to physically block the sensor to cause any major issue.

I don’t see how jaywalking being a crime could absolve the driver/operator from any guilt. Unless by the same logic I’m allowed to beat to death any driver I see using a phone or breaking the speed limit.

I don’t see how jaywalking being a crime could absolve the driver/operator from any guilt.

On a personal level, neither do i. But it happens with depressing regularity. Even in Europe.

It’s possible that the US takes a different view than the UK, however we saw in the Charlie Alliston case last year that it’s not acceptable under our laws to disregard the safety of pedestrians, even if the pedestrian had initially put themselves in danger by their own actions.If the driver could reasonably have been expected to have been able to take action to avoid or mitigate the collision then they can be held to account for their failure to do so. In my view, from the evidence I’ve seen, I think something could/should have been possible to mitigate the outcome but wasn’t due to a combination of a completely ineffective safety system in the vehicle along with the almost complete inattention of the person that is supposed to supervise those systems. If it turns out there was any corporate awareness of the potential ineffectiveness of the system in those sorts of situation then we could be looking at manslaughter type charges (again, as was the case with the Charlie Allliston trial i.e. where a vehicle known to have its safety compromised was driven in a way that it would be reasonable to assume that it was risky).

On the other hand – if this was a normal car being driven by a person without (or possibly even with!) the video evidence being available then the police would probably just be chalking it up as a tragic accident due to the victims actions, even if they thought (but couldn’t prove) that driver inattention was a factor.

Regardless of the technical reason, it’s all about speed.

If you cannot stop after sighting an obstacle, you’re going too fast for the conditions. That’s regardless of whether you’re an auto car or a driver.

The problem is most drivers in poor visibility conditions operate on the probability of the road being clear rather than the possibility there may be an soft squishy human there.

So it’s that simple, if you can’t stop, you’re too fast.

https://arstechnica.com/cars/2018/03/police-chief-said-uber-victim-came-from-the-shadows-dont-believe-it/

uber auto was at fault IMO, as was the ‘driver’. they’ve been playing catch-up WRT the tech. taking too many risks?

jeez edukator, i hope you’re a nicer person in real life than you come across on this forum.

BTW, there is a decent, informed, review of the events here:

http://ideas.4brad.com/it-certainly-looks-bad-uber

by someone directly involved in autonomy.

they’ve been playing catch-up WRT the tech.

They do seem to have taken a few different innovative approaches to that though. Hence the recent court case.

I like how someone says, after watching a video where the events were known “i’d have managed to avoid her”. Sorry, but, no you wouldn’t

I’m guessing you haven’t seen the video of a chap driving the same stretch of road at the same time of night?

It appears that visibility was no where near as bad as it seemed to be in the actual video of the incident…