Hacker News new | comments | show | ask | jobs | submit login
Autonomous braking: 'The most significant development since the safety belt' (bbc.com)
161 points by ZeljkoS 8 months ago | hide | past | web | favorite | 176 comments

Autonomous braking is a huge step, and, imho, more exciting at the moment than full autonomous driving because it can save lives right now, and is unlikely to take them (shots fired at Uber).

But buyer beware. Even within the IIHS safety standards, there is considerable variability. I love my Toyota Corolla (2017), but it's braking will only take a few mph off after warning you. I can't wait until my lease expires and I can upgrade to the Subaru (edit: or maybe the Volvo from TFA). Look up the videos, they are fully capable of stopping without any collision up to ~40 mph (disclaimer: never rely on these safety features, it's still your responsibility to be safe).

Do your research, happy and safe driving!

I've also read enough complaints about inexplicable emergency braking activation to say that all drivers should be aware these are on the road. You need to consider that now random new cars may execute a "brake check" maneuver in situations where you would never anticipate a human to have done the same thing.

You should be expecting any driver to do an emergency stop right in front of you.

I've learned the hard way not to expect anything from other drivers on the road.

Make sure your following distance is such that you have enough time to react to completely oblivious driving behaviour from other drivers.

> Make sure your following distance is such that you have enough time to react to completely oblivious driving behavior from other drivers.

While I fully endorse this ideal, the reality of driving in high-traffic cities is that it's impossible. Every time you adjust to leave proper space, someone will come along and appropriate your safety buffer. When you slow again to allow enough space, the pattern repeats.

Tragedy of the driving commons. In such situations I just drive slightly under the distance I would normally keep, but then just back off those that do barge in.

I didn't say we should take these braking systems off the road, but I do think drivers need to adapt to their presence. Personally, I am happy my 2017 vehicle does not have it and the ones on the market will probably be much improved by the time I consider a replacement.

As others discussed throughout this thread, there are many situations where drivers violate the safe following distance rules. Aside from the lack of attention scenarios, I think many drivers take calculated risks to temporarily reduce their following distances while performing some other maneuver. They should incorporate this new idea of neurotic automatic-braking systems into this risk calculus, so they don't get caught in a situation they cannot handle.

While we have a mixture of different vintage cars (and drivers!), I am afraid we will remain living in interesting times...

I’ve thought about this and and I think the best thing to do here is to relentlessly keep a long, safe distance from vehicles ahead. There are just things you can’t possibly see that the driver before you can.(a puppy/squirrel/whatever that he wants to avoid killing, you can’t possibly know). You can’t anticipate what a human will do here.

You also give yourself a buffer when vehicle ahead have an emergency stop - imagine there’s a truck behind you.

I guess what I’m saying is these ‘brake checks’ are welcome, even if they only serve to educate people.

Problem is everybody else have to behave similarly, otherwise other cards just take the place of that safe distance.

It might be annoying, but even if a few cars get in between you usually don't lose much by getting a safe distance again.

I second this.

There are three scenarios if I'm not overtaking:

1) I'm going faster than the vehicle in front. In this case, it doesn't matter if another vehicle pulls in between us because I'm about to overtake anyway.

2) I'm going slower than the vehicle in front. In this case, it doesn't matter if another vehicle pulls in between us because I'm falling back and the gap is ever-increasing.

3) I'm going approximately the same speed as the vehicle in front. In this case, there tend to be two ways a vehicle pulls inbetween us:

a- it's merging from an on-slip-road (on-ramp?), in which case this doesn't happen often, and I'll just fall back or overtake b- it's just overtaken me, then slotted into a gap that's too small for it anyway. If the car has overtaken me, it's mostly because it wants to go faster (in which case it will probably vacate the space again soon) or it wants to pull off (in which case it will definitely vacate the space again).

If I am overtaking, then yes, someone may pull into the gap, but I'm still overtaking the vehicle I want to get past.

If you stop worrying about going 2mph faster than another lane of traffic, then leaving a safe gap is mostly pretty easy and stress-free. It will only take you 15 minutes longer to drive 200 miles at 65 than at 70.

Here in the UK, where we drive on the left and overtaking is only allowed to the right of slower vehicles under normal conditions, there are some other variations:

c- a vehicle in the lane to your left that you were going to overtake has itself caught up with a slower vehicle and wants to pull out to overtake it, moving into the gap in front of you

d- a vehicle with an impatient driver is undertaking traffic (passing to the left of slower vehicles) and then moves into the gap in front of you.

The first of these is a normal situation, but still results in a vehicle moving into the space in front of you, sometimes without accelerating up to your speed first. Fortunately, it's usually easy to anticipate this situation, and many drivers will helpfully drop back a little to allow more space for the other vehicle to move out.

The second of these is a result of aggressive and probably illegal driving, and is more of a problem because the driver cutting in may well be going too fast, move out into a space that isn't really wide enough, and then brake suddenly.

Still, in my experience these don't cause much delay if you're allowing a sensible gap in front. I find people who try to keep closer to the car in front to deter others from pulling into "their" space seem to get far more upset about these situations than I do.

(also in the UK) Completely agree with your closing remark. By and large, by deciding to not care about stopping cars getting in front of me, it doesn't bother me when they do. Any delay is completely negligible.

When this happens to me on my commute, I find it incredibly hilarious. Sometimes I wave at the driver through their rear-view when we are inevitably stopped at the same stoplight not 2 minutes later.

I find myself on both sides of this equation, sometimes "what a waste, why did this guy bother passing?", and sometimes "Phew, finally in front of THAT person"(even if we are stopped at the same light, but I'm now in front).

Reminds me of the George Carlin line: "anybody driving slower than you is an idiot, and anyone going faster than you is a maniac"

Agree. In Europe, if you follow the recommended safety distance, you always have the time to brake if anything happens to the car in front of you. The problem is that people do not follow them.

Following the recommended safety distance (2-3 seconds of travel distance if I recall correctly) is quite difficult in moderate to heavy traffic.

If you leave a large enough gap with the car in front, then people will merge into that gap. If you then slow down to create a new gap - people will overtake and merge in again. You end up travelling slower than the rest of traffic, with people overtaking and merging, which creates dangers in itself.

> I think the best thing to do here is to relentlessly keep a long, safe distance from vehicles ahead.

In Germany, this is legally required. However, the safety distance is often abused by other vehicles when passing.

This is just generally good advice and always has been. Don't drive too closely to stop safely.

Sudden curves in the road when there are parked cars is the biggest issue for mine. I think it plots a path based on how my wheel is turned. When there's a 10-15 degree kink in the road, I can't turn early or I'll cross the center line, so it thinks I'm going to continue straight into a parked car. Hasn't brought me to a complete stop, but it could certainly mess with a tailgater.

> but it could certainly mess with a tailgater.

Unless the tailgater also has this system!

Now I've got a hilarious image in my mind of a bunch of modern cars taking turns to accelerate, abruptly brake, accelerate, brake... The future is now.

That’s called a phantom jam, happens right now on 101 with humans;) Won’t happen if only people don’t tailgate.

[offtopic]Yep. I'm usually in the "slow lane" bemidst truckers, amusing myself by watching the Giant Car Worm on the "fast lane".[/offtopic]

Actually, if everyone perfectly tailgated out of a rolling jam, then they'd all be fine. Instead you get people that let a large enough gap open for them to accelerate faster than the car in front, then they have to brake when they reach that lead car again. This slinky effect is what causes those rolling traffic jams and why they can break them by slow-rolling in from of the jam to force everyone to stop the slinky effect.

I imagine phantom jams will pretty much be a thing of the past once we reach a certain percentage of cars running auto-braking cruise control. Have there been studies on what that percentage is?

You have to characterize the control behavior to model whether the system will dampen its response or create a sustained or amplifying feedback loop as all the agents interact. These traffic jams are essentially standing waves propagating between agents.

You don't erase the standing wave by having everybody maintain a fixed following distance or mimicking the speed of their leader. You need to predict future changes and allow the space to contract to help smooth out the transient disruptions and slowly rebound without over-shooting. You want each following car to react less than the previous and have a weaker deceleration/acceleration curve.

Unfortunately, I have noticed increasingly absurd feedback loops in long distance routes here in CA. As a user of conventional cruise control, I can't help but notice how frequently I now have to intervene to adjust speeds in relatively light traffic conditions. On long drives like LA-SF via I-5, these yo-yo drivers are my new nemesis. I don't know if it is due to their use of adaptive cruise controls or a more basic loss of attention or self-awareness. For example, I see packs of cars with speeds oscillating between 55-80 MPH when conditions should easily allow a steady cruise at 75 MPH.

Our two cars both have decent autonomous breaking tech. They did activate in actual risks a couple of times. At least once one prevented a certain crash. They act up rarely. Just some anectodes;

One of them thinks I will drive into a wall when I’m approaching a 90degree turn with a wall. Does the break check from time to time. Nothing weird at high speeds for now.

Other had a single break check at high speed, there was nothing in front so the sensors somehow messed up I guess. It braked for something like 100ms, at least it felt like that. Then kept driving normally. But even that was scary.

> You need to consider that now random new cars may execute a "brake check" maneuver in situations where you would never anticipate a human to have done the same thing.

This is to persuade you to also buy a car with new brake tech ;)

My car does get triggered by lines on the street in a specific area. However, Dutch law says you're supposed to drive at a distance where you can anticipate that the car in front of you can go full stop at any time, so I'm not sure why this should cause people to do that and not the fact that they already should.

> However, Dutch law says you're supposed to drive at a distance where you can anticipate that the car in front of you can go full stop at any time, so I'm not sure why this should cause people to do that and not the fact that they already should.

You've just explained why autonomous braking is completely unnecessary. People are perfect, never make mistakes and never assume that the car in freely-flowing traffic ahead of them won't stamp on the brakes for no apparent reason.

The car will never emergency break as long as you keep your foot on the throttle. My Merc has beeped at me couple times where it thought I was going to collide with something, but because I kept my foot down it didn't actually start breaking - the system only activates if you let go of the gas.

The Subaru actually interprets this as you accelerating into a wall and takes over. You have to let off and then reapply for it to say, “ok, they clearly know something I don’t.”

I just bought a 2018 Subaru Impreza with Eyesight (their suite of ACC, braking, lane assist, etc). I was driving on the freeway into work less than a week after I bought the car and had someone pull out in front of me without signaling or slowing down. The car stopped itself before I had a chance to react and saved me from what could have been a really bad accident.

even warning few seconds before impact is useful, of course there are better systems but Toyota implemented Safety Sense to all cars which is great.

Sorry, but i do not see how the warning is useful, as it first of all distract your attention from driving, and secondly happens too short a time before the would be accident for you to react in my experience. So as a stand alone warning feature without automated breaking, i dont see the use case that you're praising?

A lot of accidents happen because people don't see the car ahead of them.

This could be either because they are distracted (eg. fidgeting with the radio or their mobile phone), or because vision is poor (dense fog).

A system that warns you in those cases that there's a vehicle in front of you sounds useful to me.

This is all too easy. On Friday, I was taking the off-ramp to join a different highway, on cruise control (because I knew the road layout ahead), at the speed limit (70MPH). I glanced down to adjust the heater and when I looked up, a BMW had been sitting on the ghost islands separating the highway and the off-ramp (apparently missed the turn), and the driver had just decided to join the lane I was currently steaming down. Fortunately, the excellent brakes (as well as the loud horn...) in my Supra made up for my lapse in concentration.

I'm in two minds about vehicle automation. Most of it is to compensate for human idiocy (such as the above example). The same gains could be made by humans not doing illegal maneuverers like that. Yes, there are occasions where the electronics can react appropriately to a rapidly developing situation much faster than a human could (e.g. child/animal running into the road), but I agree with the article's caution on complacency. All it needs is for a driver used to auto-braking to get into a car without it and cause an accident, and the technology will suddenly become mandatory.

At the other end of the spectrum, as I've often said, it's technology developed by humans, based on our own understandings, so there will be flaws. Having experienced poor ABS systems that would activate at low speed to the point that I feared the car would not stop at the junction I was approaching, I am not confident auto-brakes will react appropriately in every situation they're called upon.

A lot of vehicular automation is compared to aircraft systems (especially fully autonomous technology) but there is a critical distinction - you never get tailgaters in the air (or if you do, something has gone terribly wrong). Aircraft can rely on huge amounts of automation because in normal use, there are huge amounts of open air in every direction to allow the pilot time to take back control should the systems fail, and still have space and time to correct the immediate problem. Vehicles on the ground are tightly grouped together. Not only does this leave precious little time for humans to react to a bad situation, it also means that any automation systems are similarly constrained. They have to work correctly in a span of seconds or milliseconds. There is no time to 'nope' out and pass control back to the human when they exceed their programming, as is frequently the case with Tesla. Failure is thus very difficult to analyse and account for.

Automation is generally a good thing, but I can't help thinking we might end up automating ourselves into a hole on this one.

Yikes, early ABS brakes were terrifying, especially on snowy or icy roads. I had a couple vehicles that you'd have to start braking a hundred yards earlier than if you were using normal brakes, or else you'd go sailing past where you planned on stopping. Fucking dangerous, if you weren't expecting that behavior.

Another problem with early ABS is that it was widely misunderstood. Many drivers thought it would allow them to stop in a shorter distance under very slippery conditions, when in fact the opposite was often true (but you could steer more effectively). Thus they didn't slow down as much as they should for the conditions and didn't have enough time left to stop.

Also a thick A-pillar will block vision for a few degrees slightly to the sides. Enough to give me a couple of "crap, did not see that (car/pedestrian)" close calls in full daylight.

I feel you there. My partner (who doesn't drive) just bought a second hand Corolla. She'd had it just a few days when I nearly t-boned a flatbed Toyota Hilux because it was hidden completely behind the pillar and just happened to have the correct speed relative to me to allow it to match my turn and remain completely hidden. Luckily, I'm quite cautious at intersections.

I've since learned to move around whenever I'm looking for other vehicles.

Yep, that's a classic case of an engineering trade-off. The thick A-pillars in modern cars are a real detriment to visibility in a place where we need it most, and can not only cause more crashes due to low visibility, but also more impacts with pedestrians since they're frequently hidden in that spot. However, they're also a real boon for crashworthiness: thicker pillars are structurally essential for surviving a rollover accident, and they're also good for holding airbags which can prevent deaths or maimings that the front airbags are insufficient for.

It's a pretty common mistake to rear-end the person in front of you if they stop unexpectedly while you're shoulder-checking for a merge. A warning would probably prevent some of those collisions.

You mean like the mandatory lane change warning light which you should engage much in advance of the manoeuvre already?

Why not do so before the merge check?

I think you're misunderstanding. Consider a merge like the one below. Each line is a lane, and the direction of travel is upwards. This is sometimes how an on-ramp ends, for instance.

    C|| \B
     ||  \
Cars A and B are in the \ lane, and Car C is in the leftmost | lane. The \ lane has a yield sign.

The driver in Car A is a new driver, and is uncomfortable merging. The lane they would be entering is clear, but they're spooked by the car in the far lane and stop unnecessarily at the merge point.

The driver in Car B behind them is following closely and is preparing to merge by doing a shoulder check. They are not looking forward when the car in front of them stops. If the way had been blocked they would have been expecting the car ahead to stop, but they had seen it was clear.

The driver of Car B is at-fault in that collision. They were depending on the driver ahead of them to behave predictably, and they were not driving defensively enough. A warning noise would probably have triggered them to slam on the breaks and might have prevented the collision. Automatic emergency breaking would do the same, perhaps more reliably.

(This was a real accident. I was a passenger in the lead car.)

I was always taught (in the UK) to check your mirror before signaling, and then make a final check before the maneuver.

Absolutely disagreed, the car starts beeping if you're approaching a slower car in front of you but don't start slowing down - it has captured my attention more than once on long motorway drives - it does it long enough in advance that you can go "shit I didn't realise we were slowing down".

> as it first of all distract your attention from driving,

They didn't do tests with drivers inexperienced with the system?

I was very happy with the system in my 2014 Cadillac ATS. When paired with the cruise control system, I could set cruise when I got on the highway, and the car would keep distance properly, including coming to a complete stop (and starting up again) for traffic jams.

I drove into a dust storm in Central Washington in a rental Volvo several years back and autobraking saved me from rear ending a car. I made it out, but a huge pileup ended up happening just behind me. https://www.kiro7.com/news/massive-crash-closes-eastbound-i-...

I've read about a couple of accidents like this, and I always wonder why people don't slow down or stop before they drive into dense dust or fog?

Maybe because of the worry of getting rear ended themselves? And I'm sure a lot of people do; but then those drivers don't make the news.

Slow down, put on the rear fog light. That's how you're supposed to handle it, but either people have never been taught to do it, or they've forgotten.

You can't do that. We do not have rear fogs lights in America.

Well, you should fix that, then. They're quite useful.

It's utterly impossible for us to change this. We didn't invent rear fog lights in America, so because of this it would be completely impossible for us to adopt them. We're only able to change the standards for our cars when we invent them first.

Rear fog light?

All cars sold in Europe must have at least a rear foglight to be road-legal. Imported models must have one retrofitted.

Sometimes, rules and regulations actually do improve things :)

It's because the purpose of slowing down is primarily in terms of safety margin: if something goes wrong, you give yourself time to safely stop. If the occurrence of low visibility is infrequent enough for people to not have experienced an accident, but frequent enough to have experienced it more than a few times, it quickly settles into a habitual "well, I've had no problems going 70 in 0-visibility fog the last dozen times, so it can't be as unsafe as I'm told." What makes it more annoying is that actually driving a safe speed is also immensely unsafe, because the other cars are going 70 and can't see the car going 25 until it's too late.

One of the most unnerving drives I had to do was traveling on the Ohio Turnpike in super-dense fog. The traffic was heavy but moving more or less at speed. Cue the toll plaza... and the need to reform two lanes of traffic, with slowly accelerating truck traffic, cars whizzing from the right to pass the accelerating trucks, all in dense fog with almost no visibility (I don't think you could see the rightmost toll lane from the leftmost). And one idiot didn't have his lights on.

Final safety tip: if you can't see the road very well, turn your lights on. Even if they don't help you see the road (which they don't in fog), they absolutely help everybody else see you.

Probably arrogance. It's been many years since I've driven through fog thick enough to truly limit visibility, so light fog I might only drop my speed by 5-10MPH. Taillights are still visible at a safe distance in light fog. I can just about remember how easily they disappear (almost evaporate) into nothing in thick fog. Assuming the fog isn't that thick and driving through with excessive speed, it's astonishing how quickly the car in front with rear lights on full can appear out of nowhere.

A combination of wanting to get to your destination on time, inexperience and overconfidence are probably the key reasons.

Example - you're driving at 4am on an empty road with patches of heavy fog - so most of the time you can maintain the legal 70-80mph with good visibility, but then there are bit where the fog gets heavier for half a mile or so. So you slow down the first few times, but as there is no traffic at this hour, you just don't after the first dozen times or so, it's annoying and boring to be slowing down constantly. At then accidents happen because the 13th time there is something on the road.

Not directly related to your comment, but I was driving downhill a few months back when some halfwit came the other way in a car that clearly wasn't roadworthy. Must have been leaking a huge amount of coolant into the oil, because the cloud of steam (or smoke) behind it was so thick that half a dozen cars ended up just parked in the middle of the road, waiting for it to clear before moving off.

Turned my lights on to avoid being hit, seemed to work - or maybe I was just surrounded by cautious drivers that day.

People tend to underestimate the real speed. After driving for long time at 60mph, slowing down to 25mph feels like crawling while it's actually still pretty fast.

Seems very common in the US to just keep going at considerable speed with even limited visibility.

And admit that I can't see???

It's a weird piece. The describe the XC90 as the safest car they ever tested, and that it hadn't had a fatality since 2002. Then it attributes it to AEB.

The problem is that the XC90 got its AEB in 2015. This cannot be the reason for the impressive safety levels since 2002.

Yeah. The Volvo XC90 is a big car and most of the excellent safety record for it can be attributed to the inherent higher safety of larger cars:



It's just physics. When large a object meets a small object then the large object wins because it has more energy and ends up pushing the small object backwards.

Also the fact that it's a Volvo, which among other things will introduce a lot of selection bias. For several years, there were no deaths in a Volvo 240 either:


Fatalities are quite rare. 2011-era midsize luxury SUVs have 13-15[1] driver deaths per million registered vehicle years in the US (UK has half the death rate per km, vehicle or inhabitants).

It is not that improbable that midsize luxury SUV that sold just 50,000 cars in the UK since 2002 had no driver deaths.

AEB was first installed in the 2015 model year. If we assume[2] that the same number of cars were sold each year, the XC90 had 20,000 registered vehicle years with AEB. We should expect 0.03 deaths.

It would be more surprising if there was a driver death in a Volvo XC90 equipped with AEB in the UK.

[1] http://www.iihs.org/iihs/sr/statusreport/article/50/1/1 (first table)

[2] It's probably false, adjust with your favorite fudge factor.

Also, surely that needs to be more accurately described as 'safest car ever tested /for its occupants/'.

I recently bought a WV Atlas with adaptive cruise control and front assist. It also has a variety of other sensors and assists. It is no way an AV, but these features IMHO add a lot of value. If every car had them, I bet road safety would increase significantly. I agree with the article. I think it is a big deal.

> If every car had them, I bet road safety would increase significantly.

If you look into the data on fatal accidents and examine them even for a few minutes you'll easily see that this is a foolhardy bet. The causes of accidents and fatalities are highly variable and not what you would expect. There's also extreme variability between the individual states; for example, Texas has more _total_ fatalities than California. There's extreme variability between the sexes and for different age groups within those sexes. Finally, there are motorcycles.

AI/Driverless, AV and all the attendant sensors and inputs will have an impact, just not nearly as large of one as many people unfortunately expect.

I haven’t looked at any data, but I’m willing to bet non-fatal vehicle related collisions out number fatal ones by a significant factor.

Would a decrease in some types of collisions be a reasonable assumption under the grandparents premise?

What’s a good way to look at and explore this data? Where did you learn about the patterns you mention above?

Use NHTSA's FARS: https://www.nhtsa.gov/research-data/fatality-analysis-report...

Personally I downloaded their database and then converted into an SQLite database that I could run queries against. They do have an online tool, but I have no experience with that.

Cool! Just a hobby for you or do you have other motivation to explore the data?

I think safety includes non-fatalities. Where I live I can safely say that nearly every day during rush hour there is a rear ending collision causing lots of delays for lots of people. Rarely does this kill someone but it’s still a safety issue. I bet nearly all of these would be mitigated if every car had such technology.

If you look into the data on fatal accidents and examine them even for a few minutes you'll easily see that this is a foolhardy bet.

Quoting the linked article:

At the moment, most AEB systems are "forward-looking". In other words they are designed to prevent crashes where one car runs into the back of another.

Such crashes account for around a quarter of all accidents, but these are the incidents which can cause some of the nastiest injuries, such as whiplash.

AEB has been shown to reduce such collisions by 38%, according to research by the safety organisation Euro NCAP.

> Finally, there are motorcycles.

Absolutely, I wish every prospective motorcyclist would study the stats (more than 35 times increased mortality rate per driven km [1]) and discuss this with their loved ones.

Those who insist driving a motorcycle afterwards deserve their genes to be removed from the gene pool.

[1] https://en.m.wikipedia.org/wiki/Motorcycle_safety

Perhaps you can also provide a reference to how risk-taking is inherently an undesirable genetic component that should be weeded out.

I'd prefer to give a reference to how risk-taking while involving others is an undesirable trait. When I mortally hit a motorcyclist, even without a fault of my own, I might feel guilty the rest of my life. So indirectly, the irrational risk-taking of others is involving those who are trying to act responsibly.

So, I suggest to take the risk-taking elsewhere: go climb a mountain, do base-jumping or take the motorcycle to a racing track (bonus: real competition!). But leave other motorists out of your game.

Before wishing death or removal from the gene pool on anyone, maybe you should look at why the accident rate is high, especially in the US.

Severely deficient mandatory motorcycle training, riders who refuse to wear helmets or any safety gear at all, a tendency to just go for the biggest most powerful engine right away, and of course shitty drivers, who either deliberately antagonize riders, or simply don't give a shit.

Squids and outlaw bikers heavily skew the numbers.

Perhaps. The number 35 is a US factor, but similar orders of magnitude can be seen on the Dutch roads (factor of 22 if I recall). On Dutch roads, we have mandatory motorcycle training, obligatory helmet wearing and relatively save drivers.

All I am saying is: it's fun to enjoy a sport, I get it. As a top-roping climber, I also empathise with the thrill of adrenalin. But I just did the numbers: an average motor-cyclist will have a 6% chance to be involved in a serious motor vehicle accident in his lifetime, causing extreme trauma. 50% of those accidents involve other drivers. Think of them. They did not choose for cyclists driving 200 kph while overtaking through the middle. Did that cyclist take that other driver into account when he lays there on the tarmac? Did he take into account the years of psychological trauma he caused by his reckless behaviour?

Ah, so the biggest factor is behavior? People also drive recklessly in cars, but they more often end up getting other people instead. So we should encourage them to ride bikes.

The main factor is mode, not behaviour. The point is: the probability of paychological trauma involuntarily caused to others is much higher on a bike than on a car, regardless of behaviour. Seatbelts and other minimum driver safety are obligatory for the same reason.

"driving 200 kph while overtaking through the middle" is reckless behavior, irrespective of the mode of transport.

You don't have a bike license, do you? Your line of reasoning is very consistent with car drivers who are a bit afraid of traffic, and very afraid of "murdercycles".

Too often I see motorcycles being missed, because of their limited visibility and unpredictability. Due to that, and their higher vulnerability, they ask more attention of other road users. If I hit a car, well, it’s likely we both survive. If I hit a motorcycle I might well have a death on my consciousness.

That is because of inattentive and unskilled drivers. Most reasonable riders try to be as visible as humanly possible (and ride as if they're invisible). A motorcycle isn't unpredictable, it goes forwards or turns somewhat, just like a car.

Nobody deserves eugenics.

I never proposed for others to consciously select for certain genes. Those who drive a motorcycle can inform themselves of the risk they take and as such, choose for the higher probability of their genes not making it to offspring. Especially considering that it is predominantly 16-24 yo who die in motorcycle accidents.

I recently bought a T-Roc and ACC is pretty awesome. The issue I have is on the motorway, when changing lanes, it has a tendency to lose speed and with it being a 1 litre, it takes a while to build up the speed again.

It also has front sensors which has alerted and braked for me a couple of times when someone has almost stopped dead at 60MPH (even with stopping distance, it was a quick stop).

Maybe Changing lanes isn’t really a use case for cruise control that doesn’t require extra user attention.

In my car you can manually give it the gas when changing lanes and accelerate faster than by ACC without disabling ACC. Maybe your car allows that too?

yep, got a VW Alltrack last year. the adaptive cruise is really neat.

I'd like to see more variable brake lights to go with this:

i.e. a strip of light across the rear of the vehicle, that progressively lights up according to how hard the vehicle is braking (or anticipates braking, if it's autonomous).

Some cars have a flash-brake-lights-under-heavy-braking, but I think it would help traffic flow if you can more easily distinguish a touch of the brakes from a press of the brakes.

I've thought this would make sense too and I feel like I'm missing something as to why it hasn't been implemented. Complexity perhaps?

Interesting to note, though, that we do have weak forms of braking that don't light up the tail lights - heavy engine braking in the case of manual transmissions, lighter engine-braking for automatics, and in the case of electric vehicles (at least a Tesla in my experience) the lights don't kick on from regen unless it's passed a certain deceleration level.

There's no tangible benefit big enough for the added complexity and maintenance cost.

What exactly do you get from incremental brake lights?

You get clearer communication from the car in front of its actions and intentions, in an intuitive way. This assists traffic flow and might enhance safety.

I think some sort of variable lighting would be nifty, but in my book it's the wrong solution to what is really a problem of keeping distance. For highway driving the answer for me there has been adaptive cruise control – it's amazing! What I'm missing though is for a way to have the car keep distance, but still leave control over the gas to me. This would be very useful in high traffic situations where setting a cruise control speed isn't really usable, since you'll likely be disengaging it now and then due to braking because someone is cutting in.

I put a lot more faith in sensors to judge the distance to the car in front of me, than my visual perception of brake lights. At a certain medium distance, I'm sure I wouldn't be able to judge the difference between 25% brake and 75% anyway, but my car's sensors would be able to detect that I'm on a collision course, and could adjust gas/braking accordingly.

I dunno. I personally like the non-variable brake lights because I can gently tap on my brakes to make tailgaters give me space.

Volvo's stated goal is:

"Vision 2020 is about reducing the number of people that die or are seriously injured in road traffic accidents to zero. "[1]

As sceptical as I am about corporate statements, you can see that Volvo is steadily working on this. They don't do splashy announcements or announce revolutions in driving, and yet they bring more and more changes and improvements to their cars. From assisted braking to lane assist to blind spot information to city collision avoidance to many many other small and big improvements.

[1] https://group.volvocars.com/company/vision

Anti-lock brakes never yielded the accident reduction expected, primarily because drivers used the improved braking performance to drive faster in poorer conditions.

I guess the AEB works at reducing accidents because it IS autonomous and does not "improve performance".

BTW the KPI is reduction in insurance cost.

On the other hand death rates have plummeted in recent decades thanks to all the safety advancements.

I wonder how many minor incidents it has prevented though. I was raised in the country where roads are often wet or icy in winter. Without ABS if you go around a corner and hit some black ice there is a good chance you would spin out and end up in a hedge or a tree. At the low speeds you need to drive (<20mph) it’s unlikely that the passengers would suffer any serious injuries.

As someone who drives rather frequently in atrocious winter weather conditions. I often wonder how AEB will handle black-ice and deep snow where tapping the break gives you a chance for loosing traction and ending up in the ditch.

First rule of winter driving is don't touch the brake.

No, the first rule is "be smooth". No sudden sharp inputs, give the car time to settle and adjust to what you're doing.

You can absolutely brake, you just have to be smooth and gradual.

> First rule of winter driving is don't touch the brake

That doesn't sound right. Pretty much all cars today have ABS, which should make sure that you don't lose traction.

When I first drove a car with ABS in winter I was amazed how well the breaks worked in bad conditions -- of course the car would take a lot longer to come to a stop, but it did come to a stop.

The key feature of ABS is that it prevents the steering wheels from locking up, allowing the driver to steer around the obstacle. On loose surfaces, ABS can cause braking distances to increase; locked wheels can dig into gravel and slush, causing it to stack up and increase friction. ABS is also useless on sheet ice (but then, no braking system is genuinely at an advantage there).

These are edge cases though, and ABS does generally improve braking distances. At the very least, it gives you a much better chance of avoiding the obstacle in your path.

You really shouldn't have to brake hard enough that the ABS comes on. Be smooth, don't jam the pedal, brake right on the edge of ABS activation.

Having just escaped from the snowpocalypse of Minneapolis tonight, I can tell you this is right.

The other thing is to accelerate through turns - it helps maintain traction.

I guess AEB will engage the brakes like a human would so the other safety systems like ABS and ESC so the car wouldn't behave worse than if a human would break.

Depending on the conditions that can be fairly acceptable, for purposes of drifting a modern car's breaks suck a lot thanks to those two (the handbrake is free game tho)

> drivers used the improved braking performance to drive faster in poorer conditions

That's not the ideal outcome but it's still a net win, no?

A German taxicab company did a study, pitting half its fleet with ABS against the other half without it. The accident rate stayed the same (in fact, was insignificantly higher) in the half with ABS because the drivers felt overconfident in the braking system:


Yep! But they got to drive more aggressively without a substantial increased risk to their safety! Being able to go faster in the rain is useful in and of itself!

The problem with this combination is that it may reduce the number of crashes/injuries, but increase the number of deaths at the same time

Faster = higher probability of an injurious and/or fatal accident.

So from a safety standpoint, that is a not a win.

A lesson from the ABS experience has influenced the design of AEB systems. The problem isn't so much that humans abuse the ABS to drive in harsh conditions as that they don't understand that they're relying on the ABS at all. They go around a bend too quickly and they don't think "Man, I nearly came off there, good thing ABS intervened to keep the car going where I point it" but instead "I'm great at this driving thing, a less skilled driver would come off there".

In AEB systems the emphasis is on dramatic interventions when calamity is imminent. So for example rather than gently brake pre-emptively because you seem to be getting a bit close, AEB is focused on braking very hard at the last moment, this doesn't feel like your skill saved you, it feels (correctly) as if the automated systems kicked in to avoid a crash.

I'd agree anti-lock brakes haven't had much effect. It will only help people who haven't learnt how to brake properly, but we all get plenty of practice at that during everyday driving so I expect the difference is marginal. But the story is different with a related technology, ESP, which is mandatory where I live.

No so long ago I was caught out not paying enough attention approaching a roundabout at a creek crossing on a rainy day. It dawned what I hadn't allowed for the slippery road conditions on that day when the car started to slide towards the creek rather than go the direction I was steering. I'm by no means an expert driver, but I've spun a car off a dirt road into the bush a few times in my miss-spent youth, enough to know when I've lost control beyond recovery. I only had a few meters before a collision, but another thing I learnt during that miss-spent youth is if you can arrange to slide into a guard rail sideways so the load is spread along the entire side, if your speed is low enough you might get away with no panel beating at all. So I eased off the brakes to get maximum steerage.

Then the car noticed the steering wheels had lost traction, and the ESP cut in. ESP boils down to steering with the brakes and engine - sort of like a handbrake turn, but the brakes are applied to a front and a rear wheel on opposing sides and the engine powers the other two. There is no way to do it manually. Turns out it's amazingly effective, as it both slows you down and gets the car pointing in the right direction.

There was no accident that day.

I've stomped on the brakes quite a few times to see what ABS in action feels like. I reckon I can do a better job manually. But there is no way I'd drive so crazily that ESP cut in on anything but a skid pan. Besides, there is no point. I know I could never do as good a job, because as I said, there is no way to manually replicate what ESP does. Having said that, a good racing driver will do better without ESP because they get the car pointing in the right direction before it starts skidding - but that requires paying more attention than I was at the time.

I certainly hope that people do not get practice braking as hard as possible without locking up the wheels during everyday driving. Everyone 'learns' it when they first learn to drive but without frequent practice it is not going to be your immediate response when you get in a situation that requires sudden very hard braking to avoid a collision. Modern ABS is probably better at braking than 95%+ of people that aren't race car drivers ESPECIALLY in the situations where it's actually needed and people are probably panicking.

What you refer to as ESP is usually known as ESC where I live and it's also been mandatory here for a while, agreed that it is pretty incredible. I can definitely see it giving people a false sense of security especially in low traction conditions like snow/ice/rain.

Related to this, I wonder if insurers are going to demand access to an insured car’s autonomous-braking data?

Your airbag already records that info, if it fires.

don't worry, it's going to be siphoned off next time you get your car serviced anyways

Another reason the XC90 has a great safety record is that it’s a 4500 pound car with a 4 cylinder engine. This isn’t safety enhancing in itself but does ensure that it is only bought by people with the most sedate driving habits.

> does ensure that it is only bought by people with the most sedate driving habits.

I would disagree with that statement!

As an urban cyclist the two most terrifying vehicles on the road for me are the XC90 and the VAG PL71 ( Q7, Touareg, Cayenne ) not only because of their size but also because they are predominently driven by distracted parents on the school-run.

Two tonnes of SUV, poor outward visibility, stressed driver looking for a parking spot on the kerb, kids bickering in the back == danger.

I can understand parents buy them to keep their little darlings safe from the other nasty cars but I'd much rather jostle with twice as many normal-sized sedans.

The articles says that

> not a single person has been killed while driving it, or as a passenger.

It doesn't say anything about no one killed by the XC90, which is quite, ahem, surprising for an article touting AEB.

It's a 4400lb car that will do 0-60 in 6.4 seconds in top trim. It's 4 cylinder engine is both supercharged and turbocharged. The demographics determines the safety, not the number of cylinders it has.

The Audi Q7 3.0T is $7k cheaper and does 0-60 in 5.7 seconds. The XC90 is a completely drivable and enjoyable car, but people who care about performance are likelier to choose a different SUV and likelier to get into an accident in that SUV.

> The Audi Q7 3.0T is $7k cheaper and does 0-60 in 5.7 seconds.

The XC90 T8 was tested at 5.0s here, which is pretty good for being a review test and not a manufacturer figure http://www.motortrend.com/cars/volvo/xc90-plug-in/2017/2017-... Now, the T8 is expensive as hell obviously (~$70k and up) and competes with other big hybrids, not the base spec Q7 3.0TFSI. Most of thesre aren't sold anywhere near msrp, at least around here they usually go for nearly $100k with extras.

So the XC90 has a pretty reasonable 315hp T6 version that competes directly with the Q7 (But is around 1 sec slower 0-60) and one 400hp T8 version that is going to look pretty good next to any of the big X5/Q7 suvs in performance. I think it's a pretty reasonable set of trims.

If you consider 6.4s 0-60 slow, especially for an SUV, you really need to recalibrate your expectations.

Those 4-cylinder engines deliver up to 303hp and the hybrids get another 87hp from the electric drive train. So they're not exactly weak cars...

Drive one. The throttle is leaden.

That's programmed for fuel economy. It has to combine 3 impossible things: a pretty small engine, a heavy car, AND has to do X mpg in "normal" driving. The solution is to make the mapping from driver input to actual throttle/fuel be pretty lame. It's the same Usually there is at least a mode called e.g. "sport" or "power" that maps the pedal to the engine in a better way (except for economy). The XC90 T8 has a "power" mode where you get from 0-60 in 5s flat.

Have you driven it? I didn't, but my current car is a similarly-sized Seat, also 4 cylinder. It's quite snappy for its size, snappier in fact than all US cars I personally drove (though, tbh, in the US I've never been in a hurry, and I tend to drive 100% legally because I'm not in the mood of familiarizing myself with foreign/US law)

The SEAT Ateca is almost 1000 pounds lighter than the XC90.

Alhambra, but you're right, XC90 seems to be about 700pounds heavier (because it actually weighs almost 5000 pounds, from what I read). It's also more powerful. I call that a tie :) - but, tbh, the gearbox & suspension probably matter a lot.

It's easier to build a safe car if it's big and heavy - but the number of cylinders isn't really a measure of safety (anymore). All manufacturers will go to smaller petrol engines and then to no petrol engines. The new hybrid is 400hp, which is as much as they used to take from the V8 version of the XC90. It might be that the XC90 is bought by safer drivers on average - but the performance probably isn't the difference. 3-400hp is plenty even for the biggest suv if you want to go fast. To me 0-60 in 10s is plenty quick, and these cars are 5000lbs and do 0-60 in 5 to 8s which is extreme (for everywhere but the US)

> it is only bought by people with the most sedate driving habits

Aside from Jeremy Clarkson of course, who's bought like every generation of the thing

And i think this is a good thing. Public streets are not a racing track.

More people should have "sedate driving habits". The road isn't your personal racetrack.

Slow down, stop stressing, live longer.

I was wondering about the AEB when the Uber accident happened. It should’ve been equipped with AEB but it still hit a pedestrian. Did they just pull out the whole software and replace it with their own navigation logic?

The company that makes the standard safety equipment said it had been disabled[1].

[1] https://www.bloomberg.com/news/articles/2018-03-26/uber-disa...

Yes, that was something that was mentioned before by Volvo. Uber disabled the AEB that came with the vehicle and according to Volvo it likely would have detected the pedestrian.

The skeptic in me thinks that this story is Volvo trying to counter the bad news of their car killing a pedestrian with good news of how many drivers haven't died.

Although it's not their fault for the crash, celebrating the safety of an SUV, when it's size and extra weight means that crashes with pedestrians will more likely end in death, seems odd.

Volvo's automatic collision braking sure has improved since 2010


The BBC story, also reported in multiple other places, is a nice bit of Volvo marketing g, but is nonsense.

Volvo introduced Aeb in 2007 on the XC60. The XC90 only got it when they introduced the new generation a few years ago. Therefore claiming that the exceptional safety record of the XC90 is in any way related to AEB is just rubbish.

The reason why XC90s are associated with so few passenger injuries (note, no claims are made for injuries to other road users by XC90s) is that they are large, heavy and chosen by safsr-than-average demographics.

Sure, AEB is a great thing. But it's odd to see "since the safety belt". Air bags have saved more lives than safety belts, haven't they?

Also, I can imagine additional advantages of AEB. If someone's tailgating, just hit your brakes enough that their AEB will trigger.

The first google hit (http://citeseerx.ist.psu.edu/viewdoc/download?doi= says

> in the period 1991-2001 [...] about 109,000 lives were saved by belts and 8,000 by airbags

so I guess seat belts were a bigger improvement.

I guess the question then is: why bother?

Also airbags depend on seatbelt use to be effective, and some of them have killed people.

? why bother saving 8,000 lives?

In what world is it better to NOT have airbags?

If the cost of installing airbags could instead be used to install some other safety feature, say aeb in all vehicles, and if that feature saved more lives than airbags do, then it would be better to not have airbags.

While 8k is a substantial value, other technologies might be more effective, less costly and possible safer.

Remember the takata airbag defect killed 15 people in the US. https://www.consumerreports.org/car-recalls-defects/takata-a...

8000 vs 15, and that's not a problem with airbags, it's a problem with defective industrial components and corporate nonsense.

> Air bags have saved more lives than safety belts, haven't they

Not by a long shot - seat belts reduce the chance of death by 60-70% across all crashes. Airbags reduce the chance of death by 15% in frontal collisions, and very little for other types[1]. The airbag isn’t going to much good if you’ve been ejected from the car.


Airbags also have the misfortune of having killed more people than safety belts. Especially kids. Or people leaning forward.

That’s why modern cars try to avoid firing them unless absolutely necessary. You can get into a pretty serious crash and they won’t fire if the computer decides its better to let you gently tap the steering wheel than to run an expkosion into your face.

Fun example of an old and new Renault Espace crashing into each other: https://youtu.be/vRxLlFm3VUA

> Airbags also have the misfortune of having killed more people than safety belts. Especially kids. Or people leaning forward.

Not to mention the current situation with Takata, where they have had instances of their airbags directly killing people due to a failure of the metal housing, which turns to shrapnel.

> Air bags have saved more lives than safety belts, haven't they?

That depends on how you count: no life saved by an airbag would be saved without airbag AND safety belt. An airbag alone saves no lives.

And I still doubt the airbag+belt combination made a bigger dent in the statistic than the introduction of the safety belt did (Although it's hard to compare due to completely different time periods)

This will save countless pedestrians' lives too.

OK, that was certainly a thoughtless comment :(

I did mean seat belt + air bag vs seat belt alone. But even then, air bags apparently improve seat belt protection by less than 20%.

I was thinking of dramatic high-speed frontal collisions with ~immovable objects. Where forces from seat belts alone can cause damage.

This feature is also available in "cheaper" cars. My Kia Niro has is, and although I didn't get a chance to make "full" use of this (thankfully), I did incur the "too close" beep which prompted me to break one time I was not paying the road the attention it deserved. As a side note, with auto distance keeping cruise control, lane assist that actually moves the wheel to keep me in lane, side radar that alerts if a car is coming as I try to switch lanes and the AEB, this is great entry into autonomous driving (as far as some core systems that are actually in commercial use already).

I had this engage in my car once. It was the first time I was driving with glasses on - I must have misjudged the distance between me and the next car. Scared me, but not as much as the guy following me a little too closely in his E46.

It does beep randomly sometimes - usually in heavy rain. But that one time it turned on pretty late, so it's a good last resort.

Thats some really impressive stats.

I think this is a combination of great safety equipment, the safeness of the car itself (crash tests and so on) and the people who buys it. Its not the most hardcore drivers who buys a Volvo, even tho the 2017 and newer models are really good looking. Volvo has always been a pioneer in security as well

Define "hardcore driver". Is it someone who drives aggressively and takes dangerous risks?


My relative owns one. The Adaptive cruise control only picks up moving cars in front of you and tries to kill you when there’s a stopped car in front of you, say at a red traffic light. I’ve never been brave (or stupid) enough to see whether the AEB would counter that especially in marginal road conditions.

Anybody know how well these handle water? I imagine a situation after/during heavy rain with giant puddles of water. A car in the lane next to me hits a puddle at high speed throwing a lot of water in the air in front of my car. Does the car slam on its breaks?

If the splash is large enough to cause visibility issues, I'd say it must slam the brakes.

There has been a similar system on Mercedes cars (Collision Prevent Assist) for years: https://www.youtube.com/watch?v=h5ia5e07BqU

Don't use that if you're leading a bicycle race. It happened in the Dubai Tour earlier this year: the organizers car (a Mercedes) autobraked and the bunch crashed into it. They disabled the sensor after that. Obviously it's not a normal use case.


Aren't all of them just using Mobileye?

Isn't autonomous braking standard in most new premium-ish cars? My not-so-premium VW has it. Any car that has a distance-sensing cruise control should have it.

Perhaps Tesla should consider licensing this tech from Volvo.


Serious question: Would the Volvo system have picked up and appropriately responded to the gore point involved in the recent fatal Tesla crash? Any time anyone raised the question of Tesla's AEB's reliability, the responses were along the lines of "no AEB is perfect", "AEB only works on cars", etc.

> Would the Volvo system have picked up and appropriately responded to the gore point involved in the recent fatal Tesla crash?

I'd rather ask, would the Volvo system steer towards the barrier like Tesla's AP may have done [1] [2]?

"IIHS research shows that AEB systems meeting the commitment would reduce rear-end crashes by 40 percent." [3]

AEB on its own may save lives. Whether autosteer systems do or not is an open question.

[1] https://youtu.be/VVJSjeHDvfY?t=37s

[2] https://youtu.be/6QCF8tVqM3I?t=28s

[3] http://www.iihs.org/iihs/news/desktopnews/u-s-dot-and-iihs-a...

Both are good questions.

1) Do other autosteering systems (such as Volvo's) share this failure mode? (I don't know much about Pilot Assist but it seems to require a lead car to follow, is that right? Apparently Pilot Assist 2 doesn't, though? Currently reading http://forums.swedespeed.com/showthread.php?348321-Auto-Pilo... and it doesn't sound great.)

2) Should AEB stop the car if there's something solid in front of it, regardless of what the autosteer system is doing? (I would have thought so, and it's disappointing that in this case it didn't!)

Still, I can't help but feeling sad, as a human, to be taken out of the equation and not be needed anymore.

As a programmer becoming obsolete is the greatest possible achievement.

oh man my Datsun is a coffin compared to these things ha. lap belts are designed to cut you in half <:0

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact