As the article points out, a user doesn't yet know what to expect when it comes to the interface of an autonomous car. I doubt I could tell if a car I was driving had this feature because I wouldn't expect to 'see' anything - the definitely doesn't need to be an "on" button for it. If there's a logo somewhere (like an airbag) I wouldn't know to look for it. When you drive an unfamiliar car you need an UX that makes sure this sort of thing is effectively impossible. At the moment I imagine that means that the car needs to tell you what possible features it doesn't have as much as what it does have.
> a user doesn't yet know what to expect when it comes to the interface of an autonomous car.
I don't think this is a good enough excuse, honestly. Both the law and common sense say the same thing about operating a vehicle: the ultimate responsibility of that vehicle and what it does lies on who is sitting behind the wheel. Until we have truly automatic driverless cars (enter an address in the Maps app and let the car take you there - no pedals, no brakes, no turn signals), human beings need to still be aware that they are still driving the car, regardless of assistant technologies that help them along the way.
Any person who knowingly takes their hands off the wheel of a moving car they are piloting is going to ultimately be at fault when it plows into a pedestrian or into oncoming traffic, regardless if the computer tried to prevent it from happening.
The only User Experience problem here is that the users have No Experience with the new technologies available in these cars. If anything, that alone should ask for more caution before engaging these features, not less. The biggest outcome I'd expect in the short term is more bleeping and bonging from car speakers if the user does something completely insane like take their hands off the wheel while the car is moving above a certain speed to film themselves performing a Darwin Award attempt for YouTube.
>the ultimate responsibility of that vehicle and what it does lies on who is sitting behind the wheel
While the law may still hold that position, reality is starting to disagree. When you have a car laden with 'assistive' technologies which are able to take control of brakes, throttle, and even steering - sometimes without human input, things start to get complicated.
We have plenty of precedent already. Toyota's billion dollar unintended acceleration criminal liability case, the 'remote control' Jeep Cherokee case, the recent FBI advisory regarding security and safety of connected vehicles.. I find it very hard to agree with a blanket statement that the driver is responsible for everything a modern car does.
The notion of performing such a test with a real human target is about as idiotic as testing the safety mechanism on a loaded gun by pointing it at someone and pulling the trigger.
That's exactly the root of the problem, if your car runs over your dog because you didn't had firmware 3.17.4b installed or because your license has expired isn't really something any user should be dealing with.
I see Autonomous driving like iOS vs Android when you buy an Apple device you know what to expect (or atleast used to before they decided to fragment their SKU's to hell) each device comes with the same features and the software works exactly the same. The Android ecosystem on the other hand is fragmented to hell and each devices comes with completely different features, software and user experience.
This can easily kill autonomous driving because it unlike poor app performance self driving cars can actually kill people if the self driving experience I get on a BMW is drastically different than what i on a Honda or a Volvo I won't be confident in using any of those, not to mention that at the moment you can't even get reliable experience on the same make and model and it's a feature not a bug in most cases because car manufacturers want to monetize self driving so for every (even basic) feature you'll have to pay extra.
Self driving cars need to get a standardization and they need it fast a fixed set of features that all cars have to properly support to be able to market their self driving cars in the US and Europe and this has to be a dichotic standard you either support everything in that standard to a satisfactory level or you don't get to turn any of those features on.
The last god damn thing I want to worry about is if my car's pedestrian avoidance app was reactivated or not after the weekly firmware update, but sadly enough i foresee some one getting run over because the driver of a "smart" car did not accept the new EULA that was pushed over the weekend before the industry gets any shred of sense.
People are notoriously bad at statistics and will do what they can to interpret this in any way that serves them.
"Human killed by robot car" will definitely be lapped up and even if statistically it may be better than "Human killed by human driver" the one is the exception, the other the rule and so the one will get airtime.
Marketing this kind of thing properly is going to be extremely hard once the casualties start mounting (and they will, no matter how good the software is).
That's axiomatic in that we let cars kill a million people, and injure millions more every year and we haven't yet done much about that. Because we suck at statistics and we are also convinced it won't be us because we are in control of our vehicle.
But that's still a compelling number. Just as most people thought malaria was some quaint disease that inconvenienced the British Empire until Bill Gates got on it, I think road deaths are potentially at least as compelling a target, especially once people look outside Europe and North America.
Yes and no, that may hold true for specific mission autonomous applications e.g. metro transportation, sanitation, cargo transportation etc.
In these cases the the vehicle and to some extent the environment is much more controlled and it configured for optimal operation.
On the consumer side it's a completely different story because not only does the autonomous driver needs to outperform the human driver it needs to do so in a case when you have compound errors if the driver of the car does not trust what the auto pilot is doing you'll get into more accidents not less, and in a market environment where you have a very broad range of features you can inconsistencies which lead to more accidents.
And this is where you hit the UX issue, all autonomous drivers have to present a reliable and consistent user experience to avoid compound errors that arise for the involvement of the human (jeez that sounds eerie too natural these days) driver.
When I step into a car with an auto pilot I need to know that it will get me from A to B with minimal supervision and will on it self avoid obstacles no matter what car I step into.
If car maker A is allowed to sell it's auto pilot without collision avoidance for example while car maker B has it it will create confusion which leads to accidents, if car maker C then decides to sell an auto pilot feature which is nothing more than a glorified cruise control it can lead to even more accidents.
This is why some standardization is really needed with this tech getting off the ground.
The standardization also has to go beyond sheer features it has to provide a consistent user experience across the board so things like acceleration, breaking distance, driver alerts etc. need to be more or less consistent across different cars from different manufacturers or again you'll can get an increase in accidents due to driver errors.
> if your car runs over your dog because you didn't had firmware 3.17.4b installed or because your license has expired isn't really something any user should be dealing with.
Any victim should be dealing with.
Either a car is fully autonomous and carries the same liability and associated penalties as a car driven by a human or it might as well not exist.
Doing this half-assed is worse than doing nothing in my opinion, drivers that get to point at their car and say 'it wasn't me' is not an option.
Emergency brake should not be compared with fully autonomous drive. Emergency brake is there to reduce the risk in case of driver error, given that the driver has already made an error. You should not need to know whether it's installed or not because you should not be making errors in the first place.
Let's say for sake of simplicity that the system can reduce the risk with 50%. Under normal driving conditions when the risk of hitting somebody is say 0.0001%, you instead get 0.00005%, that can't be bad. If you provoke the system by actively aiming at a human you by definition have 100% risk of collision without the system, with it you have 50% of 100%=50%, I wouldn't want to be that pedestrian, even if the system could reduce down to 1%.
Autonomous drive on the other hand, you give the car control, as opposed to the car taking it from you (because you have been a bad driver). If you give the car control then you should know the systems limitations and which responsibilities you are giving away. What is control anyway? Is cruise control a half-assed version of autonomous drive or not?
That article contains some pretty worrysome language from the Volvo rep:
"But even if it did have the feature, Larsson says the driver would have interfered with it by the way they were driving and “accelerating heavily towards the people in the video.” “The pedestrian detection would likely have been inactivated due to the driver inactivating it by intentionally and actively accelerating,” said Larsson. “Hence, the auto braking function is overrided by the driver and deactivated.”"
I interpret that as 'this feature will likely not work when you need it most, even if you've paid for it'.
If you read their description[1], it is filled with warnings and caveats. It does not sound like something you should rely on, but something which reduces chance of an accident by a few percent.
"He hasn’t replaced driving with, say, watching a movie or relaxing—instead, he’s replaced the stress of driving with something worse. He looks at the road, he looks at the wheel, he looks at his hands."
It was much the same when I bought my parents a Roomba. It was meant to do housework for them, and let them do other things. Instead, what used to take 30 minutes of pushing a vacuum cleaner now takes 50 minutes of watching a robot.
Claims that we will all read books and join video conferences from our self-driving cars are unlikely to be realized in the first generation. Of the users, not the machines.
I've seen a roomba in action, and it does a terrible job at cleaning a place. The feeling I got is that it replaces 30 minutes of cleaning with 50 minutes of watching a robot move around, and 20 minutes of cleaning yourself all the spots it missed.
You missed the 10 minutes of rough tidying so it doesn't scoop up your kid's puzzle pieces, the 5 minutes of cleaning the Roomba itself with a real vacuum and the 2 minutes avg of finding and freeing it from the place or string it's been stuck in. Our Roomba isn't running anymore since we have kids - it's literally more work than not doing so :P
There is a great Econtalk podcast about this very issue. David Mindell has studied human interaction with automated systems and found that in practically every case there is some vestigal human supervisor of these systems. Lots of really interesting detail is given in the show:
Quick gist: The author surveyed 40 systems which were chartered as fully automated which subsequently incorporated human involvement before the systems were deployed.
Humans get frustrated quickly with systems intended to be fully automated.
There's no such thing as "no human involvement" in an automated system, at most you are moving the human involvement to the development phase.
Automated systems add a lot of value when they surface important decisions to the user.
Finally, somebody is thinking about this properly.
I've been railing against the "deadly valley", automatic driving systems that are good enough that the driver stops actively paying attention, but not yet good enough to be trusted unwatched. Tesla's "autopilot" is in that category, or at least it was before they put more restrictions on it. VW is trying to figure out how to operate safely in the deadly valley.
This matters, a lot. Misunderstandings over who's doing what between cockpit automation and pilots are a major problem.[1] There are lots of modes, often too many. This will not work for cars; drivers don't get procedure training, simulator time, and check rides. The number of modes must be small, and it must be very clear to the driver what mode is engaged. VW's slightly retracted steering wheel makes this very explicit. That's clever.
Modeling the design after riding a horse on a loose or tight rein (the proper term is "collected") is interesting. But it's probably too much for a car. As a longtime rider who's had some good horses, I know what that feels like. I had a good Thoroughbred hunter teach me that. Trotting him up a switchback trail, I could either give him a loose rein, and let him work each tight turn his way, or collect him up, bend him around my leg, and control the turn myself. But I had to do one or the other; if I was unclear approaching a tight turn, he'd get through it, but not cleanly. Then he'd toss his head and snort, annoyed that I'd made him screw up. Some people might like that kind of relationship with car automation, but a lot of people won't be able to handle it.
(Military aircraft, though... DARPA is trying to figure this out for the F/A-18E Super Hornet, which is a single-seater with a serious cockpit workload problem. Sometimes the pilot needs to focus on weapons and targeting, and sometimes on flying the aircraft. Doing both at once causes at least one of those tasks to suffer.)
The "deadly valley" is very real and it won't go away. It can be sidestepped by taking human drivers out of the picture entirely.
Imagine what would be involved in doing that.
Some analysts think that trucking would be an easy foot in the door for this technology. Freeways present fewer challenges in driving. There could even be dedicated lanes for autonomous trucks. The problem is political. A lot of peoples' livelihood depends on trucking. Those people won't go quietly into the night.
Uber and Lyft want to use this technology for commercial passenger service. They've already disrupted taxis. Dropping drivers would be less of a political hurdle for them. I imagine Uber setting up driverless cars in pool routes like buses. Ride services are also the best hope for completely driver-free tech because no one is going to want to assume driving liability as a passenger.
Completely driverless is where this technology needs to go and I think that taxi/ride-service replacement is the lever. Once that works you'll see people buying cars that don't let you drive.
I think you will see pilotless planes long before completely driverless cars gain any popularity, because of the problem of human-automation handoff latency.
In planes, when the autopilot disconnects because of some problem there's usually plenty of warning (including a sound like https://www.youtube.com/watch?v=8XxEFFX586k ) and the pilot(s) have some time, seconds or more, to assess the situation and take control. Despite how fast planes fly, the rules of separation and general vastness of airspace make it much easier to smoothly transition control. It may not sound like much time, but consider that in a self-driving car, the latency required to avoid a collision could be in the range of milliseconds. By the time the driver realises (s)he has to take control, it's too late.
Let's also think about the diminishing skills that we'll experience once automation takes over the majority of our driving. Imagine taking 3-4 weeks off from driving and then being asked to take the wheel under a difficult circumstance.
You already basically have them. The pilot mostly just gets involved in emergencies at this point. I'd assume we may see something similar with large buses soon... large bus with an operator paid more to keep attention than "drive", taking passengers on longish routes (Ex: LA to SF), with local metro rail handling second leg, and automated small passenger cars and/or walking closing last gap.
If I have to do anything, it's not self-driving. It's driving assist.
A legal definition of self-driving could help reduce consumer confusion here. I feel like the goalposts were moved as companies rushed into this space.
The 2008 Knight Rider pilot had a similar UI. When considering a risky maneuver, the car displayed the intended route, and asked the driver if it was okay.
A key difference, of course, is that KITT considered its driver a partner to assist, whereas companies like Google consider the driver a flaw to be corrected.
34 comments
[ 4.5 ms ] story [ 89.0 ms ] threadI don't think this is a good enough excuse, honestly. Both the law and common sense say the same thing about operating a vehicle: the ultimate responsibility of that vehicle and what it does lies on who is sitting behind the wheel. Until we have truly automatic driverless cars (enter an address in the Maps app and let the car take you there - no pedals, no brakes, no turn signals), human beings need to still be aware that they are still driving the car, regardless of assistant technologies that help them along the way.
Any person who knowingly takes their hands off the wheel of a moving car they are piloting is going to ultimately be at fault when it plows into a pedestrian or into oncoming traffic, regardless if the computer tried to prevent it from happening.
The only User Experience problem here is that the users have No Experience with the new technologies available in these cars. If anything, that alone should ask for more caution before engaging these features, not less. The biggest outcome I'd expect in the short term is more bleeping and bonging from car speakers if the user does something completely insane like take their hands off the wheel while the car is moving above a certain speed to film themselves performing a Darwin Award attempt for YouTube.
While the law may still hold that position, reality is starting to disagree. When you have a car laden with 'assistive' technologies which are able to take control of brakes, throttle, and even steering - sometimes without human input, things start to get complicated.
We have plenty of precedent already. Toyota's billion dollar unintended acceleration criminal liability case, the 'remote control' Jeep Cherokee case, the recent FBI advisory regarding security and safety of connected vehicles.. I find it very hard to agree with a blanket statement that the driver is responsible for everything a modern car does.
I see Autonomous driving like iOS vs Android when you buy an Apple device you know what to expect (or atleast used to before they decided to fragment their SKU's to hell) each device comes with the same features and the software works exactly the same. The Android ecosystem on the other hand is fragmented to hell and each devices comes with completely different features, software and user experience.
This can easily kill autonomous driving because it unlike poor app performance self driving cars can actually kill people if the self driving experience I get on a BMW is drastically different than what i on a Honda or a Volvo I won't be confident in using any of those, not to mention that at the moment you can't even get reliable experience on the same make and model and it's a feature not a bug in most cases because car manufacturers want to monetize self driving so for every (even basic) feature you'll have to pay extra.
Self driving cars need to get a standardization and they need it fast a fixed set of features that all cars have to properly support to be able to market their self driving cars in the US and Europe and this has to be a dichotic standard you either support everything in that standard to a satisfactory level or you don't get to turn any of those features on.
The last god damn thing I want to worry about is if my car's pedestrian avoidance app was reactivated or not after the weekly firmware update, but sadly enough i foresee some one getting run over because the driver of a "smart" car did not accept the new EULA that was pushed over the weekend before the industry gets any shred of sense.
"Human killed by robot car" will definitely be lapped up and even if statistically it may be better than "Human killed by human driver" the one is the exception, the other the rule and so the one will get airtime.
Marketing this kind of thing properly is going to be extremely hard once the casualties start mounting (and they will, no matter how good the software is).
But that's still a compelling number. Just as most people thought malaria was some quaint disease that inconvenienced the British Empire until Bill Gates got on it, I think road deaths are potentially at least as compelling a target, especially once people look outside Europe and North America.
I don't think that's even remotely accurate.
In these cases the the vehicle and to some extent the environment is much more controlled and it configured for optimal operation.
On the consumer side it's a completely different story because not only does the autonomous driver needs to outperform the human driver it needs to do so in a case when you have compound errors if the driver of the car does not trust what the auto pilot is doing you'll get into more accidents not less, and in a market environment where you have a very broad range of features you can inconsistencies which lead to more accidents.
And this is where you hit the UX issue, all autonomous drivers have to present a reliable and consistent user experience to avoid compound errors that arise for the involvement of the human (jeez that sounds eerie too natural these days) driver.
When I step into a car with an auto pilot I need to know that it will get me from A to B with minimal supervision and will on it self avoid obstacles no matter what car I step into.
If car maker A is allowed to sell it's auto pilot without collision avoidance for example while car maker B has it it will create confusion which leads to accidents, if car maker C then decides to sell an auto pilot feature which is nothing more than a glorified cruise control it can lead to even more accidents. This is why some standardization is really needed with this tech getting off the ground.
The standardization also has to go beyond sheer features it has to provide a consistent user experience across the board so things like acceleration, breaking distance, driver alerts etc. need to be more or less consistent across different cars from different manufacturers or again you'll can get an increase in accidents due to driver errors.
Any victim should be dealing with.
Either a car is fully autonomous and carries the same liability and associated penalties as a car driven by a human or it might as well not exist.
Doing this half-assed is worse than doing nothing in my opinion, drivers that get to point at their car and say 'it wasn't me' is not an option.
Let's say for sake of simplicity that the system can reduce the risk with 50%. Under normal driving conditions when the risk of hitting somebody is say 0.0001%, you instead get 0.00005%, that can't be bad. If you provoke the system by actively aiming at a human you by definition have 100% risk of collision without the system, with it you have 50% of 100%=50%, I wouldn't want to be that pedestrian, even if the system could reduce down to 1%.
Autonomous drive on the other hand, you give the car control, as opposed to the car taking it from you (because you have been a bad driver). If you give the car control then you should know the systems limitations and which responsibilities you are giving away. What is control anyway? Is cruise control a half-assed version of autonomous drive or not?
"But even if it did have the feature, Larsson says the driver would have interfered with it by the way they were driving and “accelerating heavily towards the people in the video.” “The pedestrian detection would likely have been inactivated due to the driver inactivating it by intentionally and actively accelerating,” said Larsson. “Hence, the auto braking function is overrided by the driver and deactivated.”"
I interpret that as 'this feature will likely not work when you need it most, even if you've paid for it'.
[1] http://support.volvocars.com/en-CA/cars/Pages/owners-manual....
It was much the same when I bought my parents a Roomba. It was meant to do housework for them, and let them do other things. Instead, what used to take 30 minutes of pushing a vacuum cleaner now takes 50 minutes of watching a robot.
Claims that we will all read books and join video conferences from our self-driving cars are unlikely to be realized in the first generation. Of the users, not the machines.
I think the switch will be near instant once the technology is there.
http://www.econtalk.org/archives/2015/11/david_mindell_o.htm...
Humans get frustrated quickly with systems intended to be fully automated.
There's no such thing as "no human involvement" in an automated system, at most you are moving the human involvement to the development phase.
Automated systems add a lot of value when they surface important decisions to the user.
I've been railing against the "deadly valley", automatic driving systems that are good enough that the driver stops actively paying attention, but not yet good enough to be trusted unwatched. Tesla's "autopilot" is in that category, or at least it was before they put more restrictions on it. VW is trying to figure out how to operate safely in the deadly valley.
This matters, a lot. Misunderstandings over who's doing what between cockpit automation and pilots are a major problem.[1] There are lots of modes, often too many. This will not work for cars; drivers don't get procedure training, simulator time, and check rides. The number of modes must be small, and it must be very clear to the driver what mode is engaged. VW's slightly retracted steering wheel makes this very explicit. That's clever.
Modeling the design after riding a horse on a loose or tight rein (the proper term is "collected") is interesting. But it's probably too much for a car. As a longtime rider who's had some good horses, I know what that feels like. I had a good Thoroughbred hunter teach me that. Trotting him up a switchback trail, I could either give him a loose rein, and let him work each tight turn his way, or collect him up, bend him around my leg, and control the turn myself. But I had to do one or the other; if I was unclear approaching a tight turn, he'd get through it, but not cleanly. Then he'd toss his head and snort, annoyed that I'd made him screw up. Some people might like that kind of relationship with car automation, but a lot of people won't be able to handle it.
(Military aircraft, though... DARPA is trying to figure this out for the F/A-18E Super Hornet, which is a single-seater with a serious cockpit workload problem. Sometimes the pilot needs to focus on weapons and targeting, and sometimes on flying the aircraft. Doing both at once causes at least one of those tasks to suffer.)
[1] http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.571...
Imagine what would be involved in doing that.
Some analysts think that trucking would be an easy foot in the door for this technology. Freeways present fewer challenges in driving. There could even be dedicated lanes for autonomous trucks. The problem is political. A lot of peoples' livelihood depends on trucking. Those people won't go quietly into the night.
Uber and Lyft want to use this technology for commercial passenger service. They've already disrupted taxis. Dropping drivers would be less of a political hurdle for them. I imagine Uber setting up driverless cars in pool routes like buses. Ride services are also the best hope for completely driver-free tech because no one is going to want to assume driving liability as a passenger.
Completely driverless is where this technology needs to go and I think that taxi/ride-service replacement is the lever. Once that works you'll see people buying cars that don't let you drive.
In planes, when the autopilot disconnects because of some problem there's usually plenty of warning (including a sound like https://www.youtube.com/watch?v=8XxEFFX586k ) and the pilot(s) have some time, seconds or more, to assess the situation and take control. Despite how fast planes fly, the rules of separation and general vastness of airspace make it much easier to smoothly transition control. It may not sound like much time, but consider that in a self-driving car, the latency required to avoid a collision could be in the range of milliseconds. By the time the driver realises (s)he has to take control, it's too late.
Let's also think about the diminishing skills that we'll experience once automation takes over the majority of our driving. Imagine taking 3-4 weeks off from driving and then being asked to take the wheel under a difficult circumstance.
This is also called "living in New York"
A legal definition of self-driving could help reduce consumer confusion here. I feel like the goalposts were moved as companies rushed into this space.
https://www.youtube.com/watch?v=FlpPR_GgwEc
A key difference, of course, is that KITT considered its driver a partner to assist, whereas companies like Google consider the driver a flaw to be corrected.