US to Allow Cars Without Steering Wheels
spectrum.ieee.org
spectrum.ieee.org
E.g. somebody that wants to rob you could just stand in front of your car (or put some small obstacle) knowing that the car will stop (to avoid damaging itself and/or whatever is in front of it).
Even in the case of simpler tech which is generally available already today some unexpected problems might arise once people become really aware that it exists on 99% of the vehicles.
For example, in the case of an at least partially selfish person, why should s/he wait to cross the street if s/he knows that the car that is nearing has an anti-collision/auto-break system? I admit that I might even start doing that myself whenever I'm late for something, have to catch the bus, etc... :)
Might end up being a funny experience sitting in a car that has to do emergency breaking every other minute... :)
As opposed to placing some rocks on the street like they do now?
> why should s/he wait to cross the street if s/he knows that the car that is nearing has an anti-collision/auto-break system?
Why do them wait now? Current cars stop too.
If I see rocks I can e.g. drive around them, judge if they're really too big for my car, drive on the sidewalk, etc.. => all these options are really complicated to program.
Current cars don't stop, at least not in Europe - they might break/go slower but it's going to be mixed with steering to avoid whoever is there causing all that fuzz. And if I think that whoever is in front of me is a threat that I might just run over him/her (or "it" for "animal" - a different kind of complexity - e.g. do you stop for 1 frog? 10 frogs? 100 frogs?) , at least partially.
It should be able to program a self-driving car with a) a database of known intersections and their conditions as a fallback and b) to evaluate the conditions of an upcoming intersection for safety even without the requisite signage.
Maybe it'd run it if there's no one else around, but a human probably would too in that case.
Another imperfect analogy is power outages and stop lights. For the most part, it doesn't result in full-speed collisions; instead, folks treat it as a stop sign.
Self-driving cars are going to need a "huh, that doesn't look right..." system. Detecting missing/defaced/misleading signs should be a part of it.
That's what you're supposed to do. In some cases, the traffic lights fail to a mode where one direction is blinking red, which is treated like a stop sign, and the other is blinking yellow, which is treated sort of like a yield sign (i.e., proceed, but with caution).
I believe the law (and common sense) states that if the traffic light it out, you treat it like a 4-way stop sign intersection.
They'd also have to disable any manual override controls.
In the least dystopian-surveillance-state future scenarios, there will be a giant array of cameras and sensors pointed at anyone who crosses this way. How hard would it be to identify this person and ramp up enforcement as necessary? You'll never completely stop it, but I imagine it could be lowered to acceptable levels.
I would highly recommend visiting Cairo in Egypt and crossing a busy road: With a swarm of cars going left and right, you simply 'pick a car' by looking directly at it and step out into the stream of (multi-lane) traffic. It's freaky at first, but it slows, moves and bends around you as you walk across the road. I've seen mothers with bags and children doing this at even busier places.
It seems that here in the west, when we're driving, we have a 'get out of my way' mentality that seems out of proportion with how we walk down a pavement. ie, As we walk, we move and flow around slower/oncoming people -- we certainly don't shout at fellow pedestrians to get out of the way and then barge them out of the way when they don't.
Edit: A video of cairo street crossing https://youtu.be/qYjiO_kLYKY
Traffic might come to a stop in a big city (as they are today)? Or will "good habit" prevail? Or, will it be a grey zone?
I like to joke that it's why people wear those big conical hats in Vietnam -- you can't see the cars, so it's their job to avoid you :-).
https://en.wikipedia.org/wiki/List_of_countries_by_traffic-r...
If they get advanced enough, they may pull a Terminator 2: slow down enough to hurt you, but not kill. This fulfills their directive not to kill while disincentivizing jaywalking.
The first link in the article's body points to "https://www.nhtsa.gov/DOCUMENT/ROAD-AHEAD" which redirects to "https://www.nhtsa.gov/sites/nhtsa.dot.gov/files/documents/12... => I did search for the strings "level 4" and level4" but did not get any hits => which page(s) is/are used as foundation for the article's subtitle mentioning "level 4"?
EDIT: None of the links that I posted work directly (they're accessible only through https://spectrum.ieee.org/cars-that-think/transportation/sel... in the area at the top which states "blah blah... said today in an 80-page report") - don't know why the direct links aren't working... .
https://www.wired.com/2017/01/bad-audis-swanky-new-headlight...
Additionally, accepting a request from a specific country might trigger a chain-reaction from other countries, multiplying what stated above by the nbr of countries (and I personally don't think that it would be a linear curve as at the same time the costs for coordination between the multiple lines would rise).
Because of the regulations, the burden is on the manufacturer to show the matrix light technology is actually safer than the traditional one.
I will never understand this. In the case of people who sometimes use signals but not always, it must take more mental energy to evaluate the situation and decide than reflexively signalling every time you turn or change lanes. In the case of people who never signal, it must be much harder to navigate a city when nobody knows what you're trying to do.
>The Model State Policy confirms that States retain their traditional responsibilities for vehicle licensing and registration, traffic laws and enforcement, and motor vehicle insurance and liability regimes while outlining the Federal role for HAVs.
Surely, liability for Level 4 autonomy vehicles would be borne by the car's manufacturer, rather than the operator. Is that maybe one of the more controversial aspects separating the definitions of Level 4 and 5 autonomy? I'm surprised this would remain unregulated at the federal level.
[1] https://www.nhtsa.gov/sites/nhtsa.dot.gov/files/documents/12...
I certainly understand the owner/operator having some form of liability insurance based on usage, but an at-fault collision shouldn't be the responsibility of the owner or operator, nor should it affect their rates. That fault lies entirely with the manufacturer, or maybe eventually the subsystem supplier.
Thinking about it, I'd not be terribly surprised if it plays out with manufacturers successfully lobbying to avoid any liability, and insurance gets pooled among owners of specific Level 4 autonomy subsystems. Maybe the best we can hope for would be an equitable split of liability between the manufacturer (who controls the behavior) and owner (who controls the level of usage).
What about a rental vehicle?
What about a vehicle you borrowed from someone?
I really don't see why the manufacturers would not be the ones to be liable, and totally agree that none of the people in the vehicles should be liable, even better: they should be able to sue the manufacturer too since they were made part of the accident.
Probably a moot point considering that the operator, owner and manufacturer will be the same company.
This is one of my top reasons why I will not willingly own a self driving vehicle.
Manufacturers can make unsafe cars right now. Regulators and consumers seem pretty good about stopping most of that from happening.
of what? The entertainment? If having two screens reduces the safety of the vehicle, it isn't a vehicle I want to be in.
I don't want a large void space inside the passenger area of my car at all.
There should be a standardized control that pops out of the dash using a fixed icon, so that anyone can enable an override.
Nuts.
> Basil: Well I'm sorry it wasn't wide enough for you. A lot of the English cars have steering wheels.
Oh yea this is great idea. Lets all just not own cars. I can't wait to wait 10 min for a automated taxi to arrive, where the interior is left gross and nasty by the previous riders who were doing God know what inside. Also if you ever need to leave in a hurry (get a call that a loved one is the hospital, etc.) and need to leave right the fuck now, too bad. Also when your shitty taxi arrives it won't go any faster, it is programmed to never break the speed limit.
Sounds like fucking HELL. NO WAY, NEVER, EVER, EVER. I will fight to the death to keep manually driven cars.
[1]https://www.supermotors.net/getfile/801863/fullsize/7824_103...
Seriously though, why was this ever a law in the first place?
In that case, likely a lot of people would want to not have to invest in the hefty price of owning, maintaining, and insuring a vehicle, let alone driving which everyone knows is more dangerous than most things a person does in a given day.
This is simply untrue if you're considering average commute distance of drivers in major cities.
> hefty price of owning, maintaining, and insuring a vehicle
These costs are highly variable, and many reasonable options are quite cheap.
I don't see self-driving cars catching on any time soon in the US. Possibly in Europe. I think they will be limited to smaller vehicles that operate in glorified bike lines, only on known routes.
I'm sure that'd be correct if you'd said suburbs and rural areas, but according to this[0], the average commute for almost every major city is under 10 miles.
[0] https://www.marketwatch.com/story/here-are-the-typical-commu...
All of the costs are worth it, as long as I don't have to share transportation with anyone else.
Or, god forbid, public transport with an actual bus driver...
I don't see this happening any time soon. There might be a contraction in highly speculative investments (i.e., VC cash) aimed at autonomous driving systems. However, even if tech never gets beyond level 2, this will still be a high-growth and high-wage field for the next couple of decades as automobile manufacturers and their tier one suppliers incorporate existing ADAS across the lineup.
And that's to say nothing of the many limited domains where level 4 is definitely doable, including closed-environment mining and manufacturing sites. As well as adjacent industries (e.g., maritime) where even levels 1 and 2 could help a lot.
Self-driving is a scorching hot fireball right now. It might cool down as people realize that level 4-5 is not happening any time soon, but a prolonged winter is hard to believe.
The definition of level 5 being "under all roadway and environmental conditions that can be managed by a human driver" means that whatever mechanism being used would have a better "brain" (at least in the area of "driving") than the human one, correct?
If yes: kind of scary, as such capability would probably have to include the comprehension/understanding of e.g. "context", the ability to "abstract", be able to generalize from indirecly related informations, etc.. to understand potential upcoming dangers.
For example when driving behind a truck which has its trailer that wobbles continuously left and right (but which can still keep its lane) because of a partially flat tire would probably need a very advanced AI to recognize the potential danger.
Unfortunately, the politics around autonomous cars insists that human beings are basically idiots behind the wheel, so the problem seems more trivial than it is. After all, how hard can it be to design an AI smarter than an idiot?
Or, say, driving in Mumbai or Beijing.
Yeah, I have no doubt about special purpose level 4 in limited environments. My comment was about seeing general purpose 4 and 5.
A quick Google search for average car age hints that it's nearly 12 years [0]. Even if everyone buying a new car bought an EV, we're probably 12 years away from 50% of the cars on the road being EVs. And since the age trend is actually increasing, we're probably looking at 13-15 years in actuality. And again, that assumes 100% of new vehicles purchased are EV, which definitely isn't the case.
This doesn't even begin to look at self-driving. We're still years away from reliable level 4 automation.
[0] https://www.cnbc.com/2017/05/28/car-owners-are-holding-their...
In many places, there are no lines drawn on the road and you have to imagine the lanes. It could be a space big enough for 3 or 4 lanes. It makes me wonder if cars would recognize the general direction of the road and not drive diagonally through the imaginary lanes when there's a turn.
Some lanes may also have horrible holes that seem like they could easily take your wheel off if you fall on them with speed. Sometimes they're sinkholes, other times they're part of a construction job that was left midway for months seemingly until someone has a horrible accident. Some lanes look like they're at risk of becoming large sinkholes, and you'd rather avoid them lest the whole car suddenly falls meters below the ground. For both of these, you know they're there, and you know they're basically unavoidable once they become visible. How would you communicate these risks to the car?
Jaywalkers on high-speed, high-traffic highways might be common due to lack of bridges or any other alternative to crossing the road. They coordinate their movements with the incoming traffic and the drivers also coordinate their movements with them. Behaving unexpectedly, like simply changing lanes, even at a distance, could be fatal because it changes the shape of the incoming traffic the jaywalkers depend on once they decided to start crossing. Emergency stopping might worsen the situation when tailgaters are common.
Pedestrians might also be suicidal and you might be able to discern their intent from a distance by watching their behavior, but the car is not going to interpret that and it'll get close enough for the pedestrian to throw themselves at the road when it can't avoid them.
Advancing when a streetlight becomes green might generally be unsafe in zones where it's common for cars to cross at high speeds when the light is about to or just turned red. Would cars be on the lookout for high speed traffic coming from the left or right at a distance?
Simply put, there's lots to be on the look out for, especially when pedestrian city infrastructure, vehicle city infrastructure, traffic law enforcement, driving education, etc. is lacking. If self-driving cars become the norm in developed countries, I think it'd be in great part because they're not lacking in any of these things, and the car software can deal with a somewhat consistent environment. That wouldn't be the case elsewhere, though.
As to what my opinion is on the issue at hand, as a software developer, I wouldn't put so much trust in software as to not have a way to take manual control when this software has the ability to bring physical harm or death to me and others. Even disregarding the possibility of malice through malware and assuming whatever code that executes was written with the best intentions, there's just too many ways for things to go wrong in this problem domain and the consequences can be deadly.
Good design lies in simplicity. I cannot imagine the behemoth of code that must be required to implement safe automated driving. That's a lot to put faith in when lives are on the line.
That could happen, but if it did, it would probably require a generational shift in public perspective.
This would definitely be the minimum requirement, but it would be consistent with history (cigarettes, firearms, climate change, etc.) for people to deny any safety advantages or simply not be persuaded by them long after research suggests safety benefits.
The lawmakers would have to ban motorcycles first.
What was that old saying again... "Those who give up freedom for security deserve neither."
If that's the case then self-driving cars will never take off. What are they only aware of other self-driving cars? How about children? Or a tree that fell in the road? They would have to be aware of any object.
You sound like someone who would have said you would never ride in a carriage that didn't have a horse.
The user isn't against the horse-less carriage out of fear, he's against the horse-less carriage because the option to control the horse has also been removed.
"jstaney you are under arrest - remain calm your vehicle will deliver you to the nearest processing facility. Do not attempt to control the vehicle. Do not attempt to exit the vehicle. Any action deemed as attempting to exit the vehicle or prevent it from arriving at the detention facility will result in your immediate admission of guilt and the appropriate sentencing will commence. Have a nice day, thank you for using GovCloud Services."
Similarly, we could easily have mobile phones that do not track us. It's just in none of the manufacturers' interest to design them, as the vanishing epsilon of people who care enough to pay extra do not offset the economies of scale from going with the current surveillance-based designs.
Having said that, one of the themes in discussion of mass surveillance is the quantitative difference of it being easier than a police department having to expend the effort of physically following someone around. Government (de jure or de facto) controlled cars will make the act of arresting someone similarly effortless.
Historically, it seems not to matter. The advancements in technology let societies that adopt them support a much higher population level. Those societies then go and kill off or enslave everyone who didn't adopt the new technologies, once a few generations have passed, their military supremacy is obvious, and there's enough cultural distance to view the adherents of the old technology as savage, subhuman primitives.
...which they stole from Jaguar.
Oddly enough, Mercedes has gone to column shifters in the last few years, which have previously been associated with American trucks.
Have you ever been at the Las Vegas airport at the start of CES? I’ve waited an hour in line for a taxi. Of course they should have built the metro to the airport.
If you commute by train, this type of car would be perfect to get you to and from the station, especially if it could drive itself back home during the day.
If the car runs down a pedestrian in a crosswalk while said pedestrian has the right of way, on whom, exactly, does the civil liability fall? The manufacturer, or me, the occupant. Even more importantly, on whom does the criminal liability fall?
If the answers to those questions are not spelled out, explicitly, in the regulations congress puts out, then I agree, it would be hard for me to get into one of those things.
Especially if they said you were at fault because you told the car, "Take me home", or something. I can hear it now, "Your honor, the occupant was obviously exercising control over the vehicle!"
It may fail in practice, but the claim is all that should matter in court.
Yet, I'm sure somebody somewhere is going to be thrown into the lions for protecting some rich company. So I'm certainly not going to be one of the first to buy one.
I don't think so (never heard of) but I might be wrong.
I can imagine that they would be very complex to create.
Later in the pub he told someone who worked at BAE, and was fairly high up, and the guy said, "yes, I heard about that, you need to contact them."
He did, and he got the full "men in black" treatment with several interviews/interrogations to determine exactly who he was and what his intentions were, and threatened with jail time. Eventually they believed his story and he even got his drone back, although the camera was obviously missing.
Apparently it had landed next to something "extremely sensitive", which may or may not have been the cause of it drifting off in the first place.
[1] https://www.wired.com/2014/06/infiniti-q50-steer-by-wire/
Observing how these things are usually handled I would say the liability would fall on no one. Unless the driver is intoxicated car crashes, even those resulting in deaths, are treated as unavoidable accidents with charges rarely being filed.
When a cab maims a pedestrian in NYC the typical response is the NYPD releasing a list of pedestrian safety tips.
The downside to this is that these vehicles would likely make even the most cautious myopic octogenarian look like speed racer.
However, if the manufacturers are smart, they'll make the self-driving hardware record video footage they can use to defend themselves in court. You wouldn't want people to be able to jump in front of a moving car to try to get a payday after they get hit. This already happens and is why many people have dash cams. See https://www.youtube.com/watch?v=ibjWz9ehrJ8 for some good example.
What bothers me most about this tech is we have a pretty long history of auto manufacturers knowing about problems involving safety issues that they evaluate with a cost/benefit methodology as opposed to a safety first and always approach.
The Ford Pinto [1] is a good example of that.
I still trust me more than an autonomous machine but I do look forward to a day when I don't.
An elevator can go up or down and has inertia brakes among other safety mechanisms. Most of these mechanisms are regulated and inspected and can not be disabled by software. I have watched videos of elevators behaving badly, but it takes a comedy of errors and many inept people for this to be a thing.
An autonomous car on the other hand, can do anything a car can do and anything that the buggy or compromised software tells it to do. I will not be victim to the first angst filled 12 year old that is abused by his/her parents daily, or by some burnt out developer, or a serial number typo in an assassination attempt. The serial number of my car could be one digit off from some politician or spy.
By all means, have fun in these things. Nobody will convince me to use one of these cars. Not now, not five hundred years from now.
Look at windows in cars today and being able to see. It drives car designs.
But with SDC it is not longer needed to make seeing how necessary and therefore would allow cars to be made completely differently.
That is what I am most interested to see. Today Waymo and others are just using "regular" cars.
Waymo has purchased up to 82k cars which are Fiat Mini Vans and Jaguar i-Pace.
Would hope Waymo at some point buys the capability to make their own cars and start designing cars optimized for SDC.
Waymo parent now has over $100B in cash with less than $5B debt.
Ten years ago Jaguar and Land Rover combined sold to Tata forr $2.8B. So would not be any problem for Waymo parent to buy equivalent.
Or even outsource a lot of it like the i-Pace was outsourced to Magna.