> According to the Chronicle, the preliminary investigation found the Uber car was driving at 38 mph in a 35 mph zone and did not attempt to brake. Herzberg is said to have abruptly walked from a center median into a lane with traffic. Police believe she may have been homeless.
I know this isn't the most pertinent issue here, but I'm surprised the self-driving AI is allowed to speed, though with further research, this seems to be the case with Google/Waymo software too:
Though in that BBC article (which is from 2014), the stated scenario is when the self-driving car is surrounded by traffic exceeding the speed limit. It doesn't sound like this was the case for the Uber vehicle.
Of course a self driving car should be able to speed. My understanding is that speed limits are set by measuring the speed people are driving on the road and then determining a speed at which roughly 15% of the drivers are exceeding that limit. There is an expectation that people will speed, but by setting the speed limit a little bit lower almost all traffic will fall under safe bounds. Driving 38 in a 35 zone is completely reasonable. I probably would have been going 45 - 50 if the road was empty. If self driving cars aren't allowed to speed, they're going to be a nightmare for passengers and other drivers. Obviously different people have different feelings about the issue of speeding, but personally I lose my mind when I get an Uber driver that's driving to slow. Try and think about what it would be like if a computer followed every traffic restriction to the letter of the law rather than driving like an actual human.
People will vote me down because the idea of a speeding self driving car sounds scary, but I personally can't wait until the software has advanced to the state that it can confidently drive 200 mph on the 5 from LA to the Bay Area. (Presumably they'd have to build a much bigger better 5, with special lanes dedicated to autonomous vehicles)
> My understanding is that speed limits are set by measuring the speed people are driving on the road and then determining a speed at which roughly 15% of the drivers are exceeding that limit.
Talking about a future in which AVs are allowed to go as fast as physically possible seems irrelevant when we're talking about current policy and infrastructure, in which AVs have to live in a world with majority manually-operated vehicles, driven by people who do speed, but are also punished for going over the speed limit.
An autonomous vehicle should also be pulled over and ticketed for operating unsafely. At this point in time the driver is still liable for the actions of the car and is required to monitor the autonomous system and takeover if necessary. However like a human it should have some discretion in interpreting the law. Should it go 120 in a 60 zone... definitely not but if it thinks it can do 38 in a 35? That sounds fine.
>> People will vote me down because the idea of a speeding self driving car sounds scary
Or because they don't think that speed limits are set so that everyone can keep under them- but so that keeping under them means you're driving safely.
It's one thing to state that it's common and simple to, say, overclock a CPU, even if the processor specs have a speed a few percent lower. Typical headroom, operator requirements, and a decent cooler suggest that everyone could probably bump clock speeds by a few percent without issue.
But you wouldn't expect a plane crash investigation to mention offhand that the flight computers were mildly overclocked.
A test drive of a new self-driving car should not speed, even if it's common or normal.
Going at speed limit NEVER creates a hazard. It is always the drivers going over or under it. I am not getting a speeding ticket for peer pressure. If you want to speed then you can pass me at your own risk.
Going the speed limit OFTEN creates a hazard. Technically speeding is not "exceeding the speed limit" but "driving too fast for conditions." In a torrential downpour with slick roads and poor visibility, the legal speed limit may easily be too fast for conditions, and you can be ticketed--to pick one scenario that refutes your "NEVER." And driving the speed limit in the passing lane creates an obstacle to faster traffic trying to get by, a frustration and a hazard. There was a video of some idiot "pranksters" with your exact attitude that created a rolling roadblock on a freeway by coordinating five cars to drive abreast at the exact speed limit, proving ... something. They were ticketed for hazardous driving. cthalupa is correct, though more diplomatic than I. It's more hazardous to be out of step with the surrounding traffic flow than it is to "speed."
>Going at speed limit NEVER creates a hazard. It is always the drivers going over or under it.
Please be in the right lane if you are going slower than the flow of traffic. Regardless of whether or not you feel the people speeding are ultimately at fault, if everyone is speeding, it is a more dangerous to have a car that is going a dissimilar speed than it is for you to be speeding.
What speed would you drive? Would you take close corners at 120km/h? People (on average) drive at the speed their actually very well tuned survival instinct tells them to.
I disagree that the average driver has well-tuned instincts. Accidents are rare enough events that you can drive poorly for a long time and still not end up in a seriously bad situation, especially in areas where inclement conditions are also rare.
The first time you lose control of your car might be in icy conditions, and might be not just losing control, but spinning out of control in traffic. The line between in control and not in control can be very thin, and so hard to know how close you are to it.
I mean self-preservation instincts, not driving instincts. I agree totally with the later!
By self-preservations instincts what I mean is that most people will try to be as safe as possible on the road, meaning drive slower on low-visibility situations, when there are many pedestrians around, etc. You cannot account on unforeseen circumstances or lack of knowledge, but otherwise people will try very hard not to get into harm's way.
Without a speed limit, I'd take i-70 from KC to STL at about (160~mph) 250km/h for 95% of it. I'd make day trips just to get white castle at white knuckle speeds.
I-70 is a mostly straight highway with a divider, it'd be pretty safe as long as it's not crowded. Autobahn speed limits are certainly doable in many of America's highways.
I don’t think that’s true at all (though I only have my own experience to go off of). I tend to use the speed limit as a point of reference, and then drive a little above it. Prior experience has told me that going the speed limit plus some percentage ‘feels’ safe. If there were no speed limit I’d actually have to pay more attention to the road and surrounding environment to determine what a safe speed is. Right now most people’s ‘survival instinct’ is basically speed limit + some delta in clear conditions. That delta may change from driver to driver, but in general everyone drives within 10-20% of the speed limit.
As a fun thought experiment, imagine what would happen if your residential speed limit were raised from 25mph to 70mph. People may not hit the 70mph limit, but I can almost guarantee that the average speed would increase significantly, perhaps even dangerously in some cases. I don’t have much faith in people’s ability to judge space and speed accurately, particularly when they aren’t used to doing so. And this includes myself.
Ah but that is a whole different issue, the problem is that for the last few decades residential areas have been designed more and more like high-speed areas, and that had a larger impact on driver speed than the actual posted sign. I cannot seem to find the original articles but here are a couple of related ones:
This is well-studied. It turns out that the 85th percentile speed (the speed where 85% of drivers are below that speed, 15% above) is nearly insensitive to speed limit. Road design is the biggest single determinant of travel speed. http://www.lsp.org/pdf/troopc85thSpeed.pdf
> Won’t raising the speed limit cause people to
drive faster and cause more crashes?
> The Federal Highway Administration studied nearly 200
roads in 22 states where speed limits were raised, lowered
or left unchanged. Prior to the speed limit change, 55
percent of drivers exceeded the posted speed limits. After
speed limits were raised or lowered as much as 20 mph,
there was a slight change in speed, but generally less than 1
mph. There were no significant changes in crashes,
although crashes tended to decrease where speed limits
were increased to realistic levels. Also, there was little
effect on speeds or crashes on intersecting or nearby
roadways.
This is for federal highways, traffic in a city should not go faster than 25mph in most shared spaces. Motorways is a different subject, but there should only be overpasses or resegined roads and traffic lights for such places.
I would like the speed limit on my interstates and highways to go up to 70(they're all straight. 55 is just the only speed limit around here). Driving at 65-70 feels safe in normal conditions. Driving above 70 feels a bit too fast.
Speed limits should be more variable depending upon car size and braking power though. I hate having to fear that I might be pulled over for going a bit above the speed limit when there are so many more threats than just speed. I don't believe speed limits should be anything more than a suggestion with dangerous driving being the focus.
I find this is extremely dependent on car. I've had cars that felt like you were about to die at 55 mph, and I've had cars that you couldn't tell you were going fast until you were really moving. For laughs, as my wife and I were driving home one afternoon earlier this year on a completely empty stretch of I-5 south of Portland, I casually accelerated up to about 110 and left it there for maybe a minute, then eased back down to normal speed as we came up into traffic. She never even looked up, didn't notice. But then again, we were in my brand new Camaro SS and I'm not positive it even got over 2000 RPM for that exercise, so I guess that's not a surprise.
So I guess I'm saying I agree. Speed limits are pretty arbitrary, there are a lot of things drivers do every day that are far more dangerous than merely speeding. Closure rate, unsafe passes, tailgating, cutting people off, etc.
/don't get me started about how people drive around you if you're towing a travel trailer, what the heck is it with people?
On most roads around where I live, if you are driving exactly the speed limit, you are a slow moving obstacle that is probably more of a danger due to the extreme speed differential with the other drivers who are 10-20 MPH over the limit. I’d hope self driving cars were programmed to deal safely within the bounds of reality. If there is nobody else on the road, sure drive the speed limit. If cars are whizzing past you on all sides you’re a hazard even if you’re obeying the law.
Depends on the conditions. With moderate temperature and dry conditions every single corner on a highway can be taken at 120km/h on any modern car, the overwhelming majority of them can be taken far in excess of 200km/h depending on the car's characteristics.
Then again under bad conditions on a 90km/h roads my survival instinct like you mentioned hasn't allowed me to go past 20-30km/h sometimes.
Prevailing traffic plus 10-15 mph. Or 135-ish with good conditions and no other cars. My days of real speed runs are behind me, and 135 is fast enough to keep things interesting and make good time without getting too crazy.
I would say they should be able to drive up to 15% higher than the speed limit at their discretion. So in a 35 zone they could go up to 40. But that's not a question for me to answer.I just know that the answer isn't that they follow the letter of the law all the time.
It may not be programmed to speed per se, but if you’ve ever had cruise control turned on you’ll know it doesn’t keep you exactly at 55mpg. It might be 54 with the throttle applied to speed up, or it might be 57 which doesn’t trigger the brakes yet but is coasting to a slower speed.
Imagine the car in front of you cycling rapidly between the gas and brake pedal to keep the speed exactly at 55mph. One, that’s completely impossible with how cars are designed, and two that’s really illogical. Cars have to follow the laws of physics too.
There is no brake involved with cruise control. What you are describing is bad drivers who do not use cruise control. It's actually extremely pleasant and predictable to be following a driver who is using cruise control. Following a well-regulated self-driving car would be similar.
My car has cruise control with brake functionality. Not sure if I've ever seen it actually do it - I only really use cruise control on the motorway, where I expect wind resistance is a major factor - but apparently it will apply the brakes to maintain the set speed when necessary.
At town speeds I would expect it to use the brakes a lot more readily. At 30mph in 3rd (2000rpm) or 4th (1500rpm) there's not really all that much engine braking, and wind resistance won't slow it much either.
(Mine is a 2010 model, but I think this stuff was introduced in 2004-6, something like that... it's not exactly new technology.)
My car has a CVT (Continuously Variable Transmission). With cruise control enabled, going downhill, it will actually change the effective gear ratio and "downshift" automatically.
(Once or twice when going downhill without cruise control, just using the brake pedal, it "downshifted" automatically while I had the brake pedal pressed.
CVTs are computer controled and better than manual/standard for gas mileage, but with the downaide that you can't tow heavy loads.)
I've never heard of that phrase before, so I went to look it up. Apparently it has different meanings to different people when it comes to CVTs. I haven't experienced any of those different issues.
When it comes to accelerating and high RPM, I just don't mash the gas pedal down. By pressing the gas pedal on my car just enough, I can accelerate past other cars without high RPM just fine.
Gas pedal management is definately different with CVTs and other types of transmissions if you want a good driving experience.
My car absolutely uses the break when required to maintain the set cruise control speed, I've watched it happen. The car I owned before my current car didn't use the break to keep to the set cruise control speed though, so I get why you may think there is no break involved. It largely seems to depend on make and cost of the car you're driving.
Our Passat uses brakes for cruise control, our older Kia Ceed did not. I actually hate when it uses the brake. The non-brake method is much better imo. Feels like such a waste to heat the brakes on a steep incline instead of just rolling with it.
Yeah true. Seems like the more feature complete assistance systems are making their way down to even some of the more low-end brands these days. Level 1 autonomy is getting pretty old hat now I guess.
All SDVs that I know of operate some kind of longitudinal controller that has access to throttle and brakes at a minimum. That and steering wheel angle (or torque, depending) are the two essential control primitives. All the motion planning stuff happens on top of this.
I don't think that's true at all. There was an article posted about the subject a while ago, but the thrust of it was that researchers think they need to do more to make cars behave like (fully attentive) human drivers rather than being weird anomalies that behave in surprising/irritating (to humans) ways if they're going to actually go on the road, and I think that's completely right.
"(fully attentive) human drivers" probably cause deadly accidents every day, and they certainly behave in surprising/irritating ways. Anyways, a human driver going the speed limit is no hazard, so I fail to see how having self-driving cars follow the limits could be a problem.
Why not? The range between 1-10 MPH over the speed limit is statistically the safest. (Yes, I know, good luck finding a citation for a study I read 20 years ago...) It's hard to argue against the notion that you're safest when you keep up with prevailing traffic. That means that self-driving cars shouldn't be programmed to go slower.
Speed limits are usually too low. Often they have nothing to do with engineering or safety considerations, but are primarily revenue-oriented. If the advent of autonomous vehicles forces regulators to listen to the engineers for a change, that will be an unalloyed Good Thing.
You have to remember that many people on here probably don't own a car, or necessarily even hold a driver's license. They most likely either live at home or at school, or take a bus to work (NTTAWWT).
As far as I understand your comment, driving at the speed limit is unsafe because everybody is speeding. Sounds like a self-fulfilling thing. Maybe it would be even safer if the majority would obey the speed limit?
Did you read the link I posted? Theres lots of empirical data out there that says that drivers drive the same speed on a road, regardless of the posted limit. That speed is based on their judgment of the road - sight lines, shoulder width, sharpness of curves, etc.
I remember early reports about self-driving experiments saying that some of the issues where because the cars were driving at the speed limit, which is usually not the case in the US with the human drivers.
If they just track human driver habits then we're no more safer! What's the point left about self-driving cars then? Free up drivers of even more responsibility when they kill people?
> What's the point left about self-driving cars then?
The point of a self-driving car is unburdening the user from the task of driving it, allowing them to engage in other activities if they chose to do so.
That it has the potential of being safer is a nice perk but definitely not the primary goal. Right now the challenge is whether we can make them as safe on average as human drivers do, that way we can avoid effectively sacrificing people for the convenience such a system would provide.
> Free up drivers of even more responsibility when they kill people?
That's a strange way to put it. Surely if someone is not driving their car, they can't be responsible if the car kills people (assuming of course the car was well-maintained and the model was legal). I don't see how that's a bad thing, since they have zero control over the situation.
Do you mean that it could make car users feel less responsible for deaths, even though cars will presumably still kill people?
Responsibility isn't being absolved, it's being shifted to the one that actually operates the vehicle - probably either the car vendor (who either manufactured or licensed the self-driving car software), or the fleet operator depending on what kind of sales and operations model self-driving cars adopt. When human escalator operators were replaced with automatic escalators, was there are problem with shifting responsibility over to Otis and other escalator manufacturers?
As with automatic escalators, there are two primary benefits to society: greater aggregate safety[1], and freeing up people's time. You point to the lack of a human operator to hold responsible as a detriment, but the statistics point the other way around: human drivers are horribly irresponsible and even if we can punish them when they do wrong that doesn't substantially change their behavior. Freeing up people's time is self-explanatory.
So I think you have a great analogy, and I generally like your perspective. The problem is that I think this accident clearly demonstrates that the suggestion that AV's would be safer than human drivers is now suspect. A post elsewhere compared the deaths per mile for human drivers to that of this pilots' and it's worse by orders of magnitude.[0]
That's my point. I 100% agree, if self-driving cars did make things better, than the safety aspect is a good thing.
I think the freeing up people's time is okay too but only if all of society has access to AV's. However, like many innovations of the past, advancements which are not accessible to all people just free up the time of some people while offsetting burdens onto others (ie., those who can't afford AV's). BUT, that is a separate issue.
[0] It is just one data point, but it's a bad sign to be so quick to kill a pedestrian so soon out of the gate.
Worse by an order of magnitude is only the case if you exclusively measure fatalities - and as you point out there is exactly one data point and not enough miles logged to produce any meaningful conclusion. The data on non-fatal accidents for Waymo indicates that self-driving cars are an order of magnitude less likely to be at fault for an accident that human drivers: https://www.huffingtonpost.com/entry/how-safe-are-self-drivi...
As another driver on the road, I'd far rather that autonomous vehicles maintain speed with the flow of traffic around them, in lieu of strictly sticking to the posted limit.
There is no traffic "around" them, just the car behind. And if the car behind is speeding it's up to them to slow down if the car in front is obeying the speed limit.
This kind of argument is just used to justify speeding.
You don't seem to drive. If you don't follow traffic flow, people will cut off in front of you, tailgate, zigzag, and do other dangerous lane changes/maneuvers much more frequently. And that's when the trigger-happy ones don't yell obscenities at you.
LOL. I'm not sure if you're trolling, rude or just don't pay attention to the cars around you when you're driving, but I'll guess it's a combination of all 3.
>If you don't follow traffic flow, people will cut off in front of you, tailgate, zigzag, and do other dangerous lane changes/maneuvers much more frequently.
From my 45 years of observing people's driving and driving myself for about 30 years, I find that most people generally obey the speed limit or don't go much faster. Probably about 90% of people in general. At rush hour, it drops to about 50%. Some of the other 50% certainly do do all those fucked up, rude and dangerous things you mention, but it doesn't mean that the 50% of people obeying the speed limit are at fault. It's the impatient speeders who are at fault.
It's sometimes a bit of a Darwin award scenario. I live on Vancouver Island, and there is only a single highway over the mountains to Victoria. During commute times, about 50% of people speed significantly, and they do some of the dangerous and rude shit you mention. A few times a year one of them drives too fast for the conditions and ends up killing themselves, which then shuts down the entire north-south highway (the only fucking route) for 4 hours. The government is currently spending a shit ton of money to put barriers in the centre of the highway to try and prevent the excessively speeding idiots from killing themselves so often.
So no, I don't have much patience for the rude people who tailgate me, go 50km/h over the limit even in rain, and then get themselves killed and hold up the entire island for 4 hours because of their stupidity and lack of knowledge of basic physics.
(And I'm not calling out everyone who speeds, I do it myself sometimes, it's only the ones who go way over the limit with no consideration for weather conditions or safety, and the rude idiots who are in too much of a hurry to extend basic courtesy to other road users).
> I find that most people generally obey the speed limit or don't go much faster.
Regional culture definitely plays into this. I grew up in the midwest and almost everybody speeds. Now I live in Seattle and I'm the fast guy in the left lane going 5 MPH over the speed limit.
You are ranting about reckless speeding, which I agree 100% with you that it's a stupid thing to do. But the comment you replied to was talking about following traffic flow, which is about minimizing speed differentials between vehicles so as to minimize triggering fits of impatience in drivers around you.
[1] It's not my problem if others get irate because I'm at the speed limit.
[2] It's not dangerous to be at the speed limit, it's dangerous to be over it. And you're not likely to hit someone going at the speed limit if you're dangerously speeding, you're likely to hit the median or an oncoming car, and more likely to get killed doing so (much greater speed differential than between the speeder and the person going at the speed limit).
[3] There is only a differential if you are passing, and that is easily managed by waiting for a sufficient gap in traffic.
It's not like in kindergarten where kids point at each other yelling it was the other who started it. If someone does something dangerous around me, it absolutely concerns my family's safety, regardless of what triggered the dangerous situation. Thinking two steps ahead to prevent these situations from occurring in the first place is at the core of defensive driving philosophy and it's why it's often recommended that one pick going with traffic flow instead of going at the posted speed limit if forced to choose given traffic conditions.
I agree that in many roads, going above the posted limit is a very accurate measure of recklessness, but that isn't the case everywhere and it isn't at all uncommon that the flow of traffic is 10 or even 20 km/h over the posted speed limit on many freeways or countryside roads.
I disagree that going at the posted speed limit when someone behind me is going faster is dangerous, or that the person going at the speed limit is at fault. If anyone in this scenario is creating danger it is the person speeding. Generally it is not "everyone" who is speeding, only a percentage. If everyone is going over the limit and it creates a dangerous situation if someone drives at the limit, then there is some serious problem with that road and it needs fixed urgently (either by increasing the limit, or by enforcing the existing limit). It's unhelpful and silly to suggest that people must break the law to be safe, and it completely defeats the point of speed limits and the rule of law, not to mention common sense.
Defensive driving means looking ahead and predicting when someone is going to do something stupid, not braking hard if someone is behind you, being very careful when changing lanes, driving in a parking lot, etc. It definitely doesn't mean driving above the speed limit to keep people behind you happy.
> It's unhelpful and silly to suggest that people must break the law to be safe, and it completely defeats the point of speed limits and the rule of law
The law doesn't necessarily codify what's safe. You could go at 40km/h on a highway and be technically entirely within the law, but that's certainly not a safe thing to do. Similarly, you could go at exactly the posted speed limit on a snowstorm when everyone is driving slower and you would be entirely within the law, but you're probably going to get yourself killed.
Maybe if I always drive in the rightmost lane, being pedantic about posted limits is probably safe for a variety of reasons, but if I'm on the left lane of a 100km/h 3-lane highway, I wouldn't exactly feel comfortable going at 100 if everyone else is going at 125 (including cops).
>How about when merging, or changing lanes? The speed limit is not _always_ safe or practical to obey strictly.
In the case of overtaking, you just need to make sure you don't cut someone off in the overtaking lane. No need to go over the speed limit just because you're too impatient to wait for a proper gap in the traffic.
I've never seen a situation where you're merging onto a freeway and everyone in the right-most lane is speeding excessively. That would be a very dangerous situation, and if you yourself go significantly faster than the speed limit to merge you're just adding to the dangerous problem.
This is probably less the case with self-driving cars, as I imagine they're using pretty new or well-maintained vehicles, but there's usually a 5mph leeway with speed limits due to slightly mis-calibrated speedometers. I anecdotally remember hearing this being the reason for speed limits that end in "5" rather than being multiples of 10.
Wouldn’t that be mitigated by the GPS and the map? It would seem that an automated car using maps as an important input would no its exact position and therefore speed.
GPS can be remarkably bad in urban environments. You can't really trust it for second-to-second speed or positioning. It's only really good for long term/high level planning. For speed, the wheel sensors are going to be way more reliable and consistent.
The typical approach is to fuse GPS, inertial, and wheelspeed sensors to track a vehicle's motion in space. They all have their shortcomings, and all compensate for each other.
GPS is, as you said, wildly inaccurate, disconcertingly often. Wheelspeed is great when you're moving, but at low speed it tends not to have the granularity you want. Inertial sensors are pretty amazing nowadays, but they still drift.
To illustrate, if your wheels aren't moving, then you can ignore the IMU telling you that you're sinking into the ground at 5mm/s. Watching the monitor of a mapping vehicle as it depicts the car returning to mother Earth is pretty hysterical though. Until you have to fix it.
Actually, with a good receiver you can use GPS for very precise speed measurements by using the doppler shift of the carrier signal. It's just that most receivers don't give you access to that so the only other option is using the current and previous position and the timings of them.
That could easily be the case, sure. Some other people mentioned that it could be just to more closely emulate a human driver. I don't think we'll have a good answer without knowing the implementation.
I know this isn't the most pertinent issue here, but I'm surprised the self-driving AI is allowed to speed, though with further research, this seems to be the case with Google/Waymo software too:
http://www.bbc.com/news/technology-28851996
Though in that BBC article (which is from 2014), the stated scenario is when the self-driving car is surrounded by traffic exceeding the speed limit. It doesn't sound like this was the case for the Uber vehicle.