This is mostly about sensors and geometry. Machine learning has a role, but only in target identification. That's how Waymo does it. The fake it til you make it crowd had the fantasy that you just hook up some cameras to a machine learning system, train it, and you have self driving. Doesn't work. Machine learning is way too dumb.
You can maybe identify "traffic light", "car", "pedestrian", and "deer" with machine learning. That's just used to guess what they're going to do next. It's not used to decide if they're an obstacle. Obstacle detection and avoidance is all sensors and geometry.
Also, "self driving car", "electric car", and "transportation as a service" are all independent. All will be available, but not from the same companies.
I am curious why you believe this to be the case? Driverless cars have yet to even come close to being workable in adverse conditions, to my knowledge, please correct me if I am wrong.
> This is mostly about sensors and geometry.
And mostly about the limitations of sensors, and the limitation of available algorithms to overcome geometry. A algorithm a driverless car uses to overcome some geographical feature might expect something to be spherical, but in reality, it ends up being elliptical, only it is too late for the car to adjust, for example, or vice versa. There is literally hundreds of thousands of potential corner cases out there. The earth in itself is a good example: although many people try to do distance calculations with lat/long and some standard radius are usually off by n precision points because the earth isn't perfectly spherical, only approximately. Driverless cars have to be precise. A couple points off might spell the difference between safely driving in the lanes, and veering into a ditch, off a cliff, or into a median or obstacle.
Sensors will suffer some of the same problems (or different problems) as our biological sensors do. How will road signs, lines, and likewise be detected in adverse conditions? Snow? How will radar overcome natural geographical corner reflectors?
Driverless cars seem to becoming along just fine in perfect conditions.
Not trying to be defeatist, but I believe if we want true driverless technology, the infrastructure will have to support the vehicles for these edge cases that we cannot overcome. Or we can continue to be idealists. I for one wish we never gave up on trains -- they are the perfect "driverless" technology candidate.
Not to mention, any snow on the road is going to make it supremely hard for a computer to determine where their lane is, or even the road is.
Waymo has demonstrated solutions for many of the edge cases you mention. This video is from 3 years ago:
Regarding snow, it maybe a difficult problem, but if cars can't see road markings, neither can humans, so how do they do it? Sometimes by following a car in front. Though far from real autonomy, my Tesla doesn't need to see road markings for autopilot to work, it will fall back to following the car in front.
And.. Why does Waymo have to solve this inclement weather problem now? Surely it would be better to demonstrate safety in good weather, then light rain, then heavy rain, then light snow? They can just refuse to operate in bad weather unless they are confident about it.
I think I've seen at least a prototype Waymo car (it was exhibited in the Computer History Museum) that was designed without the steering wheel and pedals for a human to use. That wouldn't work so well given that the car might pull over and be unable/unwilling to drive anymore. (Maybe you can put an expected upper limit on how long the weather will remain bad ... Nah, storms can last for days, by which point a human who didn't bring food and water could be in bad shape.) I guess you could rely on being able to call AAA. Though a bad blizzard might be exactly the case where AAA would have a hard time reaching you. The human being able to take over eventually seems like an important feature. I guess that is the way all the cars in the field (that I've seen) work.
Does anyone happen to know if self-driving cars are practicing "pull over and stop" maneuvers? For some reason I've never heard of it. I think it is the only potentially reliable fallback mechanism if the car encounters, say, weird terrain it knows it can't handle. (I don't think you can assume the human will be able to respond within n seconds.)
It'd be fine - you just need to have some backup option. A small joystick or two behind some panel, or screen control, or perhaps you control the car with a phone app. It's like having an emergency spare tire - you just need a minimal driving UI that can be used in case of emergencies but users otherwise don't need to see or deal with. It can be inconvenient and speed-limited, since you don't expect to use it much.
Though if there's decent connectivity an obvious intermediate option is remote operation. Some human who is really good at driving might take over for a bit, driving your car from the comfort of their own home or office if you're not comfortable doing so yourself using the backup control pad.
Maybe we need driverless cars that can show panic faces when they don't know what to do and be coached by a friendly citizen what to do? It would be super cute.
Sadly, no one seems to have figured out a way to economically motivate human drivers to do this.
- human drivers in demanding situations
- sensor and compute package planned for autonomous driving
From there it is mostly a question of
1. getting permissions to collect the data
2. finding some efficient way to store -> sift for interesting situations -> magic happens here -> lots of regression testing -> self driving car -> profit!
I would think?
I think it comes down to being able to model the world around us, having a model for human behavior, animal behavior, etc. If we see a baby deer, we look for the mother deer.
It feels like we're also really good at creating multiple open-ended stimulations in the back of our mind about how things might play out, where they can easily be paused or fleshed out depending on which became more probable or if we want to learn something. It's kind of a running background process.
30,000 deaths per year suggest that humans fail at these edge cases all the time. Maybe ate their best a human outperforms a machine but as any person who drives if other drivers are at their best.
I mean yeah, compared to medieval statistics, it's great, but in comparison to the other non-medical ways to die you face as a 21st century American? automotive fatalities are quite high, and the problem deserves a lot of focus.
You could also make it harder to be qualified to drive.
I view autonomous cars as the 'lazy technocratic' solution to a problem caused by total government mismanagement.
(I mean, I agree that reducing miles driven is also a good goal. I'm just saying, making safer cars is a lot easier than talking people into allowing higher density construction and cities built around transit and walking rather than around cars)
A minimum 14 years of "human learning".
Anyone who's seen a Waymo car drive around the valley knows they're much safer than human drivers. They're annoyingly good, because for instance they just never miss a bike and are careful around them, and if you can't tell why it's a bit annoying.
But I've always come away from such incidents with a feeling of "I should have seen that". I've yet to see them screw up even once.
And yet you can't drive from SF to San Jose without seeing a human driver screw up.
I don't think anyone is debating that in perfect conditions the computer cars can sometimes out perform the worst drivers.
If your concern is terrible drivers killing themselves and others, what you should be arguing for is much stricter standards for obtaining a driver's license. Probably removing 'full coverage' insurance and only allowing liability would go a long way as well. Can't keep wrecking your vehicle if nobody's buying you a new one.
Or for keeping a drivers license.
>Probably removing 'full coverage' insurance and only allowing liability would go a long way as well. Can't keep wrecking your vehicle if nobody's buying you a new one.
I disagree... a used car is pretty cheap. Cheaper than full coverage, if you have a bad record.
I think this should be attacked by increasing the minimum liability coverage. right now, in California, you can drive with homeopathic levels of liability insurance... $35K isn't going to cover very much hospital time/time off work if you hit someone who has a good job. The minimum liability coverage should be 2-3 orders of magnitude higher.
Of course, for either of these things to be politically possible, we first need a society where it's reasonable to get from A to B without a car. An idea that is fought against daily even in relatively dense places like where I am now.
Self-driving cars might be able to beat that, but most of the self-driving car models aren't even close to 100 million miles driven.
(from page https://en.wikipedia.org/wiki/Transportation_safety_in_the_U... )
Furthermore self-driving cars are at this point ONLY driving through areas with a lot of pedestrians. You'd expect them, therefore, to have much higher accident fatalities than normal vehicles. Why ? They're always around pedestrians. Instead, in the first years of operation, they're safer than humans.
Even with Uber's less than stellar safety practices, and Tesla's ... well ... how do I put this politely ?  seems a good link. Let's just say Tesla owners could be more careful, especially since they signed a contract stating that they would be more careful (and youtube has much worse than that video). Given that that's how people use self-driving cars I would argue that's pretty damning for human driving skills.
HD mapping allows cars to know precisely where they are on the road even if they cannot see the lane markings below. In fact, georeferenced HD maps can pinpoint the car's location in the world down to less than 10cm of absolute precision (as opposed to relative precision)
It's just as easy to come up with any sort of contrived situation for which a machine would outperform a human.
- They aren't as easily distracted as humans,
- you can program them to stick to the rules even if slowing down results in problems for the passengers
- you can program them to care about any neighborhood, not just the ones we live in
† CA driver's handbook page 37: http://driveca.org/cms/assets/uploads/2014/09/dl600.pdf
Great strides have been made in using machine learning to filter out obscurants such as snow. Perception for autonomous vehicles is effectively a solved problem.
The biggest technical challenges are related to planning. So say you're approaching an intersection. There is a pedestrian about to cross, a cyclist in front of you and another vehicle waiting to turn left across your path. As a human, you understand that if you behave one way, it will cause the pedestrian, the cyclist and the other car to respond a certain way, but if you respond to the situation another way, it will cause all three to respond differently.
Our ability to game out scenarios like this is intuitive, but for AI to predict how it's behavior as an agent will influence the behaviour of other agents on the road is a daunting undertaking, particularly when taking into account the full scope of scenarios that need to be mastered before an autonomous can reliably safely navigate anywhere.
Great strides is not 'perfected' and I don't think perception for autonomous vehicles is a solved problem whatsoever. What is a driverless car to do if it slides out of position on the road and is now facing the wrong direction? Does it have the ability to know whether it needs to call for help, can safely maneuver, it's occupants are in immediate danger and should exit the vehicle (or not exit the vehicle)?
I think the most likely scenario is self driving cars won't drive in those conditions, hopefully those conditions don't develop while you're on the roadways or you'll be parked somewhere cold.
Will similar team exist for weird intersections in the middle of nowhere as well?
Will all these teams solve the problem of all the truck drivers rendered unemployed by "self-driving" vehicles?
Will all America's vehicles just wind-up instead being driven by tech support using joy sticks when the whole system goes south?
The goal isn't perfection. The goal is to be as good as humans.
Eg: An autonomous vehicle driving through snow backed by millions of miles of experience driving in snow is better than someone from the tropics driving through snow for the first time.
Which particular humans?
>> An autonomous vehicle driving through snow [..] is better than someone from the tropics driving through snow for the first time
Better than the worst human we can imagine?
Better than the average human?
Better than the average human who drives on that road at that time of year?
The problem is that you can't separate "perception" from "cognition". A human could drive a car reasonably well via a webcam feed, so does a webcam count as "solved perception"?
What GP means by perception being a solved problem is that existing sensors and algorithms produce high quality and reliable models in a wide range of environment conditions.
A webcam is a sensor device, the human observing the webcam feed most likely cannot construct a good model of the car and its surrounds, as they can't do things like shoulder checking or looking in the rear view mirror, which a good driver does frequently to maintain a model of where other cars are.
But then "reasonably well" for a human driving a car remotely via webcam feed is likely to be a far lower standard than we're holding self-driving cars to.
Why does the self driving car have to restrict itself to extremely subtle human nonverbal communication? Just put a loudspeaker on the outside of the car and have it announce its intensions clearly, yielding to pedestrians/bicycles according to the laws of the land.
There's some video here for:
>“The Yandex.Taxi autonomous car safely navigated the streets of Moscow after a recent snowstorm managing interactions with traffic, pedestrians, parked vehicles and other road hazards on snowy and icy streets,” https://www.theverge.com/2018/2/16/17020096/self-driving-car...
Presumably it's not perfect but it seems to be coming along. Snow and ice covered Moscow is more adverse than what I usually drive.
Given how people drive in the rain and snow in the Northeast, I would argue human drivers have not even come close to being workable in adverse conditions.
We drive in them anyway, despite often times magnitudes more risk of crashing, because we want to.
If you look at all the relative statistics, both IIHS and insurance, the drivers who practice safe driving techniques and do not drive distracted have something crazy like a 3-4 fold lower accident rate. Moreover, those drivers who have had an accident are much more likely to have another one. Have you ever met someone who has gone their entire life without an accident? They exist, I encourage you to speak to them about how they drive -- more often than not it isn't by...accident (sorry)
The other point I am making here -- the drivers who do account for reckless, endangering driving, or who are careless by some measure of frequency (not careless as a whole, but maybe they forgot to change to their winter tires and properly inflate them?) -- they tend to pay the price the most. Their insurance rates are higher, they get tickets and points, and eventually the worst ones aren't allowed to drive at all.
When the "worst" AI's that are responsible for some amount of crashes, who is going to pay the price? Whose license is going to get deducted? Whose insurance is going to go up? In the world of driverless ubiquity, who is going to be the check and balance? What self-selecting mechanisms are going to protect us? You think Waymo will give a shit?
There are other reasons this argument is fallacious, the most obvious one that driverless cars have not yet proven they are safer or can remain statistically significantly safer than safe human drivers. How do we know driverless cars cannot or won't be compromised? What if a foreign nation state deploys an attack that exploits a known weakness? What if a car company is infiltrated by a bad actor and uses bad training data? We were able to break into a country's nuclear reactor. Russians were able to break into our democracy. I imagine this would be a nice juicy target for a country that really doesn't like us. Ect. Ect. It is going to come with a whole host of it's own unique problems, so comparing them to current human driving problems (which have potential solutions that our government won't embark on -- different story for a different day) generally won't be fruitful.
That's quite the qualifier. Problem is I see at least 3 people with their eyes on their phone during every 30min commute to work ...so...
>When the "worst" AI's that are responsible for some amount of crashes, who is going to pay the price? Whose license is going to get deducted? Whose insurance is going to go up? In the world of driverless ubiquity, who is going to be the check and balance? What self-selecting mechanisms are going to protect us? You think Waymo will give a shit?
These are literally the least difficult issues around self driving cars.
> There are other reasons this argument is fallacious, the most obvious one that driverless cars have not yet proven they are safer or can remain statistically significantly safer than safe human drivers.
so we should stop trying?
You are either young, or extremely naive (or both) if you don't think these are going to be challenging problems.
> so we should stop trying?
Not sure why you thought that this begged that question. What I am saying is we should stop making arguments that compare driverless car habits, problems, and safety to the current habits, problems, and safety of human drivers. They are generally bad arguments. They won't bear any fruit. Statistics aren't people. Trying to sell a generation on driverless cars on some unproven projection that "they will be X times safer!" will fall on deaf ears to drivers who have made it a lifetime of safe driving.
are they show stoppers? c'mon assessing risk is old as the hills. There's an entire profession that is devoted to assessing and pricing in risk. With regard to liability, the courts will decide. It'll be messy but it'll get sorted out in time.
>Trying to sell a generation on driverless cars on some unproven projection that "they will be X times safer!" will fall on deaf ears to drivers who have made it a lifetime of safe driving.
people will get over their fear in time. People feared cars when first introduced too.
"With all the anxiety around driverless cars lately, it’s worth remembering there was a time people worried about cars exactly because they had human drivers. In fact, it was the removal of the horses—the horseless carriage—that gave some people fits.
In the 1890s, the prospect of a person driving without the aid of a second intelligence was a real concern. A horse, or team of horses, acted as a crude form of cruise control and collision aversion.
In 1896 Alfred Sennett warned, “We should not overlook the fact that the driving of a horseless carriage calls for a larger amount of attention for he has not the advantage of the intelligence of the horse in shaping his path, and it is consequently incumbent upon him to be ever watchful of the course his vehicle is taking.”
Also, the lack of support for harder environment isn't because they're impossible, but rather because they are focusing on getting it perfect in normal environments before expanding. The person said they are "coming along fine", not that they are done.
I beg to differ.
I claim that adverse environments are very hard, even pathological problems and placid driving in well laid out suburban road with generally well behaved drivers are merely hard problems.
The thing is that an engineering team working on a problem like self-driving in adverse conditions is always going to be making some progress. The main question is whether they are going to make the kind of progress that justifies the hefty expenses involved.
Much more improvement has yet to come before we get to being safe enough with level 5 autonomy.
But in the meanwhile, we can also still celebrate the intermediate stepswise improvements that have been made.
anyone have any idea of how frequently horse-drawn carriages crashed in dense urban inner city environments 120 years ago normalized against the incidence of crashes for contemporary cars on the same streets at the same speeds?
Maybe this could form the basis for a kind Turing test for level 1-4 autonomous Vehicles.
The other thing people forget is the bar isn't that high: self driving vehicles don't have to be perfect, they just have to be better than humans are. All of the "whatabout" edge cases people proffer as examples of areas an AI would have trouble with, people have trouble with too. The difference is that once an AI learns to solve that edge case, it doesn't have to relearn going forward.
How are more sensors going to tell the difference between a plastic bag and a cat? How does lidar tell you which car is parked on the side of the road and which one is in your path? Relying on a dozen sensors seems like the crutch to me. Nothing about our current world has been designed for lidar and radar. It was designed for binocular vision. Sure, if your only goal is to not hit things at low speed in perfect weather, lidar is great. But you can't get much further than that.
"Nothing about our current world has been designed for lidar and radar. It was designed for binocular vision."
Really? Was the world designed for binocular vision, or is it just three dimensional?
Having more sensors, especially when they have different failure modes seems like the only possible way to create a reliable system. LIDAR isn't super dense, but generally has accurate returns. Binocular vision sucks on untextured objects, like the side of a white truck in the fatal Model S collision. Why wouldn't you want the crutch of both types of measurement?
No it wouldn't. The whole problem with avoiding stationary objects is that they are _everywhere_. Do you really think radar and vision didn't "see" the barrier? Should it have known that it was in it's path? For sure. But that has nothing to do with sensors. Stationary objects are in our paths constantly while we drive, but we tend not to hit them because they are usually only temporarily so. At 80 mph the difference between a parked car being in our path and not, at the stopping distance of over 300 feet, is only a degree or two of wheel turn. We have enough information. Acting on it appropriately every single second of operation is the problem.
> Really? Was the world designed for binocular vision...
Yes. That's why we have road signs, stripes on the road, reflectors, lights on cars, lights hung over the road which light up in different colors to indicate right of way, etc. LIDAR is useless with all of the most important signals on our roads.
Ermm, actually yes, in the sense that anything human made was designed for and by those with binocular vision themselves.
What human has advantage of is utilizing knowledge in unrelated domain to another domain.
Evolution suggests that you can, in fact, create systems that are more intelligent than what came before them.
Does binocular vision even matter at such distances (i.e. beyond a few meters)? I thought it was only effective for relatively close objects.
Back-of-the-envelope calculation: assuming 1920x1080 camera sensors with a 90-degree horizontal field of view, each pixel covers a visual angle of approximately 0.001 radians. If you have two cameras separated by 5 feet horizontally, and your measurements are accurate to within a pixel, then you can measure the distance to an object 500 feet away with roughly 10% accuracy. That's a reasonably conservative following distance at typical highway speeds.
And that's just from a single parallax measurement. By using other visual cues (e.g. apparent size of an object whose physical dimensions are known) and averaging measurements across multiple frames, you can get a more accurate measurement.
I found a page  that summarizes the level of visual acuity required to obtain a driver's license in each US state. It appears that all states will potentially issue a license to someone with vision in only one eye. Some states impose additional conditions, such as requiring a favorable report from an ophthalmologist, limiting the maximum speed, or restricting the driver to vehicles with outside rear-view mirrors (to compensate for the loss of peripheral vision on one side).
Infrared transparency? Radar cross section? There are so many ways. Sensor doesn't just mean "webcam".
Honestly, getting them to work to this basic level is the "easy" part. Real world driving conditions in urban environments are suboptimal at best. You're not only going to need a system that can navigate around and not hit things. You need it to be able to dynamically route around road-construction, do clever things to find parking.
Roads in most places aren't that well standardized. It takes a lot of small judgement calls for a taxi to actually do its job. Uber and Lyft still have trouble dispatching people to pick you up at your actual door!
For the amount of money it would take to completely overhaul the country's entire transportation infrastructure to accommodate self-driving cars in this way, you could have built a phenomenal train and bus system with bike/scooter lanes everywhere.
So what exactly is the value add of cars then?
The guy who started iRobot, Rodney Brooks, doesn't believe we'll see a true driverless car operating throughout a city on normal roads until 2035 at the earliest: http://rodneybrooks.com/my-dated-predictions/
As software developers, we all know how big the difference between "a demo the boss is excited about" and "reliably working in production for all users" can be. Given how complex the domain is and how many edge cases it has, I can easily believe Brooks is right here.
> A driverless "taxi" service in a major US city with arbitrary pick and drop off locations, even in a restricted geographical area.
> Not Earlier Than 2032
This is odd because isn't Waymo doing just that in Phoenix with a general release date of 2019? I checked the source because I wanted to verify that it says "a city" and not "any city".
What I've seen about Waymo (and maybe I've missed something) is very hazy. And it's not out yet. I'll be very surprised if they go from "operating in secret" to "completely competitive with Lyft" in one jump. My guess is that there will be significant limitations for whatever their first release is.
That is much much closer than 2035.
The best a teleoperator can do is suggest a path to a car that is confused, and the car decides if the path is a good idea. We also don't know how often that's used. If it's rare enough, does it even matter? Does it not make it driverless if once in a while they need a little nudge?
That's like saying your car isn't a real car because 1% of the time you need to take it to the mechanic.
But yes, if a human has to get involved, I think it's correct to say it's not self-driving. In particular, what Wired describes is either SAE Level 3 or Level 4: https://en.wikipedia.org/wiki/Self-driving_car#Levels_of_dri...
Until it's Level 5, it's not really driverless.
Could be correct. I think likely it is correct. BUT, everyone thought we still wouldn't be winning games of Go by now. So could be wrong.
What you wrote makes a lot of sense to me. However, I would also hope that in a matured self-driving environment, traffic lights would be mapped ahead of time. It's always possible that a traffic light shows up in an unexpected place, but in my mental image of the future, this raises a flag for system management to modify the map of permanent objects.
There is a chance that a powered non-permanent traffic light is found somewhere, a dude walking along the side of the street with a stoplight on his back, and connected batteries. In that case, the police need calling. In machine learning, there are just so many corner cases where best to not try to fully interpret, just call out "dude, look at this thing!"
Temporary traffic lights are extremely common, at least here in Australia. They are used in the case of roadworks or other short-term disruptions that requires traffic flowing in both directions to share a single lane, in conditions not suited to manual traffic control (i.e. overnight, weekends, long-term works on rural roads).
Since evolution didn't provide us with our personal lidars, human drivers are an existence proof that it is possible to learn reliable driving performance from vision alone. Which isn't to say that machine learning based on vision alone is already close to human-level or competency or that human-level competency is sufficient for a automated driving system or that sophisticated sensors don't make the problem a bit easier. But I don't know how you can be so confident that vision alone (or primarily) won't be route that eventually succeeds.
The underlying assumption I'm making is that AI driving improvement is accelerated by being on the real road. If that's true, then we want the cars on the road as soon as reasonable possible. Because I would gladly trade hundreds or even thousands of AI driver caused deaths in the short run if I am reasonably convinced that it will prevent the tens of thousands of human driver caused deaths every year. And I am convinced of that. I also acknowledge I'm in the pool of people who might be killed by the AI driver. Just as we let teen drivers on the road with the expectation they will improve over several years, so should we accept a similar risk from AI drivers. The return is far better for the "inexperienced" AI driver because the AIs will continually improve forever, but as you noted we get a new batch of bad human drivers every day.
Could this/should this also apply for autonomous vehicles? If so, how?
Instead, how about:
All new autonomous vehicle configurations (let's call that the algos + sensors + vehicle) have to take some kind of actual driving test, just like us humans do.
Maybe the public could even help design a good test? "Not driving at speed into a stationary fire truck which is parked on the highway right in front of you" would be one element I'd want to see tested.
If an autonomous vehicle is involved in an accident, and the algo/sensors/vehicle are found to be (partially) at fault the configuration earns penalty points.
If that configuration earns enough penalty points over a period of time, the entire configuration loses its certification, plus a fine, plus a mandatory re-test.
This method appears to work reasonably well in dealing with us not-always-perfect human drivers, and ought to concentrate the minds of the designers/developers/managers behind autonomous vehicles.
To be fair, nobody has been killed inside any driverless car either.
You wouldn't vet a system that is susceptible to high false negatives and likely to kill people- no matter how good the accuracy is.
iow tesla needs to beat the average 2014+ 3 series death per mile.
At some point the right balance of sensor quality and manufacturing processes will converge to be an affordable solution to highway driving. (broad daylight, then dusk, then maybe rain, etc etc).
If history is any guide (airlines anyone?) This will come along piecemeal, and the CEOs being forced to make ridiculous promises to appease shareholder anxiety will be given nice scapegoat packages along the way.
Self-driving car: ZF
Electric car: ZF
Self-driving advertising services: ZF
I think the hard part of the problem is the infrastructure. Everyone wants to fix the cars, but they'll leave the hard part to the government- which is the development of autonomous-ready infrastructure.
For your professor, yes assistive technology like lane keeping assist and collision avoidance will save lives, but are no reason to go driverless.
Ultimately the faster we bring driverless cars to fruition, the fewer people will die, simply because it's inevitable that they will quickly exceed the safety of humans.
Then we will have turned automobile safety into an engineering problem. It was partially that before, we could package the victim better to improve their survivability, but now we can modify the driver ... every driver.
There are about 7 billion instances of those learning systems though.
Will this be replicated with machines in the next 10 years? I have no clue.
Is this doable at all? Sure.
The problem with the human brain is one of inattention. CNNs / Artificial Neural Networks can remain at attention 100% of the time due to their artificial / machine nature.
But CNNs, despite being at 100% attention the entire time, still have issues determining if that splotch on the screen is the sky or an 18-wheeler.
Artificial Neural Networks / Convolutional Neural Networks have a very long way to go before they reach human equivalence. In contrast, sensor systems or LIDAR brute-forces the problem. LIDAR can see things human's can't see, and advanced sensors can tell you (at least, in clear conditions) the location and velocity of virtually every object around the car.
Fake-it till you make it camera-only driverless cars are clearly hype that relies upon a fundamental misunderstanding. Just because CNNs are kinda-sorta like the visual cortex of the human brain doesn't really mean that it works like one.
CNNs have really cool visual learning properties. But I've yet to see one 100% successfully tell you background vs foreground in pictures like a human brain can do. Even in clear weather conditions, the CNN can confuse a truck for the sky and still run full speed into an 18-wheeler.
Surely the field of "machine learning" includes things which are not even invented yet, the same way maths comprises of theorems not yet discovered.
I was just saying that I think one could "just hook up some cameras to a machine learning system, train it, and you have self driving", maybe not today though.
Even calling it "learning" is kinda overselling itself. There really are only two camps of "Learning": auto-optimization (against a trained dataset), and auto-categorization (aka self-training).
Its auto-optimization: the algorithm self-corrects itself to try and look more like the training-set's "ground truth". Or auto-categorization, as the algorithm looks for patterns and tries to draw its own categories.
"Learning" implies model finding. Which... strangely enough... I'd argue that 3-SAT solvers are more "learning" based, at least with colloquial use of the word. Those things really do craft new theories and test them through the process of elimination / resolution / etc. etc. "Neural Networks" explicitly DON'T do this however.
Have you looked into Hubert Dreyfus yet?
Just go through this recent paper I found on DDG.gg and see the amount of effort it takes to parse foreground / background data that a self-driving car needs.
You gotta figure out if there's a still-object on the road. And whether or not its a shadow (shadows don't move after-all but its safe to drive over). Like, neural networks can't do that stuff 100% reliably yet. And it may never happen.
Or some researcher next year might come out and discover a method to parse background / foreground / shadows out of pictures. But then there are a whole host of OTHER issues involved.
I think self-driving technology has potential. But you need more than just neural networks hooked up to cameras. I really like Waymo's direction with advanced super-human sensors. Avoid the shadow / background / foreground problem entirely and just have LIDAR give you the precise coordinates of all objects within 100-feet of the car.
IMO, if self-driving technology ever happens, it will be because of advances in advanced LIDAR or other kinds of sensors. Stuff that can avoid the research-problems that the "Deep Learning" community hasn't been able to solve for the last 50 years.
The "until we can program computers to..." is a when question, not an if question. There's nothing about driving, in any situation, that doesn't fall in the face of "assuming infinite computing power and infinitely good sensors". Driving isn't a creative act, it's a responsive one.
My main issue is that a large number of people seem to think that cameras + machine learning are enough to solve this problem. And while I'm not an expert at machine learning, what I know about it makes me a pessimist. There's just too many unsolved problems in the machine-learning community to apply machine learning to the car-driving problem.
Machine learning probably can solve weird cases people don't expect. IIRC, CNNs are better at recognizing blurred or garbled text than humans these days. So CNNs can read speed limit signs, road signs, and other texts and and at least process that.
Even figuring out if its a speed limit sign, an address, or a route-number probably can be solved by CNNs. But higher level reasoning (is that spraypaint messing up the signpost?? Which was common in some of the areas I drove through) seems like an unsolved problem.
Anyway, Machine Learning + Cameras are IMO, at best... a partial solution to some of the problems. Anyone who thinks that cameras + radar + machine learning is sufficient is probably just a TSLA long who wants to believe in the success of their stock. Otherwise, I think most people are reasonable and recognize the importance of experimenting with a ton of different methodologies to solve this problem.
Computer vision has been improving rapidly in the last 10 years, I think it's too soon to rule out the viability of a camera-based solution entirely. Though I do hope improved lidar technology can improve on humans.
Humans are way, way better than current CNNs on this field. We can talk about cats, shadows, and boulders when CNN-based methods stops crashing into concrete barriers, parked fire-trucks, and 18-wheelers making a left turn.
I don't want to dismiss the work of Deep Learning / Machine Learning specialists. I just want to point out that the problem is incredibly difficult. It is very far away from being a solved problem.
> Traffic-Aware Cruise Control cannot detect all objects and may not brake/decelerate for stationary vehicles, especially in situations when you are driving over 50 mph (80 km/h) and a vehicle you are following moves out of your driving path and a stationary vehicle or object is in front of you instead
This is a known issue, a known pattern and has happened multiple times this year. Its repeatable. CNNs today are not working in this case, and fixing it will require a research effort of mammoth proportions.
Comma.ai is doing vision+radar right now and driving cars with it.
Are you referring to the midnight jaywalker in Uber's case? I still feel like most of the blame goes to the pedestrian.
I would bet in less than 10 years you'll have camera only fully autonomous systems.
>Perception is a game of statistics.Crudely speaking, if we have three independent modalities with epsilon miss-detection-rates and we combine them we can achieve an epsilon³ rate in perception. In practice, relatively orthogonal failure modes won’t achieve that level of benefit, however, an error every million miles can get boosted to an error every billion miles. It is extremely difficult to achieve this level of accuracy with a single modality alone.
>Different sensor modalities have different strengths and weaknesses; thus, incorporating multiple modalities drives orders of magnitude improvements in the reliability of the system. Cameras suffer from difficulty in low-light and high dynamic range scenarios; radars suffer from limited resolution and artifacts due to multi-path and doppler ambiguity; lidars “see” obscurants.
There are scenarios it’s possible to encounter on the road that humans don’t even necessarily know how to handle immediately. In proper, brutal, real-world testing, one very quickly enters the realm of problems that you’d need full-fledged general AI to reason through.
Fog on a straight road caused an 87 car pile up after a trailer jackknifed. The point in the article I remember at the time was a mother recounting watching her 14-year old daughter burn to death. Mom got out, but the daughter had been pinned by her leg.
The quote from the time was ""Mama Sheila, please don't let me die. I'm only 14," Marceya McLamore begged."
I suppose my point is that neither human or machine are good enough yet. and not to treat this as an abstract academic topic, it's real people.
I work in machine learning in a different area (healthcare) so my perspective may be incomplete, but what I see in ML/AI is that models are really good at memorizing, not so good at understanding. What I mean by that is that a human navigates the world by knowing what a car is, what fog is, what a person is, what a plastic bag is, all sort of things. Object detection can get much of the way there, whether it can get enough of the way there is an open question, but ok we'll grant that. Whether AI/ML can actually make the step from knowing what the object is to knowing how that object interacts with all the other objects in the world is another.
In other words identifying an object is just step one, understanding what that object means in context is the next absolutely required step, and it's way harder. We have a leg up as humans so we can rely on visual alone, but machines almost certainly will need other sensors that report additional information because they don't "get" context in the same way. And I'd say it's an open question whether even with those sensors, they'll be able to get there. I hope so, but the game is far from won.
Just because thousands of people roll the dice every day driving in unsafe conditions does not mean that we should tolerate machines doing so.
I assume the car will give you an ETA, so you'll check the clock and go back to your book/social media. I'm okay with "slower, but smoother".
Assuming that day is infinitely far away, then you're right.
If there's magic in the human mind at all, it's the creative side of the mind. Driving isn't a creative act.
I really really hate the concept of a driverless car. It's an incredibly difficult problem space; and for the same cost, America could build up municipal rail and bus infrastructure to where they were in the 1940s/1950s. We need more cities like Seattle with its rail expansion and fewer New York City where the infrastructure is finally getting money so it doesn't fall apart.
There are already so many tax breaks going into driverless companies. If Alphabet or Here want to do this on their own, go for it; but governments around the world should stop giving tax breaks and municipal incentives for this technology.
I can see it being more useful in Europe, where so much of the country is connected and it'd help sold the last leg problem. But in most of America we need to get back to the point where cars are no longer a necessity, not just for those who can afford to live in the city, but all the people who are barely making it who's lives fall apart if something on their car breaks.
While I completely agree with you, the cultural barriers to this are petty much a non-starter in much of the country. I love not owning a car and being able to rely on mass transit where I live (Austin, TX), but Texas is so culturally opposed to anything like this.
A lot of people think that mass transit is for the poor, and the car represents independence and freedom. Go where you want, when you want (but you're gonna be bumper-to-bumper most of the way there.)
I hate that mentality, but driverless cars represent the "have your cake and eat it too" solution. It more closely aligns with the culture of driving here in the US, while also claiming numerous commercial and infrastructure benefits.
That's because it is. I had to take the metro in Montréal to get to work this week, hadn't taken it in months because I usually bike, it's such a shitty experience. It smells bad, it's hot, it's filled with tired people that don't feel like going to work (you feel it in the air), beggars harass you, it's slow (it can take 3x as much time as riding a bike to get where you want to) and it's operating past maximum capacity during rush hour. I haven't even gotten started on when service is interrupted and you show up over an hour late to work or class. Anybody who can afford a bit more for reliability and comfort will spend it without looking back.
New York, Chicago, Tokyo, Taipei, London, Berlin, Munich come immediately to mind.
And as sibling commenters have pointed out, in some places even smaller cities and towns have excellent public transit networks that pretty much everyone uses (maybe not always, but when convenient). Freiburg, Germany is one example I've seen.
NYC... altright, better than 15 years ago but still not great at all. Not as good as Germany or DC.
Tokyo... lol, no! While clean and efficient, it’s a joke that you have different rail stations owned by different companies, different cards that may work on each other but probably not. The ticket machines are almost all cash only. And instructions in other languages might have well used Google translate. Tokyo was one of the more difficult public transport systems I’ve used and I’ve been around the world a bit.
I was surprised that the machines only took cash, but it was annoying as hell that we were 10yen short and the nearest ATM was 2 blocks away under the street then back up again.
The rush for the 11-11:30pm trains are nuts. I've never seen so many black suits filled by unsmiling faces than the last train out near Tsukiji.
Did you just claim DC public transit is good? Wow. Coherent bus lines are few and far between. Light rail is non-existent. And the Metro is falling apart at the seems... Prone to excessive delays. Entire lines are brought down for extended periods to perform decades of deferred maintenance. There's no ring routes - to get from Dulles to Rockville, you have to go all the way downtown and back out again. The system was built without consideration for express lines. And neither of the DC airports are on the same line as Union Station.
My wife tried to use Metro for her Reston->downtown commute. It was a disaster - rail delays made it completely untenable. Then she tried the bus. It was more consistent, but overcrowded. She ended up driving 4/5 days because it was faster, cheaper, and more consistent.
That's probably only hold true for (Western) tourists. And I don't think it is fair to compare on that. Tokyo (and Japan in general) rail system can be really intimidating but that's just because its sheer size, but once you know your way around it, it's gotta be among the best rail system in the world.
So for the two of us, a Muni trip is
- total cost: $5
- usually takes 30-60 min to get wherever we want to go in the city.
- we need to to figure out the optimal route, schedule, and entry & exit stops.
- we need to make sure we will arrive at the stop on-time.
- we need to walk 5-10 min to closest stop on either end
However... a low-cost Uber or Lyfts pool ride for two is
- total cost: $5-12 (same serice area as Muni)
- duration: 15-45 min
- we don't have to plan the route or schedule
- we don't have to worry about going to a remote pick-up stop on-time
- we get dropped off directly at our destination
Of course, here in SF ride share vehicles are definitely not allowed to use the special public-transportation-only lanes of some of the major roads, which remain reserved for the exclusive use of Muni buses and taxis to improve their transit times.
(edited to fix indentation)
Sure there are worse systems that don’t have any coverage (El Paso for example) but they are not functioning so I don’t count those.
Problems with muni include:
- To many stops (like every two blocks).
- Hardly any dedicated bus lanes, bus gets stuck
in traffic too many times.
- Slow and un-intuitive routes.
- Small and overcrowded busses.
- No transfer to BART or ferries.
: Sometimes too frequent, I sometimes see the 14 being stuck in congestion of other 14s
If public transit i.e. Muni serves 10x or 100x riders each day vs ride sharing than I see your point... I think.
Wait. Maybe you could spell it out?
On the other hand, 99% of vehicle trips in SF are either private cars or TNCs.
Neither of these was my point, though. My point was that comparing cost between a per-trip price and a per-person price will always favor the former when there are multiple people. But most trips don't involve multiple people, which is why such an analysis is besides the point.
I think what’s actually happening is Uber drivers are paid peanuts and provide and maintain their own vehicles. There’s a technology innovation (eg the efficiency of knowing if and where someone needs a ride) coupled with a labor innovation in paying drivers less than minimum wage.
Huh. This is interesting. The "Learn More About TNCs" button on the page you linked to leads to some more facts, including:
> "On a typical weekday, TNCs make more than 170,000 vehicle trips within San Francisco, approximately 12 times the number of taxi trips, representing 15% of all intra-San Francisco vehicle trips." - https://www.sfcta.org/tncstoday
So: 1,333,333 intra-San Francisco vehicle trips each day, including Transportation Network Companies (TNCs), and from your other links, 1,165,000 MUNI+BART trips per day, for a total of 2,498,333 rides per day
The population of the city during business hours is ~1,100,000... Oh of course, most people make round trips so it makes sense for the total number of rides to be >2x the city population.
> "On an average weekday, more than 5,700 TNC vehicles operate on San Francisco streets during the peak period"[sfcta.org/tncstoday]
So if TNCs provide ~30 trips a day, and all 1,165,000 MUNI+BART passengers switched to TNCs, there would have to be an additional ~40k TNC vehicles in the city :).
Maybe it would work if they were really, really small.
In cities that were designed or entirely remade to support the automobile, the automobile works better than everything else. This is almost a tautology, but it's remarkable how often it's overlooked in this discussion. If cities were re-designed around transit (with more density, smaller roads, and more space for alternative transportation modes), then transit would work better and cars would be worse. †
In summary, yes, obviously cars work better in cities designed for cars! But cities that work well for cars aren't a natural feature of the universe. It took a lot of work to get them to look like that.
† An important implication of this is that transit will never overtake the automobile in low-density, sprawl-heavy cities. Transit use is high where the transit infrastructure is good and driving is painful. Both things have to be true. ††
And the tension here is direct. Sprawl-heavy areas can't support good transit infrastructure and high-density, pedestrian-friendly areas are awful to drive in. A choice for one is a vote against the other.
†† This is why arguments about the superiority of cars based on revealed preference are usually spurious. If you put a light rail system in a town built for cars, then people will keep using the thing the town was built for. But those people are revealing their preference for cars in an environment built for cars. Where the built environment is different, people behave differently.
It largely depends on the country or even city. Me and my gf went to Switzerland by car a couple of years ago (we live in Eastern Europe) and I’d had expected that we’d visit different parts of the country by car. What happened once we got there is that we “forgot” the car in the hotel’s parking lot for a week and we did all our travel by train, it was wonderful. Even traveling inside the city itself (we were based in Lausanne) was a very nice thing, I’ve started building a soft spot in my heart for the city’s trolley-buses (electric public transport rocks, btw, always has, alwsys will).
While the electrification of the mass transit systems is super nice, unfortunately it came about because of the World Wars. Switzerland was hit really hard, and no one wanted to trade their precious coal away (Switzerland had no coal mines). Thus out of sheer necessity they began electrifying all the trains. So unfortunately there's not replicable policies other countries can use to achieve what Switzerland has achieved in terms of electrification, so it'd take some creativity.
Well that's because it's actually true for a lot of people. Sure, there's no freedom in commuting between SF and south bay by car, or navigating the most dense downtown environments. But I definitely use my car on a daily basis for things that would be impossible or incredibly time consuming by transit. And yes, many of the transit patrons around here would probably prefer a car if they could afford it. Most places are not like NYC where the subway is often the most effective option regardless of wealth.
I don't think it's really feasible for more spread out suburban environments to implement effective mass transit. The anti-car idealists often seem to forget that there are a lot of people living outside of dense city centers.
If self driving cars are terrible at coping with the chaos humans grit their teeth and deal with, maybe software for engineering cooperative bumper-to-bumper traffic jams involving clogs of slow-moving self driving cars, that reduce the road speeds to constant 20 MPH (the /constant/ part being important), and eliminate chaotic lane changing to near zero for the group, is one way to bootstrap self driving cars.
In other words, low-performance self driving cars might work fine, and since people can lounge about, and disengage from the guidance and navigation tasks, why not go slower?
Then, add swarming to the mix, such that cars going to the same place, all travel slowly in a single group.
It's against the law to obstruct traffic, or violate minimum speed requirements, so certain laws would need to change. For example, zero emission vehicles can idle indefinitely, and software-controlled vehicles can interface with an external control network, approved to operate in such ways, with points of contact at operations centers for resolving problems, technical or otherwise.
So change laws to permit idle electric vehicles (in more places and controlled, desginated places, not just anywhere), and permit large slow moving groups to occupy certain corridors, at certain times, when jammed traffic is a known quantity, that driverless systems aren't the cause of. Driverless networks can cut into a slice of the jam, and possibly improve outcomes for members only at first, and eventually boost efficiency within the jammed system by augmenting flow through expert participation.
Mass transit is expensive, boring, and thankless. How many bus companies have ever applied to YC?
Does Remix (https://www.remix.com/) count?
The thing is, if there were a bus, it would have to stop to pick up / drop off passengers along the way, which would make it so much slower than driving that it really would be for the poor. I've been on such buses as a kid. Even when traffic is bad, bumper-to-bumper in a car (taking the fastest route) is going to be faster than bumper-to-bumper in a bus (taking a more circuitous route and stopping frequently along the way)
I think the only way buses will stop being for the poor, is when they stop become faster than more expensive alternatives. (Or at least nearly as fast and significantly cheaper.) If you can save a lot of time by spending a little bit of money, which would you choose? What if you could spend a lot of money but it wouldn't save you any (or much) time?
In cities where congestion is a major problem, dedicated bus lanes/roads can help balance the equation in favor of buses. Usually there aren't enough dedicated bus routes though. (I might point out that the people deciding how to balance bus/car traffic are not the people who ride the bus.)
I still think self-driving cars are far in the future, but I find them appealing because they would make "buses" more efficient. A lot of the cost of operating buses is paying the drivers. To maximize economies of scale, buses are large, which means they have to make a lot of stops to pick up and drop off passengers, which in turn makes them slow. Self-driving buses could be a lot smaller and wouldn't need to make as many stops. If they could tell that no one will be boarding or deboarding at a given stop, they could skip it entirely and take a shorter path instead. At some point, the lines between bus/ridesharing/taxi get pretty blurry.
> If they could tell that no one will be boarding or deboarding at a given stop, they could skip it entirely and take a shorter path instead.
Everytime selfdriving tech comes up people say things similar to this, this problem is 100% solvable now. Why are we waiting for self driving cars to solve this problem?
What the hell is going on in Austin?
Bumper-to-bumper traffic happens mostly at commute hours in commercial zones and during special events. It's very much not the norm for driving at other times. To wit, I live in San Francisco, one of the densest and least car-friendly cities in the U.S., but the last time I returned from a road trip, I traversed the whole city corner-to-corner -- from crossing the Bay Bridge to arriving at my home near the zoo -- in less than fifteen minutes. Why? Because it was midnight! Cars provide unparalleled mobility for irregular trips outside of peak commuting times, even in the largest cities.
I wholeheartedly support deprioritizing car infrastructure and working towards a norm where most people do not drive to work, but I nonetheless believe that car ownership provides a lot of independence and ought to be accessible to the modal citizen, at least outside of the very dense coastal cities (SF/LA/NY). Cars work really well for irregular trips outside of rush hour. They work terribly when large amounts of people need to go to the same place at once.
I can count on 1 hand (ok, maybe both hands) the number of times I've been in bumper-to-bumper traffic in my life. One of those was in Austin. There's lots of places to go that aren't big cities.
<Insert obligatory Portal reference.>
Maybe it will take 15 years for it to be everywhere, but I'm still fully expecting a deployment in Phoenix this year or next.
I'm not really clear what tax break you're complaining about, AFAIK there is no extra tax break for Waymo.
Plus driverless car tech is global. The US infrastructure problems are irrelevant to me as a Canadian or the billions of people in Asia.
Toronto has horrible traffic too.
Instead of fantasizing about the day modern nation states are marginally functional at delivering services, like they were half a century ago. Basically everything they touch, especially under the false guise of 'private' partnerships ends up a decade late and 2-4x the cost (in billions). And often delivering only half of what was originally planned.
The less we gamble on such a trainwreck the better IMO. Unless, of course, something changes. But every year and every election it seems to get worse.
You seem to think self driving cars are magnificent devices that just glide over roads and don't need bridges or tunnels. For some reason, we never include these costs with cars when compared to mass transit. That is simply unfair.
I'm not against self driving cars by any means. But people misrepresent them all the time.
Many of the carriages were used as scrap metal for WWII, and once cars took off the demand wasn't there to rehabilitate them. But it would be so much cheaper to build out than even just replacing the signals in NYC would be. All the right of ways are still there, many of them turned into bike trails (which is great, but I'd rather have a $5 ride to the next town over)
The carriages back in the day were pretty slow, maybe 20mph, so I get why cars won, but rebuilding this infrastructure with modern equipment would be a really nice situation, and if gas was 10$/gal we would probably see the demand, but as long as gas is cheap we lack the ability to plan ahead...
Not to mention a multibillion dollar, unnecessary tunnel linking O'Hare with downtown which will almost certainly cost multiples higher than Elon Musk and Rahm Emanuel claim. We'd rather spend money on vaporware than build something that would actually help.
That project is privately funded. The Block 37 station was the expensive vaporware.
This is why US passenger train service is so poor compared to Europe.
> The picture for freight is different. According to Panorama 2009 , 46 percent of EU-27 freight goes by highway while only 10 percent goes by rail, while in the U.S. 43 percent goes by rail and only 30 percent by road. (In both cases, nearly all of the rest is waterways and pipelines.)
> So, it isn’t so much that Europe decided to move people by train rather than by automobile. It is more that Europe decided to use its railroads to move people while the United States decided to use them for freight. America moves almost six times as many ton-miles (or tonne-kilometers) of freight by rail as Europe, while both move about the same number of tonne-kilometers by road. While Europe moves about twice as many tkm of freight by waterway as the U.S., we move six times as much oil by pipeline. 
- area of Continental US: 3.12 million sq mi 
- area of Europe: 2.306 million sq mi 
- rail network length, US: 141,808 mi 
- rail network length, Europe: 157,667 mi 
- goods carried, US: 1.558 trillion lg tn mi/yr (long ton-miles per year) (world rank: 1st) (2015 estimate) 
- goods carried, "Europe": 0.32604 trillion lg tn mi/yr (long ton-miles per year)
I guess it feels like the distances involved in the US are more advantageous to freight, and in the EU are more advantageous to passengers
Since then railroads have to be completely self-supporting in the US, they focus on how they can achieve the highest productivity: giant, slow trains on shoddy track.
It also means the infrastructures are competing with one another, so there are redundant, shoddy, competing rail networks, rather than a network for slow and heavy freight trains, and a network for fast and light passenger trains.
City driving is a hard problem space, and all we'd achieve is cooler taxis.
The real game changer is instead driverless trucks.
Shipping is both a trillion dollar industry, and the problem space of "highway driving when it is sunny out" is much much easier than the consumer usecase, while still being extremely valuable.
I think you underestimate the impact of ubiquitous, free valet parking on urban design. We surround everything with parking to minimize walking distance to/from our vehicles, but you could organise things very differently if everyone was picked up and dropped off.
You could also park cars bumper-to-bumper, because you don't need trusted human drivers to move the cars boxing you in. Removing isles from the parking lot would nearly double the number of cars that could fit in a given space.
It's counter-intuative, but I think driverless cars may eventually make cities more walkable.
Driverless. Technology of the future. Always was. Always will be.
But seriously, how much would it take to mostly handle driving trucks on limited access highways between exurb mega-warehouses? By "mostly," I mean, be able to handle 98% of all situations, and be able to flag attention from a remote control human supervisor for the other 2%. I suspect we're pretty close.
And honestly any number is simply just 'pulled out of the ass' anyway. It's not as if they can see daylight on all the problems that need to be solved or predict various soft issues (the government only one example). So this is not that someone is building the atomic bomb and can realistically judge more or less the scope of the problems that need to be solved and then build in a bit of leeway.
It has has attained this status as revealed truth because it is indeed roughly true -- more people together in a vehicle done right will indeed use less energy per person and less road space. But the "done right" is very important as it is commonly done quite wrong.
As I have studied robocars, this has led me to the discovery that some of our old assumptions are wrong. In particular, more sharing is not always good, and the styles of sharing (including the vehicle sizes) of current public transportation are almost certainly not the optimum sizes, and that smaller vehicles are likely more optimal once we eliminate the need for drivers and move to a highly communicating world.
I believe there are strong arguments that while shared travel is beneficial, we actually have too much of it in most transit systems, and not enough in private cars. That the 'shared' future is one of van-sized group vehicles with a mixed fleet of more personal cars with 1-4 seats."
1. I would theorize that the economic benefits of transportation are directly proportional to how fast the service is, and how much people it can move. As an estimation, I'd say cars are limited to 60 mph while high speed rail would be limited to 200 mph.
2. With fixed-path mass transit, the problem of traffic is greatly simplified. Self driving cars may solve a labor issue but they won't solve the traffic problem: cars routing themselves selfishly will not result in the lowest travel time possible. You can argue that a central coordinator can exist to coordinate routes, but if there's an incentive for a person (not necessarily a driver) to deviate from their set path, why wouldn't they take it?
There’s no efficient way for busses to get people close to their door. It makes no sense. Mass transit should relieve congestion on major routes, it’s not an efficient solution to the last mile problem.
How is one supposed to move 3 horses across the state using rail and buses?
What about my parents driving to surrounding lakes to kayak? Just cary that on a train that for some reason actually goes to these lakes in the middle of nowhere?
What about hauling larger personal watercraft? Just don't do that? Store it at the body of water you've chosen to exclusively use your boat/jetski on?
Rely on Amazon to handle the logistics of shipping hay for your livestock?
Maybe these are all just cultural things that need to disappear and be replaced by magical super-efficient factory farms or not replaced at all. But there are other ways of life than just living in an apartment and commuting to an office.
This conversation is about the 62% of Americans that live and work in 3.5% of the land area. (Source: https://www.census.gov/newsroom/press-releases/2015/cb15-33....)
There’s no reason for rural America to act like these ideas are crazy, nor for urban America to act like these problems are universal. So often these two sides act in opposition to each other when they really have no conflict with each other - except perhaps that both side has a hard time relating to the other lifestyle.
FWIW those 62% live in "incorporated paces" or cites, which would include my town (as long as I'm understanding that correctly.) Down under "new incorporated places" it mentions Sandy Point, TX, population 200. It's crazy that even including little towns like that you're still only talking about 3.5% of the land.
I think it also speaks to how relatively solvable some of the urban problems are in the sense that we could be intentionally building new cities (not just attached suburbs) if we wanted to, and that would - if it worked - take some of the pressure off the places that are struggling right now.
For whatever reason America just stopped doing that after WW2. Maybe it’s time to start again?
Regional transportation is an entirely different question. Is there a rail line running through the small town in Texas where you grew up? I bet there is, and I bet there is the remnants of a passenger station too. What would it mean to your small town if people could easily commute by rail to the bigger city 45 minutes away? I bet people would be more interested in living in that small town.
And that is just going to the nearest larger city... What if you want to go to a different city? Or a smaller town? Or a random place that doesn't have a bus station?
I predict very few people would participate.
Despite being slow and expensive, those trains are packed, day after day, with commuters who would rather live out of the city, and not spend 15-20 hours of their week sitting in traffic on the 401. They recently extended two of the trains with additional carriages to handle the capacity.
We've been clamouring for hourly bidirectional service for over a decade and it's only in recent years that the relevant authorities have finally gotten their act together to build out the necessary infrastructure to make it possible (twinning a bunch of track sections, building a bypass for US-bound freight, and electrification).
People do carry their horses to work with them. They aren't just luxury goods.
Maybe people need to take personal responsibility and not spend money as if the government will make things easier for them and solve all problems. I am not claiming this is possible for everyone there are truly people who need help. But the vast majority of people can try to live in a way that they can have personal transportation (which has been around how long now?) and is reasonably enough priced currently.
Separately back before autos wouldn't people have had the same issue with horses or with even machines that they needed to run their lives or their business? Isn't it part of being human to plan for breakdowns of various things that you need that you can plan around failure?
Whenever someone says this I assume the opposite; America is not a place of uniquely irresponsible people.
People have had cashflow issues since the dawn of cash, often mitigated by complex family informal credit systems.
I don’t think the killer market for driverless cars in America in the first place. Places like China that have all that (commuter rail, subway) but lack places for new roads and parking would benefit greatly from optimizing their limited road infrastructure. They already have people who are used to taking taxies everywhere, now make the ring roads automated only to triple their capacity.
Europe, Asia, can be benefit greatly from driverless cars because they have a different situation from America.
I think driverless has the potential to be both - we can have a bus system that runs dynamic routes based on the users that are going to ride it and runs closer to ideally efficient vs. a fixed route that is likely inefficient.
This doesn't seem to require driverless busses. I can see an argument of why self driving taxis would be helpful -- cost of labor is near zero -- but with busses, the cost of labor is amortized across passengers.
c'mon, who is gonna dream about that ? </cynism>
We also need to focus on the actual problem we're faced with. Cities are hitting limits to growth due to congestion and we can't build more roads. The state of tech right now is capable enough to solve that problem by increasing the throughput of the major arteries of LA / DC / etc.
But my understanding is that at least as far things like subways go, driverless trains aren't a difficult problem.
because a world in which a greater proportion of people travel via multi-passenger vehicles like the buses and trains of public transport networks is a better world because these transportation solutions are 1) more efficient per passenger (does this help w/ reducing carbon emissions?); 2) economically more accessible because passengers pay per trip instead of purchasing vehicle outright; 3) potentially less expensive to operate due to economies of scale (purchasing & maintaining a large fleet of vehicle vs a single vehicle) and higher equipment utilizationl; 4) perahaps encourage more social connection amongst passengers; 5) passengers do not have to focus on driving and can "recover" the time that would have been spent driving a car doing something else.
I ask because I think the endgame for self-driving car companies is not so much to sell cars (self-driving in this case) to individual consumers, but rather to to win the zero-sum-game of becoming the largest on-demand "elastic" self-driving vehicle transportation platform; i.e. 80% of the cars Ford makes in 2030 it reserves for its automated fleet, which it dispatches on a per-ride bases to subscribers/customers for a single trip, a day, a month, etc. The ultimate technological opportunity here is effectively "packet switching" for physical transport via the highway network.
While it's possible (and desirable, in my opinion) the fleet of vehicles enabling this automated transport network could be a composition of myriad different vehicles, from many different vendors, contributed dynamically by everyone from families (or groups of families) that own a single vehicle and rent it out when its not in use to the fleets of large companies like ups, hertz, ford, etc...
it seems more realistic that a few giant companies (Waymo, Uber) will be the ones that can scale up a working network fastest and cheapest by building, owning, and operating huge fleets of vehicles made up of just a few vehicle variants.
Anyway, imagine 60% of the cars on the roads today were capable of participating in an automated-vehicle-on-demand transportation network. This scenario seems like it has the same benefits I attributed to public transportation. What do you think - is it an equally desirable solution, or are there other particular benefits to a world with significantly more (conventional) public transportation / less cars that I failed to articulate?
- Substantially reduced traffic on surface streets (a bus or subway is MUCH more densely packed with humans than a road full of cars).
- Encourage more walking and overall fitness
- Improved safety for non-mechanized road and sidewalk users (reduced traffic will do more for pedestrian safety than automated traffic will, I promise).
- Less space required to be devoted to parking and streets, so better land utilization.
So... this got really long. It was just supposed be a quick back-of-the-napkin sanity check estimate of the economics of some rural autonomous vehicle service scenarious and it got a little out of hand. If anyone else also finds this interesting, drop me a line or comment!
Lets also consider the merits of these two transportation scenarios when depoloyed to serve regional/rural populations too. I don't think any of these benefits are significant for the other much less dense half of the US population (but by 2033 or something this fraction will fall to like 1/3 I've heard).
I grew up in Northern Michigan, where 99% of all daily trips were between 5-40 miles and required a personal car because there was no availability of taxis (too expensive per ride) or buses (because population density was too low to marshal enough passengers per bus route). In cities, where density is high and distances are low, buses/trains that batch passengers into one vehicle that's part of an interchange network works pretty well. But in rural areas where the opposite is true, passengers can't be batched together as easily, so buses don't work and an interchange takes too long to transit because of the distances.
A personal car is a necessity for most people living in rural areas because there is no other practical way to get to a job, store, movie, friends house, hospital etc when these locations may be 10-40 miles away. The time spent traveling to these places may be about the same as the commute time for a city-dweller, but rural residents have to do the driving themselves. Aha, so the critical question is: what, if any, features of personal transportation via on-demand self-driving-cars-as-a-service in rural areas might make it competitive with or more desirable than owning a personal car?
If costs per year are roughly equal, then I would love a self-driving car that I could work in. Or self-driving land-yacht office RV for 4-8 people.
Say a personal car costs $3000-6000/year in gas, maintenence, payments, etc. If a person travels 200 days a year and takes 2 trips a day, that's 400 trips a year. This is hard to estimate. To be conservative lets say self-driving-cars-as-a-service would be competitive if it allowed a customer to take 400-800 trips a year for the price of owning a car. That works out to $3.75-$7.5 per trip for the ex-owner of a cheap car, and $7.5-$15 per trip for the ex-owner of a mid-range car. I guess if these are electric self-driving cars, the fuel cost might be $0.50-$2 per day. Hmmmm. If the vehicle cost ~$50,000 and the profit per ride was $4, then it would take ~12,500 rides to break even. If a vehicle can get 10 average fares per day, 350 days a year, thats 700 fares per year.
So actually, the economics are not toally insane. At $4 profit per ride, it would take ~2 years to break even on a $50,000 vehicle assuming 10 rides a day that that take 1 hour, have a rider only one of those two ways, and only have 1 rider at a time. These vehicles might be operating at 20 hours a day though. Not sure they would last 2 years. So... what startups are working on self-driving Land Yacht Office co-working-commuting vehicles?
Now I see a significant portion of the fleet should be something that could pick up at least a few passengers along a route, otherwise there would need to be basically as many vehicles in the fleet as there are riders to support solo rides during commuting hours. $200,000 land yacht, 5 riders, $5/ride: 8000 trips to break even, 10 rides a day = ~2 years. But for a region of 40,000 riders, that might require 5000-8000 land yachts that cost $200,000; so $1-1.6 billion in vehicle costs. Aw man. What's ubers valuation? 70bn? And goog is 820bn? hmm.
Ah so here's the problem. Because of the distances involved in rural trips, it might take 15-60 minutes after requesting a ride for an automated fleet vehicle to travel to the pickup location. That might be ok for occasional travel, especially if trip times are 3+ hours, but it wouldn't be convenient for daily travel. So rural passengers would probably prefer owning their own car, or sharing an automated land-yacht with a couple local other folks, because they need it to be physically nearby most of the time to minimize the time-to-pickup.
In conclusion, InitialLastName's comment leads me to imagine a future in which the dense core of urban areas is served by high-capacity public transit, a significant portion of the central surface streets are closed and used for other purposes, with small-vehicle parking and pick-up/drop-off hubs are sprinkled throughout and around the city that link transport throughtout the wider region and state via privately owned vehicles (for daily commuters) and automated on-demand vehicle fleets (rolling offices) for the ocassional traveler.
Personally, I find it hard to do productive work on public transit, largely because transfers interrupt continuous work and the passenger compartments prioritize density over work surfaces. I think I would be ok paying 2x the fare of public transit for a trip that took 1.5x as long if the vehicle was outfitted with normal office furniture (desks) ~6 passengers. A relatively low-density vehicle for its size, scheduled a couple of hours in advance, providing a hub-to-door trip with no transfers. Ha. Yeah. Self-driving-land-yacht-office-as-a-service company?*
* we could validate it right now... SF <-> Palo Alto, 6 round trips per day per vehicle, 6 passengers per trip, $19 fare, $200,000 vehicle cost, $19/hr for driver, and lets ignore fuel costs. So say revenue of $600/day, that leads the vehicle break-even within a year. hmmmmmmmmmmm.
Well, yeah, it exists, it's called a train, it's magnitudes longer than any "autonomous truck convoy" ever will be, and instead of the billion dollar AI machine learning big data system we have to keep them close together, we use a fricking mechanical linkage.
no the real solution is get over this mythological idea that America is too car heavy. everywhere that people get freedom of travel they go their own transport. what automation will bring is freedom to those who cannot drive so we best better figure out how to get cars talk to each other and which areas will be reserved for automated driving only.
funny you mention Seattle, a city that blew over their recent rail budget by half a billion dollars . plus is it also a city deferring maintenance to hide the costs of their folly.
If you look at how many people go in debt to own cars or the amount of credit on cars, you would see poor numbers there as well. Cars are incredible money sinks. There is a reason why Uber doesn't operate its own fleet.
If public transit billed by value it would easily be kept afloat. But the value proposition in most cities in America just isn't there, because they simply aren't dense enough. It's not a mythological idea. It's simple math.
> funny you mention Seattle, a city that blew over their recent rail budget by half a billion dollars . plus is it also a city deferring maintenance to hide the costs of their folly.
Funny you mention things going over budget because nearly all infrastructure projects in this country do - yes, including roads. This is a problem with our government, not mass-transit solely.
In your very own article, the reasons stated for the price increase have nothing to do with mass transit inherently. They would be true of any infrastructure project.
By way of contrast the Millau Viaduct cost approx. €400 million and the Øresund Bridge cost approx €2.6 billion — and represents an order of magnitude more engineering challenge.
Civil engineering in the US just seems to cost a lot more than in other developed countries.
They have freedom to travel, but not freedom to live where they want. The United States has entirely banned dense housing in everywhere that people want to live, and this consequently makes transit impractical. Get rid of zoning restrictions on housing, and we'd be in a much different place.
The second disregard, that well... there are airplanes.
The first, that driverless busses that creep at 30km/h and do emergency brake every every 10 minutes thanks to overattentive collision avoidance system also been around for decades.
But then, you effectively get a copy of a rail system, only without rails. Saudi experimental city once played with the concept
And the same parallel for driverless car, except for even when well executed it is kinda of arguable usefulness/value.
Are regular cars useful? A self-driving car is just as useful for someone who can't drive: minors, many seniors, many people with disabilities - or people who just don't have a driving licenses, like me.
I expect them facing same issues like trams and trolleybuses - terribly bad experience in traffic.
You only see risky people aggressively trying to overtake and scratching their paint, while leaving the driver no opportunity to move an inch.