Back then, as now, the big issue is that driverless vehicles have to deal with human drivers. Remove humans from the road, and the challenge gets exponentially easier.
First we remove human drivers. Then we add electronic signalling systems, to make the driverless vehicles work in all environmental conditions. Then we add central control. Then we add X. Then we add Y.
Finally, we have reinvented the railroad. Amazing. I feel like the fact that we don't have automated railroads (in practice) should tell us something about the viability of the self-driving car. The problem for autonomous rail isn't even technical anymore, it's regulatory.
You forgot the part about spending trillions of dollars on dedicated tracks that connect everyone's home, workplace, and shopping and entertainment destinations. Fortunately, we already have those tracks; they're called "roads."
Your incredulity at the fact that trains still have human operators involved in any capacity at all, on the other hand, is spot-on. I'm with you 100% there. Every time somebody wrecks a train because they were asleep or distracted or texting or otherwise had better things to do, there is always a giant thread on HN full of postings from train experts explaining why the world would end if humans weren't involved in fixed rail operations.
It is indeed ridiculous that the same arguments -- whether they're valid or not -- never seem to apply to autonomous road vehicles.
I just don't understand, if the vehicle is totally autonomous how is it going to get down my friends mile long driveway that is little more than two gravel tracks with grass in the middle and find a place on his grass to park while missing all the ruts.
I hear all this talk about better infrastructure, however what about the places in the middle of nowhere.
I see more of the Tesla, with self drive being in the future. Kind of like the movie demolition man.
"I just don't understand, if the vehicle is totally autonomous how is it going to get down my friends mile long driveway that is little more than two gravel tracks with grass in the middle and find a place on his grass to park while missing all the ruts."
We could do that in 1985. We used to test our DARPA Grand Challenge vehicle (a Polaris Ranger ATV) at a horse ranch, and watched it climb dirt hills with ruts and avoid ditches. You profile the terrain ahead with a LIDAR and choose a driveable path. You can detect potholes, dropoffs, and ruts. You build a height field from the sensor data and run a planner.
The hard part is out-driving the LIDAR range and going faster. You can't profile terrain more than tens of meters away; you're looking at the ground at an oblique angle and can't see holes and drop-offs. We didn't have an answer for that. Stanford's team did; when the near part of the road profiled OK, and the far part of the road looked like the near part of the road, they'd assume the far road was drivable and speed up. We overdesigned the off-road capability and went too slow. But we didn't crash. The vehicle was eventually donated to UC Santa Cruz, where it was used for various projects for years.
The remaining hard problems all involve other road users.
> Your incredulity at the fact that trains still have human operators involved in any capacity at all, on the other hand, is spot-on. I'm with you 100% there. Every time somebody wrecks a train because they were asleep or distracted or texting or otherwise had better things to do, there is always a giant thread on HN full of postings from train experts explaining why the world would end if humans weren't involved in fixed rail operations.
I figure it's probably a low-rent version of the reason that the Space Shuttle had to be landed by people: just like astronauts wanted to feel useful, so too do the rail engineers.
> I figure it's probably a low-rent version of the reason that the Space Shuttle had to be landed by people: just like astronauts wanted to feel useful, so too do the rail engineers.
I assure you that is not the reason the Space Shuttle, Trains or Planes require a human operator. It is the same reason cars do -- automated control is not sufficient that anyone is willing to put lives at risk to the automation. Especially when something abnormal could happen (most likely in a landing).
That's a good point. When you have full control of design from the ground up it is reasonable to automate. I meant trains on existing complex tracks that were not designed from the ground up for automation. There are not trains navigating legacy systems, because that requires more uncertainty handling than our current automation is capable of. When the track and switching system and track vicinity are all controlled and designed for automation then it is a significantly less complex problem with respect to automation.
Easy. Bike roads. Separate roads for bikes. Layout is much less than for cars, can overlap with pedestrians for last mile, and takes up signifigantly less space and can be used duel purpose with other slow vehicles.
It's also possible that cyclists can adapt to sharing roads with robots. Perhaps by RFID-tagging their bikes, or wearing something LIDAR-visible. At least the autonomous vehicles aren't deliberately hostile, and conflicts between cyclists and drivers over road rules designed around cars, not bikes, might be restructured.
The knock-on effects are more interesting, in some ways. A nation where shared, autonomous vehicles are normal would likely have a totally different set of considerations in urban planning. Cars and spontaneous roadtrains can take the place of light rail and urban transport. Inner-city parking might disappear, and commutes could be much longer. You can imagine two effects: suburbs could really spread out for those who want space and don't mind a couple of hours each way reading the newspaper or snoozing, and cities can concentrate and build way up as you no longer have to build parking space into apartments. A more densely populated city with fewer dedicated drivers might have more political reasons to become bike-friendly.
I don't know a lot about the topic, how well do the current autonomous vehicles handle roads not currently on GPS , I.e private drives and such.
I now understand it can handle the terrain, thanks for the explanation that is pretty neat they could do it thirty years ago, just curious when I tell it to go to a house in the middle of a 20 acre property with multiple private roads, and I use the term road lightly, if it is actively going to search for a path or if it will have to have manual intervention.
I have seen google cars handle well mapped streets, just curious.
The humans are the issue today because they are still the dominant force on the roads. When these autonomous cars start seeing other autonomous cars, then it gets interesting. Android and Apple cellphones don't play nice together. I wonder what happens when the Google car looks to overtake the Apple car.
Will people pay more for a more aggressive car, for one that perhaps gets you there that little bit faster by jumping stop signs and squeezing through smaller gaps in traffic. Will the Apple car refuse to let the Google car merge? These things will find new enemies.
I remember growing up believing flying cars were just around the corner. As today's children get their driver's licenses a decade from now, they'll think about when they grew up believing self-driving cars were just around the corner.
There is a significant difference between self-driving cars in 1960, flying cars and self-driving cars today.
Flying cars are doable, and there are several prototypes of a flying car. However, flying cars aren't safe and aren't cost-effective.
Self-driving cars in 1960 require a significant infrastructure to operate, and can't operate with human drivers. (Edit: Not to mention they don't exist)
Self-driving cars today work, and are cost-effective, and it's likely that they will be safer and create a much better traffic ecosystem than humans ever will.
"Self-driving cars today work, and are cost-effective, and it's likely that they will be safer and create a much better traffic ecosystem than humans ever will."
Self-driving cars ONLY work on the immaculate freeways of California that they are almost always (unfairly) tested on. Any precipitation renders the LADAR sensors all but useless, and many everyday driving scenarios like "left turn onto traffic" and "no lane markers" are still are far from being solved yet.
Any real improvements in the technology will be the result of fundamentally different techniques than what the state-of-the-art is currently using (since this is Hacker News: state-of-the-art really is just an "ad hoc" pipeline that looks for things like "lane markers" using things like Canny edges along with a PID controller for the steering wheel actuator, with some other "cheats" like an over-reliance on human-compiled map data provided "a priori").
An end-to-end deep learning approach seems promising, but the current results aren't even usable at this point.
Five years from now, every car (even my Jetta, not just luxury cars) will have the equivalent of what Tesla's "auto pilot" looks like now, but a human in the front seat will still very much be a necessity.
The first step to good self-driving is full profiling of the terrain and road ahead, along with any obstacles. We did that in the DARPA Grand Challenge, and so did almost all the other teams. That's essential to off-road driving. That tells you where you can physically go. That was figured out by 1985. It's mostly a sensor problem.
Next comes recognizing where the road wants you to go (lane markings and such), which works reasonably well now, especially if you have mapping information. That's automatic lane keeping. All the major manufacturers have that working.
Then comes dealing with other road users. Google is putting a lot of effort into this, with some success. See Urmson's video from SXSW.
Looking ahead, we need somewhat better and much cheaper sensors. The rotating Velodyne thing is still too clunky and too expensive. LIDARs that deal with rain and fog can be built; you need to get back more info than just the first return. "First and last" return is helpful; you'll get a solid "last" return from a hard obstacle in the rain, while "first" looks like noise. That technique is used in aerial LIDAR scans to get both the top of vegetation (the first return) and the ground surface (the last return). It's also possible to range gate through fog. Here's some range-gated imagery.[1]
Solid-state flash LIDAR may be the way to go. Units today are still about $60K, but that's a consequence of low production quantity. The custom imaging ICs aren't inherently expensive. There's a startup claiming to do this, but their web site is all hype, no shipping products.[2] (Pro tip: Calling yourself "The leader in 3D sensing" when you haven't shipped makes you look fake.)
Automotive is now using mostly 77 GHz radars. That's almost good enough if you have scanning in both elevation and azimuth. Even at 77GHz, you can see bicycles and people. Really good 3D radar plus vision might be good enough for serious automatic driving. Existing low-end 2D narrow vertical angle radar just keeps you from rear-ending the car in front.
You're making a lot of claims here about the current state of the art of self driving tech. Presumably google has the best right now, so you must be talking about them. Do you have insider information or are you just guessing? Because it seems to me that the google self driving cars are capable of doing much more than working on "the immaculate freeways of California."
edit: Also to be clear the driver assist features seen on cars like the model s are not what most people are referring to when they talk about self driving cars here.
"Also to be clear the driver assist features seen on cars like the model s are not what most people are referring to when they talk about self driving cars here."
Just saw your edit/addition, so I'm going to add more as well:
When did I even conflate them -- even if the state of the art is actually "driver assist" as you've described?
Google’s self-driving cars aren’t “fully autonomous”. Without human drivers, they’d lose control approximately every 1,500 miles, according to Google’s own reports.
Can you guess how many driverless miles their "self-driving cars" have driven?
Hint: It's zero.
If you're curious, here is an example of a much more grounded academic talk (read: not a TED talk) from 2015 by a renowned (in the autonomous vehicle research community) Boston (not California) researcher who worked with Thrun that corroborates the issues I've mentioned earlier:
"Do you have insider information or are you just guessing?"
That's a loaded question.
I don't feel comfortable sharing my credentials here (even though they would very much lend credibility to what I am saying), but I can also assure you that I am not divulging any insider information.
However, I claim that my argument still stands without the support of "insider information" -- given that self-driving vehicle research's domain is primarily academia; all of Sebastian Thrun's work (THE founder of the Google self-driving car project and Google X itself), for example, is more or less out in the wild (even if behind walled garden academic journals).
"Presumably google has the best right now, so you must be talking about them."
California roads and freeways are actually not that great, at least compared to other western countries. But regardless, we will see self driving cars adopted first where they are needed the most, cities like Beijing and Shanghai that have lots of cars, no where to put new roads, and a need to optimize the roads they have. The killer app is not road rich America, but the rest of the world that is dying in traffic.
If you noticed, one thing I didn't dispute at all was the "cost effective" part -- LADAR sensors are becoming cheaper and cheaper every year. I remember back in 2008, they were prohibitively expensive (around the same price as a Model S is now).
Now, you can get ones anywhere from $200-$5000. That's why I am sticking with my 5 year prediction for an auto pilot Jetta.
In 1956, GM made a promotional film for self-driving cars.[2] It's a musical. There's also GM's "Design for Dreaming".[3]
The modern version of such ads, from Volvo.[4]
[1] https://youtu.be/xKOdux6Gjno?t=720 [2] https://www.youtube.com/watch?v=F2iRDYnzwtk [3] https://archive.org/details/Designfo1956 [4] https://www.youtube.com/watch?v=bJwKuWz_lkE