Slashdot Mirror
Self-Driving Cars Aren't Going To Be So Great Until We Make Our Maps Better (theverge.com)
Uber is debuting its self-driving cars in Pittsburgh this month, a move that has many taxi drivers upset. The Verge's Nilay Patel argues that this move should change the way we think about maps and addresses. He adds that Uber is currently unable to pinpoint his home, and often ends up at the door of a "widely different address." Citing the CEO of a "large ridesharing company", Patel writes that this issue is known as the "egress problem" -- "the way we locate buildings on a map doesn't really describe how people move in and out of those buildings." Though there are workarounds and inventive ways to pinpoint your exact address, Patel argues that when we grow reliant on self-driving cars, things will get far more complicated and futile if we don't make our maps and navigation services better. He writes: Driverless cars are one of the ultimate signifiers of the future -- the real Jetsons stuff. And we're so close to making them happen: tons of cars have sophisticated adaptive cruise control that can basically keep you going on the highway, prototypes of true self-driving cars from Google and others are already on the road, and the momentum is only increasing. But maybe we shouldn't hand control of how we get somewhere to the machines until we're entirely sure the robots know where we're going.
This should be something that is only an issue once, or if it is an issue simply have you go to the pick up spot and press a button so they can find it
I live on a fairly busy highway and for some reason Google Maps has the street numbers swapped. I should be odd numbers on the west side and even on east, but they have it swapped. I sent them a message a couple years ago and they replied "You're right, it's incorrect. We'll send an email when it's corrected". It's never been corrected. I've sent messages since, but no reply and no action.
SJWs are the new boogeyman. -Me
There are two ways to solve this problem.
A) Map every single building including ingress and egress.
B) Rely on the user being able to pick a route from the road to the building entrance.
B) doesn't even require the user to actually drive from the road to the entrance. They can simply pick an entrance on a map, or high resolution ariel imagery.
Plus, entrances remembered from other users may be usable.
Umm this article is about AUTONOMOUS Uber cars..
Sort of. That's only half of the problem. The other half is:
* Bad data
* Incomplete or missing data
In California and Washington I've seen incorrect map data both with Google Maps and Apple Maps. Sure a better driver _could_ (and should) be able to work around that but are they forced to in the first place??
The road data should both be:
[ ] complete
[ ] accurate
Obviously construction will cause some of that but it shouldn't take _years_ for a non-moving road to be added to the system.
--
GIMP v2.8 still sucks compared to Photoshop. Where are any of the Layer Effects???
There will always be something not on the map. The AI on the cars will need to be good enough to figure out what to do in many cases or else allow the passenger to manually maneuver the car. That is why many of the plans from google are about driving from a known location to another known location, as it may be decades or more before they'll be able to figure out how to get into and out of a condo garage, negotiate mountain roads, deal with temporary obstacles (the dog is in the middle of the driveway so there's no way to recalculate a route), etc. But pickup from a street corner and get dropped off at a street corner, that's much more doable.
No, the article is about human drivers of a cab company who can't get him to his front door. He then launches into a worthless rant about how self-driving cars need better maps.
If a human driver, using the same map and their set of eyes, can't get the guy to his front door, what makes him think a car programmed by humans will be any better, especially by humans who have never seen the place you're going to?
We will bankrupt ourselves in the vain search for absolute security. -- Dwight D. Eisenhower
Maps aren't going to solve your 'pseudo' problem. The ability for the self driving car to ask questions however, is.
I have had plenty of taxi journeys to obscure places where the cabbie has asked wtf are we going, even in my home town. It's not unusual to ask, and being able to reply will help loads.
Add mapping to that and it should be a bonus on knowing where to go. The article makes it that databases should be queryable for anything, and that's not practically possible. Asking questions back is, and generally gets you places (please excuse the pun). any system that can do this is basically going to earn the $$$$
That's not his point at all, but if a dumb human following software maps and the real world can't do it then an automated car can't either, unless you want it driving on the side walk and through houses.
I think one of the frustrations people will have with autonomous cars is the lack of mind-reading for their specific preferences. Do you prefer to park in a particular section of the parking lot? Do you want to park in the shade today? Which entrance to the store/mall/school/etc do you want to go into today? Even with perfect maps it is not possible to fully know, or easily get at the subtle desires of the occupant.
The autonomous future might be rather frustrating as HAL drives past the parking spot you would have chosen, or that you have to wade through infotainment selections to pinpoint which level of the parking garage you want to go to. How about the joy of finding a way to tell a fully autonomous car to dart into a gap at the airport arrivals/departure scrum? Even having to sit in the drive way for just 1-2 minutes to key in your destination will drive a lot of folks batty. Most of us don't fire up the navigation for any routine bit of travel. Infotainment systems largely suck today, why should be expect them to get better in the future?
Call me a skeptic, but I see this as a novelty that will mostly get shut off in-town where most of the traffic is.
Recently featured on The Last Ship, lol.
No, my point is if the Uber Cab Company wanted to get people to their front door all they had to do in the first place was use GPS. Not the generic coordinates but the actual GPS coordinates. They have to be correct.
But instead of stating the obvious this self-important dreg tried to make himself sound like he knew what he was talking about when he explicitly states he was "talking to the CEO of a ride sharing company*" which is code for, "This is nothing but a surreptitious ad for a cab company".
* Sharing a ride does not involve calling a company to have someone come pick you up to go to a destination of your choice and you paying that person for the ride.
We will bankrupt ourselves in the vain search for absolute security. -- Dwight D. Eisenhower
I was helping a startup in the next town over, who had a package that FedEx couldn't deliver, so I agreed to drive over to the local FedEx office to get the package.
While I was there, the person at the counter pointed out that the delivery person couldn't find the address, and I explained how to go into the parking lot, down and around the building, to the front door of the startup.
It was indeed a weird situation where you can't see the front door from the road, and you had to know beforehand where to go to make deliveries.
The point here is, the FedEx person at the counter typed in my instructions in the "notes" section of their database and then assured me that further deliveries should go through OK.
Will it *really* be that hard to do something similar with self-driving cars? By which I mean, report an error to the company along with the correct data, or manually direct the car to the correct destination and note the error, and similar work around.
I'm not sure this is a terribly important issue. I mean, it sure *seems* like there's a simple solution and the problem will quickly be self-correcting.
When I was in college, and for the first couple of years of grad school, I drove a taxicab. Not a fruity Uber car, but an honest-to-god hack. For part of that time, I drove an actual Checker Marathon, which may have been the finest automobile ever built.
Decades later, I can still find my way around that city (Chicago), to any address and give you the best route. If you dropped me blindfolded anywhere within the Chicago city limits, I could find my way (as long as I was allowed to take the blindfold off after you dropped me). If there are Uber drivers who can't find their ass with both hands, it has nothing to do with maps. Maps? Pshaw. Learn your town and don't rely on the goddamn Google maps for everything. Learn how to navigate by the stars like we did back in my day (only partly kidding).
You are welcome on my lawn.
Let's get real.
Driverless cars make sense if you live in a neighborhood full of drunk adults, or a retirement community, or if you're severely handicapped.
And that's it.
Now stop playing Pokemon GO on your smartphone and running over neighborhood kids.
-- Tigger warning: This post may contain tiggers! --
The system doesn't have to be perfect, only be equal to or better than typical human taxi drivers. Human drivers make mistakes and/or have bad maps also.
Table-ized A.I.
It means the cab company has incompetent drivers.
That's not exactly the point of the article. Like he says in the very article:
For the moment, it’s a pretty minor issue - the easiest solution is just for the drivers to call the rider, and it works itself out.
And then in the very next sentences he points out that this solution isn't possible for autonomous cars.
Technology is not the solution to human incompetence. Better humans are.
In a sense, you're right. The solution to this problem isn't necessarily better mapping, it's better addressing that is less arbitrary and error-prone. If you want to go to a certain door then that door should have an address that makes sense based on its location on the street.
"Our two-party system is like a bowl of shit looking at itself in a mirror." - Lewis Black
Sure, self-driving cars will frustrate us in new ways and it's always great to address potential issues prospectively.
However, just because you are used to all the frustrations and inconveniences of the current system and you're just thinking about the annoyances of self-driving cars for the first time doesn't mean the new annoyances are worse than the old.
Those of us who live on obscure streets or on divided streets have had to talk taxi drivers, friends and delivery people to our houses for years. GPS has made much of this easier but the whole complaint here is that google doesn't know exactly where every address is.
Yes, self-driving cars remove the option of talking the driver in despite the GPS error. But in the long run we are better off if we are forced to learn once that "Address X" will bring any visitors to our home rather than having to give them all directions. Not only is this an easy simple fix it actually forces us to do what we should have done a long time ago.
If you liked this thought maybe you would find my blog nice too:
"This thing that already exists will never exist!"
Never change, Slashdot.
Reliance on maps is never going to cut it. All maps have errors and the cost of updating to maintain all current details is high and never done. What we need are many advances in AI algorithms like the ability to ask for directions from the occupants and translate that contextually into where the vehicle needs to go, perhaps using street view information and vision algorithms.
Thousands of little problems like this will plague autonomous vehicles and it will likely be two at least decades before they function the way popular perception depicts them.
There is already a solution to this problem. Whether it's the best solution remains to be seen, but it seems like a great step in the right direction to me.
This was posted on slashdot many months (over a year?) ago.
http://what3words.com/
Every 3x3m square in the world has its own unique 3 word code. So you can give people the 'address' of your garage, or your front door, or the lampost outside the park down the road, without having to provide the exact GPS co-ordinates.
While that is a start there are still numerous problems.
1. Where is the install script? I guess someone was too lazy to include:
2. Why isn't this functionality a plugin and not native to GIMP like Photoshop??? Proof that the GIMP devs are out of touch with reality.
3. Importing PSD files with Layer Effects are still broken.
What's the point in learning how to orient yourself to your environment without using your short-battery life phone? You're joking, right?
When the zombie apocalypse comes and EMP weapons have wiped out your GPS, it will be fun watching sweet summer children like you stumble around blindly.
You are welcome on my lawn.
So that the vision systems of the cars can help them stay in their lane.
You know what is an easier solution? Be prepared to walk the last few yards.
You do realize that these anonymous responses, which have cropped up fairly recently, give me a chuckle and charge of energy, yes?
I must be doing something right :-)
Why not the Russian ones?
http://www.popularmechanics.com/military/news/a16530/these-soviet-maps-were-the-top-secret-google-maps-of-their-day/
Sort of. That's only half of the problem. The other half is:
And one you missed, correct data for an inappropriate purpose. For example, data that has the location corrected to take you around the back of a building to the carpark entrance isn't very appropriate for an Uber driver/self-driving-car that needs to pick you up from the front door.
If a driver brings you from point A to B in exchange for cash, it is a taxi service.
How hard can this be... when you request a pickup, the self-driving car begins driving to you, and a pin is placed on the map where it *thinks* it should pick you up. You review the pin, and move it to whatever door / street corner you want. The self-driving car then re-routes to pick you up where you placed the pin.
On future visits through the same provider, it will remember where you dropped the map pin and default to picking you up there. If you choose to move the pin again (say, you're coming out of a door the other side of the building) it then obeys that, and the app prompts you to give "friendly" names to those two points. Next time you choose to be picked up from the same vicinity, you are presented with a simple dropdown that offers you three options:
"The Green Door"
"The Other Side With The Cafe"
"Choose Another Location"
Am I missing something that makes this problem unsolvable with today's tech?
"And we're so close to making them happen". No, we're not. Stop believing the hype, we're literally still in the infancy of self driving, the problems that still need to be addressed are numerous and not easy.
"What NASA could teach Tesla about the limits of autopilot" is a great read. http://www.pbs.org/newshour/rundown/time-seem-fly-faster-age/
Good point!
Yes, context is important.
I work in a building where the entrance faces a pedestrian walkway. It is a half-block walk to either of two streets. The problem is generic to the whole multi-block walkway.
With Uber I can position the cursor on the map to identify the pickup location. This is not that hard, and there is no reason for the ride-sharing company to not learn from it, and sell information to or trade information with a mapping company.
No, Mr Patel, a significant number of your potential market/ audience will not become reliant on your product until after map and navigation services are better.
And incidentally, some of us are used to spending time where you don't have electrical power or any mobile phone/ data signal (Iridium excepted, all 9600 bps of it) , and the magnetic field is sufficiently variable that your compass is decidedly dodgy. We carry multiple different location-determining technologies that won't be disabled by simply running out of battery life. Because, that like kills people, you know?
Birds are not dinosaur descendants;birds are dinosaurs, for all useful meanings of "birds", "are" and "dinosaurs"