6 ms·
Australia will update its latitude and longitude after moving 1.5m
- mmahemoff 10y agoWouldn't this also be relevant to boring old human-driving cars and loads of other applications?
- maxerickson 10y agoHumans don't really navigate at a meter scale when they are driving. A nav system giving directions 19.9 seconds before the turn isn't going to be noticeably different than one doing it 20 seconds before the turn.
- varjag 10y agoOne would really hope the steering in self driving cars is derived from sensory input and not nav system waypoints.
- ceejayoz 10y ago"Other applications" is general enough to be obviously true, but what would "boring old human-driving cars" need sub-meter-precise GPS for?
- sparky_z 10y agoI would think that normal GPS maps intended as a navigation aid for humans don't need ultra-high precision. Even an offset of a few meters wouldn't be enough to confuse it about which road you're on and the direction you're headed. What's the worst that could happen? It tells you to turn right in 100 ft when it's really 105 feet? A human driver wouldn't even notice.
- wongarsu 10y ago>Self-driving cars, for example, must have infinitesimally precise location data to avoid accidents. With GPS having at best three metres accuracy? It's nice that they keep their data up to date, but I'm not so sure self-driving cars have anything to do with it.
- tachim 10y agoDifferential GPS can hit ~10cm level accuracy.
- djsumdog 10y agoThose are also $3k ~ $5k and have to be configured to use a GPS base station. A professor once let us borrowed his. It had a full backpack and was supposedly accurate to 3cm using the airport as our base station.
- extrapickles 10y agoDGPS is much cheaper than it used to be as it is a feature of most GPS chipsets, though its poorly documented how you enable it. You also need to provide a serial data link between the two GPS units (for most applications this is over RF). Also, from my understanding, you only know the relative position of the two stations accurately, not the absolute world location.
- ghaff 10y agoYes but I'm not sure why that's relevant to most autonomous vehicles or typical commercial/consumer GPS use.
- planteen 10y agoThat doesn't make any sense to me either. To avoid accidents - do they mean so the car doesn't drive off the road when making a turn? Surely the car uses visual feedback as well.
- asafira 10y agoEven with a +/-3 meter accuracy, a systematic 1.5 meter shift still makes a difference.
- the_duke 10y agoIf a self driving car can't deal with a 1.5 meter misalignement by relying on other sensors (camera, radar, laser), then there are way more serious issues...
- planteen 10y ago>Self-driving cars, for example, must have infinitesimally precise location data to avoid accidents. I'm sure surveyors, machinists, and physicists would be interested in how "infinitesimal" location data can be obtained. This article doesn't really spell out what is going on. Did Australia use some other datum than WGS84 and now they changed it?
- SilasX 10y agoI think they meant to say (or should have said) that SDCs need more locational precision than GPS provides. I was going to justify their phrasing as being easier to understand, and accurate enough for a lay audience, but ... mine is still better, IMHO. I would have said: "Self-driving cars, for example, need more precision than GPS has to avoid accidents."
- verandaguy 10y agoIsn't this why many self-driving cars employ cameras, IR, LIDAR/radar etc.?
- jonknee 10y agoYes, I don't know of any self-driving car that uses GPS alone. That sounds problematic for many reasons (enter a tunnel and you're dead!).
- komali2 10y agoJust take a look at your google fitness tracker map after a run to see why this could be bad. Oops, we're suddenly a block north for some reason? Let me just veer sharp right to fix this!
- verandaguy 10y agoClearly, you can just avoid tunnels -- nobody really _needs_ to leave Switzerland when they're already there, right? :P
- djsumdog 10y agoReminds me of the great Melbourne earthquake of 2012. And by great I do mean my flatmate came to my room and said, "Was that an earthquake?" "um....maybe. I .. I think so?" My fizzy drink fell over. It was a mess. Never forget!
- Analemma_ 10y agoHeh, the East Coast of the US had a similar experience in 2011, leading to great jokes like http://wp.patheos.com.s3.amazonaws.com/blogs/friendlyatheist/files/2012/08/Never-Forget.jpg http://wp.patheos.com.s3.amazonaws.com/blogs/friendlyatheist...
- Stratoscope 10y agoLucky you! I was in Los Gatos in 1989, near the epicenter of the Loma Prieta earthquake. Two dozen homes in our neighborhood knocked off their foundations, major damage to the downtown buildings (from the way some of them looked, it was a miracle they didn't collapse), and this was nothing compared to what happened on the Cypress Freeway in Oakland. I suppose you get used to it, because now we hardly notice anything below a 4.0 or so.
- jjp 10y agoSlightly more detail - http://www.abc.net.au/news/2016-07-28/why-it-matters-that-australias-coordinates-are-moving/7668014 http://www.abc.net.au/news/2016-07-28/why-it-matters-that-au...
- sctb 10y agoThanks, we updated the link from http://www.cnet.com/news/australia-has-moved-1-5m-so-its-updating-its-location-for-self-driving-cars/ http://www.cnet.com/news/australia-has-moved-1-5m-so-its-upd....
- taneq 10y agoDON'T USE GPS ALONE FOR SELF DRIVING CARS. Not even DGPS. Really. Don't. A ton of metal moving under its own volition really needs to be looking at the things around it, directly, using some kind of vison. Cameras are good.
- pmiller2 10y agoI don't think any sensible self driving car system would use GPS alone. Collision avoidance is pretty important, and GPS only stops you from crashing into buildings and mountains. :P
- posterboy 10y agoNot to mention reports of tracking in the mountains, where position was off several hundred meters, because of signal echoes.
- MiguelHudnandez 10y agoIt would be criminally negligent to use GPS for any control decisions in a self-driving vehicle. Aside from wishy-washy long term tasks like what route to take, or which street has traffic, GPS simply cannot be used. GPS can fail, be jammed, or lose accuracy. I can't imagine there is any risk of any production vehicles using GPS alone. Certainly GPS matching physical reality will be helpful, but no car is going to drive on a sidewalk and plow through fire hydrants because GPS was off by 1.5 meters. Before a car can enter a physical space, it should verify with at least two local, independent, physical sensors that the intended space is unoccupied and that no detectable objects are on a trajectory to occupy that space. Our brains do these calculations very easily and with minimal conscious thought. However, our brains' attentiveness and accuracy leave much to be desired.
- notahacker 10y agoI think the potential issue with GPS being a couple of metres out issue is less cars driving through fire hydrants and more the car driving down what GPS thinks is the correct route and looks like a nice open stretch of road to the computer vision, but is in fact the wrong lane...
- perseusprime11 10y agoIn a few million years, it would have moved so much that it will cut down commute times from U.S to Australia. Obviously riffing here but I am curious where it will end up in a few million years.
- gkya 10y agoSee this comment for a video simulation: https://news.ycombinator.com/item?id=12189301 https://news.ycombinator.com/item?id=12189301 It'll seemingly hit Asia in the guts while part of siberia leaves with Korea moving east. And my poor little city's gonna get squashed and move north to replace some cold Russian city...
- guelo 10y agoHere's an animation showing Australia crashing into Indonesia in 25 million years and into China in 50 million years. https://www.youtube.com/watch?v=2It3ETk2MGA https://www.youtube.com/watch?v=2It3ETk2MGA
- ungzd 10y agoHow is it related to GPS? Usually GPS devices return coordinates in WGS84 which will not be changed. They're just updating old maps which was based on their local datums?
- nraynaud 10y agothey are basically changing the official WGS84 coordinates of a number of famous Point Of Interest. The way they do it that they measure the position of some reference point in WGS84 on every tectonic plate, then they measure the relative position of a bunch of point of interest from the local reference point. At a regular interval, they re-measure the actual coordinates of the reference point in WGS84, and that automatically moves all of its dependent points of interest in WGS84. Tectonic plate movement is why WGS84 is not always the best way to store your points of interest, keeping them in the local reference also helps propagation when an update occurs.
- 205guy 10y agoWow, 7cm/yr seems fast. I thought I read once it was 1-2 cm/yr, about the same as your fingernails grow (it was a rule of thumb). Probably different plates are moving at different speeds. I think it's exciting to see examples of continental drift on the human scale. Usually, all the models talk about Pangea and hundreds of millions of years, and our brains have trouble seeing the context. But I wonder if we don't need a more flexible system, maybe like leap days and leap seconds that get added much more often. Or just have a global reference and an offset for the local datum that can be updated yearly.
- gkya 10y ago"1-2 cm/yr" IIRC that's the speed South Africa and South America are moving away from each other.
- a3n 10y ago> Australia will update its latitude and longitude after moving 1.5m I imagine a guy in a Commodore's outfit climbing a ladder to some bridge (but on the land), making an observation, swatting the dust off a log book with his hanky, making an entry, closing the book, looking around with satisfaction, and then climbing back down the ladder.