Approximately 5 hours later, just as I was about to arrive, the car finally managed to figure out my correct location.
Exciting trip, not a huge fan of Teslas, but their charger planning is really nice. It was very unpleasant to suddenly lose it.
I just genuinely wonder how such a bug can actually occur, surely you'd update the GPS fix more often than every couple of hours. Hard to imagine the car just suddenly couldn't get a GPS fix for hours either. But if it did somehow totally lose the ability to use GPS, the car must have a pretty good dead reckoning system given how well it was responding to my changes in direction.
On a vaguely related note, driving 3000 kilometers through Europe in an electric car was surprisingly nice. Certainly didn't affect the length of the trip nearly as much as I'd have expected, but it was certainly super annoying to try and figure out the optimal rate of travel on the Autobahn. Charging at Tesla's supercharges was vastly more expensive than I expected, the "fuel" costs weren't much lower than what you could easily reach with a diesel car.
It amazed me that Volvo programmed an SUV to disbelieve that it could ever actually leave a road.
I don't think I could have ended up there if I tried in the Golf we were in. Nice try.
My kids thought it was the funniest thing, but it's a good technology lesson.
This seems like an entirely different level of craziness, though.
The explanation I found online at the time was that a GPS receiver needs to download data about the exact orbits of all GPS satellites from time to time. Satellites slowly lose altitude and change their orbits. Up-to-date information is constantly broadcasted by every satellite, but it takes about 15 minutes for a device on the ground to download this dataset.
Most GPS devices do this automatically whenever they get the chance. But if your GPS is somehow unable to stay online for 15 consecutive minutes (bad firmware, faulty memory, tunnels, underground parking lots, etc), it will be relying on increasingly outdated info and drift far off its actual location.
But it's true that neither of those factors accounts for miles of error. That has to come down to either poor sky coverage/signal strength, poor software, or (more likely) both.
There are also private paths that aren’t public roads but are still intended for vehicles.
Google Maps gets it right: it tried to keep you on road, but only for a few tens of seconds. After that, if you are in the middle of uncharted territory, it'll show the marker there.
(This is probably because Google Maps can be used for walking/biking too)
Not that I mind too much, I know how to get around without navigation.
Wish we could put it into a manual mode where you just reset it's position once and then it updates based on wheel encoders & snapping to roads.
You've got it backwards. I was explaining why GPS should take priority over mapping data. Not the other way around.
Please don't do that. The map is simply not good enough and does not have enough context (road quality, terrain, trail difficulty) for anything but very causal activity. Even then I highly recommend to use a proper map, electronic or paper.
It has a lot more map data accessible and you can even overlay National Park Service maps, land ownership, accurate cell service grids, mountain biking trails, weather conditions and things like that.
Disclaimer: Just because you see a route on a map, digital or paper, does not mean it is passable today. Or it may be passable but at an extremely arduous pace.
Only for footpaths in residential areas.
I wonder what class of vehicles is least likely to go off roading?
1. Car entered ferry, loses GPS
2. Car entered dead reconning mode used for tunnels and such
3. Car left ferry, acquired GPS
Then either:
4a. Location via dead reconning vasty disagreed with GPS because the car doesn't know about the ferry's movements, triggering some kind of failsafe.
Or:
4b. There's just a plain old bug in the condition to switch back to GPS and maybe people haven't noticed because you don't get as badly desynced in a tunnel.
>the car must have a pretty good dead reckoning system
Yeah all the pieces are there: accelerometers and gyros for stability control, compass for navigation, and the wheel speed sensors give you exact distance traveled.
Doesn't usually take 5 hours to figure out where it is though. At least not on my vehicles, even the one that's always getting confused.
And maybe the system sucks to get the GPS almanac if badly desynced
What helps is that tunnels don't usually branch, so once you're in it, your path is usually quite predictable.
TomTom maps also have a statistical model of what speed is expected along each stretch of road by hour and weekday/weekend (not 7 individual days, but 2 kinds of days). But I don't know if it uses that to help estimate your expected speed when dead reckoning, it's actually for route planning.
One of my co-workers came up with the great idea of gamifying driving: maintain a real time speed leader board, showing the top ten speeders along any stretch of road! So on every road in the world you could compete with other TomTom users who drove it. Kind of like checking in with 4Square, but more fun and dangerous! TomTom legal did not approve.
I suggested gamifying and monetizing driving with TomTomagotchi, a virtual pet that gets depressed if you don't drive it around enough, begs you to visit interesting landmarks and sponsored points of interest, like driving through McDonalds to feed it, or through the park to let it take a shit, or driving fast enough to make the leaderboard to entertain it. I'm sure Bandai's lawyers wouldn't approve.
I drove in tunnel where the asphalt had undulations today, I'd like to see them try to figure that out!
OTOH creating a dumb user experience for a fairly common scenario, to try to preempt a hypothetical scenario is probably not a great idea. A fun problem to think about, though it may be completely unrelated to this issue.
The Phenom 300 jet can't fly straight without a valid GPS signal: https://news.ycombinator.com/item?id=12519629
Every swept-wing jet has a tendancy to enter a positive feedback loop of rolling and yawing, called a dutch roll, when flying at altitude, so even when not using autopilot, an active system is needed to fly straight and level. If the Phenom 300 doesn't have a valid GPS signal, that system fails.
This won't prevent jamming, that requires a different mechanism, but spoofing should be prevented.
Having done a number of multiple-thousand km trips in Europe in an EV (not a Tesla, nobody buys those anymore) — it's amusing how non-EV muggles think this is somehow an ordeal. It's just fine! There are drawbacks: you do have to use your brain and plan ahead more than you do when burning dinosaurs. But I found that the 20-30 minute stops every 2h really improve how I feel after a day or two of driving.
Agreed about prices: there is gouging going on with some crazy margins. When charging at home, an EV is 2-4x less expensive per km than a gasoline-powered car, but when fast-charging on a road trip the cost of energy is nearly the same.
I honestly started considering this a feature. I am a huge believer in "productive friction" - where some things are intentionally made annoying or hard so that humans avoid them - and this is a really good example.
Range anxiety is a fact, only recently electric vehicles have started to have more acceptable and practicable ranges and also, the charger network is evolving more and more. In winter the ranges are also reduced by about 10-20%.
So I dont think people cant see the "magic", the mass market is just risk averse and doesnt want to get stuck in the middle of the highway with 2 screaming kids in the back seat.
Also now that most European countries (and also the US) have stopped subsidizing EVs, the real costs are shining through, so maybe that 1.9L Diesel engine looks more attractive now again.
And God forbid you come across inclines!
> breaks windows
> sets cars on fires
>> nobody buys those anymore
surprised pikachu
However, about those 20 minutes — this, again, is a misconception. Stop at a busy rest area or a gas station and actually start your stopwatch. If you are a single male, it might take you shorter. If you are with a family, 20 minutes is pretty much the minimum.
Such minimum reasonable rest is what keeps you from killing innocents with your 2-ton weapon/tool.
Since it was offline, the bug was obvious although a bit frustrating - you had to put in multiple waypoints to make it forget its urge to send you on the ferry to Hull when you were trying to get to parts of the South Shore.
I did 2 cross country road trips here in the US (~5000mi/8000km total) and had a similar experience. The nav's automatic charger routing did a great job, and we had 0 issues with charging.
https://teslamotorsclub.com/tmc/threads/map-location-and-rou...
I always figured it has some code to get a quick location, and it assumes that location is very close to the last time it got a sat lock.
I just ignore it for the first day, then it comes good.
The story arc is about a lost airtag that is living it's secret life in Mexico.
The link is for a post where Apple decides the distance to it's location in Mexico is greater than the circumference of the Earth.
Sometimes I just want to eyeball distances, and not plug in addresses and request routes for everything.
I know not everyone always wants it, but a scale is a fundamental feature of mapping! Scale almost defines mapping.
Edit: One reason I ask is because I suspect some technical limitation.
With map simplification and symbolic markers (roads are the same thickness at multiple scales) it can be really hard to figure out distances. I was once going to a less populated area, based on my local intuition I thought the distance between two locations was a maybe 30m walk max if I missed the bus, but it turned out it was a 3h walk.
The post author, Dan Piponi, clearly knows about fractals, but his post raises the question of whether asking Fermi questions in interviews is actually effective. I am skeptical that such questions would have prevented this type of bug.
I suspect the issue stems from small measurement imprecisions accruing over long distances, which is—in my view—tied to the fractal nature of roads traversing natural landscapes.
However, as others have pointed out, it may also be tied to road closures: if closed segments are set to a higher length internally (to discourage routing), these values might be getting summed up blindly over longer distances.
None of these issues would have been prevented by being good at estimating quantities alone.
Apologies again for the unconstructive tone of my previous comment.
> One of my AirTags is inexplicably 1600 miles away in Texas. It's one of my floating ones that I throw into bags when traveling as needed.
> Maybe it dropped out of my bag only to be found by someone who thought it'd be fun to confuse the owner. Or maybe it became e-waste to be dumped in another country.
[0] https://www.reddit.com/r/MapPorn/comments/a2mcjq/21212_km_fr...
I have seen API calls you people wouldn’t believe.
Requests hanging off the edge of the load balancer, Ethernet tubes glittering in the dark.
Latency logs reporting some API calls that took longer than the age of the universe.
Their means and percentiles burning on the shoulders of weekly reports.
Plot points dying at the edge of y axis
... like tears in the rain.
I also love that Find My says my partner was "last seen" at home "8,912 days ago".
Apple just doesn't give as much of a shit any more.
A geostationary orbit is ~26,000 miles. Dollars to donuts, that's where it is.
It intensely frustrating to use maps for direction finding while walking or cycling when the app is telling you are facing one direction when in fact you are facing another.
Even if it had, the airtag location service is not accurate enough to have detected that.
The "fix" is to smooth those details as the straight line distance grows bigger.
Dan Piponi, in his post, called for scale questions in interviews, suggesting that whoever worked on the feature at Apple would not have been hired if they had been asked those types of questions.
I found that position non-constructive and wanted to counter it by mentioning that fractals could hint at why the length was so large.
It is hard to predict how others will read my comment beforehand, and I apologize for not meeting the friendliness bar here on Hacker News.
If you put all of your blood vessels end to end they would go to the moon and back.
The moon orbits too quickly to make this practical
Yes, the distance according to roads can be different from the distance as the crow flies. No, it cannot realistically be 10x the distance when the crow's distance is 2500 miles.