[HN Gopher] Visualizing Ship Movements with AIS Data
___________________________________________________________________
Visualizing Ship Movements with AIS Data
Author : bookofjoe
Score : 78 points
Date : 2024-09-22 20:33 UTC (1 days ago)
(HTM) web link (www.beautifulpublicdata.com)
(TXT) w3m dump (www.beautifulpublicdata.com)
| efitz wrote:
| Ok that is pretty amazing. I followed the link to the repo of
| tools and it's crazy that not only can you use automation to
| track and identify ships but also look up their cargo manifests.
|
| The visualization is beautiful but the time scale is too large
| and the maps too zoomed in to identify much in the way of
| patterns.
| TrackerFF wrote:
| I work in this field. Some notes here
|
| "Unlike ADS-B in a plane, AIS can only be turned off in rare
| circumstances."
|
| AIS can and does get turned off all the time. Different countries
| have different laws on this, so you will see very different
| patterns on where AIS goes dark. Sometimes vessels will turn off
| their AIS as soon as they leave some EEZ for international
| waters, sometimes they just turn it off if there's little risk of
| getting caught. As the article says, AIS is primarily used for
| safety measures - and is what it was designed for. But most
| developed countries have since started to use AIS for maritime
| surveillance, and have in the later years codified in laws /
| regulated how vessels of certain size must use their AIS
| transponders. VMS, which has a much greater capability in sending
| out data (catch logs, for example), has become more and more
| common.
|
| It is also difficult to know whether or not a ship deliberately
| turns off its AIS transponder, or if it is in an area with poor
| coverage. AIS data is collected via satellites and land based
| antennas - some data providers also provide signal strength to
| give a clue. When presenting AIS data in court, we never, ever
| say that a vessel has turn off their AIS transponders - only
| present the data and some probabilities.
|
| Bogus AIS tracks is a massive, massive problem. I frequently
| analyze areas where as much as 100% of the "traffic" turns out to
| be junk. They _can_ be pretty easy to filter out if there is some
| pattern of randomness, and a lot of the junk is just that - the
| AIS messages are just scattered uniformly around the globe,
| leading to impossible speeds. 99% of junk tracks are either
| registered as Chinese, or some micro-nations with little to no
| maritime activity.
|
| Spoofing is also a problem. In my work, we've been analyzing the
| Russian tanker movement closely for a couple of years now, and
| some of the spoofing has been really sloppy - so sloppy that the
| public quickly noticed. This is my favorite, because it was so
| low effort: https://gcaptain.com/blink-and-youll-miss-it-russia-
| deploys-...
|
| (if you look at the data, you'll see that they just spoofed the
| AIS messages with a constant speed for two week straight, going
| through the exact same positions multiple times).
|
| With that said, the average of the data gives you nice patterns.
| There's a ton of variance due to all the fake data, missing data,
| faulty data, etc. - so in a day-to-day job a lot of your work is
| to figure out whether or not the dataset is a valid one.
| brk wrote:
| What's the motivation for the bogus data? Just adding noise to
| obscure real vessel activity?
|
| Similarly, the entire AIS tracking system needs a layer of
| authentication. I ran an AIS station for several years that fed
| data to Marinetraffic, and I could have stuffed whatever data I
| wanted into their database with no apparent restrictions or
| checks.
| mschuster91 wrote:
| > What's the motivation for the bogus data?
|
| Some people just want to see the world burn. Vandals and
| trolls basically. IIRC there have been multiple incidents of
| people spoofing genitalia shaped tracks (on top of an even
| larger multitude of _actual_ pilots steering air and water
| craft to draw genitalia). Easy enough to do, all you need is
| a transmitter (which are getting very cheap these days
| thankfully) and something that can output a stream of fake
| GPS positions in NMEA format on a serial port.
|
| Others run broken or misconfigured equipment and either don't
| know, don't care or do it intentionally with plausible
| deniability (Chinese ships in non-China waters stealing fish
| are a _huge_ problem [1]).
|
| At least a lot of the dumbasses and breakages can actually be
| filtered - AIS Class A transmitters on large ships can reach
| up to 100km, AIS Class B transmitters (everyone else) about
| 40km, ADS-B on aircraft about 500km. That means even as a
| solo operator of an AIS/ADS receiver, you can safely discard
| anything more than these distances away from your
| coordinates, it's likely garbage, just like stuff that moves
| at impossible speeds - you won't get a boat at more than 100
| km/h, a small prop aircraft like a Cessna at more than 500
| km/h (unless it's nose-dive crashing), or a civilian airliner
| at more than 1.200 km/h.
|
| [1] https://www.seaaroundus.org/industrial-fleets-operating-
| indi...
| mcculley wrote:
| Some do not want their competitors to know where they are.
| Some vessels are fishing in illegal waters or violating
| sanctions.
| causal wrote:
| Thanks for sharing!
|
| I'm trying to understand what bogus AIS tracks means - are you
| saying these are AIS messages injected into the network
| somehow, or actual AIS radio signals that somehow ended up
| transmitted from the wrong place?
|
| I'm assuming the former, which sounds more like an
| authentication problem?
| TrackerFF wrote:
| Sure, I'll give you a scenario.
|
| Let's perform a search for vessels that have been inside some
| search box / geometry in the past 72 hours. In this case I'll
| do a search south of Africa. The search area looks like this:
|
| https://i.imgur.com/z2LTC7k.png
|
| Which will return the following list of vessels:
|
| https://i.imgur.com/cKm1YlO.png
|
| And if you plot the tracks for that area, there won't be
| much:
|
| https://i.imgur.com/Max72J4.png
|
| Now, let's save the ships (by MMSI number), and search where
| those ships have been in the past 72 hours?
|
| Here's the resulting tracks:
|
| https://i.imgur.com/SqS8GH9.png
|
| ---
|
| So, what happened here?
|
| 1) As you can see from the list with vessels, some of the
| rows only have a MMSI number - nothing more. A MMSI number is
| not static/constant, whereas an IMO number is - think of the
| IMO number as the VIN number of a vessels, it will following
| the ship from build to scrapping.
|
| So a big problem is that we're not just looking at one track,
| but every AIS message that has that MMSI number associated.
| And generating those can be trivial.
|
| Anyone with some AIS transponder can send out AIS messages
| with some MMSI number can achieve that
|
| 2) Because these AIS messages are being spammed at random,
| they will still appear in geometry / area searches. Usually
| just as one point / message, but you still have to filter
| them out.
|
| Now, with that said, junk like that is easier to spot -
| because they usually do not have much more data than some
| MMSI, and maybe a flag state.
| causal wrote:
| Nice, thanks for sharing such a complete example. So I take
| it there isn't really any kind of verification of AIS
| message validity. I suppose that would be difficult to do,
| so it makes sense that AIS would be easy to spam.
| TrackerFF wrote:
| Indeed, here's a good and short article on the topic:
|
| https://globalfishingwatch.org/data/spoofing-one-
| identity-sh...
|
| Global Fishing Watch is one of the orgs that publish good
| and accessible stuff on this very topic.
| thenthenthen wrote:
| Another possible example are the 'Shanghai Crop circles' [1],
| but these are attributed to gps jamming, why not ais jamming?
|
| [1] https://gpspatron.com/gps-spoofing-circles-in-china/
| scirob wrote:
| wish those images had individual captions instead of us needing
| to do work to figure out which image goes to witch caption
| djoldman wrote:
| data: https://hub.marinecadastre.gov/pages/vesseltraffic
| CalChris wrote:
| > AIS can only be turned off in rare circumstances
|
| This is SOLAS convention which has no enforcement offshore.
| Within the navigable waters of the US and up to 12 miles
| offshore, there is enforcement [1]. But elsewhere, there are
| plenty of reasons to turn AIS off [2].
|
| The US Navy generally operates with AIS off, although they've
| started to broadcast AIS reports in high traffic areas [3].
| Private boats can turn AIS off at will in international waters
| [4]. It's also possible to spoof AIS reports [6]. Then there's
| avoiding piracy in Gulf of Aden or Strait of Malacca.
|
| [1] https://www.marinelink.com/news/us-coast-guard-alarmed-
| marin...
|
| [2] https://www.darkshipping.com/post/ais-off-dark-shipping
|
| [3] https://news.usni.org/2017/09/19/deadly-collisions-navy-
| will...
|
| [4] https://www.theguardian.com/world/2022/may/28/revealed-
| russi...
|
| [5] https://archive.is/5lvTC
| mcculley wrote:
| I am personally aware of many commercial vessels that turn off
| their AIS to avoid leaking information to local government and
| competitors. My experience is limited to the U.S. Atlantic and
| Gulf coasts and the Caribbean.
| entaloneralie wrote:
| It's very eery to sail through a 100 strong fleet of chinese
| squid ships and seeing them all turn off their AIS as they see
| you approaching.
| is_true wrote:
| where did that happen?
| causal wrote:
| Huh, that doesn't really make a lot of sense to me. Why turn
| them off after being sighted? Because AIS tags aren't
| matching reality perhaps?
___________________________________________________________________
(page generated 2024-09-23 23:01 UTC)