[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)