AIS Position Error?

[Larry]

I agree with the sentiment though, NMEA 0183 is old and slow. I would
prefer WiFi. BT networking is pretty primitive and the port-count
limitations on most devices make a it difficult to create a flexible
multiport system. BT still has the power advantage though, and I hate
wasting Amp-Hours (thus the Pocket PC for my logging application
instead of an always-on laptop).

Wifi Ethernet transceivers are so cheap there's 4 of them coming out built
right into tiny telephone handsets to hook Skype-in-EEPROM directly to the
router, eliminating running Skype on desktop or laptop completely. I bet a
fully Ethernet-compliant chipset and radio transceiver now costs OEMs less
than $20, maybe even less than $10. Amazing what Asians can do with slave
labor.

The Chinese company making Ipods for Apple was investigated for child labor
abuses. The investigators didn't find anything wrong. They were paying
the help nearly $100/month for six, twelve-hour days a week.

 

[Kees Verruijt]

450ma x 3.3V = 1.485w at 13V = .114A x 24 = 2.74AH which isn't much a
significant part of a small 330AH house battery, even smaller on the two
banks of beasts we drag around in the stern. 2.7AH a day isn't much of
an issue, is it?

It wouldn't if it were the only WiFi network hw on the boat;
unfortunately one is not very much use by definition you need two.

The power draw starts to add up if you are going to replace all
NMEA/Seatalk/whatever-bus drivers with WiFi. 3 sensors, 6 displays, 1 PC
adds up to 27 Ah... that's no longer neglible.

 

[Kees Verruijt]

The only problem is that Wifi is only cheap if you buy 100k chipsets or
more. Most, if not all Wifi chip vendors won't talk to you or send you a
datasheet if you don't sign a contract first for at least 100k units. So for
low volume applications, Wifi is very expensive. The Wifi modules I use cost
around $100 each, even in 100's.

Meindert

Meindert,

I assume you use these modules in your non-marine products right?

To me having a WiFi multiplexer doesn't make sense, as there is no
"serial profile" defined on top of Ethernet, by my understanding. Even
if there were in my opinion it would more sense to have an (wired)
Ethernet version of your multiplexer instead? A lot of people
considering this type installation might already have a PC/access point
installed, and it would mean higher reliability for users willing to run
wires. People that really want wireless you can sell a $50 access point
to... Ethernet modules are surely a lot less expensive?

The new integrated systems (Furuno, Raymarine, Garmin) also use wired
ethernet, so the infrastructure is getting installed already.

What we need now is a standard for transmitting NMEA and NMEA-2000
(like) data over Ethernet/UDP. Guess that won't happen for a while...

 

[Bill Kearney]

If they can't figure out how to hook up the new starting battery to a

Yanmar, there's little danger they're going to break the WEP code in the
wireless lan any time soon....(c;

No, but someone onshore near their boat certainly can. Use WPA and you're
done.

Some of the lawyers can't load a flashlight. I've never figured out why
society allows lawyers to make so MUCH money with so little brains. How
stupid.

You've obviously never needed effective legal service. But hey, play dumb
until you do, then $350/hour to keep you out of prison will seem CHEAP.

 

[Matt Colie]

Miendert,
You are obviously in the position to know.
I just installed all the stuff and it sure looked like BT, but come to
think of it - it did not ever claim to be.

That is all good to know. I'm sure that the owner will be back wanting
something added that will be an issue.

Matt Colie

 

[David Rinnan]

Has anyone seen a ship position as reported by AIS being off by over one
nautical mile?

I was recently sailing from Hawaii to San Francisco, and encountered a
freighter that eventually passed about one mile north of us. We were both
heading east. The strange thing is that their position as reported by their
AIS transmitter showed them passing about one mile to the south of us! I am
using my own AIS program, so I assumed at the time that there was a bug in
my code. I captured the raw NMEA data (an option in my program), and after
looking at it, and running it through some commercial programs, it seems
that my code was OK and the ship was reporting its position incorrectly.

I've seen bad data show up in the "static and voyage related data" messages,
but until now haven't seen bad latitude / longitude data. The reference
position offset fields only allow for up to 63 meters of beam offset, so
that couldn't account for the position error I was seeing. Since the
"position report" message should get it's position directly from a GPS unit,
I can't understand how an error of this magnitude could occur. I have not
seen any similar errors with other ships. I will be digging deeper into the
captured data to look at the "position accuracy" and "time stamp" fields to
see if perhaps the ship was in some sort of dead reckoning mode or was
reporting other position-fixing errors.

I guess that this underscores the need for a visual or radar confirmation
during a close AIS encounter! Obviously the calculated CPA was quite
different from the actual.

FYI, here is the ship data my program saved:

Time: 8/4/06 3:01:52 PM
Name: LADY MADONNA
Callsign: 3EKW8
Latitude: 40.728833 deg
Longitude: -152.034833 deg
SOG: 12.2 kt
COG: 86.0 deg
Destination: CEDROS_MEXICO
Ship Type: Cargo ship
Ship Status: Under Way using engine
MMSI: 352730000

Any thoughts?

Thanks,
Paul

I beleive I have read somewhere that professional systems are able to
utilize AIS targets as input for target tracking on a conventional radar
system/screen. Has anyone been able to test anything like this on
consumer equipment? Would it even be technically possible? As Paul
mentiones in his post above it can be somewhat dangerous to rely only on
AIS data, even if one is not using Pauls own software
Being able to push the AIS data into the radar screen would be a nice
way to get both on one single screen (why not chart as well).

I understand that on a platform like Garmin 3010, able to provide chart
+ radar and AIS, the above is somewhat possible since I can get a good
visual check if the radar and AIS target are inline or not. But the
tracking I dont know about. Given the poor AIS support on the Garmin
units at this stage I would guess not.

Cheers
david

 

[Larry]

To me having a WiFi multiplexer doesn't make sense, as there is no
"serial profile" defined on top of Ethernet, by my understanding.

It's terrible, but please don't tell my captain. He thinks it's wonderful,
like I do....(c;

 

[Dave Baker]

Has anyone seen a ship position as reported by AIS being off by over one
nautical mile?

Since the
"position report" message should get it's position directly from a GPS unit,
I can't understand how an error of this magnitude could occur. I have not
seen any similar errors with other ships. I will be digging deeper into the
captured data to look at the "position accuracy" and "time stamp" fields to
see if perhaps the ship was in some sort of dead reckoning mode or was
reporting other position-fixing errors.

I discovered a problem over a year ago with Furuno GPS units interfaced into
AIS systems. I reported it to the US Coast Guard & they published some info
here:

http://www.uscg.mil/hq/gm/moa/docs/1-05.pdf

More info on the fix here:

http://www.uscg.mil/hq/gm/moa/docs/furunosafety.pdf

So, next time you chat to them, ask them if they have a Furuno GP80 or GP90
GPS, and if they do, let them know there is a firmware upgrade. I found the
problem initially when installing an AIS base station & seeing about 10% of
the vessels transmitting positions on land when I could see them out the
window. The bust was about 250 metres in my case, and the offset to WGS-84 on
the paper charts that everyone cheated off was 250 metres. I went onboard a
couple of the vessels to check out their systems & needless to say, the
Captains were pretty shocked when I plugged my notebook into their AIS pilot
port & showed them their boat 200m up the shoreline.

Dave

 

[Larry]

Captains were pretty shocked when I plugged my notebook into their AIS
pilot port & showed them their boat 200m up the shoreline.

Dave

You lawyers reading this....Stop drooling on your keyboards....

 

[Paul]

I discovered a problem over a year ago with Furuno GPS units interfaced
into
AIS systems. I reported it to the US Coast Guard & they published some
info
here:

http://www.uscg.mil/hq/gm/moa/docs/1-05.pdf

More info on the fix here:

http://www.uscg.mil/hq/gm/moa/docs/furunosafety.pdf

So, next time you chat to them, ask them if they have a Furuno GP80 or
GP90
GPS, and if they do, let them know there is a firmware upgrade. I found
the
problem initially when installing an AIS base station & seeing about 10%
of
the vessels transmitting positions on land when I could see them out the
window. The bust was about 250 metres in my case, and the offset to WGS-84
on
the paper charts that everyone cheated off was 250 metres. I went onboard
a
couple of the vessels to check out their systems & needless to say, the
Captains were pretty shocked when I plugged my notebook into their AIS
pilot
port & showed them their boat 200m up the shoreline.

Dave,

Thank you for this information. I don't know what equipment was used on the
ship I encountered, but this is a completely plausible explanation for what
I was seeing. Perhaps I will try to locate the ship owner and send them a
message -- it couldn't hurt!

What port were you in when you discovered this problem?

Regards,
Paul

 

[[email protected]]

Well, let's define one!
It would be interesting though to see how the major brands transfer NMEA
over ethernet. But indeen, UDP is the best way since it is also a
connectionless/broadcast protocol just like NMEA0183 and NMEA2000.

Agreed. I use UDP in my stuff. It makes it easy to keep the network
"open" in that any device on the subnet can grab the data and use it.
For that reason, I just send the NMEA sentences one at a time, with no
mods or higher level constructs that a listener would have to be aware
of. The only "gotcha" (and it's irrelevant to the vast majority of
uses) is that all the routers I've looked at block UDP broadcasts from
going off the subnet, for good and sufficient reasons. If you want to
broadcast to another subnet you have to target the UDP to a specific
address on that subnet and let it repeat the message as a broadcast on
that subnet. At least, that's the way I solved the problem in my
stuff.

Unfortunately, I've found that most of the potential listeners to the
NMEA data I'm making available are so busy making everything
proprietary that this is less of a benefit than I hoped.


-----------------------------------------------------------------------------------------------------------
Glen "Wiley" Wilson <usenet1 SPAMNIX at world wide wiley dot com>
To reply, lose the capitals and do the obvious.

Take a look at cpRepeater, my NMEA data integrator, repeater, and
logger at http://www.worldwidewiley.com/

 

[Paul]

Agreed. I use UDP in my stuff. It makes it easy to keep the network
"open" in that any device on the subnet can grab the data and use it.
For that reason, I just send the NMEA sentences one at a time, with no
mods or higher level constructs that a listener would have to be aware
of. The only "gotcha" (and it's irrelevant to the vast majority of
uses) is that all the routers I've looked at block UDP broadcasts from
going off the subnet, for good and sufficient reasons. If you want to
broadcast to another subnet you have to target the UDP to a specific
address on that subnet and let it repeat the message as a broadcast on
that subnet. At least, that's the way I solved the problem in my
stuff.

Unfortunately, I've found that most of the potential listeners to the
NMEA data I'm making available are so busy making everything
proprietary that this is less of a benefit than I hoped.

This thread has drifted in a direction that I like a lot. Since my original
AIS-specific question has likely been answered, I believe that the NMEA /
WiFi / BT topic deserves its own thread so it doesn't get lost. Of course I
still look forward to further AIS discussions.

I agree that WiFi and wired ethernet are the direction to go if you want to
build an open nav-electronics system (and I certainly do). BT is probably
better for the extremely power-sensitive applications, such as handhelds and
of course cellphone earphones. Meindert, I would definitely be a customer
for a wired ethernet mux. Of course, I've just got one boat, so I would
only buy one or two muxes.

I remain vigilant when it comes to power drain on my boat. I have three
100W solar panels on board, but depending on the cloud cover, and my point
of sail relative to the sun, I need to run the engine to charge batteries
for about an hour a day -- more if I am using the B&G hydraulic autopilot.
At anchor, the panels essentially keep up with the lower power drain. This
is why I am avoiding running a laptop full-time, and worry about the power
requirements of the electronics.

Of course if I shut off all the junk and just sail, I only have to power the
tricolor and compass light. But I do like my toys (and cold drinks).

-Paul

 

[Bill Kearney]

Agreed. I use UDP in my stuff.

Raymarine's E-80 and their Sirius weather interface also appear to be using
UDP for all traffic. The E-80 also sends other data, I saw a fair number of
packets before I even had the Sirius unit attached. Didn't bother sniffing
the packets with ethereal though. Come winter when I've got the units out
of the boat I may try decoding what gets sent around.

It makes it easy to keep the network
"open" in that any device on the subnet can grab the data and use it.
For that reason, I just send the NMEA sentences one at a time, with no
mods or higher level constructs that a listener would have to be aware
of. The only "gotcha" (and it's irrelevant to the vast majority of
uses) is that all the routers I've looked at block UDP broadcasts from
going off the subnet, for good and sufficient reasons. If you want to
broadcast to another subnet you have to target the UDP to a specific
address on that subnet and let it repeat the message as a broadcast on
that subnet. At least, that's the way I solved the problem in my
stuff.

Yep, UDP is great for all the reasons you mention, including isolation on
the local subnet.

Unfortunately, I've found that most of the potential listeners to the
NMEA data I'm making available are so busy making everything
proprietary that this is less of a benefit than I hoped.

Two steps forward, two steps back. Yeesh.

-Bill Kearney

 

[Dave Baker]

Thank you for this information. I don't know what equipment was used on the
ship I encountered, but this is a completely plausible explanation for what
I was seeing. Perhaps I will try to locate the ship owner and send them a
message -- it couldn't hurt!

If they are using the Furuno GP80 or GP90 it's quite likely. I found about
30% of vessels around here use those GPS models and about 1/3 of them were
"cheating" by putting their paper chart offsets into their GPS, thereby
causing the problem.

What port were you in when you discovered this problem?

It was Labuan, a small island off the coast of Borneo.

Dave