AIS - CPA to ship or antenna ?

[paulburn]

Following a interesting point raised on a previous post, I emailed Garmin for clarification.

It was concerning how CPA is calculated- is it from the target ship's GPS antenna position or from the nearest point of the ship to you ? (which could be over 400m apart)

Reply from Garmin below-

I have been discussing your question with my colleague and can now provide the answer.

The calculation of the CPA is done by the position of the antenna itself, not the dimensions of he boat as I previously suspected.

As the accuracy of GPS sometimes could sometimes be up to 20 meters and the GPS antenna cannot accurately determine the boat heading it cannot calculate CPA in relation to the dimensions of the vessels itself. Therefore the CPA and TCPA are based on the position of the GPS antenna.

If there is anything else I can help you with then please let me know. Alternatively, you can search for a solution here: https://support.garmin.com

 

[sailorman]

It can only be from the ais tx to ais antenna Rx

 

[RichardS]

It can only be from the ais tx to ais antenna Rx

If you search for the earlier thread you'll see why that is not necessarily the case ...... but the reasons why manufacturers do it that way was thoroughly explained.

Richard

 

[maby]

It can only be from the ais tx to ais antenna Rx

Not really - the calculations are done with reference to the coordinates encoded in the transmitted message and the coordinates at which the receiver display unit believes it is at. In most cases for small boats, these are within a few feet of the transmitter and receiver, but there is no reason why that should be the case. There are an increasing number of buoys and other danger areas marked by "virtual AIS" where a ground station that may be several miles away transmits an AIS message with the coordinates of the target in question encoded in it. A boat within the coverage area will receive the message and process it in exactly the same way as it would it the transmission had actually come from the target.

Equally, there are AIS receive applications that you can fire up on your smart phone even though it does not have an AIS receiver in it. They get their AIS messages over the internet from AIS receivers dotted along the coast line. I've just fired up Marine Traffic on my smartphone, selected a ferry in Southampton Water and it has given me an estimate of its distance from me and time to closest approach. That is using the GPS in my phone, not a GPS in the AIS receiver that picked up the message.

 

[sailorman]

Not really - the calculations are done with reference to the coordinates encoded in the transmitted message and the coordinates at which the receiver display unit believes it is at. In most cases for small boats, these are within a few feet of the transmitter and receiver, but there is no reason why that should be the case. There are an increasing number of buoys and other danger areas marked by "virtual AIS" where a ground station that may be several miles away transmits an AIS message with the coordinates of the target in question encoded in it. A boat within the coverage area will receive the message and process it in exactly the same way as it would it the transmission had actually come from the target.

Equally, there are AIS receive applications that you can fire up on your smart phone even though it does not have an AIS receiver in it. They get their AIS messages over the internet from AIS receivers dotted along the coast line. I've just fired up Marine Traffic on my smartphone, selected a ferry in Southampton Water and it has given me an estimate of its distance from me and time to closest approach. That is using the GPS in my phone, not a GPS in the AIS receiver that picked up the message.

You are now mixing fish with fowl
the question was ship based not sitting at home under a tree

 

[pvb]

It was concerning how CPA is calculated- is it from the target ship's GPS antenna position or from the nearest point of the ship to you ? (which could be over 400m apart)

Slight exaggeration there, perhaps! The GPS antenna is usually somewhere in the middle of the ship, so a lot less than 400m.

 

[paulburn]

Slight exaggeration there, perhaps! The GPS antenna is usually somewhere in the middle of the ship, so a lot less than 400m.

What is your evidence for this?
And 200m isn't a lot better?!

 

[maby]

What is your evidence for this?
And 200m isn't a lot better?!

My AIS has a hissy fit if the target is going to pass within half a mile of us - a couple of hundred yards discrepancy is not going to make much difference

 

[RichardS]

Slight exaggeration there, perhaps! The GPS antenna is usually somewhere in the middle of the ship, so a lot less than 400m.

Not based on any prior knowledge but I would have guessed that the a tanker/freighter/cruise liner antenna would be based in the bridge area which would be 400m for a mega-tanker and perhaps 300m for a mega-cruiser.

Richard

 

[pvb]

What is your evidence for this?
And 200m isn't a lot better?!

What sort of evidence might you believe? It's what most of us call "common sense".

And with your undoubted knowledge of AIS, you'll know that the overall length of a target vessel can easily be seen. As there are only half a dozen or so ships in the world over 400m, I don't think you need worry too much.

 

[Mistroma]

It can only be from the ais tx to ais antenna Rx

It has nothing to do with the Tx/Rx antennae as AIS isn't an RDF system like the old systems from years ago. I imagine that you meant to say GPS antennae as these provide the respective AIS systems with their position information.

You are now mixing fish with fowl
the question was ship based not sitting at home under a tree

My AIS configuration software asks for information about the location of the GPS antenna used by the AIS system.

A: Distance from bow to GPS antenna
B: Distance from stern to GPS antenna
C: Distance off centreline on port side (at max. beam) to GPS antenna
D: Distance off centreline on starboard side (at max. beam) to GPS antenna

This information could be used to improve CPA by adjusting it for nearest part of the vessel. No idea if it actually does that on my system but the parameters are there and the math is simple enough. I'd be surprised if class A systems don't use the information. However, I'd be reluctant to rely on it as you never know if it has been set up correctly. I've seen lots of vessels transmitting fairly silly information due to poor configuration.

Errors due to time delays in updating CPA are probably much more important anyway. I always add a safety margin to allow for GPS antenna error plus time delay errors. Plus watching the bearing on radar in fog or using eyeball. AIS is great but you wouldn't want to plan a close pass in front of a 400m vessel charging towards you based only on AIS. Frankly, I wouldn't plan a close pass anyway.

 

[GHA]

........ The GPS antenna is usually somewhere in the middle of the ship..... .

Opencpn can draw the ship real size indicating where the transponder is, from unreliable memory the position of the reference point as indicated by the ais message is more often near the stern in the big boys.

 

[paulburn]

Opencpn can draw the ship real size indicating where the transponder is, from unreliable memory the position of the reference point as indicated by the ais message is more often near the stern in the big boys.

That's great! I wish my garmin had similar features (tho' it's good for everything else e.g. Tide stream arrows)

 

[Sandy]

AIS - CPA to ship or antenna ?

Its not really worth worrying about! The Mark I eyeball can be use when you get so close the ship blots out the sun.

 

[sailorman]

It has nothing to do with the Tx/Rx antennae as AIS isn't an RDF system like the old systems from years ago. I imagine that you meant to say GPS antennae as these provide the respective AIS systems with their position information.



My AIS configuration software asks for information about the location of the GPS antenna used by the AIS system.

A: Distance from bow to GPS antenna
B: Distance from stern to GPS antenna
C: Distance off centreline on port side (at max. beam) to GPS antenna
D: Distance off centreline on starboard side (at max. beam) to GPS antenna

This information could be used to improve CPA by adjusting it for nearest part of the vessel. No idea if it actually does that on my system but the parameters are there and the math is simple enough. I'd be surprised if class A systems don't use the information. However, I'd be reluctant to rely on it as you never know if it has been set up correctly. I've seen lots of vessels transmitting fairly silly information due to poor configuration.

Errors due to time delays in updating CPA are probably much more important anyway. I always add a safety margin to allow for GPS antenna error plus time delay errors. Plus watching the bearing on radar in fog or using eyeball. AIS is great but you wouldn't want to plan a close pass in front of a 400m vessel charging towards you based only on AIS. Frankly, I wouldn't plan a close pass anyway.

I use opencpn

 

[savageseadog]

These days conventional GPS accuracy at sea with WAAS corrections is a mean 3m. The ship might be using a DGPS position at an accuracy of around 1m(?).

 

[RichardS]

Opencpn can draw the ship real size indicating where the transponder is, from unreliable memory the position of the reference point as indicated by the ais message is more often near the stern in the big boys.

And near the bow in the big passenger boys (see post #9).

Richard

 

[GHA]

And near the bow in the big passenger boys (see post #9).

Richard

Zoomed in on marinetraffic looks to show the data position, and yes, ferrys seemed to have the reference point near the bow & cargo looks to be generally towards the stern. No that it matters much, 12 seconds or so

 

[Mistroma]

I use opencpn

I was talking about my AIS transmitter settings to make the point that Maby was correct in his post. Technically, the AIS transmitter doesn't even need to be on the vessel. That was his point regarding navigation marks.

I was simply pointing out the error regarding the remark about location of Tx/Rx antenna.

It seems that OpenCPN also confirms his viewpoint as shown by other posts. Pictures show the AIS location offset from the centreline as well as usually being close to the stern on cargo vessels.

 

[Mistroma]

These days conventional GPS accuracy at sea with WAAS corrections is a mean 3m. The ship might be using a DGPS position at an accuracy of around 1m(?).

Although the GPS data is likely to provide a very accurate position it may be out of date. AIS transmitters can clash over time slots in congested areas with significant delays between updates. The vessel might have altered course and/or speed many minutes after your plotter has interpreted the vessels last packet of data. Speed and heading might also be incorrect at source for a variety of reasons. It is therefore perfectly feasible for the predicted CPA to vary from the real CPA by quite a large margin.