Thoughts about AIS

[colemj]

By law, English must be used on the helm of commercial ships for communication. DSC rings directly onto the bridge and is very loud regardless of volume setting until answered. It cannot be turned off. There is no confusion of hailing name, uncertainty over position and vessel, or not hearing a call because the volume was too low.

At 10 miles out, if one did find themselves in a cat/mouse over and over and over, then it is very easy to show a big change if necessary.

However, if some young WK (I don't know that acronym, so may be using it incorrectly) IS slewing the ship around in a cat and mouse, I'm sure a more senior person is going to be getting involved very quickly.

Are you suggesting that from far out, one makes unnecessarily large changes of course simply to underscore with gusto your intentions?

I really don't understand the logic and reasoning behind all of these "seductive power" of electronics threads.

Mark

[colemj]

Originally Posted by Pelagic

The situation I am envisioning is closing where the other vessel is seen approaching at low relative speed from about122° on your port side.

We don't change course for vessels overtaking us at such a large angle. That is the exception to our practice of always considering ourselves the burdened vessel and taking action.

Changing course when being overtaken is dangerous regardless of how much you show your intentions.

In your example, why would a relatively slow closing be more dangerous than a fast one? And how did you come to the strangely precise 122*?

Mark

[Pelagic]

Sorry Mark.. WK means watchkeeper.
What was emphasized during simulator runs was to avoid making minor course changes early on. (If we did that then examiner would inject new small fishing targets to make our life hell )

The reasoning being that both vessels maintain courses until the conflict and multilateral developments are very clear, before taking appropriate action.

At 10-12nm out I believe too many things can change so I monitor closely and prefer to show clear intentions at about 5nm when I should have the other vessels' attention.

Re com's....in my area in Asia... No one talks except on approaches to main ports and many do not have DSC.

[Stu Jackson]

For Dockhead & Mark:

Nice discussion thanks. Coupla questions:

1. Could we simplify the discussion by eliminating "offshore" single meetings from the discussion? It appears that regardless of the length of the lines on the screen, those situations become pretty obvious.

2. Can we define "close quarters" as say crossing the English Channel from S to N or N to S, and also eliminate close quarters in harbors?

I think these would help to better define the discussion and avoid wandering.

I also am not sure if "offshore" counts the English Channel. My understanding is that it does, but is different in my mind than #2 above, hence, my request.

I am also still trying to understand Dockhead's difficulty with the length of the lines. Based on an earlier post (sorry, I didn't go back to find it, but I DID read it), Dockhead said he could adjust both lines based on time. So, if we put aside the "arguments" of what the extent of this time period is, and without any course changes, it seems to me that the length of the lines will show where each vessel is after that period of time. Wouldn't that then be a rather "simple" solution?

Given that, what I'm having difficulty understanding is Dockhead's repeated and quite understandable issue with multiple targets.

Is the issue changing the "target" vessel after the first one has been confirmed? Is it just the difference between the Zeus screen and the Vesper examples? Please help me understand. I've read all of these posts twice now, and mybe I'm just not getting it.

Thanks again for a very illuminating and helpful and polite discussion.

[colemj]

OK - I was only considering commercial shipping with the radio part. That should also be true in your part of the world for those ships.

However, at 10nm out, why not make that 1-2* change? For a small sailboat (even yours is small at this stage), it will just be in the noise of things navigational on the bridge of the other ship and they would not be engaging in multilateral moves at this distance.

And you seem to be approaching this from the point of view of the large ship (simulator runs). I agree that their practices should be different than those of a small, slow, highly maneuverable sailboat. This underscores that they will not be engaging in the cat/mouse thing at those distances, so you (the small sailboat) are free to make small changes that makes everyone's life easier.

Mark

[colemj]

Originally Posted by Stu Jackson

For Dockhead & Mark:

Nice discussion thanks. Coupla questions:

1. Could we simplify the discussion by eliminating "offshore" single meetings from the discussion? It appears that regardless of the length of the lines on the screen, those situations become pretty obvious.

2. Can we define "close quarters" as say crossing the English Channel from S to N or N to S, and also eliminate close quarters in harbors?

I think these would help to better define the discussion and avoid wandering.

I also am not sure if "offshore" counts the English Channel. My understanding is that it does, but is different in my mind than #2 above, hence, my request.

I am also still trying to understand Dockhead's difficulty with the length of the lines. Based on an earlier post (sorry, I didn't go back to find it, but I DID read it), Dockhead said he could adjust both lines based on time. So, if we put aside the "arguments" of what the extent of this time period is, and without any course changes, it seems to me that the length of the lines will show where each vessel is after that period of time. Wouldn't that then be a rather "simple" solution?

Given that, what I'm having difficulty understanding is Dockhead's repeated and quite understandable issue with multiple targets.

Is the issue changing the "target" vessel after the first one has been confirmed? Is it just the difference between the Zeus screen and the Vesper examples? Please help me understand. I've read all of these posts twice now, and mybe I'm just not getting it.

Thanks again for a very illuminating and helpful and polite discussion.

1. OK

2. We have not crossed the English Channel and, right now, do not ever foresee doing so - so I will bow out of that and leave it to DH.

However, I would like to know how often this "English Channel" situation occurs (outside the English Channel, of course!)?

We regularly go into and out of the Panama Canal zone which literally has >100 ships anchored, circling around, moving randomly at different speeds, approaching from and leaving to all directions, going in and out full speed ahead, and coming around blind areas into tight quarters you are also trying to fit in. But this seems to be a "harbor" situation in your definition. Singapore and Hong Kong would also seem to fit the "harbor" definition even though those are also extremely busy waters with confusing navigation problems.

We have been in commercial traffic areas that are divided into inbound and outbound areas separated by a buffer zone that have much constant full speed traffic in each lane - all going in the same direction in each lane, but again, find these to not be a problem (actually, we find them MUCH easier to navigate than the Panama Canal zone and its like).

So it is possible that either we have not really experienced real AIS navigation problems, or DH's English Channel problem is a rarer local one that does require additional situational support.

Mark

[Pelagic]

I have been lucky enough to see the perspectives from both sides (Big and Small) and in both cases young WK's are often tempted to make small adjustments at 12nm when coasting.

Nothing wrong with that if the situation is clear cut but I have watched them put themselves in trouble when the conflict vessel makes the wrong assumption at 9nm and closes the gap again.

If on my own sailboat and assuming that they ignore the sailing rights.... I am happy to stay on course till 3 miles out then make an obvious course change.

I think it also helps out the issue of CPA aspect which concerned Dockhead on AIS.

[Pelagic]

Originally Posted by colemj

We don't change course for vessels overtaking us at such a large angle. That is the exception to our practice of always considering ourselves the burdened vessel and taking action.

Changing course when being overtaken is dangerous regardless of how much you show your intentions.

In your example, why would a relatively slow closing be more dangerous than a fast one? And how did you come to the strangely precise 122*?

Mark

A typo should have been 112° aspect

This is often the grey danger area of misunderstanding between a crossing and overtaking situation.

In a close quarter situation it is the only time you are allowed to alter course to port when being overtaken from around 122-222 on your stbd quarter.

[Stu Jackson]

Originally Posted by colemj

1. OK

2. We have not crossed the English Channel and, right now, do not ever foresee doing so - so I will bow out of that and leave it to DH.

............

We regularly go into and out of the Panama Canal zone which literally has >100 ships anchored, circling around, moving randomly at different speeds, approaching from and leaving to all directions, going in and out full speed ahead, and coming around blind areas into tight quarters you are also trying to fit in. But this seems to be a "harbor" situation in your definition. Singapore and Hong Kong would also seem to fit the "harbor" definition even though those are also extremely busy waters with confusing navigation problems.

We have been in commercial traffic areas that are divided into inbound and outbound areas separated by a buffer zone that have much constant full speed traffic in each lane - all going in the same direction in each lane, but again, find these to not be a problem (actually, we find them MUCH easier to navigate than the Panama Canal zone and its like).

So it is possible that either we have not really experienced real AIS navigation problems, or DH's English Channel problem is a rarer local one that does require additional situational support.

Mark

OK, I understand. So let's try this: your Panama Canal zone sounds like the same situation, or even worse, than the "regularity" of the English Channel or any evenly spaced in&out traffic separation scheme. The point you're making is the level of random-ness, right?

So if you can Dockhead can agree to not call this either harbor or offshore, then does this create the one situation you can continue this discussion around without "wandering" into close harbor and offshore areas?

I'm just trying to narrow down the situations to better understand the technology issues that prompted this whole thread to begin with.

Can you both create one scenario, that you can agree upon, and then say: here's where it doesn't work and here where it gets hairy, OR from what I seem to "take" from this is it doesn't matter much based on the differences between the Zeus and the Vesper displays.

I will go back and reread all of this, maybe I'm just thick...

Does that do it for you?

You both have written some very good stuff, thanks again.

[colemj]

Originally Posted by Stu Jackson

Mark

OK, I understand. So let's try this: your Panama Canal zone sounds like the same situation, or even worse, than the "regularity" of the English Channel or any evenly spaced in&out traffic separation scheme. The point you're making is the level of random-ness, right?

So if you can Dockhead can agree to not call this either harbor or offshore, then does this create the one situation you can continue this discussion around without "wandering" into close harbor and offshore areas?

I'm just trying to narrow down the situations to better understand the technology issues that prompted this whole thread to begin with.

Can you both create one scenario, that you can agree upon, and then say: here's where it doesn't work and here where it gets hairy, OR from what I seem to "take" from this is it doesn't matter much based on the differences between the Zeus and the Vesper displays.

I will go back and reread all of this, maybe I'm just thick...

Does that do it for you?

You both have written some very good stuff, thanks again.

I don't think any of that is necessary. I think the only difference between our two postings have been comfort level with specific tools is all.

Certainly the Vespar, OCPN, Coastal Explorer, Nobeltec, etc methods provide even more useful ways of presenting the data, but it really only boils down to I am comfortable with 30 minute COG vectors and Dockhead prefers the additional information.

I am in complete agreement with him that additional information should always be the goal with manufacturers, and that current chart plotters severely lack capabilities that computer programs have had for many years.

Mark

[Stu Jackson]

That's clear.

Thanks so much.

Stu

[Downline]

Originally Posted by Pelagic

I think one of the main reasons that we regular yachtsmen have appreciated AIS over Radar is the poor stabilization of Radar on a small yacht bouncing and slewing around.

Target data jumps around so less confidence in the MARPA.

Are there "yacht radars" that have a slower refresh rate that averages out heading or uses own ship COG to help stabilize the Target data?
If not...I think this would help.

August two years ago I was on the bridge of a big ship travelling south through the English channel. The ship was a DSV (dive support vessel). It uses dynamic positioning to hold position so the electronics needs to be really good. It was one of those stunning days. No clouds and hardly any wind and lots of traffic. I was watching all the screens with great interest. What was an eye opener was the sailing yachts. Some of them you had to use your imagination to see them on the radar screen. They disappear in the background scatter. The yachts with AIS was clearly visible.

[Dockhead]

Originally Posted by colemj

I just described how I find them useful and don't have any problems using them. So yours is not a general or factual statement.

Mark

Sorry, I must have missed that. Didn't you say you were alerted to a problem if the lines cross? But that doesn't tell you much -- objectively. They have to cross at a point where both ships are going to be at the same time. You can see that easily if the ships are going at roughly the same speed. You can figure it out by proportions of the total length of time. If you can do that easy -- I'm happy for you. I can't, and I doubt that most people can do it at a glance, and seeing that at a glance is what you need in order to do these problems with multiple ships.

[Dockhead]

Originally Posted by Stu Jackson

For Dockhead & Mark:

Nice discussion thanks. Coupla questions:

1. Could we simplify the discussion by eliminating "offshore" single meetings from the discussion? It appears that regardless of the length of the lines on the screen, those situations become pretty obvious.

2. Can we define "close quarters" as say crossing the English Channel from S to N or N to S, and also eliminate close quarters in harbors?

I think these would help to better define the discussion and avoid wandering.

I also am not sure if "offshore" counts the English Channel. My understanding is that it does, but is different in my mind than #2 above, hence, my request.

I am also still trying to understand Dockhead's difficulty with the length of the lines. Based on an earlier post (sorry, I didn't go back to find it, but I DID read it), Dockhead said he could adjust both lines based on time. So, if we put aside the "arguments" of what the extent of this time period is, and without any course changes, it seems to me that the length of the lines will show where each vessel is after that period of time. Wouldn't that then be a rather "simple" solution?

Given that, what I'm having difficulty understanding is Dockhead's repeated and quite understandable issue with multiple targets.

Is the issue changing the "target" vessel after the first one has been confirmed? Is it just the difference between the Zeus screen and the Vesper examples? Please help me understand. I've read all of these posts twice now, and mybe I'm just not getting it.

Thanks again for a very illuminating and helpful and polite discussion.

All very good questions!

So -- first of all -- for those who have never done it -- crossing the English Channel (or the North Sea shipping lanes or similar areas of dense, fast traffic) is not anything you would call "close quarters". These are large bodies of water. Aside from the very end near the Dover Strait, the English Channel is no less than 60 miles wide.

So it's very different from maneuvering in harbors or entrances to anything. It's really much more like flying a plane, I think. You have to manage a whole lot of possible spatial conflicts, and you have to manage them efficiently, because there are so many of them. You can't call on the VHF and discuss every crossing because you would be overwhelmed. In the Channel, VHF calls from yachts are mostly just ignored. Really. Even if you use DSC. You can't custom design every crossing when you may have three or four crossings in the space of 10 minutes. You can't just dodge around -- the way we do in harbors and estuaries and harbor entrances -- because others have also made calculations based on our course and speed, and sudden moves on our part ruin the solutions of the ship watchkeepers.

Ship watchkeepers want, more than anything, smooth, predictable maneuvers on our part, made in accordance with the Colregs. This is another reason why I am loath to make otherwise useless course changes, just to observe whether the CPA to a particular ship increases or decreases.

Encountering a single ship in open water is rarely any kind of problem. With AIS, you are alerted to a close crossing from far away, and you can easily unwind it from a safe distance. Before the risk of collision exists, you are not required to hold course as speed if you would otherwise be the stand-on vessel, so you are free from 10 miles out, say, to make a course change which will prevent a close crossing from ever happening, even if you are under sail. And obviously this is good practice.

The difficulty arises when you have to deal with a number of ships at once -- which is typical in the kinds of places we're talking about. Here you can figure out somehow whether you're passing ahead or behind, and figure out the alteration of course which will resolve the situation, but that changes everything with regard to Ship 2 and Ship 3. To find a course which will keep you safely clear of all three can be very complicated, maybe requiring a lot of trial and error, and the absence of the information about whether you are passing ahead or behind can make it just about impossible to figure out efficiently.

Yes, with regard to any one ship, you could change the length of your COG extension lines until you get a picture of the relative positions at CPA. To do that (on my plotter, at least), you have to go into the setup menu and change them separately for own vessel and AIS targets. And you can't set them at just any length -- I have a choice of 1 min, 2 min, 10 min, 30 min, 60 min, and 120 min. This is a hell of a lot of fiddling to figure out one crossing, and really unfeasible to do it when trying to figure out a course to keep you clear of three vessels. You'd wear out the buttons on your plotter going back and forth between this ship and that ship to try to check the result of any possible course change. Maybe some people -- like Mark, perhaps -- can work out the proportion of the line based on speed and visualize the relative positions at CPA. If so, that's a real talent which I lack.

And imagine what you will do to the work of the watchkeepers on the ships -- who are also dealing with the same situation -- if you are jinking around, changing course, just to determine whether you are passing ahead or behind. They are trying -- just as you should be! -- to work out a course which will keep you at a safe difference from everyone, not violating anyone's "personal space" of one mile, without changing course in close quarters.


The Vesper system is totally different from what we have on our plotters -- showing you at a glance with no fiddling the relative positions at CPA with every ship of interest. For all of the reasons above, I consider this vitally important in a place like the Channel.

[Dockhead]

Originally Posted by Pelagic

I have been lucky enough to see the perspectives from both sides (Big and Small) and in both cases young WK's are often tempted to make small adjustments at 12nm when coasting.

Nothing wrong with that if the situation is clear cut but I have watched them put themselves in trouble when the conflict vessel makes the wrong assumption at 9nm and closes the gap again.

If on my own sailboat and assuming that they ignore the sailing rights.... I am happy to stay on course till 3 miles out then make an obvious course change.

I think it also helps out the issue of CPA aspect which concerned Dockhead on AIS.

There is a famous case of a sailboat run down and sunk in the English Channel a few years ago which is really illustrative of this point, and actually illustrates the whole discussion:

http://www.maib.gov.uk/cms_resources...%20wahkuna.pdf

Anyone who has read the thread so far to this point should really read the MAIB report on this incident.

The sailboat was going to pass 8 cables ahead of a fast container ship travelling at 25 knots. The sailboat was nearly new and was equipped with radar, and was skippered by an extremely experienced (40 years) and qualified ocean sailor. There was fog and the vessels could not see each other.

The skipper of the sailboat, based on his estimate of the relative positions of the vessels made by bare-eyes observation of the radar screen, decided that the ship was passing ahead. Instead of behind, which was the real case. His instruments did not say whether he was passing ahead or behind, and so, like Mark, he eyeballed it (although to be fair, he didn't have COG extension lines).

Whereas the master of the ship could see from ARPA that the sailboat was passing ahead. So he held his course and speed.

The skipper of the sailboat decided it would be prudent to let the ship pass by, and he took off way by taking his engine out of gear.

This unexpected maneuver by the sailboat confused the master of the ship, who decided to take no action, because he didn't know what to expect next from the WAFI.

So in a matter of seconds (the ship was travelling at 25 knots), the vessels collided and the yacht was sunk. Miraculously (sounds like the skipper kept a cool head and did a great job of getting everyone off) no lives were lost.


The sailboat skipper was not using MARPA, but MARPA would not have told him whether he was passing ahead of behind. Perhaps if he had been using MARPA he would have seen the CPA decreasing after he took way off, but I imagine it would have been too late.

The master of the ship was criticized for being satisfied with a crossing with the yacht only 8 cables (0.8 miles) ahead. The MAIB explained that this is simply too close, and explained in some detail why. This is really worth paying attention to, because all the same reasons affect the accuracy of AIS:

"A crucial factor in P&O Nedlloyd Vespucci's bridge team's assumption that they would pass clear of the yacht, was their over-confidence in the accuracy of the information provided by ARPA. Several factors reduce the accuracy of relative velocity or triangular calculations carried out by automatic radar plotting equipment [or AIS!]. These include:
1. Where there is a large own ship vector produced by high speed where
course and speed information is reliant on own ship sensors and equipment
for course and speed information.
2. Where the speed of the radar target is small in relation to own ship speed.
3. Where the radar target or own ship is continually changing courses and
speeds.
4. Where the speed information is ground-based rather than water-based in
areas of strong tidal streams."



I think if you want to be really philosophical about it -- the really key core thing here is to know what you don't know, to avoid overconfidence, and to be sure you really have enough information to avoid a collision. Just because you crossed ships 100 times before without and accident doesn't mean the 101st one is not going to run you down -- if you don't have enough information and you're not using proper technique. Both skippers in the accident above were highly experienced and were highly confident in their ability to avoid a collision. But neither skipper actually knew as much as he thought he knew. The yacht skipper did not have information about whether he was passing ahead or behind and made, fatally, the wrong move. The ship master had confidence in his ARPA, and failed to understand its limitations. He saw a few cables of CPA and thought everything was ok. The result was disaster which would ordinarily have killed people.


In my opinion, an extremely useful lesson for all of us. I take away the following, some of which are related to points I have inarticulately been trying to make in this thread:

1. AIS is not infallible and is not a panacea, however fabulous it may be. You can't avoid collisions just by making that there's some positive CPA on your screen, doing nothing else. "I've got a couple of cables of CPA on my AIS; what could go wrong?" Haven't we heard this just recently, and in this thread?
2. You cannot be sure to eyeball the geometry of a crossing with a ship on your radar screen.
3. Spontaneous maneuver when already in close quarters is a poor substitute for setting up a crossing from a safe distance with complete information and guaranteed safe crossing distances.
4. Not knowing whether you're passing ahead or behind can get you killed.