Collision Avoidance, Cones of Uncertainty, and Appropriate CPA

[Kenomac]

I disagree with RR on most everything except on this subject. I agree five miles out is ridiculous, changing your course only seems to complicate matters for the large ships. If all small boats did this, the confusion would be incredible.

Just avoid cutting in front of the freighters, slow down when in doubt, but never speed up trying to cut across their bow.

Us people in small boats aren't as important as some on this thread seem to think we are. The bottom line... drive defensively and stay out of the way. Forget about the small stuff and needless radio chatter.

[El Pinguino]

Originally Posted by Dockhead

Heard a lot in American waters as well. And in the pages of the GCaptain forum for commercial mariners.

'Heard a lot' being used by professionals on VHF ? Not very professional behaviour I must say.

GCaptain? Just another internet forum I'm afraid...... populated by all sorts....

[Juho]

Originally Posted by Dockhead

BUT -- the question is a really good one! I agree that this is one of the basic things, which we should be able to know at a glance. Sometimes the bearing is not changing quickly enough to see this very soon.

OpenCPN (and I believe Vesper separate AIS displays) solve this problem by graphically displaying the crossing geometry.

They should give also the angle towards the CPA. Maybe best to have it as an angle relative to the direction of my boat's bow (and maybe as a true angle too).

[Juho]

Originally Posted by Kenomac

I disagree with RR on most everything except on this subject. I agree five miles out is ridiculous, changing your course only seems to complicate matters for the large ships. If all small boats did this, the confusion would be incredible.

Just avoid cutting in front of the freighters, slow down when in doubt, but never speed up trying to cut across their bow.

Us people in small boats aren't as important as some on this thread seem to think we are. The bottom line... drive defensively and stay out of the way. Forget about the small stuff and needless radio chatter.

I guess one's appropriate course changing distance depends also on if one's vessel is the slow, nimble and unstable vessel, or the fast and slowly turning one.

If I turn before coming too close to a shipping lane (with my sailboat), in order to avoid cutting in front of the freighters, would that count as a collision avoidance change of course 20 miles before the nearest ship reaches the CPA?

[ramblinrod]

Originally Posted by skipmac

You still do not answer or address Dockhead's point #2.

From 5 miles out, (sailboat at 5 knots crossing a ship at 20) no course correction may be required at all, so 1 degree is more than enough.

You claim a 1 degree correction is adequate 5 miles out then you correctly state that a sailboat could not hold a course anywhere close to a +/- 1 degree accuracy.

In a heavy sea the sailboat could be varying course by +/- 10 degrees at any instant, but an average course can be held precisely over distance.

The only way for another vessel to tell what you are doing and react accordingly is for you to make a large correction, early in the situation

The actual course change required may be quite small, perhaps only 1 degree.

It is common practice to turn off by a large amount ("show 'em the red") to indicate reaction and then turn back onto the true course change actually required to avoid collision.

Of course, this means nothing if the other vessel is not watching.

[ramblinrod]

Originally Posted by Dockhead

I only have two things to say about this:

1. Your idea of when a risk of collision exists is faulty. Talk to a commercial skipper some time and let him explain it to you, since you don't want to listen to anyone here. 5 miles out in open water is a fully developed risk of collision situation and rapidly approaching the end of the window of opportunity for normal maneuvering. Your misconception of this is an absolutely typical WAFI misconception, which is one of the things which drives commercial seaman crazy, about the way we maneuver. You cannot just make a 180, one mile out, and be sure to avoid danger, and you cannot just blithely sail on from 5 miles out to 1 mile out, telling yourself that there is no risk of collision. The reasons were explained in great detail in this and the other thread, and because you were so much more interested in arguing, than in learning anything, you missed it. The fear of appearing to be wrong, is the mortal enemy of knowledge and learning, and this is a really great example of it..

Any commercial skipper, who expects a sailboat not showing a steaming light to hold speed precisely for any period of time is a PAFI.

If a sailboat is on a collision course with a ship on a 90 degree crossing, and 1 mile from the intersect point turns 180 maintaining about the same speed, they will most definitely avoid collision to the greatest extent possible.

2. You explain very correctly that you can't hold a steady course within one degree on a small boat -- no way to do it within one degree. So how can you say that a one degree course correction in any case (not only head-on), puts you 180 feet into safety? Do you not see the contradiction here?

A sailboat may sail an average course of 275 degrees with a +/-5 degree variation and then change course by 1 degree to 276 with a +/-5 degree variation. Happens all the time.

If from some considerable distance off, one slows by some very small amount, no course change may be required at all.

We do this all the time when crossing lanes, with two vessels converging from opposite directions).

I don't change course at all; I hold course right at the lanes, but regulate speed so the two ships pass each other well ahead of me, and we intersect the lane well astern of both while they steam on in opposite directions away from each other (and us) while we cross the lanes keep our original course.

I may change course by as little as 1 degree, to end up equidistant between the 2 vessels as I cross the lanes.

[ramblinrod]

Originally Posted by transmitterdan

Speed counts. 5 miles separation from a big ship is not 1 hour. It's more like 15 minutes in many cases and even less for head on situations. So 5 miles is well inside the "risk of collision" window in my opinion. Even 10 miles is in some cases which is why some ships have already maneuvered before we see them.

5 miles from the intersect point at 5 knots is 1 hour.

[skipmac]

Originally Posted by ramblinrod

From 5 miles out, (sailboat at 5 knots crossing a ship at 20) no course correction may be required at all, so 1 degree is more than enough.

In a heavy sea the sailboat could be varying course by +/- 10 degrees at any instant, but an average course can be held precisely over distance.

Not even under power with a good autopilot in totally flat seas can a small boat hold a course accurate to within 1 degree over a several mile distance.

So suggesting a course change of 1 degree to avoid a collision under any combination of speed, distance and crossing angle is completely invalid.

On the other hand, in my experience and contrary to another statement you made, speed offshore can often be stable and constant when averaged over a relatively short time to compensate for wind and wave action. It is not unusual to hold an average speed constant within a half knot for hours

[Dockhead]

Originally Posted by ramblinrod

5 miles from the intersect point at 5 knots is 1 hour.

Only if the ship is at anchor. Do you really think this??


If the ship is making 20 knots, then TCPA at 5 miles out is 12 minutes, if you are on reciprocal courses, and 14.5518 minutes with perpendicular courses.

As I said: 5 miles out in open water is far into the risk of collision zone. If you are standing on, you should already be getting ready to make your own move, if you still have <1 mile CPA by this point.


Edit: If by "5 miles from the intersect point" you mean 5 miles from the supposed point of CPA, then you are talking about being 20.616 miles away from the ship, when he is well under the horizon and your recreational AIS can't even see him. No one is talking about this. "5 miles out" means 5 miles range from the target.

[Dockhead]

Originally Posted by ramblinrod

. . . If a sailboat is on a collision course with a ship on a 90 degree crossing, and 1 mile from the intersect point turns 180 maintaining about the same speed, they will most definitely avoid collision to the greatest extent possible. . . .

This is simply not correct, and it was explained above. Your mistake is that you assume that you have perfect knowledge about where he will be, but you don't have this perfect knowledge. At five miles out, you THINK you know where you will be at one mile out, so you blithely hold your course, even though you see that he hasn't seen you and is not changing course himself, feeling safe with the dangerously erroneous thought that you can just do a 180 at one mile out, and be safe, as you have just suggested.

But the other problem is that the closer you get, and the bigger difference between your speeds, the less power you have to get out of his way. So let's say that your course wasn't as steady as you thought, or his course was not steady, or a combination of the two (as usual), and at one mile out, you are STILL on a 0 CPA.

What do you do then? One mile out on perpendicular courses, you are 174.6 seconds from impact. Yes, 174.6 seconds! So let's do a 180 -- tack or gybe the sails around -- how long does that take? Can you do it in one minute? OK, now we have 114 seconds left. Another minute to get back up to speed -- 74 seconds. How far can you move in 74 seconds, at five knots? 1.028 cables, that's how far. 200 meters. But when you started your maneuver, 174.6 seconds from impact, did you even then know within 200 meters, where he will be?

Don't you understand now?

Many small boat sailors don't understand this, because they've never met a ship in open water. When you meet a ship in a lake or bay or in approaches to harbors, and the ship is following a definite lane, or a fairway, then it's easy to know where he will be -- even within tens of meters. You can sail right up to the edge of the channel and stop, and you'll be completely safe.

But the danger is extrapolating this situation, to how collision avoidance works in open water. The root problem is the fact that you cannot know exactly where you will meet, and the closer you get, the less ability you have to get out of the way.

My last attempt to explain this.
.[/QUOTE]

[transmitterdan]

Originally Posted by ramblinrod

5 miles from the intersect point at 5 knots is 1 hour.

I doubt you can see a big ship when you are 5 miles from Intersection with their track. It will be 15-20 miles away from you at that point. Why do you change the scenario to a ridiculous one?

[ramblinrod]

Originally Posted by Dockhead

This is simply not correct, and it was explained above. Your mistake is that you assume that you have perfect knowledge about where he will be, but you don't have this perfect knowledge.

No I don't, and sir you cannot have both ways whenever its suits your argument to change your position 180 degrees.

At any time, we can only make course changes to avoid collision based on where we assume the vessels will be some time for now.


If there are no other influencing factors, we assume that stand-on vessel holds course and speed, and the give way vessel adjusts course and speed, so that the vessels don't arrive at the same point in time, that we originally assumed they will be.

The stand on vessel may in fact change course and speed. We may be able to predict this based on knowledge of land and other boats the stand-on vessel may need to avoid (if we can see them).

Or the stand-on vessel may be making a turn for any number of reasons we have no idea about.

According to your supposition that one must account for any turn the other vessel may make when giving way, that is impossible, as I indicated before, if the are faster, no matter what you do, they may run you down.

[ramblinrod]

Originally Posted by transmitterdan

I doubt you can see a big ship when you are 5 miles from Intersection with their track. It will be 15-20 miles away from you at that point. Why do you change the scenario to a ridiculous one?

I see ships 20 miles away all the time, visually by day, by night, and on AIS.

It is not ridiculous, it is reality.

I am not changing the scenario.

There is always a variety of scenarios possible with variations in:

a) Speed of both vessels.
b) Distance between vessels.
c) Any course change any (of the perhaps many involved) vessel may need to make during approach.

To suggest that one can make a course change when 5 nm or even 1.25 nm form the intersect point with certainty is ludicrous. It is based on everyone maintaining course and speed.

Dockhead is saying that one must adjust course to avoid collision no matter what the other boat does between now and point of impact. That is impossible if they are faster, they could keep changing course and run your down no matter what series of course changes you make, even turning 180 to leave a 90 degree crossing.

[ramblinrod]

Originally Posted by Dockhead

This is simply not correct, and it was explained above. Your mistake is that you assume that you have perfect knowledge about where he will be, but you don't have this perfect knowledge. At five miles out, you THINK you know where you will be at one mile out, so you blithely hold your course, even though you see that he hasn't seen you and is not changing course himself, feeling safe with the dangerously erroneous thought that you can just do a 180 at one mile out, and be safe, as you have just suggested.

But the other problem is that the closer you get, and the bigger difference between your speeds, the less power you have to get out of his way. So let's say that your course wasn't as steady as you thought, or his course was not steady, or a combination of the two (as usual), and at one mile out, you are STILL on a 0 CPA.

What do you do then? One mile out on perpendicular courses, you are 174.6 seconds from impact. Yes, 174.6 seconds! So let's do a 180 -- tack or gybe the sails around -- how long does that take? Can you do it in one minute? OK, now we have 114 seconds left. Another minute to get back up to speed -- 74 seconds. How far can you move in 74 seconds, at five knots? 1.028 cables, that's how far. 200 meters. But when you started your maneuver, 174.6 seconds from impact, did you even then know within 200 meters, where he will be?

Don't you understand now?

Many small boat sailors don't understand this, because they've never met a ship in open water. When you meet a ship in a lake or bay or in approaches to harbors, and the ship is following a definite lane, or a fairway, then it's easy to know where he will be -- even within tens of meters. You can sail right up to the edge of the channel and stop, and you'll be completely safe.

But the danger is extrapolating this situation, to how collision avoidance works in open water. The root problem is the fact that you cannot know exactly where you will meet, and the closer you get, the less ability you have to get out of the way.

My last attempt to explain this.
.

[/QUOTE]

My point exactly. There is no such thing as a cone of uncertainty with 100% certainty. The only thing 100% certain is that no matter what you do, you coculd be ivnvopved in a collision. Any commercial seaman who believes they can or should be able to plot a course from 5 miles away that will be correct should ha e their ticket revoked.

One can only plot a course based on a set of assumptions and the parAmeters could change any which way from Sunday.

[Dockhead]

Originally Posted by ramblinrod

. . . Dockhead is saying that one must adjust course to avoid collision no matter what the other boat does between now and point of impact. That is impossible if they are faster, they could keep changing course and run your down no matter what series of course changes you make, even turning 180 to leave a 90 degree crossing.

You did not understand at all what I was saying.

Obviously, if a much faster vessel decides to turn towards you and run you down, there is little you can do to avoid it. Obviously a significant course or speed change somewhere in the middle of a crossing, will invalidate a collision avoidance solution (and that is exactly why you are obligated to stand on, when you are the stand-on vessel).

But here's the point: even when vessels are holding their course and speed as well as they can, there are variations. These variations do NOT always and automatically average out -- the average will be at the very best, in the case of a ship, plus or minus a couple of degrees, and in the case of a yacht under sail, more. A vessel under sail is, moreover, subject to sometimes very big variations of speed, as the wind varies.

So when you do collision avoidance in open water, and you are trying to maneuver to make a safe pass with another vessel, you must take account of the uncertainties of his and your own position which come from these errors. In order to be safe, you have to maneuver so that the zone of your PROBABLE position at CPA does not intersect with the zone of his PROBABLE position at CPA. These zones are much bigger than the dimensions of the vessels; their size is determined by the cone of uncertainty discussed above. If you do all the geometry based on the dimensions of the vessels ("180 feet is plenty of a safe margin"), without considering these uncertainties, then you fail.

And when we say PROBABLE positions, we mean where we will PROBABLY be IF we are all holding course and speed. This does not take account of what happens, if someone maneuvers. If someone turns or stops in the middle of the crossing, all bets are off, and once again, this is why you are obligated to hold course and speed during certain times, when you are the stand-on vessel.

As a handy rule of thumb, professional sailors are taught to set up crossings with at least one mile CPA, in order to be sure of a safe pass. One mile deals with these zones of uncertainty in almost all cases, provided everyone holds course and speed. On most ships, the bridge standing orders require maintaining minimum 1 mile CPA in all crossings, and require calling the captain in case of a closer encounter. In real life, something less than a mile might be safe enough in some circumstances, but under no circumstances, in open water, should you ever only START maneuvering at one mile -- that's already in extremis -- an imminent collision -- where already no one is standing on any more and everyone is doing whatever desperate thing he can, so that you are taking the risk of simultaneous maneuvering into each other.