Why Integrate the Autopilot ?

[s/v Jedi]

Nigel, I agree with most you write here but let's go back to that what happens with the AP when your GPS suddenly puts you 4 miles north or similar. You imply that the boat will make a hard turn because the AP assumes that it is at a XTE of 4 miles, or at least that is what most readers will think.

But it ain't happening with the AP's I've used. When the GPS comes with correct position after that hickup, a decent AP will ignore the rogue position. When GPS maintains the 4 mile error, a decent AP will beep, beep, beep and either disengage switching itself back to Standby or switch back to Auto mode (compass).

The question now is if your example of the 4 mile error is what you actually experienced with an AP or is it something you assumed might happen? If it is an assumption, it is unfortunate that you brought it up because that only clouds the issue imho.

The whole underlying idea in your story is one that I partly agree with. I agree that these crazy accidents happen, but I do not agree that the nav aids are to blame. An idiot that navigates using these aids is still an idiot. That is how I see it. True, they would have failed old-style coastal navigation exams which might have kept some of them ashore.
Overall, I think navigation has become much safer with the new aids and we would have had much more accidents without them (basically repeating here what Mark said).

This makes me think about my automated MOB system. It will direct you back to the MOB with terms like "left" instead of "port" so that even an idiot can use it. Is that bad or good? I guess good....

ciao!
Nick.

[cfarrar]

Sidmon,

We get your point about complacency with automated systems. Of course, "complacency" is a broad term. Aviation accident investigations and training programs focus on more specific things like training and proficiency, workload prioritization, crew communication, cockpit distractions, etc. Interestingly, atrophy of basic skills is not high on the list of accident factors. Many highly skilled (and non-complacent) pilots have flown into the ground simply because they lost situational awareness, and sometimes the improper use or understanding of automated systems is part of the chain of causal factors.

What does this have to do with using your sailboat's autopilot in track mode? Not much. Fighter pilots rarely use their autopilot, while airline pilots use Nav mode (ie, track mode) most of the time. However, pilots, cruise ship captains, and sailors agree on this - we need to understand our systems, we need to practice, and we need to prioritize our workload and maintain situational awareness. The beauty of sailing (as opposed to flying for the airlines) is that we are free to develop our own techniques and habits. We can learn from each other, but we're not bound by the company SOPs, thank goodness! And as long as we sail the boat wisely and look where we're going, we'll be fine, whether we're in track mode, heading mode, or hand steering. So stay off the rocks, keep the keel side down, and have fun out there!

Oh, and I just read Ex-Calif's post... +1.

[nigel1]

Originally Posted by s/v Jedi

Nigel, I agree with most you write here but let's go back to that what happens with the AP when your GPS suddenly puts you 4 miles north or similar. You imply that the boat will make a hard turn because the AP assumes that it is at a XTE of 4 miles, or at least that is what most readers will think.



ciao!
Nick.

Hi Nick, I think I put 4 cables, but thats besides the point.
I've not used the NAV mode on my boat, I saw it being used prior to buying though.

I have seen two instances of GPS error when on nav mode auto pilot, (or Track Control), both occasions on state of the art kit.
The first was using track control on an ECDIS system made by TRANSAS, and connected to an auto pilot. I noted a change in the vessels heading and the vessel symbol on the ECDIS was about 3 cables off track, but the vessel was being steered back to the line is a gentle manner, no sudden or excessive helm.
On the second occasion, with a much more expensive bit of kit, the boat made a a pretty drastic turn, it does not take too much helm to get these boats to take up a good angle of heel at speed.
I checked all the settings in the system, and could find no reason why it should try to regain the line so dramatically. We sort of simulated it later by plotting a track about a few cables to the south of our position and then selected Track Control and the boat virtually tried to get on track by the shortest possible means.
I contacted the makers, and they could offer to explanation other than to say it should not happen, as did the makers of the auto pilot, in fact, each implied it was not their bit of kit at fault, but the others.
All the settings are as per the book, and I have tried subtle changes, but still does the same.
I am 90% certain its the auto pilot at fault, but cannot prove it.
I'm still not sure what caused the GPS error in the first place, although on the first occasion it was off Nigeria, and there had been a fair bit of sun spot activity. The second time was in the North Sea. Dont think it was the military as they usually put out a nav warning when they conduct GPS jamming.

Anyway, its just my preference to use the auto pilot in a standalone mode.
My nav system on the boat, an old B&G network system has been through the repair shop a few times, and I have had two auto pilot head failures. So, just trying to keep it as simple as possible

[colemj]

Originally Posted by Ex-Calif

Well I can guarantee you all that the airline rate of groundings is the same as it has been since around 1902. Once per take off.

As we say in the aviation business, take offs are optional - landings are mandatory.

BTW - on some level it is always human error.

The peeing contests are amusing to watch but the only thing grinding my gears are the flip comparisons to commercial aviation. Each false statement makes me want to jump in and correct but this is not an aviation forum. Suffice to say no one so far has posted an accurate airline analogy so please stop. You arre giving me and probably the commercial pilot guys reading this headaches.

BTW - not bragging or "credentialling" but I have been in commercial aviation 30 years and part of the NTSB accident team on 3 commercial accidents. I have been very close to the whole cockpit integration stuff and what some are talking about here falls way short of the level of integration, sophistication and situational awareness benefits that commercial aviation has achieved.

And those tossing out references to Air France (and others) just go read up on airline safety statistics since the 50's and 60's. You will find today's numbers absolutely and convincingly eclipse the super connie days when we still had skylights to take star shots with sextants. And that 50s cockpit with 5 or 6 crew now has 2. Modern is better.

Sorry. It is irrefutable.

Just to be clear: I don't know anything about airplanes or the systems they use and did not mean to say anything "factual" or be flip about them. My entire point was that those systems could not be compared to recreational boat systems and accidents or errors occurring in aviation could not be projected onto recreational boating.

Mark

[s/v Jedi]

Okay, from what you describe, it's 100% certain that the AP was to blame, or it's settings. AP's must (should?) have a setting that defines the maximum course change that it is allowed to do autonomously. This prevents any of those problems.

If a GPS suddenly decides you are 4 cables (that's 240' right?) further north and persists in that... I have never heard of that before. Most often these things happen when:

- two GPS units are active and give the two different positions (never that much apart)
- a correction is received (DGPS but those are small changes)
- the GPS unit or part of it like antenna has failed

In your case you have to wonder which was the correct position. The one before the jump or the one after that. Or may be you were thrown to an alternate universe that is 4 cables off course

But in anycase, these are malfunctions and should not stop anyone from integrating the AP with GPS.

Let me give an example why it is good to integrate, just to offset all the negative views a bit:

My Simrad/Robertson allows me to point the bow to an entrance and then activate it in a track mode to virtual waypoint. All it needs for that is GPS feed. Upon activating it will create a waypoint far ahead and steer a straight COG to it. I like this mode because we are short handed. The first time we arrived in Curacao heading to Spanish Waters, we felt like going into a funnel of lee shore with cross currents and big waves into a very narrow boca that we couldn't even see. So I pointed the bow to where I was pretty sure the entrance must be and activated this AP mode which compensated for the waves and current, giving me time to check the heading line on the radar and adjust AP a couple of degrees. All this while the Admiral secured the sails and sheets and checked the anchor for readiness. I can now keep a good watch without having to concentrate on steering. I can even go down to pee. Without this feature, the approach would have been more stressful and thus less safe.

ciao!
Nick.

[Jim Cate]

Originally Posted by s/v Jedi

Okay, from what you describe, it's 100% certain that the AP was to blame, or it's settings. AP's must (should?) have a setting that defines the maximum course change that it is allowed to do autonomously. This prevents any of those problems.

If a GPS suddenly decides you are 4 cables (that's 240' right?) further north and persists in that... I have never heard of that before. Most often these things happen when:

- two GPS units are active and give the two different positions (never that much apart)
- a correction is received (DGPS but those are small changes)
- the GPS unit or part of it like antenna has failed

In your case you have to wonder which was the correct position. The one before the jump or the one after that. Or may be you were thrown to an alternate universe that is 4 cables off course

But in anycase, these are malfunctions and should not stop anyone from integrating the AP with GPS.

Let me give an example why it is good to integrate, just to offset all the negative views a bit:

My Simrad/Robertson allows me to point the bow to an entrance and then activate it in a track mode to virtual waypoint. All it needs for that is GPS feed. Upon activating it will create a waypoint far ahead and steer a straight COG to it. I like this mode because we are short handed. The first time we arrived in Curacao heading to Spanish Waters, we felt like going into a funnel of lee shore with cross currents and big waves into a very narrow boca that we couldn't even see. So I pointed the bow to where I was pretty sure the entrance must be and activated this AP mode which compensated for the waves and current, giving me time to check the heading line on the radar and adjust AP a couple of degrees. All this while the Admiral secured the sails and sheets and checked the anchor for readiness. I can now keep a good watch without having to concentrate on steering. I can even go down to pee. Without this feature, the approach would have been more stressful and thus less safe.

ciao!
Nick.

G'Day Nick,

Your last example is a good one. My a/p lacks that easy-access feature, but I'd sure use it if it did. I'm all for thoughtful (and watchful) use of autopilot integration with nav systems, and don't see what all the arguing is about. Having it available does not imply improper use of it, any more than having a tiller implies that you will drive the boat onto a rock.

But, by the way... 4 cables is 2400 feet, not 240 (cable is 1/10 nautical mile by most definitions).

Cheers,

Jim

[s/v Jedi]

Originally Posted by Jim Cate

But, by the way... 4 cables is 2400 feet, not 240 (cable is 1/10 nautical mile by most definitions).

that much! Wow, this is the first time I hear of a GPS stepping that far from position. Clearly something broke and must be replaced because it's now unusable.

About the AP mode: I'm told other brands do this too, so I would check the manal to make sure.

ciao!
Nick.

[Jim Cate]

Originally Posted by s/v Jedi

that much! Wow, this is the first time I hear of a GPS stepping that far from position. Clearly something broke and must be replaced because it's now unusable.

About the AP mode: I'm told other brands do this too, so I would check the manal to make sure.

ciao!
Nick.

Nick, perhaps I misunderstood what you said. What I would like is the ability to simply steer a constant COG without having to establish a GPS waypoint. As in taking a bearing on something and then sailing a constant COG towards it. There are three different pilots on Insatiable II -- Navman G3100s, AutoNav 1500 and an old Autohelm ST4000 tiller pilot. All of these will accept GPS input, none will do what I described. All the required info is there, just no way to execute the plan!

Has seemed odd to me, and i'm glad that someone has adapted the ability. Should I be buying another pilot I'll have a look at Simrad/Robertson.

Cheers,

Jim

[s/v Jedi]

Originally Posted by Jim Cate

Nick, perhaps I misunderstood what you said. What I would like is the ability to simply steer a constant COG without having to establish a GPS waypoint. As in taking a bearing on something and then sailing a constant COG towards it. There are three different pilots on Insatiable II -- Navman G3100s, AutoNav 1500 and an old Autohelm ST4000 tiller pilot. All of these will accept GPS input, none will do what I described. All the required info is there, just no way to execute the plan!

Has seemed odd to me, and i'm glad that someone has adapted the ability. Should I be buying another pilot I'll have a look at Simrad/Robertson.

Cheers,

Jim

You understood the function correctly. There was an earlier thread on CF where members listed APs that could do this, may be you can find it as I can't remember the other brands with this function. Not that you can go wrong with the Simrad pilot

Nick.

[nigel1]

Originally Posted by s/v Jedi

Overall, I think navigation has become much safer with the new aids and we would have had much more accidents without them (basically repeating here what Mark said).


ciao!
Nick.

Agreed, and as most of us been saying, so long as its treated with respect, and limitations are understood.

Maybe the GPS errors were equipment failure. Last year had two Furuno antenna's failed within a fort night, and a few months later the spares fitted also failed, possibly a bad batch, and incidentally, no alarm
We also suffer from antenna cable damage, usually through heat from the engine exhaust being blown on to the mast, 4 x 4000HP engines make for alot of heat.

I can see the advantage you describe in making for the entrance to Curacao (nice place, there last year), but you qualified that by saying that the radar was in use as well. Which is good, and the point I think we are all making, have a secondary fixing method.
On the ship, I will, in a similar circumstances use parallel indexing on the radar, and tweak the auto pilot to maintain the required course to make good

[sidmon]

Originally Posted by colemj

Well, if this statement isn't damning electronics, then I'm missing something.

Yes. You are.

The cite was yet another case of where an integrated bridge did not save a grounding.

And nowhere did I blame the electronics.

Not sure what you are missing about how I have stated this is a HUMAN FACTORS problem numerous times.

And one -no matter the poohpoohing here- WILL find its way into the world of recreational boating.

And, again, it will be a problem partially masked by a much shallower draft of most boats of the denizens that may read this, and also of the light usage pattern in which most boats get used.

But another paycheck bet is on the table that the actuaries at our insurance companies will start to notice.

If they haven't already.

[donradcliffe]

Originally Posted by colemj

How is that a reason not to integrate? In that specific example, the OP doesn't yet know what the problem is (it could be something not to do with integration) and even though he has a problem, the autopilot still works without the integration.

Man, you people see witches everywhere...

Mark

Read the thread again--his autopilot refuses to work if he tries to un-integrate it. There is too much complexity in his integrated system to easily find and fix the problem. The machines are taking control, and I hear the Terminator is looking for you...

[sidmon]

Originally Posted by Ex-Calif

Well I can guarantee you all that the airline rate of groundings is the same as it has been since around 1902. Once per take off.

As we say in the aviation business, take offs are optional - landings are mandatory.

BTW - on some level it is always human error.

The peeing contests are amusing to watch but the only thing grinding my gears are the flip comparisons to commercial aviation. Each false statement makes me want to jump in and correct but this is not an aviation forum. Suffice to say no one so far has posted an accurate airline analogy so please stop. You arre giving me and probably the commercial pilot guys reading this headaches.

BTW - not bragging or "credentialling" but I have been in commercial aviation 30 years and part of the NTSB accident team on 3 commercial accidents. I have been very close to the whole cockpit integration stuff and what some are talking about here falls way short of the level of integration, sophistication and situational awareness benefits that commercial aviation has achieved.

And those tossing out references to Air France (and others) just go read up on airline safety statistics since the 50's and 60's. You will find today's numbers absolutely and convincingly eclipse the super connie days when we still had skylights to take star shots with sextants. And that 50s cockpit with 5 or 6 crew now has 2. Modern is better.

Sorry. It is irrefutable.

The only thing "irrefutable" is Zero AGL. So make sure you ALWAYS approach it on your own terms....

I possess 3 Airman's Certificates (including the one with that little blue logo). And exercise the privileges of one daily to make a pretty fair paycheck...One that is highly dependent on keeping things SAFE (and have been on NTSB investigations too).

The reason I bring up the aviation cases is because the results of "The Vincennes Effect" are both more immediate and brutal, and more systematically studied in the aftermath. And there are direct similariies to crashes like AA at Cali or AF over the Atlantic that directly apply to what I am trying to make folks aware of here in the marine recreational setting.

So if you are what you say you are, then you should be keenly interested in what has come out of the AF447 investigation (an industry colleague I saw today just got back from Paris helping train the good folks there as they continue to transform from the European model to the practices employed at US Pt 121 carriers-But I digress)

And you must know that "Aftermath" is the most popular and first read feature in Flying Magazine. But anyway:


Air France 447. The controversy continues

Rush-Hour Read: How Automation Leads to More Aviation Accidents | INFRASTRUCTURIST

Rush-Hour Read: How Automation Leads to More Aviation Accidents

Pilots’ “automation addiction” has eroded their flying skills to the point that they sometimes don’t know how to recover from stalls and other mid-flight problems, say pilots and safety officials. The weakened skills have contributed to hundreds of deaths in airline crashes in the last five years.

[savoir]

Originally Posted by donradcliffe

Read the thread again--his autopilot refuses to work if he tries to un-integrate it. There is too much complexity in his integrated system to easily find and fix the problem. The machines are taking control, and I hear the Terminator is looking for you...

Hasta la vista - baby !

[colemj]

Originally Posted by donradcliffe

Read the thread again--his autopilot refuses to work if he tries to un-integrate it. There is too much complexity in his integrated system to easily find and fix the problem. The machines are taking control, and I hear the Terminator is looking for you...

His autopilot doesn't refuse to work - it goes into standby when he turns off his chartplotter. It works fine not connected to the chartplotter.

Your statement that there is too much complexity to find and fix the problem is unsubstantiated. Too much for whom? Can't be found and fixed at all?

And again, he does not yet know what the problem is. It could be completely unrelated to the fact that he has it connected to the chartplotter, it could be a broken autopilot or a loose connection. It could be installed incorrectly.

None of that is a valid reason not to have an autopilot communicating with other instruments.

And things do break. Back in the usenet days, there were these types of rabid arguments around whether windvanes should be integrated into the ship's steering or provide independent standalone function.

The more things change...

Mark