AIS crossing Shipping Lanes

Too dumb to solve a basic piece of Euclidean geometry, but smart enough "to boldly go where no sailor robot has gone before".

Most observed intelligent behaviour is instinctive and not the result of logical reasoning. The top Go players in the world cannot explain their moves, they just have a feel for how to win a game. The recent major advances in AI have mimicked instinctive intelligent behaviour.

Captain Kirk of the USS Enterprise or a British Airways pilot landing at Heathrow in low vis, do not feel like lesser men because they let a computer drive their craft, so why do the little Captain Blythes of YBW feel so threatened by the same?

TAs to your fear of ruining a major patent application with your random tech musings. Well I expect our friends in Silicon Valley will get over it in time :encouragement:
I suspect the market for providing septuagenarian Westerly owners with a cognitive crutch on their annual pilgrimage to St Peter Port is too small to get a Silicon Valley venture capitalist salivating.
 
Oh dear oh dear oh dear.. not he Show a neural net a gazillion pictures and it can do anything garbage.
...
The fully trained robot will screw up.
Let me think for a moment, who should I be influenced by?

  1. Some crotchety crofter from the Western Isles who has probably been in a bad mood for 3 years since the Referendum.
  2. The countless legions of investors and engineers at the largest IT companies who are currently rather excited about AI and throwing $10's of billions of investment into exploiting the recent AI breakthoughs?
Hmmm
 
It's probably far easier to automate the ColRegs than to programme up self-drive cars. No doubt the techniques developed for self-drive cars will eventually feed across.
Indeed and probably easier than getting a Ford Focus to auto park, which is a solved problem.

What must on no account happen is to allow these AI systems to ever access an on-line ColRegs debate. They might get funny ideas.
I am genuinely hopeful that within a generation the human race will be able to throw away the ColRegs as "not fit for purpose". All debates I follow here confirm this opinion.

If every vessel has a thinking auto pilot installed that was trained through AI techniques not to hit other vessels, then providing different branded AI pilots are cross trained on each other then the ColRegs no longer serve any purpose. When the ColRegs were first created I assume they were just a formalization of existing helming convention so what is wrong with fusing today's convention in the brains of AI autopilots and letting maritime convention evolve digitally from that point onwards.
 
Let me think for a moment, who should I be influenced by?

  1. Some crotchety crofter from the Western Isles who has probably been in a bad mood for 3 years since the Referendum.
  2. The countless legions of investors and engineers at the largest IT companies who are currently rather excited about AI and throwing $10's of billions of investment into exploiting the recent AI breakthoughs?
Hmmm

Yeah, someone you know virtually naff all, about or a bunch of "merchant bankers" that have a seriously invested interest.
never mind the personal crap, if you dont address the statistics I posted then you are guessing.
I had this conversation with an AI proponent about 25 years ago. Your knowledge of AI then was what? Non existent?
 
I suspect the market for providing septuagenarian Westerly owners with a cognitive crutch on their annual pilgrimage to St Peter Port is too small to get a Silicon Valley venture capitalist salivating.

All this ill-informed venom because someone asked a simple AIS vector question ...wow!!
 
Yeah, someone you know virtually naff all, about or a bunch of "merchant bankers" that have a seriously invested interest.
never mind the personal crap, if you dont address the statistics I posted then you are guessing.
I had this conversation with an AI proponent about 25 years ago. Your knowledge of AI then was what? Non existent?
A post which illustrates you cannot articulate the difference between merchant bankers who try to spread or dilute risk and venture capitalists who track down risky opportunity.

I had this conversation with an AI proponent about 25 years ago. Your knowledge of AI then was what? Non existent?
You had a chat with an AI expert 25 years ago! What a shame Google did not consult with you prior to wasting £400 million on a 4 year old London start-up called DeepMind 3 years ago, they specialise in those hopeless neural nets, well hopeless according to you.

33 years ago I attended a few BSC talks in London on Expert Systems that were related to Margret Thatcher's 5th Generation project. By the time of your chat with an AI expert hopes for AI were waning and the subject was entering one of its winters. Fortunately a British born scientist by the name of Geoff Hinton perservered with neural nets and made some breakthroughs. He is now referred to as the Godfather of Neural nets.

A thinking autopilot that could steer a yacht through the channel shipping lanes is well within the scope of current AI applied practice, whether there is the commercial incentive to pursue such a niche product given the rock star salaries commanded by leading AI experts is another matter.
 
Most observed intelligent behaviour is instinctive and not the result of logical reasoning. The top Go players in the world cannot explain their moves, they just have a feel for how to win a game. The recent major advances in AI have mimicked instinctive intelligent behaviour.

Captain Kirk of the USS Enterprise or a British Airways pilot landing at Heathrow in low vis, do not feel like lesser men because they let a computer drive their craft, so why do the little Captain Blythes of YBW feel so threatened by the same?


I suspect the market for providing septuagenarian Westerly owners with a cognitive crutch on their annual pilgrimage to St Peter Port is too small to get a Silicon Valley venture capitalist salivating.

A good number of year ago I worked for a company that produce a collision avoidance system base on the concept of calculating an area where a collision may occur. All the ship needed to do was avoid that area. On that basis an automated system programmed to avoid these collision potential areas would be easy.
 
A good number of year ago I worked for a company that produce a collision avoidance system base on the concept of calculating an area where a collision may occur. All the ship needed to do was avoid that area. On that basis an automated system programmed to avoid these collision potential areas would be easy.
This is an example of hand crafted explicit logic, such systems tend to fail as the amount hand crafted rules expands. Your suggestion would not acquire behavioural conformance with the ColRegs. A trained neural net would.
 
Last edited:
All this ill-informed venom because someone asked a simple AIS vector question ...wow!!
I am a positive person with a natural optimism about the future, I do however react to ill informed negativity frequently postulated in this forum.

No wonder the Aussies call we Brits "whinging poms", this forum provides daily evidence. I should not be surprised given the demographics of this place, I suppose YBW provided a refuge when Saga closed down its social forums due to all the inter pensioner spiteful behaviour.

The OP should be commended for quickly homing in on a functional deficiency of many AIS displays and his open question raised in this thread.
 
Thankyou for the posts that suggested OpenCPN - it does exactly what I want - shows both your own boat and AIS target position at TCPA, with the two points joined by a line which they call CPA line - really cool !

So my next problem is... where do I find UK charts that I can download for this product ? Grateful for any help. And do I need to buy a Vesper transponder too ?

I find it amazing that the big chartplotter companies can't display this simple but 'essential' info. I guess it's like displaying live current streams with arrows - it was available on specialist computer software around 2004 but didn't appear on mainstream chartplotters until recently.
 
I used to teach navigation and was regularly horrified by the blind belief of students in modern tech. People thinking they could go into a harbour entrance at 10 kn in thick fog on GPS alone believe it or not. Their problem wasnt that they were too technical but that they werent technical enough to realise the confidence limits.
You might not like it and be horrified, but I once did Cherbourg - St PP in thick fog from Cap de la Hague on DECCA. I didn't see a thing all the way down the Little Russell, I turned right when the DECCA said that I was outside the Harbour, then saw the glow of the light on the wall. The hairiest trip I have ever made, but what else can you do except to put your faith in your instruments with the currents and depths round there?
 
A post which illustrates you cannot articulate the difference between merchant bankers who try to spread or dilute risk and venture capitalists who track down risky opportunity.


You had a chat with an AI expert 25 years ago! What a shame Google did not consult with you prior to wasting £400 million on a 4 year old London start-up called DeepMind 3 years ago, they specialise in those hopeless neural nets, well hopeless according to you.

33 years ago I attended a few BSC talks in London on Expert Systems that were related to Margret Thatcher's 5th Generation project. By the time of your chat with an AI expert hopes for AI were waning and the subject was entering one of its winters. Fortunately a British born scientist by the name of Geoff Hinton perservered with neural nets and made some breakthroughs. He is now referred to as the Godfather of Neural nets.

A thinking autopilot that could steer a yacht through the channel shipping lanes is well within the scope of current AI applied practice, whether there is the commercial incentive to pursue such a niche product given the rock star salaries commanded by leading AI experts is another matter.

£400,000,000 and they have what to show for it? 30+year old concept with billions sunk into it and we have...... smile recognition?? IS that all?! OK so probably a bucket load more but how many commercial products? How many white goods run AI? Squat, thats all, just because some large companies are pouring $ down the drain does not make it work.

[crystal ball]10 years from now AIS will be on the way out, going the way of the sextant, well probably just dying, sextants will still work just fine. 10 years from now your boat computer will have satellite access to t'interweb and be part of IoT (gawd forbid!) Vessels will communicate and negotiate avoidance with trivial algorithms, Not very much more complicated than CPA is right now. The communication will not only include speed, course, change in both but capabilities and cost of manoeuvre.[/crystal ball]
 
Most observed intelligent behaviour is instinctive and not the result of logical reasoning. The top Go players in the world cannot explain their moves, they just have a feel for how to win a game. The recent major advances in AI have mimicked instinctive intelligent behaviour.

A post that shows you cannot articulate the difference between instinct and sub-conscious. The difference between a new born's fascination for circles and inspiration.
The AI mimicry is apt in as much as the programmers wont have a clue as to the details of the AI instance.
 
Thankyou for the posts that suggested OpenCPN - it does exactly what I want - shows both your own boat and AIS target position at TCPA, with the two points joined by a line which they call CPA line - really cool !

So my next problem is... where do I find UK charts that I can download for this product ? Grateful for any help. .......

If you wait until the end of the month, 2017 UK raster charts for OpenCPN can be purchased for a very reasonable price from:
http://www.visitmyharbour.com/download-unified-charts/
 
Apologies I think they do include the channel (somewhere else on their website it said something about copyright problems with the Channel)
 
Thankyou for the posts that suggested OpenCPN - it does exactly what I want - shows both your own boat and AIS target position at TCPA, with the two points joined by a line which they call CPA line - really cool !

So my next problem is... where do I find UK charts that I can download for this product ? Grateful for any help. And do I need to buy a Vesper transponder too ?

I find it amazing that the big chartplotter companies can't display this simple but 'essential' info. I guess it's like displaying live current streams with arrows - it was available on specialist computer software around 2004 but didn't appear on mainstream chartplotters until recently.

You don't actually need any charts to use OpenCPN as an AIS display. If you use it without charts it will be like a radar display AIS. The Vespermarine AIS app VmAIS which I use is excellent and is one of the "radar" style displays so has no charts. Once you are away from the coast the land masses are probably off the display anyway once you zoom in a bit so even, if you had the charts, you might well just be looking at open sea.

Of course, if you want to use OpenCPN as a plotter as well as an AIS display, then you definitely need charts. :)

You don't need a Vesper receiver/transceiver but you do ideally need an AIS transponder which will transmit the AIS data over wifi to easily enable you to get the data onto your PC or tablet. Digital Yacht and other manufacturers also do wifi enabled kit. You could use kit which will send the data down a USB/NMEA/serial connection and then use suitable converters but you do then end up tied to the location of the transponder by the cable.

Hope this helps.

Richard
 
Thanks Richard. Great info- I think I'll have go at getting this to work with my existing raymarine 650 AIS transceiver via cable to Mac/openCpn. Good project for the Winter!
 
Thankyou for the posts that suggested OpenCPN - it does exactly what I want - shows both your own boat and AIS target position at TCPA, with the two points joined by a line which they call CPA line - really cool !
The highlighted CPA line is a really useful feature of OpenCPN. It will stay constant as long as target and own vessel COG and SOG remain constant, however any change in course or speed of either results in visible change in the red and yellow dashed CPA vector.
 
Top