PC Autopilot program
- Thread starter bendyone
- Start date
snowleopard
Well-Known Member
Quite right. While I was reading this thread, Microsoft decided to re-boot my computer to install some updates. No warning, no confirmation dialog, just all my applications shutting down. Arrogant barstewards!
Anyone who uses Windows for a real-time critical application is asking for trouble!
If we decide to give it a go, here are the main areas to address:
A routine to read sensor data (compass, rudder sensor, heel indicator). An output to the motor (a reversing on/off relay). User input to select setup parameters, e.g. helm sensitivity and input course, destination waypoint, variation & deviation. Algorithms for calculating CTS, XTE etc, correcting for tidal set & leeway, sea state adjustment and rudder offset for heeling.
Interfacing to various chartplotters would be tricky but readin NMEA is straightforward.
Not trivial but not beyond the wit of a good programmer. Testing and tweaking would take a lot of time though.
I'm up for it if someone wants to build the hardware.
bendyone
Well-Known Member
I'm up for the hardware side of things
ShaunG
Well-Known Member
Me too, Not an expert in any area but a good general all rounder. IT network cosnsultant by trade with lots of dealing in RS232/NMEA data strings.
I will do whatever is needed - let me know.
how do we go forward from here?
I will do whatever is needed - let me know.
how do we go forward from here?
Jeannius
Well-Known Member
I'm definitely up for this if only as a backup to my Raymarine kit.
I've been writing software for more than 30 years including controlling stuff via rs232 using Intel Assembler. rs232 serial to usb is easy. Also write web based stuff so my skills are still up to date. Not to clever on the hardware side of electronics though.
bitman
Well-Known Member
agreed, but if you turn off auto updates then you don't have those reboots. it's better anyway since then you can control when (with internet access) to install updates.
but if you use that (re-comment) computer only for steering the boat etc you don't want internet access or running other software on it. for sure no internet browsing! to get grib files on it you can simple copy those files to it.
but if you use that (re-comment) computer only for steering the boat etc you don't want internet access or running other software on it. for sure no internet browsing! to get grib files on it you can simple copy those files to it.
snowleopard
Well-Known Member
I have always found that windoze based computers have crashed on a regular basis, not just with updates. My current laptop for example just goes blank about once a day for no apparent reason. Vista doesn't have the BSOD and is more resilient than XP but it still falls over.
ShaunG
Well-Known Member
My boat pc is only ever used for boaty stuff.
Is is Windows XP, No service pack, No Anitvirus, no firewall, all but essential services disabled, nice looking bits turned off etc. In essence the windows setup is for the Chartplotter and SSB connections only. It has not and never will eb connected to the Net.
I work in the IT, (specialised area) and a similar setup has been adopted for 24/7 systems and very very few failures of the in excess of 20000 desktops in use are found.
Windows PC's can be reliable, However they do need to be setup for purpose, if you want a multirole machine (Games, internet etc) then a boat PC is not for you.
Boat PC has never had a problem, but then again as you can see above there is so little of the windows nice bits running that there is not a lot to go wrong.
A bit like buying a turbo, fuel injected car and taking it all off, not as fancy but a damm site more reliable
Is is Windows XP, No service pack, No Anitvirus, no firewall, all but essential services disabled, nice looking bits turned off etc. In essence the windows setup is for the Chartplotter and SSB connections only. It has not and never will eb connected to the Net.
I work in the IT, (specialised area) and a similar setup has been adopted for 24/7 systems and very very few failures of the in excess of 20000 desktops in use are found.
Windows PC's can be reliable, However they do need to be setup for purpose, if you want a multirole machine (Games, internet etc) then a boat PC is not for you.
Boat PC has never had a problem, but then again as you can see above there is so little of the windows nice bits running that there is not a lot to go wrong.
A bit like buying a turbo, fuel injected car and taking it all off, not as fancy but a damm site more reliable
Anthony
Well-Known Member
I am up for helping out too, I am a Marine Electrician, with a previous background in IT and electronics.
Maybe we should move this to a seperate place, or setup a private yahoo group for it?
I think there also needs to be an initial decision as to whether it is definatley going to be PC based or not. If its PC based then I'd suggest somthing like VB.net, if its not, then one of the modern PICs could do the job, which can also be programmed with basic, dont have to manipulate indivudal registers etc.
Ants
ShaunG
Well-Known Member
I have created a group in here, if you go to community and then social groups there is a group for this project. i have invited everyone who contributed to this thread, i am guessing you will get it as a PM
Anthony
Well-Known Member
Great idea, I forgot we have a shiney new forum with new features to play with 
Phoenix of Hamble
Well-Known Member
Great idea folks....
I can see some very interesting extensions to the software that once the basic controller is written, would really make it different from a standard autopilot...
such as:
Logging of STW vs wind direction, strength and relative angles..... keep logging it for long enough and you could build up some very interesting polars that would allow you to predict the optimum course and hence VMG for current conditions if it was associated with plotter software and waypoints
The same feature could give you a read out of current 'efficiency'.. eg what percent of the boats potential you are sailing to given current wind/direction/relative angles, and hence prompt better sail tuning....
If you also have AIS, then it could have an emergency 'the skipper is asleep/distracted/disabled' function, that at certain extreme close ranges, makes some decisions... (perhaps this suggestion is one with some baggage!!!)
I'm sure there are more with thought.....
I can see some very interesting extensions to the software that once the basic controller is written, would really make it different from a standard autopilot...
such as:
Logging of STW vs wind direction, strength and relative angles..... keep logging it for long enough and you could build up some very interesting polars that would allow you to predict the optimum course and hence VMG for current conditions if it was associated with plotter software and waypoints
The same feature could give you a read out of current 'efficiency'.. eg what percent of the boats potential you are sailing to given current wind/direction/relative angles, and hence prompt better sail tuning....
If you also have AIS, then it could have an emergency 'the skipper is asleep/distracted/disabled' function, that at certain extreme close ranges, makes some decisions... (perhaps this suggestion is one with some baggage!!!)
I'm sure there are more with thought.....
ShaunG
Well-Known Member
Blimey Morgana
there's food for thought. I like the idea of some of that being part of the pilot.
i have most of it, Polars, VMg, % Performance etc in seapro at the moment and then sent to mast displays.
However i can see some massive advatages of being able to set the pilot to steer to bets % performance or VMG - Not in coastal waters obviously
feel free to join the group
there's food for thought. I like the idea of some of that being part of the pilot.
i have most of it, Polars, VMg, % Performance etc in seapro at the moment and then sent to mast displays.
However i can see some massive advatages of being able to set the pilot to steer to bets % performance or VMG - Not in coastal waters obviously
feel free to join the group
lw395
Well-Known Member
Autopilots
I believe that the hardware side of things and interfacing are surmountable. Fundamental requirement will be some sort of compass sensor. This could be hard to do cheaply.
The most difficult part technically is probably the calculation of how much rudder to apply at any given time.
One can mock commercial attempts at this, but it's not trivial and I guessing Raymarine et al have spent a few hours on it over the years!
It's easy to do a very crude solution that will oscillate, weaving along, but you have to optimise the amount of weaving you will tolerate against the ability to hold course in adverse conditions.
Likewise if you are trying to steer to the wind, you have to filter out the small variations, which tends to imply a delay in the response. If the response is not exactly right, either a windshift will cause oscillation or the steering will be too slow. A human helmsman is doing a lot of subconscious processing. The human also has more inputs, for instance feeling the boat heel will make you think a larger steering response will be needed.
I would be interested in having a go, I have some experience of control theory, and a fairly dull book or two about it!
It's a big difference whether you are looking for something to motor in a straight line while you put the sails up, get you to Cherbourg, or keep the boat from broaching in weather with the kite up.
I'm not sure a PC is the best platform. I would consider a PIC or other microcontroller, with a PC being used to display data and maybe provide numbers to the algorithm.
I believe that the hardware side of things and interfacing are surmountable. Fundamental requirement will be some sort of compass sensor. This could be hard to do cheaply.
The most difficult part technically is probably the calculation of how much rudder to apply at any given time.
One can mock commercial attempts at this, but it's not trivial and I guessing Raymarine et al have spent a few hours on it over the years!
It's easy to do a very crude solution that will oscillate, weaving along, but you have to optimise the amount of weaving you will tolerate against the ability to hold course in adverse conditions.
Likewise if you are trying to steer to the wind, you have to filter out the small variations, which tends to imply a delay in the response. If the response is not exactly right, either a windshift will cause oscillation or the steering will be too slow. A human helmsman is doing a lot of subconscious processing. The human also has more inputs, for instance feeling the boat heel will make you think a larger steering response will be needed.
I would be interested in having a go, I have some experience of control theory, and a fairly dull book or two about it!
It's a big difference whether you are looking for something to motor in a straight line while you put the sails up, get you to Cherbourg, or keep the boat from broaching in weather with the kite up.
I'm not sure a PC is the best platform. I would consider a PIC or other microcontroller, with a PC being used to display data and maybe provide numbers to the algorithm.
fireball
Well-Known Member
Interesting - perhaps something that could be factored in - just increase the rudder correction factor by a multiple of the degrees of heel - would need to be a user variable though!
Jeannius
Well-Known Member
I agree but... a PC is what most people are going to have and I think we should be trying to limit the amount of extra hardware required.
I believe we should try and do this as simply as possible. My Raymarine 7000+ is a superb piece of kit that we couldn't hope to replicate. I would be happy if we can come up with a get-you-home solution for when the main autopilot fails. The thought of my autopilot failing half way across the Pacific is terrifying. I don't have any statistics but my guess is that it is the electronics that fail most often.
If we can come up with something that just instructs the drive mechanism to move so enabling steering a course to a heading provided by a pc and a basic gps that would be a great start.
snowleopard
Well-Known Member
We met a couple in Bermuda whose autopilot computer had failed on the way from Antigua. They ordered a replacement but it was the wrong one so they took on an extra crew to hand steer all the way to the Azores. Not funny.
It's interesting though how few people are prepared to even try to rig lines from the sails to the tiller/wheel. I once sailed 800 miles after my steering snapped off at the rudder head. It wasn't as hard as you might think.
ShaunG
Well-Known Member
hey snowleopard.
have you recevied your invite to the autopilot group? getting quite busy now and nimbus has setup a fantastic new wiki as start as a devolopment start point. Are you still up for the some of the software side?
have you recevied your invite to the autopilot group? getting quite busy now and nimbus has setup a fantastic new wiki as start as a devolopment start point. Are you still up for the some of the software side?
snowleopard
Well-Known Member
Certainly interested. Not heard anything yet.
ShaunG
Well-Known Member
I am guessing that it may come in as a PM but not sure as i set it up. if not go to Community>Social groups and you will find it there. There is a join function in the top right. The wiki that Nimbus has setup is looking fantastic and moving quickly
