Discussion- Multi-purpose Onboard Computers

[masalai]

Yes "flavours" of linux will happily install on a memory stick (usb flash drive) but be warned these devices are not very reliable or robust over time. - - - the hard drives used in modern notebooks are more robust than the desktop ones as they are made for "mobility"... - - If you are still nervous, mount the "box" in a cupboard suspended by "occie straps" such that it does not bump into external surfaces or risk getting moist from humidity (the latter being the greater risk of failure).

Another option may be to make it boot from a DVD, then you can load a huge range of applications. - Ubuntu and Linux Mint will both fit on a normal CD. If you know how to burn an ISO image then you can download either, burn to CD and with your bios setting to boot from CD they will both do that...

Go here http://www.linuxmint.com/releasenotes.php then download Linux Mint 4.0 Daryna by clicking on download for the "main edition" which includes all the plugins and codecs. then when installed click on system tools end "envy" to download specific video drivers for your ATI/Raedon or nvidia card/chipset.

As a newbie to linux I strongly recommend you download the manual and read it.... things work like UNIX, which is a more secure, robust multiuser operating system.... so you will need to tell it your username (personal access) and a password (DO NOT FORGET IT) of at least 8 characters including upper and lower case letters and numbers. If everything is plugged in to your computer and turned on there is a good chance it will be found and necessary drivers installed or you will be prompted to automatically download and install if necessary....

You can find further stuff here http://distrowatch.com/ and http://www.desktoplinux.com/ and http://www.linuxtoday.com/ and http://www.osnews.com/ and http://tuxmobil.org/linux_gps_naviga...lications.html and from boatdesign net people http://openpilot.sourceforge.net/

Every time you boot up you will be asked to "log in" with your user-name then password. Have fun and learn...

[burgoynebc]

Thanks for the post. Cd boot leaves me with doubts as the flimsyness of the laptop drive concerns me. Certainly good backup in event of disc failure. I just realized that there is a screen keyboard on XP so that should allow the removal of the keyboard. I plan to remove the case, substitute some simple switches for the cheap case mounted switches etc and mount the remains on the underside of a cabinet top. There remains a longer term "fix" required and that is to find a small (say 6"' square) weatherproof display to mirror my monitor in the cockpit. I found that no matter how well I studied my inside chart display, I still had to dive below to make sure I had remembered things right. This was particularly the case when in fog. The ideal of course would be a rack mounted (and expensive) computer but even on a 38ft sailboat there are just not enough places where this could be mounted.
As a point to be made re general purpose boat computer. I do not think that a PC used for navigation should be made to run anything but the operating system, charts and nav program. There is simply too much risk in the longterm of software problems. Particularly with windows. Particulary if any external communictions (internet, wireless etc) are used. I have to completely reinstall XP every year or so just to get my desktop PC to run properly.

[gpsangler]

Hi everyone

As you can see I'm new to boatdesign.net as well as this discussion.

The overall concept is certainly intriguing to say the least!
A few things concern me though.........

a) I think that some people are way over-estimating the performance

[gpsangler]

sorry, I pressed enter and it submitted the post!
Why did it do that?
anyway, I was about to say "over-estimating the performance requirements
of the computer(s)"
A 2GHZ machine should be more than ample for this project - don't forget that sample rates of GPS/AIS/engine management/etc. should be no more
than 1 per second. So a fanless mini-ITX would be perfect.
b) 2 or 3 less powerful PCs would be better than 1 large one, thus introducing an element of redundancy into the equation.
c) If "media entertainment" is required, use another machine.
d) I think the choice of C++ as the main programming language is strange to say the least. I get the impression that there are quite a few members who would like to contribute to the software development side but are daunted by the prospect of it. IMHO, a higher level language, even some of the interpreters, would be far more suitable. Java, Perl, REXX immediately spring to mind.
e) The resilience of the hard disk drive(s) is indeed a problem. Booting from a CD/DVD drive is good - has anyone got this down to a fine art yet? maybe use the little USB stick as the paging and volatile data device?

Anyway, that's my 2 cents worth. I hope to contribute to the project in any way I can, especially on the software side, but please, please not in C++!!

[Tim B]

Well, I suppose it's about time I stuck my head above the parapit and apologised for not disseminating more information about OpenPilot. Unfortunately, 3 week holiday followed by breaking my right elbow have slowed me down a bit. However...

I gave a talk on OpenPilot on Thursday, with a little demo at the end, showing how you basically plug the widgets together. The slides are attached, it's not intended to be documentation, it's just an introduction, but it does give a fair idea of OpenPilot's progress.

Now, gpsangler, I hope I'm not going to offend you, as you have some fair points.

a) Yes, in fact I was successfully demoing OP on a rather low-spec laptop.

b) Doesn't usually work like that with redundancy, and it only works if the software will cope with it. It is something to bear in mind though, especially for bigger boats.

c) Fair enough. I wasn't going to go anywhere near media for the moment, though I do use Linux and MythTV to handle all my home PVR tasks.

d) Ok, why? I personally don't agree, but I'd like to know your reasoning. My choice of C++ was mainly driven by it being Object-Oriented and having QT to handle the graphics. It might not be the easiest language to learn, but it is quite readable. Could I also suggest that you look at OpenPilot and QT4 in detail. You might be surprised.

e) Live CDs/DVDs are commonplace and quite successful now. However, I'm not convinced that a CD/DVD drive is much more robust than a hard drive. I thought we'd concluded that some form of solid-state flash-drive was the way to go here.

There is plenty more to do with OpenPilot, and this project as a whole. Primarily, though, I want to progress on the Serial I/O widget.

Cheers all,

Tim B.

[gpsangler]

You're certainly not offending me, mate!

My point was that C++ is a steep learning curve for the less experienced
potential programmers amongst us. As we seem to agree that the speed of the software might not be of paramount importance, I thought there might be scope for coding some small programs in a language more easily understood. I think we'll have to agree to disagree on this one.

Anyway I'm still in for what it's worth... (even if I have to code in C++ !)

[Tim B]

I agree that there are easier languages to learn, but given the graphical nature of what we're trying to do, I really think it's worth the effort. It will eventually give a nice, coherent code-base for future development.

It might be worth looking at doing some analysis code, possibly web-based, where I can see PHP being very useful (upload a custom log file, then analyse it). ie. plotting speeds throughout a journey, or doing a GPS track with appropriate coastline data. Perhaps even using speed/wind data to produce a measured VPP. There is certainly a lot of stuff that could be done.

Glad you're still on board,

Tim B.

[tspeer]

You've made a lot of progress, Tim!

I'm especially interested in building polars from measured data. One thing I've learned from recording instrument data on a laptop is I almost need to be dedicated to running the laptop and have enough crew to do all the sailing. If there are only two of us, it's very difficult to collect decent data. For one thing, one really needs to be able to start & stop recording so as to only collect data when the conditions are right, and to be able to label the data files. A PDA linked to the laptop would be a useful remote control.

I agree that the serial I/O widget has to be top priority right now! Then we can start acquiring data and help to define requirements based on real experience. FWIW, my data come into the laptop via Bluetooth serial stream, and that's a very convenient way to do it.

[Tim B]

Ok, my primary task for the serial IO widget is to get it handling port settings in both Windows and Linux.

But we need to think of how we want to log the data. I would suggest logging each nmea feed to a single logging widget which can then dump all the data to a single file, with a unix-timestamp. ie.

1300555 $GPGSS,1,2,....

We can then detect the presence of a timestamp by whether there is a $ in the first column, and handle it accordingly. It should also be reasonably easy to extract the data required for whatever you intend to analyse.

A PDA might make for a useful "stopwatch" to extract this (timestamped) data, but I don't want to get into writing code for PDAs just yet!!

Tom, for a permanent OP Installation, I was expecting to use multiple COM ports, but I like the idea of using a virtual serial port which delivers all the data. Since it is handled at OS level it makes no difference to the code anyway. Would you like to become our chief tester for OP?

gpsangler, I guess there's a few things to do. I'll tidy up the Serial I/O widget, and write the logging widget, as that's fairly simple for me to do, and as Tom rightly says, it's top priority. One thing I would like to do, and until now, I haven't done it (mostly) due to time, is to develop a reporting system, so that we can dump a daily (or per voyage) "Ship's Log" from OpenPilot (or a web-based system when a log file is uploaded). I would think the following would be required on an A4 page (pdf or printed format):

Point of departure and time (City/Town plus GPS Location)
Final Location and time (City/Town plus GPS Location)
Intended Destination and time (City/Town)
Distance sailed
max boat speed
average boat speed
max wind speed
max boat speed
Coastline/Chart with GPS Plot

The charting could be done with the existing map widget, but you'll have to set it up to (optionally) use QT's printer device (QPrinter) which is well documented, and basically provides an alternative device to draw on. Let me know if you fancy doing this.

It would be good if anyone who wants to join the project could register with sourceforge (who host OpenPilot) and join both the project and the mailing list. (see website for more details.

Cheers for now,

Tim B.

[Tim B]

Ok, here's a sample log file. I have a little more updating of the widgets to do, then I will upload the new code to CVS (It will be tonight).

Right, time for some Pizza!

Tim B.

[tspeer]

The NMEA streams I've been recording don't have the time stamps. I wish they did. Is this something you're adding to the data, based on the computer's clock?

(And, yes, I'd like to test out your software. Especially if it runs under OS X.)

[gpsangler]

Tim B,

Yes mate I'll have a go at that.

Can't seem to see any "wind speed" data in your log file sample though....
Looks like the "MWV" sentence is the one we want....

Also, what's the format of the charts we're using?

Cheers

[Tim B]

Tom,

The time stamps come from the current machine time. I haven't got access to OSX at the moment, I'll see if I can borrow one of my colleagues machines at the weekend, though.

GPSAngler,

Thanks for agreeing to do the report, You're quite brave to be attempting that!! There isn't any wind-speed data in that log file. Tom posted some a while ago, here:
http://forums.boatdesign.net/showthr...highlight=nmea

The Charts are S-57 format, which is an IHO Binary standard. Currently OP uses GDAL to load the charts, then does all the rendering internally. The raw Coastlines are GSHHS. There are links to the related info on the OpenPilot site.

I'll post the code updates on Sourceforge as soon as I can.

Cheers,

Tim B.

[gpsangler]

Tim B,

Any idea where I can find libmap.a , mate?

or map.lib

or anything else that let's the "map test" link under visual studio express edition 9.0 ??

Thanks,

Graeme

[Tim B]

gpsAngler,

So far, I have used QT/GNU. I will install a copy of MSVC and see if I can get OP to compile.

Tim B.