Celestial Navigation with OpenCPN

[Andres_RG]

I have tested the pluggin and the results are not useful for navigation.
I'll try to put some light in this issue in the following days.

In some private e-mails with geckosenator I says:

"I am interested in collaborating, but first I want to know how your work is. I have seen an entry in the OpenCPN Manual Celestial Navigation | Official OpenCPN Homepage
But I am loosing something, I do not understand how this it works, is a very strange solution for celestial navigation: at sea/air the azimuth never is an entry, and the measured angle above the horizon of the body with a sextant, is called altitude, not elevation."
and
"geckosenator : I do not require azimuth for navigation, but for fun I support using it. This allows navigation with only a compass and no sextant.
me: is only for fun, the error is unacceptable for navigation. This method only works in an stationary astronomic observatory with the appropriated instruments, not at sea."
__________________________________________________ _____________
I have updated the "Celestial Navigation" item at Celestial Navigation | Official OpenCPN Homepage including a link to:
Manual of Chart WorK
Example: pdf and GPX file

Regards.

[Andres_RG]

Originally Posted by Andres_RG

I'll try to put some light in this issue in the following days.

OpenCPN, a great and valuable project for navigators!. Since I discovered it, a few months ago, is in my mind the possibility to plot the celestial or the coastal fix on a nautical chart. This prompts me to write the program Chart Work, and the GPX add-on to CelestialFix.exe. This permit to implement on an ECS the traditional work on a paper nautical chart via a GPX file input. Also, CelestialFix plots the celestial circles of position on a Mercator chart based on the WVS -World Vector Shoreline-.
Nearly all ECS lack for such applications for coastal or celestial navigation, the only one I know that has a sight reduction module is NavPack.
Since my programs only run under MS-Windows, the natural way to do it cross-platform is to rewrite the code into a plug-in for OpenCPN.

I propose two solutions:

1-The plugin plots the true celestial line of position: the Circle of Equal Altitude. Input data for each observation is:
· Date and the time UT1.
· Body name
· Geocentric equatorial coordinates of the celestial body: declination and Greenwich Hour Angle (Dec, GHA)
· the semidiameter SD
· the Horizontal Parallax, HP. All obtained from the Nautical Almanac.
· The sextant altitude Hs
· Time for the fix
· Course and speed, (supposed to be constant between the observations).
· It must be calculated:
ü Running fix
ü Correction for sextant altitude.

2-The same but with a built-in Nautical Almanac


The first is the relatively simply, and with collaboration could be implemented in a reasonable time. The second is a whole project, perhaps Astrolabe project could help.

First of all, is very good idea to include a module for celestial navigation in OpenCPN, I think so! But today the celestial_navigation plugin is very inaccurate; the error is so great that is NO suitable for navigation. In astronavigation nothing to be invented, only more accurate methods to solve the problem. The source code, Sight(), for obtaining the observed altitude Ho from the measured one with the sextant Hs is inaccurate due to some simplifications. Theoretically is very easy to obtain the latitude and the longitude from one sight if altitude and azimuth are obtained with and acceptable error, solving the navigational triangle for these variables. But in real world at sea is impracticable. The plugging does not take into account the motion of the observer. Let’s collaborate to do this useful plug-in (for offshore navigators and enthusiasms!) real!!!

Hoping I did not commit any error due to the rush,
Best regards,

References:
Corrections for Sextant Altitude. Andrés Ruiz, Navigational Algorithms
Chart Work and CelestialFix.exe are available at my Web site for free
Celestial Navigation Manual of Chart WorK and an example(pdf and GPX file for import in OpenCPN)
Coastal Navigation
Great Circle and Composite Routes
NavPak http://www.globenav.com/

[Andres_RG]

And example with the error in the altitude and azimuth solutions with the plugin:








Solution with Chart Work

Sun
GHA = 18.238797 º = 18º 14.3'
Dec = 13.659883 º = 13º 39.6'
SD = 15.795038 '
HP = 0.144745 '

Limb: Lower
Hs = 57.0170 = 57º 1.0'
Ho = 57.2246 = 57º 13.5'
Calculated:
Hc = 57.2246 º = 57º 13.5'
Zc = 210.1 º
Fix:
Lat = 43º 19.0' N
Lon = 002º W

[ad_astra]

Nice software! I'm pretty impressed with the screenshots (although I haven't had a chance to try it myself). I tried to implement this sort of calculation several years ago, particularly the fun bit with azimuth. However, I kept banging my head against the wall with calculating the magnetic variation. Do you use a lookup table or the actual magnetic modeling? I was never able to get the magnetic modelling bit to work....

[seandepagnier]

Originally Posted by Andres_RG

The source code, Sight(), for obtaining the observed altitude Ho from the measured one with the sextant Hs is inaccurate due to some simplifications. Theoretically is very easy to obtain the latitude and the longitude from one sight if altitude and azimuth are obtained with and acceptable error, solving the navigational triangle for these variables. But in real world at sea is impracticable. The plugging does not take into account the motion of the observer. Let’s collaborate to do this useful plug-in (for offshore navigators and enthusiasms!) real!!!

Maybe you could elaborate on the actual errors so we can correct them. I would like accuracy to be as high as possible.


I am doing a number of improvements on the plugin, I probably wont have it done for a few days.

[Andres_RG]

Originally Posted by ad_astra

Do you use a lookup table or the actual magnetic modeling? I was never able to get the magnetic modelling bit to work....

A table is impracticable because magnetic declination depends on the position. Deviation depends on the vessel and the magnetic course.

See: Magnetic Declination | Official OpenCPN Homepage the code is open.

Regards,

[Andres_RG]

Originally Posted by geckosenator

Maybe you could elaborate on the actual errors so we can correct them. I would like accuracy to be as high as possible.

Hi geckosenator,
the error to which I refer is the error in position.
Now as I do not have compiled the project, I cannot debug for values of variables, and I don't know if one is in error. As I have commented before, the plugin make some simplifications that lead to the result is not correct
(I'll try to compile the project the next month.)

For example:

• What about deltaT? (NASA - Delta T)
• Refraction must be corrected for T and P
• For the Moon HP and SD must be calculated with great accuracy
• Parallax in altitude and Oblateness of the Earth
• What about non simultaneous sight?

The approach is that I propose before.

Regards

[Capt.Don]

Hi, I just noticed the celestial navigation plugin. Currently I am not a user of OpenCPN so I haven't tried this yet. To me, it is not clear whether this includes the nautical almanac calculations or plotting LOPs. A couple years ago I developed a C# set of astronomical algorithims used for calculating the nautical almanac. I included these in a Silverlight browser celestial navigation site reduction program. I have the routines for sun, moon, planets and navigable stars. I'm not sure if you need this, but I'd be happy to work with you to include in your work.

Regards,
Don

[Andres_RG]

Originally Posted by dmclennan

A couple years ago I developed a C# set of astronomical algorithims used for calculating the nautical almanac. I included these in a Silverlight browser celestial navigation site reduction program. I have the routines for sun, moon, planets and navigable stars. I'm not sure if you need this, but I'd be happy to work with you to include in your work.

Regards,
Don

Hi Don,
I think the current version of the plugin developed by geckosenator includes Astrolabe: Astronomical subroutines and applications & Astrolabe | Download Astrolabe software for free at SourceForge.net for Nautical Almanac calculations.

Thanks for your offer, it could be useful. First source code must be translated from C# to C++, and after tested for some cases versus astrolabe.

Regards,

[Andres_RG]

I am doing a number of improvements on the plugin, I probably wont have it done for a few days.

Probably could help:
· an input/output file facility for sights
· A checkbutton with the option to use built-in Nautical Almanac or an user input one.

regards.

[seandepagnier]

I have done a bunch of updates, mostly user interface, and added lunar sights. Hopefully it will be available soon.

As far as accuracy, if anyone has working code for doing these calculations in any language, I invite you to post the code. I can then compare algorithms and we can discuss differences.

This way we can get the best of all our work, and I am also accepting patches to fix problems or give improvements.

[Capt.Don]

Geckosenator - I've provided a full set of C# classes to calculate the astronomical position of aries, polaris, sun, moon, planets and stars to Andres Ruiz. This reproduces the nautical almanac and potentially replaces having to use Astrolab. Due to size of the project, I can email you direct.

[philocat]

Cruisers,

May I make a contribution and request.

Firstly congrats to the developers of this plugin. In the future I may have pursued this myself so it is nice to see development.

I underscore Cagneys suggestion that this celestial plugin not be used for navigation but it is still a 'nice to have' addition for sight taking practice and as a xcheck although this is probable of limited use these days. In the real world - when and not if- the opencpn technology tool chain breaks one must be able to reduce sights manually. Or perhaps by a hand scientific calculator as I have done in the past pre GPS revolution.

The end result of sight reduction calculation is an azimuth and distance away/toward the celestial body from an assumed approximate initial position - at least the common method I have used. The azuimuth is purely for establishing the direction of the "line" of position which is perpendicular to the azimuth. Technically this "line" is an arc of circle around the earthly position of the celestial body but for a locally small area a line is just as accurate. For position at least one other sight is necessary or preferably three to give a cocked hat. The shape of this is the important bit as it gives an indication of the accuarcy of the LOP's. Apologies for the sermon but I am leading to a suggestion for a third mode to Andres suggestion. I'd like to be able to enter the azimuth distance pairs (and possible assumed location also) from standalone calculations and thus being able to view grahicallythe cocked hat on the screen. I'm quite happy to see inbult almanac method also but with the above caveat.

Magnetic declination has also been mentioned. For celestial purposes this is NOT needed and I'd maintain this should not be in the celestial plugin at all. Having said that it would be a very useful separate plugin when combined with deviation input (as a table). Thanks also to Andres NOAA source code website contribution as I had been wondering about the lack of deviation treatment in opencpn.

[seandepagnier]

I made some important updates to the celestial navigation patch. Please pull the latest opencpn git, and compile it. If I missed any of the requested improvements, I apologize. Please keep giving me feedback so I know what needs to be fixed.

I would be interested in a more detailed explanation of how we can add support for pen/paper math cross checking directly into the plugin. This will allow us to use opencpn celestial navigation as an interactive tool to teach people how to perform the calculations manually without error on their own without a computer or any electronic equipment. I do not know the best way to implement this given the constraints of the program. Any suggestions?

[capt_douglas]

Very nice add to OpenCPN!

I'd like to see the CN portion do:
- tell me the time of nautical twilight based on my current/projected position
- give me a selection of moon/planet/selected stars (all 57 would be nice) at my current/projected position. Data might include bearing and declination.
- semi permanent info storage for boat specific errors
- independent time input so I can sync CN (and maybe computer) to the correct time.
- reverse look-up of navigable bodies (put in bearing and Ho and tell me what object I've chosen) so I can learn the stars
- the ability to add/delete sextant data in case I find more or better objects
- LAN
- compass calibration
- reasonably perpetual (maybe a 50-year almanac)
- a good plot on the chart so I can see how I did compared to the GPS
- access to all the data I'd normally see in the almanac and possibly in the reduction tables (so I can use paper forms for training)
- the ability to save "reduction errors" so I can see if I'm getting more accurate results