[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Orekit Users] TLEPropagator Question



Thomas Neidhart <thomas.neidhart@spaceapplications.com> a écrit :

Dear all,
Hi Thomas,

I am using this very nice library (many kudos also for making it
available as open-source!) to calculate visibility events for a
satellite based on TLEs.
many thanks, your comments are appreciated

For this purpose I have used the already available example using a
KeplerianPropagator and a custom ElevationDetector to collect the events.

This works so far, but I was a bit concerned about the accuracy of the
resulting events, especially when comparing it with some available
tracking sites on the internet (e.g. http://www.heavens-above.com/), as
results were diverging in the range of minutes after some hours.

If I understand it correctly, the provided KeplerianPropagator is a very
simplistic propagator based on a keplerian orbit, while the
TLEPropagator is way more sophisticated and should yield much more
accurate results.
This is true, but I would not say that the TLE model is  
"sophisticated" and yields "accurate" results. This model is known to  
have many drawbacks and is clearly suited neither for long term  
propoagation nor for accurate orbit propagation. Apart from the model  
by itself, the published TLE data often has large errors. Several  
studies have shown that you may get several hundreds or even thousands  
of kilometers errors propagating TLE a few days (yes, these orders of  
magnitude are in kilometers, not in meters).
Unfortunately, you cannot simply convert a TLE into another orbit type  
using a single point position-velocity pair. The state representation  
(the TLE) and the propagation model (SGP4/SDP4) are tightly bound  
together. Doing such a conversion and using a more accurate numerical  
propagator would therefore not improve things. This is a real problem.
Operational systems use TLE only for first assessment of debris  
mitigation, simply because the available data is in this format. When  
more accurate results are needed (for further analysis if a close  
approach occurs or for control on known satellites), we use other  
types of orbits and different propagation models.

Sadly the TLEPropagator does not implement the
Propagator interface, but I created mysself a simple wrapper for testing
purposes (updating the SpacecraftState orbit with propagated P/V
coordinates from the TLEPropagator instead of propagating the orbit itself).
We are aware of this missing feature and it has been fixed. If you  
want to give it a try, checkout the current development version of  
Orekit from the Git repository (it uses the newly released Apache  
Commons Math 2.2). In this version, TLEPropagator does implement the  
Propagator interface (which has also been improved) and can be used  
with visibility events detectors.
The results are very promising and are almost the same compared to other
sites (~ 1-2s delta). Can someone confirm that my assumptions so far are
correct wrt orbit propagation?
Yes but it does not mean that the results are "better", it means the  
results are consistent with the way the other sites compute them.  
Unfortunately, if you don't have access to better orbit data, you  
cannot have a better result.
best regards,
Luc

Best regards,

Thomas
--
Thomas Neidhart, Software Engineer
Space Applications Services, www.spaceapplications.com
Leuvensesteenweg 325, B-1932 Zaventem, Belgium
Phone: +32-(0)2-721.54.84, Fax: +32-(0)2-721.54.44




----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.