Discrepancies found on EclipseDetector and FieldOfViewDetector

#1

Hello,

I have been working with the events detector of OREKIT 9.1 and I have identified the following two issues:
• I have working with eclipse detector (EclipseDetector) and I have detected some anomalies. To be more specific the Eclipse umbra event is not working properly for a LEO polar orbit (a difference of 17 seconds in AOS had been identified w.r.t. to the expected results). Please find below the exact implementation of the EclipseDetector

    CelestialBody earth = CelestialBodyFactory.getEarth();
    double earthRadius = Constants.WGS84_EARTH_EQUATORIAL_RADIUS;
    CelestialBody sun = CelestialBodyFactory.getSun();
    double sunRadius = Constants.SUN_RADIUS;
    // Create event detectors
    final EventDetector ev1 = new EclipseDetector(sun, sunRadius, earth,
                                              earthRadius).withMaxCheck(60.0);

• Furthermore trying to use the FieldOfViewDetector detector for the sensor events(AOS/LOS of the SUN, MOON and EARTH) computation another discrepancies have been identified. First of all the FieldOfViewDetector is taking into account the target’s center instead the target’ shape. Whereas the FieldOfViewDetector for the AOS/LOS of the objects (Sun, Moon, Earth) shall Be a detector which considers body’s radius with an eclipse-umbra detector inside. Please find below the FieldOfViewDetector implementation, where the sensorFOV is of type FieldOfView :
o AOS/LOS of the Sun
PVCoordinatesProvider sun = CelestialBodyFactory.getSun();
FieldOfViewDetector ev = new FieldOfViewDetector(sun,
sensorFOV).withMaxCheck(60.0);
o AOS/LOS of the Moon
PVCoordinatesProvider moon = CelestialBodyFactory.getMoon();
FieldOfViewDetector ev = new FieldOfViewDetector(moon,
sensorFOV).withMaxCheck(60.0);
o AOS/LOS of the Earth
PVCoordinatesProvider earth = CelestialBodyFactory.getEarth();
FieldOfViewDetector ev = new FieldOfViewDetector(earth,
sensorFOV).withMaxCheck(60.0);

Could you please let me know if you intend to correct this issues in the next versions or if there is any specific way to use this detectors in order to avoid this kind of issues.? Thank you in advance.

0 Likes

#2

Hi,

I have been working with the events detector of OREKIT 9.1

Note that Orekit 9.1 is quite old now. The current version is 9.3.1 and version 10.0 is expected
before the Orekit day in May this year.

the Eclipse umbra event is not working properly for a LEO polar orbit (a difference of 17 seconds in AOS had been identified w.r.t. to the expected results)

How were the expected results computed? Is it from another flight dynamics package or from an on-board observation? There are two different causes that could explain this, related to the modeling used in Orekit.

  1. we don’t apply Earth flatness
  2. we don’t apply refraction

As you mention polar orbits, I would say the first cause is the most likely one. So if your reference data comes from another package, could you confirm this package does consider Earth flatness for eclipses?

Could you register a feature request on our bug tracker at https://gitlab.orekit.org/orekit/orekit/issues?

Whereas the FieldOfViewDetector for the AOS/LOS of the objects (Sun, Moon, Earth) shall Be a detector which considers body’s radius with an eclipse-umbra detector inside.

You can already do that for Sun using the margin parameter when you build the field of view, because Sun angular radius is almost fixed. However, this workaround would not work for Earth or Moon as their angular radius will change depending on spacecraft position.

Here again, could you egister a feature request on our bug tracker?

Could you please let me know if you intend to correct this issues in the next versions

If you register these feature requests so we don’t forget about them, yes, we will most probably address them at some point. If you need them very soon (for example in the upcoming 10.0), then you can also add this feature by yourself and contribute it to Orekit so we include them sooner (and credit you for the fix). As an open-source project, contributors are welcome!

0 Likes

#3

Hello Luc,

First of all thank you very much for your answers.

Regarding your questions, please find below some answers:

  • For the Eclipse events I can confirm you that the results were obtained using another flight dynamics package.
  • Regarding the two possible causes of the generating the discrepancies in the Eclipses events for LEO polar orbits I can also confirm you the flight dynamics package that i am using considers the earth flattening for both Umbra and penumbra events (and also for the ocultation events that are used for generating a filter when sensor events are computed). Could you please indicate me if there is any possibility to modify/reconfigure Orekit model so that to take into account the earth flatness?
  • Regarding the bug tracker registration I can confirm you that i will register a feature request.
  • Could you please give me an approximate date/month of the next Orekit release 10.0? If I understood it correctly, the next release is scheduled for May 2019. On that case, is there any chance to include in this release 10.0 any improvements regarding the AOS/LOS FieldOfViewDetectors for the objects Sun, Moon, Earth once I have registered the request?

Thanks in advance!

Cristina

0 Likes

#4

Hello Cristina,

I am not an expert on the event detector part of Orekit. I prefer do not answer this question rather than give you bad informations (Even if I think that this is possible to modify Orekit to take into account the earth flatness). As you have registered the request, we will think about the possibilities to to take into account the earth flatness.

If we encounter any contraries in the schedule, the release 10.0 of Orekit is planned just before the Orekit day, which is held on May 23, 2019.

We have currently a lot of work on the new features that will be introduced in the release 10.0 of Orekit. In that respect, we can not ensure that this corrections will be include in this release 10.0.
As Orekit is an open source project, contributions are always welcome.

Regards,
Bryan

0 Likes