All
I recently had cause to play with the AngularSeparationFromSatellite detector, and I was wondering if it was possible that the math had accidentally been performed backwards. In the unit test AngularSeparationFromSatelliteDetectorTest::testCentralSunTransit(), when the code hits the requisite angle threshold of 10 deg, the positions of the relevant objects in ECI are as follows:
Sun: {113,860,438,991.12085; 90,638,739,947.96777; 39,295,688,385.25682}
Acatenango: {3,974,505.016056697; 4,732,962.172846461; 1,586,432.2869105348}
Satellite: {-4,561,348.540946262; -4,031,032.255374609; -3,456,778.802339459}
which makes it pretty clear that the satellite is on the opposite side of the world from both the sun and Acatenango. I think what’s happening is that this equation:
final Vector3D primaryPos = primaryObject .getPosition(s.getDate(), s.getFrame());
final Vector3D secondaryPos = secondaryObject.getPosition(s.getDate(), s.getFrame());
final double separation = Vector3D.angle(primaryPos.subtract(sPosition),
secondaryPos.subtract(sPosition));
is indeed solving for an angle of 10 degrees, except since the origin of both vectors is the satellite object, what you’d really want to solve for is when the angle between the vectors is 170 degrees and the satellite is therefore between the sun and Acatenango.
Is this correct, or am I misreading the intent of the code?