IBEXLo Star Sensor Software

1
IBEX-Lo Star Sensor Software

• Maciej Bzowski, Marek Hlond
• Space Research Centre PAS, Warsaw, Poland
• Morgan ONeil, Eberhard Möbius
• University of New Hampshire

IBEX Science Working Team Meeting, Dulles,
November 2007
2
Star Sensor key requirements and drivers
3
Star Sensor calibration results
4
What does Star Sensor see?
Solely objects in a strip of the sky 90 (8
1.5) from the Sun, perpendicular to the ecliptic
and including

• diffuse sky background (nuisance) , composed
  of- zodiacal light- unresolved stars and
  galaxies- diffuse light of the Milky Way
  (scattered off dust etc.)
• Stars (main target!)
• 4 planets Mars, Jupiter, Saturn, Earth (the
  latter nuisance, the former possibly valuable
  targets)
• 1 quickly moving target (Moon) (might be a
  nuisance, but might be usable as well)
• Possible transients (bright comets etc. --
  nuisance)

It does not see Sun, inner planets, distant
planets, asteroids
5
Star Sensor Software elements

• Simulation Package (SS/SP for predicting Star
  Sensor signal and its settings in flight
  support for SS/AOP)
• Axis Orientation Package (SS/AOP for checking
  on/improving the determination of s/c rotation
  axis pointing obtained from s/c Star Tracker)
• Tools, Gadgets, Appliances Package (SS/TGA
  miscellaneous software elements to support SP and
  AOP)

6
SS/SP supposed to

• Simulate Star Sensor objects
• Simulate Star Sensor on-board signal and its
  components- bright stars- visible planets-
  Moon- sky background
• Emulate Star Sensor signal downlinked to ground
• Do this for user-selected date(s) during mission
• Provide support for - identification of objects
  observed in reality- pre-planning settings of
  Star Sensor to be uploaded for forthcoming orbits
• Crash mode supplement Star Tracker

7
SS/SP takes

• SPICE kernels for Solar System time
• In-house developed kernel of IBEX orbit
• Sub-catalog of the SAO catalog including stars
  brighgter than m 6.5
• Tabulated zodiacal light model
• Tabulated diffuse stellar/Galactic background
  model (includes stars up to m 6.5)

8
SS/SP presently

• Calculates- sky background with zodiacal light
  and stars up to mag. 6.5- individual stars
  brighter than mag 6.5- Mars, Jupiter, Saturn-
  Moon (in a separate mode)
• Simulates on-board signal components - stars-
  planets
• - Moon
• - background
• Emulates downlinking process (encoding,
  compression, decompression, decoding)
• Can emulate offset of true rotation axis position
  from nominal pointing
• Can simulate random noise in the signal

9
Model of sky background used
Zodiacal light model based on a 10x10 deg mesh
as function of abs(ecliptic latitude) and (l
lSun), interpolated linearly. Component smooth
enough.
Stellar/Milky Way background sim based on a
10x10 deg mesh in equatorial coords,
interpolated linearly. Could use a model on finer
mesh.
10
Construction of background component of Star
Sensor signal
Sky background is diffuse, so Sensor FOV must be
pixelized for simulation. Adopted pixelization
scheme HEALPix.
HEALPix pixels are equal-area, arranged in rings
parallel to the equator of adopted ref. system,
diamond-shaped.
11
Construction of background component of Star
Sensor signal
We use 4096 pixels/(2 Pi ring) - 0.088 deg/pix
(for Moons sake) S/S FOV composed of 232 rings,
hence FOV includes 106 pixels.
The two background models bi-linearly projected
on the HEALPix mesh in FOV resulting pixels
treated as individual light sources.
Signals from all pixels accumulated as the
virtual split-V scans the visible strip
12
Construction of stellar component of Star Sensor
signal

• Hundreds of stars within FOV, but only a few real
  bright ones
• Each of the stars processed individually and
  accumulated in the 720-bin histogram of the sensor

13
Processing of simulated SS signal

• Signal formed on-board
• Signal downlinked
• Background subtracted
• Compared with stars simulation

14
Almost entire nominal mission simulated

• Sims for nominal pointing of s/c rotation axis (4
  deg off Sun at perigee and left alone for entire
  orbit)

• Sims for perturbed pointing of s/c rotation axis
  (12 positions around the nominal pointing

15
Conclusions from sims

• Currently we are in the process of gathering
  experiences and analysing test cases to learn how
  to best program retrieving genuine stars from the
  signal and discarding false positives. Seems like
  about half of the orbits will be usable,
  hopefully more

• Shown example case simulation for nominal axis
  pointing and a true signal with axis offset by
  1.5 deg

Peaks missing where expected axis pointing
different than adopted
clear valid signal
Busy region, many weak stars,peaks blended, will
be rejected
16
Conclusions from sims contd

• Planets excellent targets because bright and
  point-like, moving slowly
• Times of visibility identified for nominal orbit
  3 or 4 occurrences during nominal mission
• Example sole case Mars Jupiter visible
  simulatneously

17
SP Moon mode

• Moon will be seen during 68 passes during 96
  revolutions simulated (roughly every two weeks)
• Extended and quick-moving target, hence
  challenging
• Very easily visible, but needs special obs
  technique (can move across bins even during
  16-rec obs cycle)
• Simulated as pixelized object
• Some quirks need attention (center of light ?
  center of mass, apparent size variable from 0.25
  deg to 4 deg!)
• Use to axis determination expected at challenging
  orbits (when stars problematic)

18
SS/AOP supposed to

• take Star Sensor data fm telemetry
• take diffuse background (inhomogenous) from
  simulations
• subtract the background to calculate the genuine
  stellar/planetary signal
• identify valid objects
• take their positions in s/c ref. system
• identify corresponding objects in star catalog
• calculate rotation axis pointing

19
SS/AOP presently

• Can take simulated telemetry signal in own
  arbitrary format
• Can obtain stellar signal by subtracting
  simulated background
• Can identify easy objects (well defined,
  high-contrasted pairs of peaks)
• Given observed/catalog positions, can calculate
  s/c rotation axis

Challenge have the SS/AOP determine axis
orientation without inserting human
intervention/judgement
20
SS/AOP still needs to

• Improve algorithm to find valid stars and
  planets(research with simulated test cases in
  progress)
• Improve algorithm to associate found object with
  catalog entries
• Validate the method used to calculate rotation
  axis pointing

21
SS/TGA supposed to

• Provide support for SP and AOP and their use
• Do/facilitate miscellaneous tasks related to SP
  and AOP usage

22
SS/TGA include

• Tool to calculate nominal position of s/c
  rotation axis based on the nominal orbit -
  implemented in C, - using SPICE, - supported by
  a Mathematica gadget to construct Unix shell
  scripts for running simulations)
• Appliance to check visibility of Moon, planets
  and Earth during mission - implemented in C, -
  supported by miscellaneous gadgets prototyped in
  Mathematica)

23
SS/TGA include (continued)

• Gadget to perturb the orientation of s/c rotation
  axis to support development of positioning
  algorithm - prototyped in Mathematica-
  calculates offsets with respect to nominal axis
  pointing based on user-defined recipe- returns
  shell scripts to perform simulations
• Extraction tool to prepare a subset of star
  catalog with stars brighter than selected
  magnitude - prototyped in pascal- expected to
  be used just once (which has already happened)
• Appliance to prepare SPICE Kernel of s/c orbit
  based on expected s/c orbit (provisional
  supposed to be superseded by ISOC tools)
• (more to come, as needs show up)

24
Star Sensor Software Package elements from
external sources (integrated)

• NAIF SPICE (for orbital calculations, checking
  visibility conditions of celestial objets etc.)
• HEALPix (for pixelization of the sky to simulate
  the diffuse background)
• Various Open Source tools (visualization etc.,
  TBD)

25
Final message

• Star Sensor believed to be fully capable to do
  its job
• However, related software must be much more
  complex than oroginally thought