ap-6-11.dvi


Acta Polytechnica Vol. 51 No. 6/2011

Multi-functional star tracker — future perspectives

J. Roháč, M. Řeřábek, R. Hudec

Abstract

This paper focuses on the idea of a multi-functional wide-field star tracker (WFST) and provides a description of the
current state-of-the-art in this field. The idea comes from a proposal handed in to ESA at the beginning of 2011. Star
trackers (STs) usually havemore than one object-lens with a small Field-of-View. They provide very precise information
about the attitude in space according to consecutive evaluation of star positions. Our idea of WFST will combine the
functions of several instruments, e.g. ST, a horizon sensor, and an all-sky photometry camera. WFST will use a fish-eye
lens. There is no comparable product on the present-daymarket. Nowadays, spacecraft have to carry several instruments
for these applications. This increases theweight of the instrumentation and reduces theweight available for the payload.

Keywords: star tracker, horizon sensor, all-sky camera, photometry, astrometry, space variant image processing.

1 Introduction
Attitude determination functions are basic functions
in all space applications. Attitude information is
used in Attitude and Orbit Control Systems to sta-
bilize a spacecraft (S/C) in the required position and
orientation. Variousattitude reference sources canbe
used for obtaining attitude information, e.g. Earth
horizon sensors, which can be directed towards the
Sun, Moon, or other planets and stars, Sun sensors,
magnetometers, and star trackers (STs). However,
onlySTsprovideaccuracyof attitude estimationbet-
ter than 30 arcseconds [1]. The first generation of
STs acquired only a few bright stars in the Field-
of-View (FoV) and provided the focal plane coordi-
nates of these stars [2]. The coordinates were not
related to inertial space, and the attitude informa-
tion therefore had to be provided indirectly by an
external unit. This was due the insufficient power of
the microcomputers. Rapid improvements in micro-
computer power led to the autonomous functional-
ity of STswhich became the preferred attitude refer-
ence source. Basically, the secondgenerationSTcon-
sists of an electronic camera and amicrocomputer as
shown in Figure 1. The ST autonomously performs
pattern recognition of stars in FoV, compares the re-
sults with the star catalogue stored in the internal
memory, and estimates the attitude. Generally, the
ST has to operate in two modes. The first mode
solves the lost-in-space function which does not have
previous attitude information available, and there-
fore the star pattern in theFoVhas to be recognized.
The identification can be usually accomplished in a
few seconds. The othermode performs star tracking,
which assumes that the current attitude is closely re-
lated to theprevious attitude. Thismode is easier for
computation because it only tracks previously identi-

fied stars at knownpositions. The basic drivers of ST
designs take into account: accuracy, reliability,mass,
power consumption, and size. Typical requirements
are stated in Tab.1. There are also other parameters
that influence the design. They are: star light sen-
sitivity, the detection threshold, the average number
of stars in the FoV, and sky coverage.

Fig. 1: Principle sketchof a secondgeneration star tracker

Generally, second generation STs employ more
than oneCameraHeadUnit (CHU) consisting of the
baffle, the lens, and the optical sensor. It usually
has about 20◦ ×20◦ FoVThemulti-CHU concept in-
creases the average number of stars in the FoV and
the accuracy of the ST. In addition, due to this con-
cept full immunity against simultaneous blinding is
ensured

2 New concept of a
multi-functional star tracker

Second generation STs already have better parame-
ters than required, see Table 1. Nevertheless, they
use the multi-camera-head concept to increase the
average of stars in the FoV. To decrease the mass
and size of the current ST concept and to increase
its functionality we proposed to design, develop,
andverify themulti-functionalwide-field star tracker

61



Acta Polytechnica Vol. 51 No. 6/2011

Table 1: Typical requirements for autonomous star trackers and an example of the Micro Advanced Stellar Compass
(μASC) flown on the PROBA-2 technology demonstration satellite

Required μASC

Initial acquisition (lost-in-space function) < 1 min. 30 msec

Accuracy (EOL) (arcsec) 30 (3σ) 2 (3σ)

Attitude rate (deg/sec) up to 1 up to 20

Availability (%) 99.900 99.995

Power (W) < 10 < 4

Mass (kg) < 2 < 0.5

(WFST), which combines the functions of several in-
struments usually used in S/Cs.
One function of theWFSTwill be to provide pre-

cise attitude determination. This will be ensured
by sophisticated algorithmsusing thewidest possible
FoV optic characteristics, supported by innovative
image and data processing, a 3D Inertial Measure-
ment Unit (IMU) consisting of accelerometers and
angular rate sensors. For low orbit applications, the
system will also be equipped with a GPS receiver.
All available sources of information, seeFigure 2,will
be fused by the Extended Kalman Filter (EKF). A
two-speedupdating approach, as proposed in [3], will
be used for mechanizing and computing the inertial
navigation equations. The precision will be between
1 arcminute and 1 arcsecond, according the applica-
tion and the optics that are used.

Fig. 2: Available data sources for attitude evaluation

The second function focuses on having cost-
effective means to provide a high-potential informa-
tion source for scientific analyses of objects in the
FoV of theWFST,which extends the applicability of
WFST. This idea corresponds fully with ESA Cos-

mic vision 2015–2025, where the priorities are planet
andUniverse studies. WFSTwill be capable ofmon-
itoring fluctuations in the radiation of stars. High
potential will be provided by the fish-eye lens, which
will enable observations of up to 180 deg sky, and
by an innovative approach to image processing. Sys-
tems of this kind using a fish-eye lens are commonly
referred to asUltraWideFieldCamera (UWFC) sys-
tems. UWFC image data analysis is very difficult in
general, due to the wide range of distortions in wide
field images and mainly due to the spatial variant
(SV) properties of UWFC systems. Moreover, the
objects in ultra-wide field images are very small (a
few pixels per stellar object).
Precise astronomical measurements (in astrome-

try and photometry) need high quality images. Even
small distortions may lead to inaccurate determina-
tion of the position or movement of stellar objects.
The error depends on themagnitude of themeasured
star — the higher it is, the higher the error can be.
The properties of UWFC astronomical systems and
the specific visual data in astronomical images lead
to complicated evaluation of the image data. These
UWFC systems containmany different kinds of opti-
cal aberrations that havenegative impacts on the im-
age quality and system transfer characteristics. For
precise astrometry and photometry over the entire
FoV it is therefore very important to comprehend
howtheoptical aberrationsaffect the transfer charac-
teristics, howthe astronomicalmeasurementsdepend
on the optical aberrations, and how the wavefront
aberration affects the point spread function (PSF).
Precise stellar profile fitting is very important for

astronomical measurements. Nowadays, there are
two common functions for fitting stellar profiles, re-
ferred to as Gaussian andMoffat functions [4,5]. Ef-
forts are being made to match a star’s profile with
the Gaussian or Moffat profile and to remember the
parameters of the fit. If a star were ideal, the stellar
profile would be represented by a small “dot”. Due
to many different distortions, the dot is blurred all
around. The more it is blurred, the worse is the PSF
of the whole imaging system. The centre of a star

62



Acta Polytechnica Vol. 51 No. 6/2011

usually has a profile corresponding to the Gaussian
function. The more distant parts of the star are,
the closer the profile will be to the Moffat function.
Therefore, the ideal fitting function should combine
these two profiles. The properties of the UWFC sys-
tem are referred to as space or spatial variant prop-
erties, and the PSF of this system is different for
each point in the object plane. In our case, a sophis-
ticated algorithm employing special functions based
on Zernike polynomials has to be applied in order to
get a precise stellar profile fit for the entire FoV.

3 Conclusion
Theresearchanddevelopmentof themulti-functional
WFSTwill verify current technologypotentialsbased
on analyses reflecting fast current technological de-
velopment and improvements of digital optical sen-
sors (CMOS, CCD), low-cost IMU consisting of
MEMS (Micro-Electro-Mechanical-System) sensors,
andGPS technology. All analyseswill determine the
boundaries in the instrument functionality with re-
spect to S/C maneuvering parameters, various op-
tical sensor sizes, sensitivity, error sources, stability,
lifetime, power consumption, mass, and suitability
of various wide FoV optics. The proposed WFST
should use cost-effective components and thus fill a
gap in the market with a system that is novel and
powerful, but cost-effective.

Acknowledgement

This work has been performed under SCIEX-
NMSch—CHVIandGA205/09/1302“Studyof spo-
radic meteors and weak meteor showers using auto-
matic video intensifier cameras” of theGrantAgency
of the Czech Republic.

References

[1] Jørgensen, J. L., Pickles, A.: Fast and robust
pointing and tracking using a second generation
star tracker, Proceedings of SPIE. Washington :
1998, p. 51–61.

[2] Liebe, C. Ch.: Accuracy Performance of Star
Trackers — A Tutorial, IEEE Transactions on
Aerospace andElectronic Systems,Vol.38,No. 2,
2002, p. 587–599.

[3] Savage, P. G.: Strapdown analytics. Strapdown
Associates, Minnesota, USA : 2000.

[4] Rerabek, M., Pata, P., Koten, P.: Processing
of the Astronomical Image Data obtained from
UWFC Optical Systems, Proceedings of SPIE,
Washington : 2008.

[5] Starck, J., Murtagh, F.: Astronomical Image and
Data Analysis. Berlin : Springer, 2002.

Jan Roháč
Faculty of Electrical Engineering
Czech Technical University in Prague
Technická 2, 166 27 Prague, Czech Republic

Martin Řeřábek
Faculty of Electrical Engineering
Czech Technical University in Prague
Technická 2, 166 27 Prague, Czech Republic
Multimedia Signal Processing Group
Ecole Polytechnique Fédérale de Lausanne
Station 11, Lausanne, Switzerland

René Hudec
Faculty of Electrical Engineering
Czech Technical University in Prague
Technická 2, 166 27 Prague, Czech Republic

63