use of new developments of attitude control

use of new developments of attitude control
Matthias Waidmann, Christian Waidmann, Dominik Saile, Georg Grillmayer
Institute of Space Systems, Universität Stuttgart, Stuttgart, Germany
[email protected]
Viola Wolter
Steinbeis Transferzentrum Raumfahrt, Gäufelden, Germany
[email protected]
The Flying Laptop is a micro-satellite currently under development at the Institute of Space Systems, Universität
Stuttgart. The primary mission objective of the Flying Laptop is technology demonstration for the future projects of
the Institute of Space Systems. Several attitude sensors, either in-house developed or from external companies with
no previous flight heritage, are being used. Electronic boards and mechanical housings were designed for the GPS
system, the fiber-optic gyros and the magnetic torquers. The GENIUS experiment aims to increase the GPS accuracy
in orbit by using an ultra stable oscillator (USO) and includes attitude determination. The C-FORS fiber optic gyro is
a commercial product developed for aviation. With the Micro Advanced Stellar Compass made by the Technical
University of Denmark and the Magnetometer made by Zarm-Technik new developments, so far not flown, are integrated. All attitude sensors and actuators are connected to a field programmable gate array (FPGA). This kind of onboard computer offers a more accurate timing and parallel processing of the sensors' and actuators' signals. The paper focuses on the attitude sensors and actuators and their interfaces to the on-board computer.
Attitude Control System
Commercial Fiber Optic Rate Sensor
Charge Coupled Device
Camera Head Unit
Commercial Off The Shelf
Flying Laptop
Fiber Optic Gyro
Field Programmable Gate Array
GPS Enhanced Navigation Instrument for
the Universität Stuttgart micro-satellite
Inter Integrated Circuit
Integrated Bus for Intelligent Sensors
Lost In Space
Magnetic Torquer
On-Board Computer
Printed Circuit Board
Power Control and Distribution Unit
Pulse Per Second
Reaction Wheel
Synchronous Clock
Single Event Upset
Star Tracker
Sun Sensor
Ultra Stable Oscillator
Micro Advanced Stellar Compass
Micro Data Processing Unit
The Flying Laptop (Figure 1) is a 100 kg, three-axis stabilized micro-satellite currently under development at
the Institute of Space Systems, Universität Stuttgart and
is planned to be launched into a sun-synchronous, low
earth orbit. The primary mission objective is technology
demonstration [1]. For the attitude control system several new sensors and also in-house developed components
are utilized.
Figure 1: Flying Laptop
The satellite's motion is monitored by five different
types of sensors: two three-axis magnetometers, two
coarse sun sensors, four fiber-optic rate sensors, one
autonomous star tracker with two camera heads and
three GPS receivers. The actuators that rotate the satellite to the desired attitude are four reaction wheels and
three magnetic torquers. All sensors and actuators are
connected to the FPGA On-Board Computer (OBC) in a
star like configuration. The ACS hardware devices are
shown in Figure 2 and Figure 5, a more detailed description about the FPGA and the control algorithms is given
in [2].
2.1 Operational Modes
Figure 2: "SpaceCube" for ACS Development
The Flying Laptop uses different sensor/actuator combinations for several pointing modes as described in
Table 1:
The objective of the detumbling mode is to reduce the
angular velocity after launcher separation. An additional
use is rate damping in case of accidentally exceeding the
angular velocity limit.
The satellite is commanded to safe mode subsequent to
the initial detumbling phase or in later mission phases
whenever non-recoverable errors occur during pointing
modes. The sensors for this mode need to have a high
reliability, sensor outputs need to be available all the
time but only a course pointing knowledge is required.
Thus, only sun sensor and magnetometer information
will be used. The commanded torque is realized with the
help of the magnetic torque rods.
Table 1: Operational Modes
The inertial pointing mode will mainly be used for calibration and for technology demonstration of automated
asteroid detection using the star tracker. A pointing stability of 150 arcseconds is required. Attitude and rate information is provided by the star tracker and the fiber
optical rate sensors. Four reaction wheels are used for
During earth observation in nadir pointing mode the
satellite's body axes are aligned with the nadir coordinate system. The same type of non-linear controller with
quaternion and rate feedback as for inertial pointing is
used (the error quaternion now describes the error
between body and nadir frame). The controller allows
for large angle slew maneuvers to a nadir-pointed orientation from an arbitrary attitude and then stabilizes the
satellite's nadir orientation within 150 arcseconds.
No earth referenced attitude is directly available from
any sensor, but with the help of GPS data it is possible
to transform the initial quaternions from the star tracker
to the nadir frame. First the GPS-time is converted to
JD2000 to calculate the Greenwich Mean Sidereal Time
(GMST). From this, GPS position and velocity can be
transformed from the earth-fixed to the earth-inertial
In the target pointing mode the satellite remains
aligned towards a fixed point on the earth's surface. This
mode is mainly used for scientific investigations of the
bi-directional reflectance distribution function (BRDF),
off-nadir inspection of ground areas and during contact
with the ground station using the high gain antennas.
Suitable images for a BRDF target need to be taken
along ±60 ° pitch off nadir and within a maximum roll
angle of ±5 °. An image with the payload cameras needs
to be taken at every degree between the ±60 ° during the
The controller for this mode is only slightly different
from the one for nadir pointing. It uses quaternion and
rate feedback as well but different setpoints for attitude
and angular velocities. Star tracker, GPS and rate
sensors will be used to precisely follow the calculated
attitude profile within a pointing accuracy of 150 arcseconds.
2.2 ACS Algorithms
Figure 6 shows a global overview of the ACS algorithms. Each block shown represents a sub-function
with inputs and outputs. On the left, inputs from the 5
sensors are shown representing the interface variables
from the lower hardware dependent levels. The Processing section merges, converts and extrapolates the
sensor variables to the current time and also contains the
Kalman filters. In the Navigation part the reference attitude and reference angular velocity are calculated depending on the pointing mode selected. The Control
section contains the safe mode, detumbling, desaturation, nullspace, inertial, nadir and target pointing controller [2]. For the last three an error quarternion feedback controller is used. The torque output from the controllers is scaled and processed in the Command section
and sent to the wheels and magnetic torquers shown at
the right hand.
2.3 ACS Algorithm Development Process
Figure 3 shows the development cycle for the ACS navigation and control algorithms. After initial theoretical
analysis the algorithms are implemented in Matlab/Simulink and tested extensively for performance. All functions which are going to be implemented in the on-board
computer are written as an embedded m-file. In the next
step the algorithms have to be converted to Handel-C,
which is a high level language compiler generating the
binary netlist for the FPGA. Control algorithms and filters contain a large amount of calculations which require
a large amount of gates in the FPGA. In order to limit
the space needed on the FPGA almost all variables are
converted to fixed point arithmetic and the outputs of
the optimized algorithms is compared to the original
double precision algorithms. The comparison is still performed in the Matlab/Simulink environment and is
identical to the approach used for digital signal pro-
Figure 3: ACS Development Process
cessors. After compilation the netlist is loaded to a prototype FPGA board (Figure 2) for initial testing. For final verification of the algorithms a real time simulation
environment is needed. A model-based system-simulation environment represented by the MDVE (Modelbased Development and Verification Environment) is
GPS is a commonly used sensor for satellites in low
earth orbit in order to determine the position and velocity. This experiment will test the advanced usage of
GPS for orbit navigation.
GENIUS was developed to serve two purposes. The first
is the standard supply of real-time position and velocity
data for on-board usage during nadir and target pointing.
A precision of 10 m in position, 0.1 m/s in velocity and
1 µs in time is envisaged using and internal orbit
propagator and the possibility of uploading two line elements. The system is composed of three independent receivers, thus it assures a high level of redundancy for
GPS on-board navigation. The second, innovative task
is an experiment conducted in cooperation with the
DLR/GSOC (German Space Operations Center) for accurate determination of the spacecraft attitude. This will
be accomplished by the ground analysis of the measured
GPS carrier phase of each receiver that is recorded on
the satellite and dumped during ground station contacts.
An accuracy of 0.1° to 1° is envisaged. Furthermore due
to the use of an ultra stable time base navigation solutions with less than 4 tracked satellites can be studied,
considering the future use of GPS in higher altitude orbits.
3.1 System Design Overview
Each of the three Phoenix GPS receivers is connected to
Figure 4: GPS Block Diagram
Figure 5: Overview of ACS Sensors and Actuators
Figure 6: ACS Subfunction Map
a separate GPS antenna and low noise amplifier as
shown in Figure 4. The three GPS receivers are integrated in one box together with an interface board (Figure
7). The three external connections are the on-board
computer, the power control and distribution unit
(PCDU) and the ultra-stable-oscillator (USO).
modifications were made to prepare the receiver boards
for space usage and to adjust them to the needs of the
GENIUS system. The crystal oscillator of the GPS
boards was removed since the time base for the GPS receivers is provided by an ultra-stable 10 MHz crystal oscillator (USO) on board the Flying Laptop. This way the
receivers are synchronized for the carrier phase measurements.
Figure 7: GPS Modules
The three antennas are placed on three corners of the
satellite in an L-like arrangement, creating two baselines
of 44 cm and 61 cm (Figure 8). This configuration is
used to define the spacecraft attitude.
Figure 9: GPS Hardware
For the measurement of the angular rate, the micro-satellite Flying Laptop is equipped with 4 single-axis COTS
fiber optic rate sensors in a tetrahedron configuration.
4.1 Fiber Optic Rate Sensor
Figure 8: GPS Antenna Arrangement on the Solar
3.2 Phoenix GPS Receiver
The Phoenix GPS receiver is a commercial GPS receiver board with a new DLR/GSOC developed firmware
for space and high dynamics applications. The receiver
has 12 tracking channels and is able to measure phase
and Doppler shift of the GPS-L1 carrier signal. Several
The measurement principle of fiber optic rate sensors is
based on the Sagnac effect. The used sensors C-FORS
(Commercial Fiber Optic Rate Sensor, Figure 10) are
produced for terrestrial and aeronautical applications by
Litef. Due to the use of COTS parts, the sensor is much
cheaper than space qualified rate sensors, but nevertheless its performance is similar to those used for space
applications. Currently it is one of the most accurate rate
sensors not subject to ITAR restrictions. By flying the
sensor the space environment effects on the sensor are
The C-FORS [3] is designed to survive shocks of up to
250 g (non-operating, 1 ms half-sine) as well as linear
accelerations of up to 100 g and sine sweep vibrations
with an amplitude of 1.5 gRMS in the frequency range of
5-2000 Hz. Therefore the sensor is expected to with5
Figure 10: Fiber Optic Rate Sensor C-FORS with
Soldered Interface Circuit Board
stand launch without any damage or performance degradation.
Due to the hermetically sealed case of the sensor no difference in the sensor's performance both on ground and
in space is expected. Also outgassing is not an issue.
The operating temperature range of -40°C to +75°C is
not expected to be exceeded.
To reduce the total dose caused by cosmic radiation additional shielding for the sensor is necessary. Due to the
weight of ferromagnetic metals, these cases are made of
aluminium, no additional shielding against electromagnetic influences is intended as the sensor already contains a μ-metal shield around its optic fiber coil. The
wall thickness of the shielding is 3 mm.
The C-FORS is configured to measure the angular rate
around its sensitive axis. The output value is the mean
value calculated from the angle increment that is run
through since the last data output, divided by the time
elapsed since the last data output.
The properties and performance of the C-FORS according to the manufacturer's specification and the settings
used for the Flying Laptop are summarized in Table 2.
The performance values of the 4 flight models are all
within the specified limits.
Table 2: Characteristics of the C-FORS
Shocks (non-op.)
Linear Accel.
Sine sweep
Temp. Range (op.)
78 x 53 x 22
≤ 0.13
1.5gRMS (5-2000 Hz)
-40°C to +75°C
Rate Bias
Random Walk
Scale Factor Error
Magnetic Sensitivity
Axis misalignment
Initialization time
Measurement range
Data rate
Power Consumption
≤ 0.15
≤ 1000
±10 (absolute)
± 1 (stability)
≤ 0.12
±16.771 (max.±1000)
(5 – 4000)
< 2.5
The electrical interface for both power supply and communication is provided via soldering pins. Each sensor
requires three different voltages for power supply according to Table 3.
Table 3: Power Supply of the C-FORS
Nominal voltage
[VDC] +5
Max. current
The use of 4 rate sensors in a tetrahedron configuration
requires a mounting platform. Furthermore, an electronic unit for power supply of the sensors and for communication is needed.
4.2 Design Fundamentals
The design is based on single redundancy and the use of
COTS parts which are not specially space qualified.
For the measurement of the three angular rates around
the satellite's axes three sensors are needed. The designed FOG-unit uses 4 independent sensors to allow for
autonomous detection of a single failure.
For the whole unit COTS parts are selected, but taking
into account the design requirements for the space environment. Space qualified micro-D connectors are used to
minimize size.
The whole unit has to fit in a given envelope and be as
compact and light-weight as possible due to the limited
dimensions and weight of the micro-satellite.
4.3 Electrical Design
In Figure 11 an overview of the electrical configuration
is given. The dashed box comprises the FOG-unit in6
Figure 12: Power supply of one FOG (data lines and
ground not illustrated).
Figure 11: Electrical Design Overview
cluding the 4 sensors and an electronic unit (FOG electronics). Both, power supply and communication is independent for each FOG but placed together on this
common electronic board.
A small interface board is soldered to the bottom side of
each FOG (Figure 10) to provide a connector interface.
The DC/DC converters and the signal driver mainly consist of commercial integrated circuits.
4.4 Power Supply
Due to the necessity of 4 independent sensors, each of
the sensors is supplied separately by the Power Control
and Distribution Unit (PCDU). To avoid a huge amount
of wires and due to the exigency of switching all
voltages simultaneously, only one voltage of +5.2 V is
supplied 4 times by the PCDU. The operating voltages
for each of the sensors (Table 3) are generated by
DC/DC converters located on the common electronic
unit as shown for one FOG in Figure 12. Each converter
ensures the required voltage stability of all supply
voltages for each FOG and a simultaneous switch on
and off of the voltages.
4.5 Data transmission
The C-FORS provides both an asynchronous and a synchronous interface for communication.
The synchronous interface allows for a maximum data
rate of 4 kHz, as opposed to a possible maximum data
rate of 1 kHz for the asynchronous interface. Also, the
Figure 13: Data lines between FOG and OBC (power
supply and ground not illustrated)
data request is issued to all four sensors at exactly the
same time. Thus the data from each sensor is exactly
synchronized to the OBC time. For this reason the common synchronous 4-wire IBIS bus is utilized (Figure
13). The SCLK line is used to issue a 2 MHz clock signal to each sensor. On the receive frame synchronization
(RFS) line, a 500 ns pulse synchronization signal triggers the simultaneous generation of measurement data in
each FOG every 256 clock cycles. The sensors are then
read out sequentially following a preprogrammed
The IBIS lines of the FOGs are connected to the common bus on the electronic unit, where the ground asymmetric signals are converted into ground symmetric signals by a signal driver for transmission to the OBC. If
any FOG is powered down, the sensor is disconnected
from the IBIS by an analog switch. Thus, the bus is not
affected by an defective and/or powered down FOG.
4.6 Heat Dissipation
The waste heat produced by the DC/DC-converters and
the signal driver is dissipated via the mounting screws of
the printed circuit board of the FOG electronics.
4.7 Mechanical Design
The whole FOG-unit is shown in Figure 14 and within
the satellite in Figure 15. It consists of the following
main parts:
Electronic box which is both mounting interface to
the satellite and housing of the main electronic unit.
Electronic unit;
Base plate which is both cover for the electronic
box and basis for the tetrahedron alignment of the
Ramps and a mounting plate for the assembly of the
FOGs in tetrahedron configuration;
FOGs including the interface board equipped with
the interface connector;
Shielding cases for reduction of the total radiation
The system is currently being assembled. The calculated
total weight is approx. 1.7 kg at the overall dimensions
of 186 x 122 x 150 mm (LxWxH).
The star tracker is the most accurate attitude determination sensor on the Flying Laptop. The selected model is
the micro-Advanced Stellar Compass (µASC) of the
Technical University of Denmark.
Figure 16: Micro-Advanced Stellar Compass (left)
and Camera Head Unit & Baffle (right)
The µASC is the successor of the Advanced Stellar
Compass, which has been successfully used for more
than 15 missions for ESA, CNES, NASA, DLR and
NASDA. Its smaller size and lighter weight makes it a
suitable device for micro-satellites like the Flying
Laptop. So far it was not launched on a satellite.
The µASC system (Figure 16) consists of 3 separate
units; the micro-Data Processing Unit (µDPU), the
Camera Head Unit (CHU) and the baffle system.
In normal operation, the µDPU creates a digital image
of the acquired analogue data received from the CHUs
every 0.5 s and stores it in the internal RAM. In further
computing, the lens distortion is removed and the image
is sifted for adequately bright objects. The amount of
bright objects that are found in an image can be influenced by varying a software parameter, but should never
be below 16 stars, which is the minimum to guarantee
Figure 14: FOG Unit
Figure 15: FOG Unit within the Micro-Satellite
Flying Laptop
Figure 17: Image taken by the CHU with exemplary
position and magnitude of three stars
reliable information. The developers recommend an
amount of detected stars between 20 and 80. Up to 199
stars can be detected. The center and magnitude of each
detected star is determined and stored in a list. Figure 17
shows a representative image where the position and
magnitude of 3 stars is labeled. In normal operation, the
stars are matched against the star catalog. It contains approximately the 12,000 brightest stars and is based on
the Hipparcos catalog supplemented by Tycho II. If the
match quality is within a defined threshold, the new attitude is found. Under the threshold, a further procedure,
the Lost In Space (LIS) mode, is initiated automatically.
In this mode, stars that are brighter than a certain value
and their neighbours are compared to a separate star
database to find a crude attitude. This database contains
the attitude and distances between the 4,000 brightest
stars and is used either to feed the normal operation algorithm, the initial attitude determination or helps to
provide a coarse attitude in case of high rotation. The attitude is delivered to the OBC as quaternions in the heliocentric inertial equatorial reference frame J2000.0.
5.1 Camera Head Unit (CHU)
The CHU uses a CCD-chip with a size of
7.95 mm x 6.45 mm with 752 x 558 pixels to make images of the firmament. Stray light of bright objects like
the sun, the earth and the moon that could interfere with
the measurement is absorbed in the baffle system, which
also reduces the field of view for each CHU to
13.4° x 18.4°.
While the µASC is capable of supporting four CHUs,
for the Flying Laptop a configuration of two CHUs was
selected as a compromise between attitude availability
on the one hand and available space and size (specially
for the baffles) on the other hand.
5.2 Micro-Data Processing Unit (µDPU)
Figure 19: Integration of the Star Tracker System on
the Flying Laptop
64 kB PROM , an 8 MB Flash and an 8 MB RAM.
5.3 Electrical Interface
As shown in Figure 18 the µDPU is connected to the onboard computer through a RS-422 interface with
115200 baud. The 20 V power supply is provided by the
power control and distribution unit (PCDU). The nominal power consumption for the FLP configuration with
two CHUs is 5 W. For ground tests, the µDPU can additionally be controlled by a PC over a special debug line.
5.4 Arrangement on the Satellite
Figure 19 shows the integration of the star-tracker system on the FLP. Both CHUs with corresponding baffles
are placed on an optical bench and connected with the
µDPU which is located on the other side of the bench.
To avoid blinding, an analysis for the angle between the
boresight of the CHUs and the sun-earth axis has been
conducted. The orientation of the CHUs has to be
chosen in a way that the mentioned angle does not fall
below the worst case sun exclusion angle of 55° and the
The data acquired by the CHUs is picked up by the
µDPU to process the satellite's attitude. It is equipped
with a 486 microprocessor and three types of memory; a
Figure 18: µASC Interface Diagram
Figure 20: Orientation of CHUs and Baffles.
earth exclusion angle of 25° and that the system still fits
into the already tightly packed satellite design. These
studies have led to the following orientation of the
CHUs with respect to the axes of the satellite: 125° elevation from the z-axis and 45° between the two CHUs
(Figure 20). This orientation avoids blinding of the
CHUs for most of the pointing maneuvers.
The baffles are optimized for earth proximity operations
to maximize roll at the expense of the sun exclusion
5.5 Accuracy
The µASC allows the determination of the attitude within an accuracy of 1-2 arcseconds for an angular rate of
less than 0.025°/s [4]. The accuracy is disturbed by various influences. The most obvious disturbance is caused
by motion and rotation of the satellite that causes a
smearing of some of the stars on the image or even moving them out of the FOV. The µASC is much more sensitive to rotations across the boresight than around it: For
an update rate of 0.5 s, the µASC switches to the LIS
mode for a rotation of 4°/s around boresight, whereas
this mode is already reached for a rotation of 0.4°/s
across boresight. A higher update rate allows higher angular rates but reduces integration time and thus accuracy. For highest accuracy 2 Hz update rate is recommended by the Technical University of Denmark.
Furthermore, precession and nutation of the earth cause
deviations from the true attitude because the output quaternions are given in an equatorial system, leading to an
error of about 5 arcseconds per year. Orbital aberration
and annual aberration results from relativistic effects of
fast movement of the satellite around the earth and the
sun. The maximum displacement caused by annual aberration is 5.4 arcseconds in LEO. On the Flying Laptop it
is possible to correct these effects by transmitting the
GPS time and position to the star tracker.
Another error is caused by the inaccuracy of the internal
clock of the star tracker. To minimize it, the clock is
synchronized with the ultra-stable on-board computer
clock every 10 seconds.
The satellite is equipped with two 3-axis anisotropicmagneto-resistive (AMR) magnetometers manufactured
by Zarm-Technik. It is a new development that was initiated by the Flying Laptop project.
The measured vector of the earth's magnetic field is used
as input information for the magnetic torquers, for
detumbling after launcher separation, rate damping in
case of accidentally exceeding the angular velocity limit
Figure 21: AMR Magnetometer
and desaturation of the momentum wheels.
An integrated three-axis AMR sensor HMC-1023 from
Honeywell measures the magnetic field in three directions [5] . For each direction, an independent permalloy
bridge is used for magnetic measurement. The output is
linear proportional to the applied magnetic field. The
analogue sensor output voltage is converted into a digital value by a sigma delta A/D-converter. The magnetometer temperature is monitored by a temperature sensor
that is also used for temperature compensation. A microcontroller handles the communication between user and
sensors. The magnetometer is encased by a housing consisting of 2 to 3 mm aluminium alloy.
6.1 Magnetometer Interface
The magnetometer is connected to the on-board computer by a RS-422 interface with 57 600 baud. A microD connector is used on the magnetometer for communication and 5 V (0.082 W) power supply.
After each data request a response message, containing
2 bytes of data for each direction and additional 2 bytes
for the temperature, is transmitted to the on-board computer. The magnetometer runs in fast mode using an integration time of 167 ms, but data is requested only
every 5 s when the magnetic torquers are shut off.
The sensor sensitivity is about 8.5 nT. The A/D-converter resolution is 16 bits, which is equal to 5 nT for the
last significant bit. The measurement range is ±150 µT.
The manufacturer states an accuracy of < ± 1%.
Three ZARM-Technik magnetic torquers (torque rods)
with a linear dipole moment of 6 Am² are utilized. The
torquers are connected to a power switch that is commanded by two independent I2C buses from the OBC
(Figure 22). The magnetic torquer system is essential to
the satellites functioning and thus the whole system is
Figure 22: MGT Block Diagram
Figure 24: OBC Block Diagram
Figure 23: Magnetic Torquer Power Electronics
fully redundant.
The power electronics contain an H-bridge supplying
the torquers with either 5V, -5V or switched off (bangbang actuation). The torquers are pulse-width-modulated and commanded by the acsPWM subroutine in the
ACS controller via the I2C interface. Again, to reduce
mass the pockets of the aluminium case are milled out as
shown in Figure 23.
The sun sensors system consists of two three-axis
Coarse Sun Sensors from Sun Space mounted at diagonal corners of the satellite. The sensor's voltages are digitized and sent to the OBC via two I2C buses.
Four reaction wheels RSI 01-5/28 made by Teldix are
running in a single hot redundant tetrahedron configuration with nullspace control. Each of the wheels has an
angular momentum capacity of 0.12 Nms and a reaction
torque of 5 mNm over the range of ±3000 rpm. The reaction wheels are connected to the on-board computer
via a RS-422 interface at 9600 baud.
The on-board computer (OBC) of the Flying Laptop is a
new development based on field programmable gate arrays (FPGAs). The OBC computer consists of four
FPGA nodes (Figure 24), each of which can act as the
master OBC. The quadruple redundancy is necessary to
handle radiation as the individual nodes are not radiation hard. The master OBC is selected by a radiation
hardened command decoder and voter (CDV).
The FPGA OBC nodes are programmed with Handel-C,
a programming language simular to C, developed especially for FPGAs. Handel-C allows programming in the
high-level-language style with the additional feature of
direct hardware manipulation.
10.1 Sensor Control by the FPGA
All sensors and actuators are connected to the OBC.
Each of the four FPGA nodes is able to listen to signals
from the sensors. Only the master node is allowed to
send commands. If a restart of a node is necessary, e.g.
caused by a single event upset (SEU), it can synchronize
itself to the other nodes by listening to the signals from
the sensors and through synchronization signals between
the nodes.
The use of FPGAs allows the implementation of many
kinds of interfaces directly in the FPGA. Asynchronous
as well as synchronous communication is possible, binary signals like clocks can be issued with only minor
limitations. Since most of the attitude control sensors
and actuators are designed for the use with micro-processor systems, serial communication is the most common interface. It is used by the GPS, the reaction wheel,
the magnetometer and the star tracker. For the fiber optic gyro synchronous serial communication is chosen.
For sensors and actuators that require only simple and
low amount of data communication, the I2C Bus was selected like for the magnetic torquers and sun sensors.
10.2 FPGA Advantages
When serial communication is used for the sensor control, the speed is mostly determined by the serial baud
rate and the response time by the sensors. With the
FPGA real parallel transmission, reception and processing of the many interfaces is possible.
Another advantage of programs running in hardware is
the time stability. When telemetry is received from the
sensors, the time for the processing of the message is always exactly the same (no interrupts like for micro-controllers) and can also be pre-calculated as a function of
the OBC internal clock rate.
The GPS and the star tracker are the only devices that
do not require a data request or command, but send their
telemetry individually and not in sync to the ACS control cycle. The GPS therefore sends a PPS synchronization signal every second. The star tracker on the other is
synchronized directly with a PPS signal every
10 seconds. Having always a constant processing time
directly results in no time variation and thus the best
available navigation accuracy.
[5] Zarm-Technik GmbH, “Magnetometer Manual
(MM)”, IRS-ZAR-MAG-MM-01-0001, Issue 1,
The attitude control system contributes to the main purpose of the Flying Laptop, technology demonstration,
by using new developed components and COTS
devices. The critical components are redundant as a
safeguard against unpredicted failures. New concepts
like the GENIUS GPS system and the FPGA on-board
computer are used to increase the overall performance
of the attitude control system.
[1] Grillmayer, G., Falke, A. and Roeser, H.P., “Technology Demonstration with the Micro-Satellite Flying Laptop”, Selected Proceedings of the 5th IAA
Symposium on Small Satellites for Earth Observation, Berlin, Germany, 4-8 Apr. 2005, pp. 419-427.
[2] Grillmayer, G., Hirth, M., Huber, F., Wolter, V.,
“Development of an FPGA Based Attitude Control
System for a Micro-Satellite”, AIAA-2006-6522,
AIAA/AAS Astrodynamics Specialist Conference,
Keystone, CO, USA, 21-24 Aug. 2006.
[3] Litef GmbH, “Specification of the Litef C-FORS
fiber-optic gyroscope”, 144410-0000-312, Rev. B,
[4] Technical University of Denmark, “micro-Advanced
Stellar Compass User's Manual”, IRS-DTU-MA3001, Issue 1.1, 2005.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF