Proceedings of the

The International 20th World

Federation of Congress
Automatic Control
The International
Proceedings Federation of Automatic Control
Proceedings of the 20th World Congress
Toulouse, of the
France, 20th
July World
9-14, 2017
Congress
Toulouse,
The France, July 9-14, 2017 Available online at www.sciencedirect.com
The International
International Federation
Federation of
of Automatic
Automatic Control
Control
Toulouse, France,
Toulouse, France, July
July 9-14,
9-14, 2017
2017

ScienceDirect
From GOCE to NGGM: IFAC PapersOnLine 50-1
Automatic (2017) 6428–6433
Control Breakthroughs for European future
From GOCE to NGGM: Automatic Control Breakthroughs for European future
From
From GOCE
GOCE to
to NGGM:
NGGM: Automatic
Automatic GravityControl
ControlMissionsBreakthroughs
Breakthroughs for
for European
European future
future
Gravity Missions
Gravity Missions
Gravity Missions
A. Bacchetta***, L. Colangelo*, E. Canuto*, S. Dionisio***,
A. Bacchetta***, L. Colangelo*, E. Canuto*, S. Dionisio***,
L. Massotti**, C.L.Novara*,
A. M. Parisch***, P. Silvestrin**
A.L.Bacchetta***,
Bacchetta***,
Massotti**, C.L.Novara*, Colangelo*,
Colangelo*, E.
E. Canuto*,
Canuto*,P.
M. Parisch***, S. Dionisio***,
S.Silvestrin**
Dionisio***,

L. Massotti**, C. Novara*, M. Parisch***,
L. Massotti**, C. Novara*, M. Parisch***, P. Silvestrin** P. Silvestrin**
*Department of Control and Computer Engineering, Politecnico di Torino,
*Department of Control and Computer  Engineering, Politecnico di Torino,
10129 Torino, Italy (luigi.colangelo@polito.it)
**Earth *Department
*Department
Observation 10129
of Control
of ProgrammesTorino,
and
Control and Computer Italy
Computer(luigi.colangelo@polito.it)
Department Engineering, Politecnico
– Future Missions
Engineering, Politecnico di
di Torino,
Division Torino,
(EOP-SF),
**Earth Observation 10129Programmes
10129 Torino,
ESA-ESTEC, Torino, Department
Italy – Future Missions
(luigi.colangelo@polito.it)
Italy (luigi.colangelo@polito.it)
NL-2200 Noordwijk, The Netherlands Division (EOP-SF),
**Earth Observation
**Earth Observation
***Engineering ESA-ESTEC,
Department, Programmes
Programmes NL-2200DepartmentNoordwijk,
Exploration–and
DomainDepartment – The
Future
Future Netherlands
Missions
Missions
Science, Thales Division
Alenia(EOP-SF),
Division (EOP-SF),
Space Italy,
***Engineering Department, ESA-ESTEC,
ESA-ESTEC, NL-2200 Domain
NL-2200 Exploration
10146 Torino,Noordwijk,
Noordwijk,ItalyThe Netherlands Alenia Space Italy,
and Science,
The Thales
Netherlands
***Engineering Department,
***Engineering Department, Domain Domain 10146 Torino, Italy
Exploration
Exploration and Science,
and Science, Thales
Thales AleniaAlenia SpaceSpace Italy,
Italy,
Abstract: After the successful European gravity 10146 Torino,
10146 Torino, Italy
missionItaly
GOCE (Gravity field and steady-state Ocean
CirculationAfter
Abstract: the successful
Explorer), which provided European gravity missionhigh
an unprecedented GOCE (Gravity
resolution fieldglobal
static and steady-state
map of the Earth's Ocean
Circulation
Abstract:
Abstract: Explorer),
After
After the
the which
successful
successful provided
European
European an unprecedented
gravity
gravity mission
mission high
GOCE
GOCE resolution
gravity field, the European Space Agency has proposed several preparatory studies for a Next Generation (Gravity
(Gravity static
field
fieldglobal
and
and map of
steady-state
steady-state the Earth's
Ocean
Ocean
gravity
Circulation
Circulation
Gravity field,
Mission the European
Explorer),
Explorer),
(NGGM). which
which Space
Theprovided
providedAgency
NGGM an has
anmission proposed
unprecedented
unprecedented
objectiveseveral
high
high
aims preparatory
resolution
resolution
at measuring studies
static
static global
global
the for map
a Next
map
temporal of
of Generation
the
the Earth's
variationsEarth'sof
Gravity
gravity
gravity
the Earth Mission
field,
field, the
gravity (NGGM).
the European
European aThe
field over Space
Spacelong NGGM
Agency
Agency
time span mission
has
haswith objective
proposed
proposed aimspreparatory
several
several
an unprecedented at measuring
level ofstudies
preparatory the temporal
studies
accuracy, for aa Next
for both Nextinvariations
Generation
Generation
spatial and of
the EarthMission
Gravity
Gravity
temporal gravity (NGGM).
Mission
resolution. fieldThe
(NGGM). overGOCEaThe
Thelong time span
NGGM
NGGM mission
mission
technological with an unprecedented
objective
objective
heritage aims
aims at
is leveraged level
at measuring
measuring of accuracy,
as starting the
the both
whileinvariations
temporal
temporal
point spatial and
variations
defining of
of
the
temporal
the Earth
the Earthfuture
NGGM resolution.
gravity field
gravitymission The
over
field over GOCEa long technological
time
a long Nonetheless,
concept. span withheritage
an
time span withtoanaccomplish is leveraged
unprecedented
unprecedented as
level starting
of point
accuracy,
level of accuracy,
its challenging while
both in defining
spatial
both objective,
scientific in spatial and the
and
the
NGGM
temporal future
temporalmission
NGGM resolution.
resolution.missionTheconcept.
The
concept GOCE
GOCE
envisages Nonetheless,
technological
technological
a wide range to accomplish
heritage
heritage is
is leveragedits challenging
leveraged
of innovations, as
as starting
with starting
respect scientific
point
point while
while
to the objective,
pastdefining
defining
or flying the
the
NGGM
NGGM mission
future
missions,future both on concept
mission
mission envisages
concept.
concept.and
technological a wide
Nonetheless,
Nonetheless,
automatic to range
to of innovations,
accomplish
accomplish
control side. Thus, its with
challenging
its challenging respect
this paper focuses to the
scientific
scientific past or
objective,
on objective,
the guidance,flying
the
the
missions,
NGGM
NGGM both
mission
mission on technological
concept
concept envisages
envisages and automatic
a
a wide
wide control
range
range of
of side. Thus,
innovations,
innovations,
navigation and control design evolution for the European gravity missions, from GOCE to NGGM. After this
with
with paper
respect
respectfocuses
to
to the
the on the
past
past guidance,
or
or flying
flying
navigation
missions,
missions,
recalling the and
both
both control
GOCEon GNCdesign
on technological
technological
main evolution
and
design for the European
and automatic
automatic
concepts, control
control gravity
side.
side.
the paper Thus,
Thus,
will missions,
this paper
this
describe paperfrom
the GOCE
focuses
focuses
most to NGGM.
on
on
important theinnovation
the After
guidance,
guidance,
recalling
navigation
navigation
required byandtheand GOCE
control GNC
controlIndeed,
NGGM. main
design
design evolution design
evolution
such a future concepts,
for the
for the the
European
European
concept paper will
gravity
willgravity describe
missions,
consistmissions, the
frommostGOCE
from GOCE
of a two-satellite important
to NGGM. innovation
to NGGM. loose
long-distance After
After
required
recalling
recalling
formation,the by
the NGGM.
GOCE
GOCE
where each Indeed,
GNC
GNC main such
mainisdesign
satellite a
design future concept
concepts,
concepts,
controlled the will
paper
the paper to
independently consist
will
will of
describe
be describea two-satellite
the most
drag-free,theGOCE-like. long-distance
important
most important loose
innovation
innovation
The satellite-to-
formation,
required bywhere
by
required distance
satellite NGGM.
NGGM. each satellite
Indeed,
Indeed,
variations, such
suchis controlled
encoding aa future
future
gravity independently
concept
concept will
anomalies, to bebe
will consist
consist
will drag-free,
of
ofthen GOCE-like.
aa two-satellite
two-satellite
measured by The satellite-to-
long-distance
long-distance
laser loose
loose
heterodyne
satellite
formation,
formation, distance
interferometry where
where variations,
each
foreach satellite encoding
is
satellite isranging
inter-satellite gravity
controlled
controlled at 20 anomalies,
independently
independently
nm resolution, will
to be
to be be then
drag-free,
or drag-free, measured
better. Hence, GOCE-like.
GOCE-like. by
an orbitThelaser
The heterodyne
satellite-to-
andsatellite-to-
formation
interferometry
satellite
satellite now for
control isdistance
distance inter-satellite
variations,
variations,
required ranging
encoding
encoding
to counteract at and
gravity
gravity
bias 20 nm driftresolution,
anomalies,
of the will
anomalies, will orbe
bebetter.
residual then
then Hence,
measured
measured
drag-free an byorbit
by
accelerations, andheterodyne
laser
laser information
heterodyne
order to
control is
interferometry
interferometry
reach a bounded now required
for to
for orbit/formation counteract
inter-satellite
inter-satellite ranging ranging bias
long-term at and
20
at 20 drift
nm of the
resolution,
nm resolution,
stability. residual
Finally, or or
GOCE drag-free
better. Hence,
better.control accelerations,
Hence,flight an orbit
an orbitresultsand in order
andasformation
formation
well as to
reach
control
control a is
isbounded
now
now orbit/formation
required
required to
to counteract
counteract long-term
bias
bias stability.
and
and drift
drift of
ofFinally,
the
the GOCE
residual
residual
NGGM simulated results, via a high-fidelity simulator, will be provided. These results highlight the control
drag-free
drag-free flight results
accelerations,
accelerations, as
in
in well
order
order as
to
to
NGGM
reach
reach a
a simulated
bounded
bounded results, via
orbit/formation
orbit/formation a high-fidelity
long-term
long-term simulator,
stability.
stability. will
Finally,
Finally,
GOCE GNC in-flight achievements as well as the NGGM concept validity, showing that the expected be
GOCE
GOCEprovided.control
control These
flight
flight results
results
results highlight
as
as well
well the
as
as
GOCE
NGGMGNC
NGGM
control simulated
simulatedin-flight
performances achievements
results,
results,
are via
inviaagreement as with
well the
aa high-fidelity
high-fidelity assimulator,
the NGGM
simulator,
consolidated will
willconcept
be
mission validity,
be provided.
provided. showing
These
These
requirements, all that
results
results the
theexpected
highlight
overhighlight 10-yearthe
the
control
GOCE performances are in agreement with the consolidated
mission.GNC in-flight achievements as well as the NGGM concept validity, showing that the expected
GOCE GNC in-flight achievements as well as the NGGM mission
concept requirements,
validity, showing all over
that the
the 10-year
expected
mission.performances
control
control performances are are inin agreement
agreement with with thethe consolidated
consolidated missionmission requirements,
requirements, all all over
over the the 10-year
10-year
© 2017,
Keywords:
mission.
mission.
Keywords: IFAC (International
Guidance,
Guidance,
navigation
navigationFederation and of
and controlAutomatic
control Control)
of spacecraft,
of spacecraft, Hosting
formation
formation by Elsevier
flying,
flying, Ltd. All
drag-free,
drag-free, rights
pointing,
pointing,reserved.
gravity
gravity
monitoring
monitoring
Keywords: Guidance, navigation
Keywords: Guidance, navigation and and control
control of of spacecraft,
spacecraft, formation
formation flying,flying, drag-free,
drag-free, pointing,
pointing, gravity
gravity
monitoring
monitoring 

1. INTRODUCTION low orbits (down to 230 km) for mapping the mean and
1. INTRODUCTION  low
variableorbits (down
gravity to peculiar
field, 230 km)designs for mappingand unique the challenges
mean and
Automatic control in 1. the field of astronautical engineering low
INTRODUCTION variable
low orbits
orbitsgravity
(down
(down tofield,
to peculiar
230 km) designs
for and
mapping uniquethe challenges
the mean and
Automatic 1. INTRODUCTION can be recognized in 230 km) forand
the attitude mapping
orbit control mean and
systems
has been ofcontroloutmost in the field ofsince
importance astronautical
the beginning engineering
of the variablecan be
variable recognized
gravity
gravity field,
field,in the attitude
peculiar
peculiar designs
designsand orbit
and
and control
unique
unique systems
challenges
challenges
has been
Automatic
Automatic of outmost
control in importance
the field of since the
astronautical beginning of
engineering the (AOCS in brief), of the already flown GOCE mission (GOCE
space era. Incontrol
particularin thethe field of astronautical
Guidance, Navigation and engineering
Control can (AOCS in
space
has
has era.
been
been In outmost
of
of particular
outmost the Guidance,
importance
importance since
sinceNavigation
the
the beginning and in
beginning Control
of the URL),
the
oforder
can be andbrief),
be recognizedits of
recognized thethe
in
in
follow-on already
the attitude
attitudeflown
mission andGOCE
and orbit mission
orbit
concept, control
control (GOCE
systems
systems
provisionally
(GNC) algorithms have become more sophisticated URL),
(AOCS
(AOCSNext and
in its
brief),
in brief), follow-on
of the already
of the already mission
flown concept,
GOCE provisionally
mission (GOCE
(GNC)
space algorithms
era. In havethe
particular become moreNavigation
Guidance, sophisticated in order called Generation Gravityflown MissionGOCE (i.e.mission
NGGM). (GOCE
The
space era.
to exploit Inthe
particular
enhanced the Guidance,
performances Navigation of theand Control
andon-board
Control called
URL),
URL), Next
and
and Generation
its
its follow-on
follow-on Gravity
mission
mission Mission (i.e.
concept,
concept, NGGM). The
provisionally
provisionally
to exploit
(GNC) the enhanced
algorithms performances of the on-board observational concept relies on the fact that the Earth
(GNC) algorithms
instrumentation, thehave
have become
become
variety more sophisticated
moreprogressively
of the sophisticated in
in order
order observational
available called Next concept
Generation relies
Gravity on the
Mission fact
(i.e.that the
NGGM). Earth
The
instrumentation,
to exploit
to exploit
sensors the
and the the variety
enhanced
enhanced
actuators on the ofmarket,
the progressively
performances
performances andof ofthethe available called
thechallenging
on-board
on-board
Next Generation
gravitational field is produced GravitybyMission the mass (i.e.ofNGGM). The
all its parts:
gravitational
observational
observational
hence, it field
containsconcept
concept is produced
relies
relies
information by
on
on
about the
thethe mass
fact
fact of
that
that
morphology all its
the
the parts:
Earth
Earth
and the
sensors and
instrumentation,
instrumentation,
requirements actuators
the
of the on
the payloads.
variety of
variety the market,
ofNot
thebeing
the and
progressively
progressively the challenging
that critical available
available
as per hence, itofcontains
gravitational
gravitational
density the field
field
crust information
is
isandproduced
produced
of the about
by
by the
interior thethe
of morphology
mass
mass
our of
of all
all its
planet, itsand
and the
parts:
parts:
how
requirements
sensors
sensors
the mannedand of
and mission the
actuators payloads.
actuators(Woods, on the
on the 2011)Not being
market,
market,in and and that thecritical as
challenging
the challenging
providing per
reliability density hence,
hence, of
it the crust
contains
it contains and of
information
information the
changesaboutinterior
about of
the our planet,
morphology
the morphology and
andhow
and the
the
the
and manned
requirements
requirements
a very low mission
of the
ofrisk
theall (Woods,
over the2011)
payloads.
payloads. Not
Not in phases
being
being
mission providing
that
that critical
critical
(from as per the
reliability
as per
launch, the
density
mass
mass of
distribution
distribution
the crust and changes
of the
in time,
in
interiortime, of
due to the tectonic
due
our to the
planet, tectonic
and how
and
the a very
manned low risk
mission all over
(Woods, the mission
2011) in phases
providing (from launch,
reliability density
plates of the crust
displacement, andthe of the
rain interior
that fills of our
the planet,
river and
basins, how
the
theorbit
to manned mission
transfer (Woods, 2011)
till rendezvous), stillintheproviding
control designreliabilityfor platesthe massdisplacement,
distribution the rain that
changes fillscurrents
in the
time, the
dueriver basins,
to crossing
the the
tectonic
to
and
andorbit transfer
aa very
very
scientific low
and till all
risk
lowcommercial
risk rendezvous),
all over the still
the mission
oversatellites hasthe
mission control
phases
phases
to ensure (from
(from design
the launch,
launch,
mission for the mass
glaciers distribution
formation and changes
melting, in time, due to the tectonic
the
glaciers
plates formation
displacement,
plates displacement,
oceans, the atmosphere andthe melting,
rain
the rain that the
fills
that fills and
circulation, currents
the river
the river
so crossing
basins,
basins,
forth. As the
thea
scientific
to orbit
to orbit
goals and
transfer
transfer
during commercial
till
all till satellites
rendezvous),
the nominalstill
overrendezvous), has
still
(and to
the ensure
control
the possibly the mission
design
control design
extended) for
for oceans,
goals during glaciers the
glaciers
consequence, atmosphere
formation
the gravity
formation and circulation,
and melting,
constrains
melting, the
the and
the currents
motion
currents so ofcrossing
forth. Asthe
thea
the Earth’s
crossing
scientific
scientific
mission and all
and
lifetime, overwiththeasatellites
commercial
commercial nominal
satellites
very high (andto
has
has possibly
ensure
ensure the
toreliability extended)
the mission
mission
and the consequence,
oceans, the the gravity
atmosphere constrains
circulation, the motion of the Earth’s
mission
goals duringlifetime,
during all over
overwithfrom
thea nominal
veryground.
nominal high reliability
(andRestricting
possibly extended)andfield,
extended) oceans, both
the satellite, the atmosphere
natural and artificial.circulation, and so forth. As aa
and so forth. As
goals
minimum all
intervention the the (and possibly the satellite,
consequence, both natural
the gravity and constrains
artificial. the motion of the Earth’s
minimum
mission
mission
in intervention
lifetime,
modernlifetime,
times, towithwith fromaa the
missions very
very ground.
high Restricting
high
requiring preferablythe
reliability
–reliability and
and the consequence, the gravity constrains the motion of the Earth’s
– field,
the
ultra satellite, both natural and artificial.
in modernintervention
minimum
minimum times, to missions
intervention from
from the therequiring
ground. – preferablythe
ground. Restricting
Restricting ultra satellite, both natural and artificial.
the– field,
field,
in
in modern times, to missions requiring – preferably – ultra
modern times, to missions requiring – preferably – ultra
Copyright © 2017 IFAC 6617
Copyright
2405-8963 ©© 2017,
2017 IFAC 6617Hosting by Elsevier Ltd. All rights reserved.
IFAC (International Federation of Automatic Control)
Peer review©under
Copyright 2017 responsibility
IFAC of International Federation of Automatic
6617 Control.
Copyright © 2017 IFAC 6617
10.1016/j.ifacol.2017.08.1030
Proceedings of the 20th IFAC World Congress
Toulouse, France, July 9-14, 2017 A. Bacchetta et al. / IFAC PapersOnLine 50-1 (2017) 6428–6433
Along this paper, the GOCE mission and the NGGM
concepts will be presented with a focus on their AOCS
design, and with a particular emphasis on the design of the
fully-automated “loose” formation control tailored to future
gravity missions, unique in its genre for an Earth Observation
mission in low orbit.
2. GOCE Mission
2.1 Scientific requirements and in-flight performances
The GOCE design concept emerged gradually, over more
than 20 years, from a complex interplay of science drivers,
technology needs, and the sometimes tough reminders of
technical and programmatic realities.
The scientific objectives of GOCE were the determination of
the Earth’s steady state gravity field anomalies with an
accuracy of 1x10-5 m/s2, and the determination of the geoid
height with an accuracy between 1 to 2 cm, at length scales
down to 100 km. To achieve these scientific objectives,
GOCE flew in a Sun-synchronous orbit (96.7° inclination,
ascending node at 18.00 h) with an altitude in the range of
250÷280 km, and it carried out two measurements: gravity
gradients, by the Electrostatic Gravity Gradiometer (EGG),
and Precise Orbit Determination, based on GPS data.
The 1060-kg GOCE spacecraft was launched on 17 March
2009 from Plesetsk on a Rockot launch vehicle.
Altitude and drag compensation of the slender spacecraft
with small (1.1 m²) frontal cross section was realised by two
ion thrusters (main and redundant) with a force range
between ≈ 1 and 20 mN, operated in closed loop with the
payload accelerometers. Three magnetic torquers with fine (≈
36 Am²) and coarse (400 Am²) regulation modes provided the
attitude control. About 41 kg of xenon and 14 kg of nitrogen
made up the propellant allowance for orbit and gradiometer
calibration, sufficient for the planned 2.5 years lifetime. The
actual mission evolution was vastly different from the worst-
case predictions. Indeed the Solar cycle 24 turned out to
produce the lowest maximum ever measured. In turn, this
implied a level of atmospheric drag much lower than
expected, to the benefit of the mission lifetime. Thanks to the
low density environment and to the conservative pre-launch
satellite drag estimation, the entire mission was spent at
altitudes lower than the minimum planned before flight, first
around 260 km and then reaching 250, 245, 240 and even 230
km in the final months, thus enhancing the scientific return of
the mission.
The nominal mission duration was 20 months, whereas the
actual lifetime has been of 55 months, almost doubling the
expected one. Neither orbit raising nor hibernation were
necessary and the gradiometer continued taking high-quality
readings, unaffected by variations in its dynamical and
thermal environment, even when the slowly accumulating
mismatch between altitude and inclination brought the
mission out of sun-synchronicity, causing longer and longer
eclipses. Both temperature control and drag compensation,
the key elements for mission performance, achieved their
mission flawlessly.
The two-domain (active/passive) thermal enclosure provided
the gradiometer with 10 mK thermal stability over 200 s time
intervals, as specified. The ion thrusters proved extremely
6618
6429
reliable, totalling ≈ 500,000 Ns impulse over more than four
years of continuous operation in all environmental
conditions. The much-feared “beam-outs” (sudden and
temporary interruptions of the thrust) were extremely rare
and always recovered without a hitch. Further, the drag
compensation in the direction of the motion exceeded its
requirement by a factor of about 10 (Fig. 3), proof to the
quality of the control design, the accelerometer sensors, and
the ion thruster actuators. The in-flight calibration of the
gradiometer (Cesare, 2010) provided the expected
improvements of the Gravity Gradient Tensor (GGT)
measurements. Aided by this, by the longer lifetime and by
the exceptionally low orbit altitudes, all the mission
performance goals on the geoid height and gravity anomaly
measurement accuracy were met or exceeded, in spite the
random errors on the GGT were higher than specified above
10 mHz.
3. GOCE AOCS
3.1 AOCS requirements
An essential element for meeting the mission requirements
was represented by the Drag Free and Attitude control. In its
early design concept, the GOCE drag-free control
encompassed six (attitude and orbit) degrees of freedom
(DoF). This was intended to provide enhanced robustness vs.
the uncertainty attached to both environment and gradiometer
response. The corresponding design had two ion thrusters for
active compensation of the main component of the drag and
eight micro-thrusters for lateral drag and attitude control. For
the latter task, micro-machined cold-gas thrusters were the
first candidates, later replaced by Field Emission Electric
Propulsion (FEEP) thrusters due to limited on board
resources. In summer 2003 it became clear that also the
readiness of the FEEP technology was not compatible with
the planned launch date, even accepting a delay of two years.
It was therefore decided to move to a four DoFs design, using
ion thrusters for in line drag control and magnetic torquers
for attitude control, supplemented by on/off impulsive cold-
gas thrusters, operated during the mission early phase for
Gradiometer calibration purposes. The GOCE design team
rose to the challenge and rapidly effected a complete re-
design of the satellite controls and their interfaces, which
touched on practically all the on-board sub-systems and –
partly – their accommodation (Fig. 2 and Sechi, 2011).
Fig. 2. GOCE configuration at the Critical Design Review
(left pictures, May 2005) and at the Flight Acceptance
Review (right pictures, March 2008).
By all standards, the GOCE Drag Free and Attitude Control
(DFAC) has been an innovative design. Among its distinctive
features, GOCE has been the first European drag-free
mission, based on ultra-sensitive accelerometers, flying at a
Proceedings of the 20th IFAC World Congress
6430
Toulouse, France, July 9-14, 2017 A. Bacchetta et al. / IFAC PapersOnLine 50-1 (2017) 6428–6433

very low altitude, and it had the first pure magnetic attitude
control system for a medium-sized Low Earth Orbit scientific
satellite. The mission induced requirements not only on the
magnitude of the residual disturbances, but also their spectral
density in the science measurement bandwidth (MBW) of
{5,100} mHz. To cope with such requirements, the payload
measurements were fed to the control loop.
When in Drag Free Mode (DFM), DFAC had to ensure the
limits reported in Table 1, expressed both as maximum
values in time and as maximum values of the square root of
the unilateral spectral density inside the MBW. The
requirements in Table 1 are the final AOCS requirements
after the GOCE control re-design. The relaxed requirements,
with respect to the original specs, did not affect the in-fight Fig. 3. Example of in-flight DFM linear acceleration
performances in spite of a more complex on-ground data post performance. PSD of the linear acceleration in the direction
processing. Let us notice that the spectral density of the in- of motion on 2 mission days (red, black) compared with the
flight DFM linear acceleration performance (see Fig. 3), has requirement (blue). The performance is ≈ 2×10-9 m/s²/√Hz
been computed considering one orbit (about 5400 s) of EGG all over the design MBW.
measurements sampled at 10 Hz. Further, the MTR instantaneous plane control capability has
3.2 AOCS evolution: Magnetic Attitude Control been complemented by a passive aerodynamic control, which
As mentioned, three Magnetic Torque Rods (MTR) were was only effective in pitch and yaw. The aerodynamic
employed as the unique actuator for attitude, angular rate and passive control effectiveness were driven both by the distance
angular acceleration control. The advantages of a fully between spacecraft Centre-of-Mass (CoM) and Centre-of-
magnetic control are fine command quantization levels, a Pressure (CoP), and by the atmospheric density. This has led
high reliability, and a low mass. Moreover, taking advantage to a platform design maximizing the CoP-CoM distance by
of electro-magnetic field (EMF) intensity at the GOCE low proper selection of the spacecraft mass distribution.
orbital heights, small currents were sufficient to actuate the The plots in Fig. 4 are relevant to the in-flight performance of
necessary control torques. The main problem was related to the DFAC scientific mode attitude control. They have been
the reduced degree of controllability, because the MTR obtained using the satellite telemetry: the time series are
actuation system cannot produce a control torque along the relevant to one day of data sampled at 50 s, showing the
EMF direction. Because of the GOCE quasi polar orbit, this compliance, with margins, to the requirements shown in
direction rotated almost periodically in the orbit plane. This Table 1.
effect guaranteed an average controllability for the roll and
yaw axes with a time horizon of half an orbit. The pitch axis
was always controllable.
Table 1. DFAC DFM AOCS requirements

The reduced controllability leaded to attitude control
performances that were linked to the dynamic disturbances.
The satellite dynamics was dependent on the amplitude and
phase of acceleration disturbances. The disturbances acting
on the GOCE satellite were both induced by the environment
(drag, gravity gradient, etc…) and by the platform (residual
magnetic dipole and the ion thrust misalignments). As a
matter of fact, a constant gain solution has been preferred for
the MTR, due to its simplicity and an inherent high degree of
robustness. This has been also driven by computational
constraints imposed by the GOCE on-board processor.
Fig. 4. Attitude control performance.
4. NGGM Mission Concept
After the successful mission GOCE, focusing on steady-state
anomalies of the Earth's gravity field, the European Space
Agency is proposing several preparatory studies for a future
gravimetric mission. The main aim is the need of ensuring a
proper continuity of gravity data to the scientific community,
leveraging all the know-how gained through the GOCE and
the US-German GRACE (and Follow-On) experience
(GRACE URL). Hence, the Next Generation Gravity Mission
(NGGM) concept was generated. Differently from GOCE,
the NGGM mission objective aims at measuring the temporal

6619
Proceedings of the 20th IFAC World Congress
Toulouse, France, July 9-14, 2017 A. Bacchetta et al. / IFAC PapersOnLine 50-1 (2017) 6428–6433
variations of the Earth gravity field over a long time span
(namely a full solar cycle). Such an objective will ideally
enable the study of geophysical phenomena involving mass
distribution and transport among the atmosphere, continental
hydrosphere, oceans, cryosphere, and lithosphere, enabling
new applications and observations at short and long time
scales.
From the automatic control perspective, two major
advancements can be defined, concerning both the
technological level and the mission concept itself. Indeed,
due to the NGGM mission objectives, some aspects of the
automatic control, as the drag compensation, can be
considered recursive between GOCE and NGGM. In that
case the GOCE heritage will provide a reliable guarantee
about the required level of performance; however, some
major innovations can be expected, as an all-thruster
configuration, ideally responsible for a fine controllability
along and around all the spacecraft axes. From the platform
perspective, the design developed so far envisages the
opportunity of placing eight micro-thrusters, from micro- to
milli-Newton range, placed all-around each satellite.
Therefore, the GOCE drag-free design along the orbital
motion axis, will be extended to all the three axes, both
compensating the linear and the angular accelerations.
On the other side, given the wide range of differences in the
NGGM mission concept with respect to GOCE, the
automatic control system will be responsible of a totally new
range of functions, like the formation and precise pointing.
Indeed, the simplest mission scenario for NGGM consists of
a single pair of satellites flying on the same orbit, with
different true anomalies (“in-line” or “pearl string”
formation). This in-line formation samples the gravity field in
the along-track direction only. On a polar orbit, this
formation is more sensitive to gravity field variations (and
mass transport) in the North-South than in the East-West
direction. Therefore, a second pair of satellites must be
launched in conjunction with the polar orbit pair, operated in
non-sun synchronous orbits to fill in, as much as possible, the
polar cap. According to this approach, the gravity field can be
sensed through the combination of the satellite pairs, flying in
loose formation. Hence, an orbit and formation control is now
required to counteract bias and drift of the residual drag-free
accelerations, in order to reach a bounded orbit/formation
long-term stability. As a consequence, the time variable
gravity field signal shall be retrieved via the precise
measurement of the inter-satellite distance variations induced
by the gravity anomalies, through the low Earth orbit
satellite-to-satellite tracking (SST) technique, as displayed in
Fig. 5 (see Drinkwater, 2003). On the other side, similarly to
GOCE, an accelerometer suite, on each satellite, is intended
to measure the non-gravitational accelerations induced by the
atmospheric drag, which are used in the drag-reduction
closed loop, and then subtracted by the laser ranging
measurement (converted consistently into accelerations) to
retrieve the gravity signal, at data processing level. The
satellite-to-satellite tracking technology, with the need of an
ultra-precise ranging control among the spacecraft, is a major
GNC design driver.
6620
6431
Fig. 5. NGGM satellite-to-satellite tracking concept.
Specifically, one of the main challenging control requirement
concerns the non-gravitational CoM accelerations, both linear
and angular, as they must be ideally brought to zero. On the
other side, the requirements concerning the orbit and
formation control can be derived in order to counteract the
differential effects of the drag-free control residuals, which
make the satellite formation diverging. Therefore the
formation control, even if loose and not strictly maintaining
the inter-satellite distance at sub-cm accuracy level during
science phases, needs as inputs the real time positions of the
leader (or alternately, of the follower) satellite, exchanged via
an inter-satellite link (see Bacchetta, 2014 and Massotti,
2011). Finally, the attitude and pointing control system is
intended to keep aligned the satellite optical axis and to
eventually ensure an orbital roll motion for tracking the Sun
beam (Canuto, 2015b).
Table 2. NGGM mission requirements
Performance variable Bound Unit
Drag-free control
CoM acceleration (PSD in MBW) 0.01 µm/s2/√Hz
CoM acceleration residuals 1 µm/s2
Orbit and formation control
Formation distance variation 5 %
(distance)
Formation lateral variation 1 %
(distance)
Formation radial variation 2 %
(distance)
Attitude and Pointing Control
Satellite X-axis pointing along the 2 µrad/√Hz
SSL
Satellite X-axis roll along the SSL 2 mrad
The Table 2 lists the main requirements driving the automatic
control design, in the science mode of the NGGM mission.
5. NGGM AOCS design
The AOCS for NGGM is an innovative design, conceived in
a holistic way including, for the first time in Europe, an
automatic (and not operated by the flight operations on
ground) formation control for an Earth Observation satellite
pair in low orbit. As a matter of fact, the AOCS is in charge
of several control tasks, via a wise separation in the MBW:
orbit altitude control, inter-satellite distance formation
control, drag-free and satellite-to-satellite pointing control.
Specifically, the formation control must constrain the relative
position of the two satellites, without interfering with the
scientific measurements and the drag compensation (from
Proceedings of the 20th IFAC World Congress
6432
Toulouse, France, July 9-14, 2017 A. Bacchetta et al. / IFAC PapersOnLine 50-1 (2017) 6428–6433

where it is defined as a “loose” formation control).

Conversely, the non-gravitational acceleration control of each
satellite must operate without affecting the formation control
and pointing capabilities.
The control architecture and logic consists in several control
loops hierarchically organised (Fig. 6). Such multi-
hierarchical control leverages an attitude and formation outer
loop, which provides the long-term reference accelerations to
be tracked by each individual drag compensation (or, simply,
the so called drag-“free”) control. As anticipated, the control
tasks operate in different frequency bands in order to manage
– in a coordinated way – all the necessary controllers
(satellite formation, orbit control, drag compensation and the
satellite attitude/laser beam pointing control). Fig. 6. AOCS overall block diagram.
The attitude and formation are intended to be decoupled and Focusing on the satellites attitude, the pointing control has to
all the coordinates are decomposed, so to have several SISO ensure the alignment of the satellites optical axis, to enable
loops to be controlled separately. This design choice is also the measurement, via laser interferometry, of the inter-
favoured by the choice of the control design methodology satellite distance variations, i.e. the scientific observables of
(i.e. the Embedded Model Control, see Canuto, 2014) and a such future gravity mission. The formation attitude rationale
mrad alignment between control frame and orbital, since the seeks an independent pointing control of each satellite, given
early mission phases, before the science control mode proper optical sensors able to measure the satellite
activation. Further, an innovative approach was pursued to misalignments from the satellite-to-satellite line. The attitude
multi-satellite formation and orbit control based on their kinematics and dynamics equations used in the control design
integration through the formation triangle virtual structure, are based on the definition of a proper attitude control
and the satellite-to-satellite line (SSL) (Canuto, 2015a). The reference frame, whose origin is in each satellite CoM, in
SSL line is defined as the vector connecting the CoM of the addition to the other reference frames introduced up to now
satellites, aligned with the first orbital axis, whereas the (Bacchetta, 2015).
second orbital axis is orthogonal to the orbit plane and the A design aspect worth to be underlined, is the closed-loop
third one completes the right-handed triad. Another tuning of the attitude state observers gains. As above
fundamental factor driving the control design is the thruster mentioned, the pointing control must be coordinated with the
authority. Indeed, the current design envisages a fully angular drag-free control action. Hence, similarly to the
automatic all-propulsion mission, but conversely to GOCE, linear case, the drag-free sets a frequency upper-bound to the
since NGGM is flying higher and given also the improved pointing control action. Hence, a trade-off occurs, between
technological level in electric propulsion (and the expected several pointing control objectives, as: to cancel the
products availability), the fully automatic control unit can accelerometer drift/bias, to filter the optical attitude sensor
leverage a thruster maximum authority of few millinewtons noises, and to ensure the closed-loop stability versus the
only, to be split among all the above-mentioned control tasks. attitude neglected dynamics. Specifically, the current tests
The control algorithms, embedded in the box of Figure 6, are and simulations show how the actual requirements set can be
based on a discrete-time (DT) embedded model of the two- met without great margin (as shown in Figure 8), given the
satellite formation kinematics and dynamics, based on the contrasting nature of the first two control objectives as
virtual structure of the formation triangle. As a consequence, described in Canuto, 2015b.
the integrated formation and orbit control, by controlling the
5.2 NGGM AOCS design: simulated results
average shape of this virtual structure, allows the satellites
altitude and distance to be controlled in closed-loop. Such
control actions are based on an integrated formation and orbit
control law which is the combination of a feedback term and
a disturbance rejection term. The feedback term injects back
both formation position and velocity state variables. In
addition, given a proper formation disturbance state observer,
the disturbance rejection term allows the rejection of all the
non-explicitly modelled dynamics (i.e. J2 effect, etc.) and the
parametric uncertainty effects (see Colangelo, 2016).
Concerning the evolution of the automatic control design, it
is worth to underline how, in a preliminary design phase,
only the formation position variables were fed-back to Fig. 7. Simulated linear residual acceleration performances.
generate the command. In the evolution of the control design, In order to assess the control design validity as well as the
a formation rate damping control, operating at the time unit NGGM AOCS performances, some relevant simulated results
of the navigation data and damping suitably the formation have been obtained through a high-fidelity E2E mission
rates, was added in order to bolster the formation stability simulator. The science phase includes the linear and angular
(see Colangelo, 2016). drag-compensation, attitude and pointing control and

6621
Proceedings of the 20th IFAC World Congress
Toulouse, France, July 9-14, 2017 A. Bacchetta et al. / IFAC PapersOnLine 50-1 (2017) 6428–6433
force/torque dispatching to a symmetrically-arranged eight-
thruster assembly. From the environment perspective, the
first 32 harmonics of the Earth gravity field spherical
harmonics expansion have been simulated together with an
Oersted geomagnetic field model (order 18) and
min/mean/max solar activity conditions. Finally, all the
sensor and actuator noises and dynamics are given as inputs
to the simulations. The reference inter-satellite distance has
been preliminary fixed to 200 km. From the control
perspective, all the above mentioned controllers are
considered in the closed-loop simulation.
Figure 7 shows the unilateral spectral density of the linear
acceleration residuals versus the performance requirement.
Such PSD has been computed on the whole residual profile
including the transient, which explains the low-frequency
overshoot.
Fig. 8. Simulated attitude pointing performance.
In Fig. 8, the simulated PSD of the attitude tilt angles is
presented. Both satellite pitch and yaw angles PSD (in green
and blue, respectively) meet the requirement bound with
some margin. For the roll angle (in red) a larger performance
bound applies, not being the dynamics around the roll axis a
constraint, since the laser interferometry measurements can
be still performed accepting a larger roll angle, but given the
satellite-to-satellite alignment requirements, as illustrated
above (Table 2).
6. CONCLUSIONS
This paper presents an overview on the attitude and orbit
control systems conceived for two specific missions, devoted
to the measurement of the gravity field at very low orbit.
Specifically, the challenging linear drag compensation of
GOCE has been reworked and extended to the 6 degrees of
freedom of each satellite of the NGGM pair, where also a
closed-loop loose formation control enables an ultra-precise
ranging measurement (at nanometre level), via heterodyne
laser interferometry.
Acknowledgements
This study was carried out within the study “Next Generation
Gravity Mission (NGGM): AOCS Solutions and Technologies
study” funded by the European Space Agency, being Thales
Alenia Space Italy (Turin, IT) the prime contractor, and
Politecnico di Torino (Turin, IT) the subcontractor.
REFERENCES
GOCE URL:
http://www.esa.int/Our_Activities/Observing_the_Earth/GOCE
GRACE URL: http://op.gfz-potsdam.de/grace
6622
6433
Bacchetta A., Buonocore M., Cesare, S., Dionisio S., Parisch
M., Canuto E., Colangelo L., Girouart B., Massotti L.
“The Results of the AOCS Solutions and Technologies
study for the Next Generation Gravity Mission”, AIAA
Science and Technology Forum and Exposition -
SciTech 2015, Kissimmee, FL, 5-9 January 2014.
Canuto E., Perez C., Colangelo L., Lotufo M., “Embedded
model control: design separation under uncertainty”,
33rd Chinese Control Conference (CCC), vol., no.,
pp.3637-3643, 28-30 July 2014.
Canuto, E., Colangelo, L., Buonocore, M., Massotti, L., and
Girouart, B, (2015a), “Orbit and formation control for
low-earth-orbit gravimetry drag-free satellites”,
Proceedings of the Institution of Mechanical Engineers,
Part G: Journal of Aerospace Engineering, 229(7), 1194–
1213. doi:10.1177/0954410014548236, 2015.
Canuto E., Colangelo L., Lotufo M., Dionisio S., (2015b),
“Satellite-to-satellite attitude control of a long-distance
spacecraft formation for the Next Generation Gravity
Mission”, European Journal of Control, Volume 25,
September 2015, pp. 1-16,
http://dx.doi.org/10.1016/j.ejcon.2015.03.003.
Cesare, S., Catastini, G., Floberghagen, R., Lamarre, D.,
2010. “The in-flight calibration of the GOCE
gradiometer”, In: Proc. of ESA Living Planet
Symposium. ESA SP-686 (CD-ROM).
Cesare S., et al. (2016). The European way to gravimetry:
from GOCE to NGGM. Advances in Space Research -
Volume 57, Issue 4.
Colangelo L., Canuto E., Massotti L., Novara C., Lotufo M.,
“Orbit and Formation Control for the Next Generation
Gravity Mission”, IFAC-PapersOnLine, Vol 49, Issue
17, 2016, pp. 284-289,
http://dx.doi.org/10.1016/j.ifacol.2016.09.049.
Drinkwater M. R., et al., “GOCE: ESA's First Earth Explorer
Core Mission”, Earth Gravity Field from Space - From
Sensors to Earth Sciences, Proc. of an ISSI Workshop,
Bern, Switzerland, Springer, 2003, pp. 419-432.
Massotti L., Aguirre M., Haagmans R., Kern M., Silvestrin
P., “Mission Concepts Overview for the Preparation of
ESA’s Next Generation Gravity Mission”, Proc. of the
8th International ESA Conference on GNC Systems
(GNC 2011), Karlovy Vary, Czech Republic, June 2011.
Sechi G., et al. (2005). Magnitude Attitude Control of the
GOCE Satellite. 6th International ESA Conference on
Guidance, Navigation and Control Systems, Loutraki
(Greece)
Sechi G., et al. (2011). In-Flight Results from the Drag-Free
and Attitude Control of the GOCE Satellite. Proc. 18th
IFAC World Congress, Milano (Italy).
Woods David S. (2011). How Apollo Flew to the Moon.
Springer, ISBN: 978-1-4419-7178-4.