Autonomous assembly of large structures in space: a

technology review
Mathieu Rognant, Christelle Cumer, Jean-Marc Biannic, Maximo Roa,
Antoine Verhaeghe, Vincent Bissonnette

To cite this version:

Mathieu Rognant, Christelle Cumer, Jean-Marc Biannic, Maximo Roa, Antoine Verhaeghe, et al..
Autonomous assembly of large structures in space: a technology review. EUCASS 2019, Jul 2019,
Madrid, Spain. �hal-03316188�

HAL Id: hal-03316188

https://hal.archives-ouvertes.fr/hal-03316188
Submitted on 6 Aug 2021

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 8TH EUROPEAN CONFERENCE FOR AERONAUTICS AND AEROSPACE SCIENCES (EUCASS)

Autonomous assembly of large structures in space:

a technology review
Mathieu Rognant?† Christelle Cumer? Jean-Marc Biannic?
Maximo A. Roa• Antoine Verhaeghe+ Vincent Bissonnette
?
ONERA - The French Aerospace Lab, 31055 Toulouse, France
•
German Aerospace Center (DLR), 82234 Wessling, Germany
+
CSEM, 2002 Neuchatel, Switzerland

Magellium, 31520 Ramonville Saint-Agne, France
†
mathieu.rognant@onera.fr · ? christelle.cumer@onera.fr

Abstract
Autonomous assembly of large structures in space is a key challenge to implement future missions that
will necessitate structures to be self-deployed as a single piece. This paper presents a mission analysis of
existing concepts for in-space assembly of telescopes, provides a survey of relevant robotics technologies
and introduces the expected contribution of the PULSAR (Prototype of an Ultra Large Structure Assembly
Robot) project to this challenge.

1. Introduction

The European Commission has set up the Space Robotics Technologies Strategic Research Cluster (SRC) in Horizon
2020, with the goal of enabling major advances in strategic key-points of this domain. To fulfill t his o bjective an
European roadmap composed of three successive calls (2016, 2018 and 2020) have been defined by the PERASPERA
consortium which is composed of the main European space agencies. The first a ctivities i n t he 2 016 c all h ave ad-
dressed the designing, manufacturing and testing of reliable and high performance common robotic building blocks for
operation in space environments (orbital and/or planetary). The specific objective of the 2018 call is to integrate the
previously prepared common building blocks into demonstrators, on ground, towards applications of space robotics in
the field of orbital and planetary use.
The robotized assembly of large modular orbital structures (i.e. Operational Grant 8, OG8 of this call) has been
identified as one of the key challenge of the orbital use which should be addressed. Future missions implementation will
necessitate structures too large to be self-deployed as a single piece. The James Webb Space Telescope9 has reached
this limit and the next generation telescope expected by astronomers, like the LUVOIR,5 will therefore require new
assembly technologies, in particular autonomous robots. The need for large structures in space goes beyond telescopes
and concerns also solar arrays for power plants, light sails to reach outermost regions of the solar system or heat shields
to land on Mars.
The PULSAR (Prototype of an Ultra Large Structure Assembly Robot) project is related to OG8. It aims at developing
and demonstrating the technology that will allow the on-orbit precise assembly of a very large primary mirror by an
autonomous robotic system.
A segmented space telescope needs multiple mirror tiles and multiple interfaces and therefore a precise way to assem-
ble them. This is fulfilled by perception and planning algorithms that make use of extended mobility for very large
structures, which requires a controlled, stable spacecraft during operations and a spacecraft structure that provides
attachment and housing for arm and tiles.
This paper reviews the state of the art on technologies relevant to the achievement of the PULSAR objectives.
In a first section a mission analysis, investigating previous and near-future similar missions, is achieved to derive high
level recommendation for PULSAR-like missions and propose a system architecture. In a second section a technology
review, analysing in-depth the different building blocks foreseen for PULSAR through a review of state-of-the-art
technologies and near-future developments, is provided. Finally, the project PULSAR (Prototype of an Ultra Large
Structure Assembly Robot) is introduced as the latest European effort to develop and demonstrate the technology that
will allow the on-orbit precise assembly of a very large primary mirror by an autonomous robotic system.
AUTONOMOUS ASSEMBLY IN SPACE

2. Mission Analysis

Different approaches have been proposed to achieve in-space assembly of telescopes. In general, the concepts include a
modular deployable structure, satellites flying in a coordinated fashion, or a mission including a general purpose robot
with advanced autonomous assembly capabilities. An overview is provided by M.Roa and al in.21 This section presents
the process followed in the context of the PULSAR project to define a system architecture.

2.1 High level Science needs and context

Science observation from ground is limited by a significant part of the electromagnetic spectrum being absorbed by the
atmosphere. Concretely, Earth’s atmosphere is nearly opaque, enabling visibility only within two very specific ranges
of wavelength: first, visible spectrum (1 µm wavelength); and a second interval from microwaves (1 cm wavelength)
to radio and TV (10 m). Outside of these intervals, almost nothing could be observed from ground.
Space telescopes however offer improved capabilities. Being able to observe the whole electromagnetic spectrum
without the atmosphere limitation, the information that can be collected is much more varied. To this purpose different
instruments have been developed, enabling observation from ultraviolet to infrared, thus multiplying the possibilities
for the scientific and astronomical communities. With the information collected from different spectral bands, scientists
are able to determine the composition of distant bodies, the distance to the object and even the date when the celestial
body was created. These features are not possible with the capabilities of ground telescopes, even if the latter can
benefit from bigger collecting surfaces, more complex equipment, and in general any other advantage derived from its
implementation on ground. The future of astronomy and science of the universe needs necessarily space telescopes,
and such telescope needs to be more and more performing in order to satisfy the increasing needs of this community.
At the same time, some technological barriers that were limiting factors some years ago tend to disappear or decrease
their influence, enabling significant improvements of the observation capabilities.
The scientific roadmap can be summarized highlighting the following goals:

• Understanding the universe life which needs versatile space observatory in ultraviolet wavelength with factor
50-100 greater than existing facilities (Hubble Space Telescope).

• Planet formation and emergence of life (exoplanets) which require UV field within 50-100 times higher sensitiv-
ity for observation of key atmospheric ingredients of Earth-like exoplanets

• Study of the solar system.

Figure 1: Deployment of the James Webb Telescope (launch planned for 2021) (Courtesy of NASA).

Based on this roadmap two missions are already scheduled:

• The James Webb Space Telescope (JWST)9 will feature a deployable 6.5 m diameter primary mirror and is
planned to be launched in March 2021. The telescope has a nominal life-time duration of 5.5 years, and a best
estimate of 10 years. It could thus be considered that the needs of the scientific community in terms of space
observation would be covered up to 2030 in the best case.

• The USA are already anticipating the next generation space telescope after JWST with the Large UV Optical
Infrared Surveyor (LUVOIR)5 proposed to be launched in 2039. LUVOIR-A will feature a 15 m deployable
primary mirror.

2
 AUTONOMOUS ASSEMBLY IN SPACE

In the meanwhile there is an opportunity for an European mission if this mission arrives not later than 2035. PULSAR
has to target this timeframe in order to cope with basic profit objectives. In addition this context allows us to define the
primary mirror size which should be deployed. The targeted value is a diameter of 10 m (75 m2 ). At least a diameter
of 8 m (50 m2 ) is mandatory to outperform JWST (20 m2 ).

2.2 Orbit and Launcher Constraint

The target orbit and launcher capacity are key sizing parameters of such on-orbit assembly missions:

Orbit selection Three possibilities -a LEO orbit, a geostationary orbit and the L2 Lagrange Sun-Earth point- were
analyzed and compared according to three main indicators: the mission feasibility/performance, the mission availability
and the lifetime.
Space telescopes impose stringent constraints on pointing accuracy and stability to ensure that collected images are
usable. The first indicator therefore focuses on the selection of an orbital environment allowing the required level of
precision. Disturbances in terms of torques and forces acting on the platform could penalize the interest of the mission
or require a more complex design of the platform and actuators. For this first criterion, the low Earth orbit environment
has high gravitational forces, while the L2 point has the advantage of lower forces and lower disturbance torques.
The second indicator focuses on the impact of orbit selection on the spacecraft’s ability to collect data. In other words, it
is necessary to quantify the number of mission hours without interruption or the ratio between data collection intervals
over the total duration of the mission. Maximizing availability is an essential aspect of space missions, where lack
of availability can lead to a loss of opportunities to observe rare or sporadic phenomena. This criterion penalizes a
telescope in LEO which, for an orbital period of about 100 minutes, has a duration of observation limited to about ten
minutes. In GEO, the observation possibilities last longer, but the Sun or the Earth ends up obscuring the observation
target. The main advantage of GEO is the continuous visibility of a ground station at the same longitude as the
telescope. This is an important parameter for sizing data collection capabilities, with the possibility of a permanent
data link between ground and spacecraft. L2 missions allow the long-term observation of targets without interference
from the Sun or the Earth in a large cone opposite to the Sun-Earth-Spacecraft direction. The availability of the mission
is no more limited from the point of view of space dynamics, but from the ability to transmit the collected data.
The third indicator is the optimization of the life time of the spacecraft. This last criterion favors LEO missions because
they have the advantages of a short transfer from launch to operation and a less demanding environment in terms of
radiation, which can affect critical electronic equipment. L2 orbits are clearly the least interesting as state-of-the-art
missions in L2 present only 5 to maximum 10 years life-time duration.
As the most important objective is feasibility and performance of the mission, L2 is the most suitable choice. Moreover
the choice of L2 is supported by current and future similar missions with comparable objectives (e.g. scientific missions
for the study of universe).

Figure 2: Sun-Earth Langrangian points (Courtesy of NASA)

3
AUTONOMOUS ASSEMBLY IN SPACE

Launcher selection The USA can envision missions with large deployable telescopes like LUVOIR because they
are counting on future powerful and big launchers, such as SLS,7 that Europe does not have. In order for PULSAR to
compete (in terms of performances and also economically) with American telescopes the use of a consolidated launcher
over one under development is a key feature. In this context, development of Ariane 6 is in advance compared to SLS
and will be fully operational in the time scope of PULSAR launch.
To improve the size capabilities of the primary mirror, a double launch scenario (considering a master vehicle holding
the instruments and the platform functions and a cargo vehicle launching full of mirror tiles) may be considered. This
option presents two main difficulties. First, the rendez-vous operation between master and cargo vehicles which have to
meet in space. The second difficulty concerns the assembly of the mirror itself, which would need a long robotic arm.
Tiles would have to be taken from the container and assembled into the primary mirror without counting on additional
support other than the spacecraft structure.

Mass limitations To evaluate the faisability of a single launch scenario with the Ariane 6, a preliminary mass budget
is established by using JWST as reference and identifies the additional items present in PULSAR spacecraft. JWST
is a good reference as it already implements most of the features that PULSAR needs. So, to proceed this sensitivity
analysis, the mass increase implied by the addition of the robotic arm and the larger primary mirror is evaluated. Based
on state-of-the-art of mirror structure and tiles interfaces, the evaluated primary mirror density range are set between
20 kgm2 and 40 kgm2 . In similar way the RAS mass range is set between 0.5 T n and 1.5 T n.
Attending to launched mass, PULSAR design target must not exceed 8.4 T n in order to guarantee compatibility

Table 1: Mass budget sensitivity analysis

with Ariane 6 (by using A64 version with four boosters) for the kind of orbit that is considered (Sun-Earth Lagrange
L2 point). A first mass budget sensitivity analysis summarized on Table 1 shows that the 8 m diameter mirror is an
achievable goal and the 10 m diameter mirror would require specific mass optimization.

Size limitations Figure 3 presents the concept for primary mirror assembly using independent hexagonal tiles. These
tiles are taken from a container or dispenser and are manipulated by the robotic arm during assembly operations.
In order to check if the number of tiles needed to form the primary mirror can effectively fit within the launcher
capacity, a preliminary sizing has been done supposing a very basic shape of the space vehicle. This design considers
a cylindrical shape vehicle, optimized to best fit launcher’s internal fairing. Attending to this, the maximum diameter
of the cylinder is not to exceed 4.5m and the main modules forming the spacecraft are stacked vertically (figure 4).
It is composed of a service module, including all basic functionalities (power supply, on-board computer), a payload
module, with all the electronics and data handling needed to support the instruments, a tiles container and a support for
the secondary mirror.
To maximize filling of the container, two specific tiles disposition have been evaluated: 6 tiles of 1.5 m diameter
per floor and 3 tiles of 2 m per floor. The both configuration would fit within the dedicated volume of 6 m height
assuming that the vertical dimension of the tiles is lower than 50 cm. However a trade-off should be found concerning
the tile size. Bigger tiles would reduce their number, reduce assembly operations and the mass needed in structure and
interfaces. On the other hand, smaller tiles offer better performances once assembled thanks to a finer resolution.

2.3 System architecture and requirements

The previous analysis demonstrates that a single launch scenario is an achievable goal for deploying an orbital telescope
which complies the science needs. However this solution involves strict requirements on the Robotic Arm System
(RAS), the Segmented Mirror Tiles (SMT) and the spacecraft stability. To fulfill this requirement, the PULSAR

4
 AUTONOMOUS ASSEMBLY IN SPACE

Figure 3: PULSAR Primary mirror concept Figure 4: Sizing configuration

project proposes to integrate and improve the common building block developed in the PERASPERA context. The
proposed architecture using these components is presented on figure 5 and is composed of the following three high-
level elements:

• A functional assembled mirror with independently adjustable tiles:

Scientific and operational requirements dictate that the primary mirror must be composed of a large number of
segmented tiles, autonomously assembled one by one, with high precision, in orbit. The individual mirror tiles
must thus interface to form a parabolic section. Additionally, such a large segmented telescope needs to have an
active and precise correction of the position of each mirror surface.
The large number of segmented mirror tiles further drives the need for a standard interface to join them, manip-
ulate them, and ensure their final functionality (provide power, data and thermal networks).

• A Robotic Arm System (RAS):

This RAS should be able to manipulate SMT in a large work-space and realize also precision assembly. These
two opposite goals require extended mobility functionality (walking robot or translation rail).
On the software side, specific low-level arm control functions is required to perform a rendezvous between
interfaces and ensure a final positioning within tolerances. High-level planning functions is also needed to safely
execute the arm movements. Moreover, to further increase the robustness of this precision assembly, the RAS
should include vision and LiDAR-based localization functions.

• A stable spacecraft:
The critical aspects of the spacecraft stabilization are to simultaneously manage the dynamically changing inertia
of the system induced by the tiles deployment and the disturbance torques induced by the robotic arm movements.
An additional challenge will be to ensure that the real-time implementation of designed Attitude and Orbit
Control System (AOCS) satisfies the mission requirements and the stringent on-board computational constraints.
Delays and limited sensors bandwidth, as well as sampling times strongly affect the AOCS performance.

In the following section, a review of the relevant technology are presented.

3. Technology review

3.1 Segmented Mirror Tiles

On ground, many big telescopes have a segmented primary mirror. The biggest monolithic primary mirrors are 8.4m
wide and equip the Large Binocular Telescope (LBT in Arizona USA). Such monolithic mirrors are as large as the
current technology permits. Segmented primary mirrors technology allows the construction of wider mirrors and
also provides significant advantages: Segments are easier to manufacture, transport, handle, install and maintain than
a monolithic mirror; Segmented mirrors can be made thinner and thus lighter; Mechanism behind segmented and
monolithic mirrors are used to correct the mirror shape under changing gravity orientation. Such mechanism are
simplified and lighter behind a segmented construction.

5
AUTONOMOUS ASSEMBLY IN SPACE

Figure 5: PULSAR architecture, based on previous components

In the context of PULSAR, each SMT is independent and integrates three sub-components enabling its assembly
and precise adjustment: a base structure, a mirror and between them a positioning system. The base structure carries
several standard interfaces to allow the manipulation and the connection with the other tiles.
The first function of the SMT positioning system is to compensate the residual positioning error after the assembly
task to obtain the required primary mirror shape. The second function is to realize wavefront corrections by adapting
the mirrors shape and position. Even if space telescope will suffer from no atmospheric perturbation, some wavefront
correction system is still needed to compensate thermal effect, as well as the gravity variation between earth-mounting
conditions and on orbit mission.
The second function is the most restrictive. To respect the primary mirror maximal acceptable wavefront error, which
is part of the overall telescope wavefront error budget, the SMT positioning systems shall have nano-metric positioning
capabilities.24 For this purpose, hexapod manipulators are more suitable as illustrated on figure which present the
JWST tile positioning system.

Figure 6: JWST tile positioning system24

3.2 Standard Interface

The standard interface is a robotic device providing mechanical, data and power transfer between different components
of the system. During assembly of the telescope, they are used as end-effector of the robotic arm to manipulate the

6
 AUTONOMOUS ASSEMBLY IN SPACE

individual segmented tiles. Once assembled, during operation, through the mechanical connection, they provide the
mechanical integrity of the segmented telescope. At the same time, the integrated connectors enable data and power
transmission between the different tiles and with the OBC.

Figure 7: SIROM interface design and compo- Figure 8: HOTDOCK interface concept before Figure 9: iBOSS
nents and after coupling iSSI Interface

In the context of PULSAR three interfaces have been considered:

• The SIROM interface is the product of the OG5 activity,10 one of the building block of the first SRC call. The
SIROM interface is a standardized robotic interface consisting of mechanical, electrical data, thermal and control
interfaces. It aims at connecting payloads to payloads or payloads to structures, with the capability to transfer
mechanical loads, data power and heat. With an androgynous design, it can match and couple another SIROM
interface on one side. On the other side, it can be connected to an Active Payload Module (APM) or to the
end-effector of a robotic manipulator.
• HOTDOCK is a standard robotic interface supporting mechanical, data, power and thermal transfer. Under
development by Space Applications Services, its design has been initiated to provide an answer to the highlighted
weaknesses of SIROM design in the context of the second call of PERASPERA roadmap, while keeping the same
targeted applications.
• The iSSI interface12 has been designed by the iBOSS consortium and is commercially available from iBOSS
Gmbh for Laboratory version. The multifunctional interface iSSI combines the four subassemblies for power,
data, heat and mechanical load transfer. The main parts of the mechanical interface include the coupling and
guiding elements as well as the positioning pins that are arranged around the data interface. Around the mech-
anism, the thermal interface consists of a carbon-nanotube copper-alloy composite material (the mechanical
interface can be implemented without the thermal).
The table 2 provides high level feature comparison between the different envisaged interfaces.

3.3 Robotic Arm System

Autonomy requirements Assembling a large space structure implies putting together modular components in an or-
dered fashion, dictated by a high level master plan that indicates the relative positioning of each part. Common robotic
systems in space applications have a small degree of autonomy. The execution of tasks usually relies on remote opera-
tions, which require an appropriate feedback channel for the operator, typically affected by substantial time delays. The
concept of shared autonomy increases the dexterity of such systems and reduces the effort for the operators in difficult
tasks. Nevertheless, remote operation approaches are not suitable for assembling a complex structures. Because of the
fine granularity of assembly tasks, classical remote operation becomes unfeasible as it consumes substantial amounts
of time for the synchronization of operator commands and manipulator actions. Therefore, a robotic assembly system
should be capable of performing a sequence of operations or even the complete assembly task autonomously.

Current technology Autonomous operations in space are still very challenging and there have been a limited num-
ber of demonstrations on-orbit. The first successful demonstration of an unmanned spacecraft to conduct autonomous
rendezvous and docking operations was done by NASDA in 1999 on ETS-VII. It was the first satellite equipped with
a robotic arm that allowed ESA to conduct the VIABLE experiment demonstrating computer vision support for au-
tonomous robot control. Several robotic arms are now present on board the ISS, including Canadarm, Dextre and
Kibo, but for now they are all teleoperated. Autonomous robotic assembly of space systems has been demonstrated on
ground for planar truss and beam structures, but their test in orbit and with more complex structures still remains a chal-
lenge. Among the missions under development, NASA’s Restore-L and DARPA’s RSGS are expected to demonstrate

7
AUTONOMOUS ASSEMBLY IN SPACE

Table 2: Standard Interfaces Specifications Comparisons

in the near future the autonomy and dexterity required for on-orbit assembly. An overview of current technologies for
in-space assembly is provided in.20

Figure 10: Canadarm2 Figure 11: JEM-RMS Figure 12: ERA

Extended mobility Given the size of the mirror envisaged for a space-based telescope, the robotic arm would need
some kind of mobility in order to reduce its required length. A very long robotic manipulator would make very difficult
to meet the accurate position and orientation requirements needed for assembly; a preferred solution is using a smaller
robotic manipulator (typically 1 to 2 m length) but with the ability to move within the structure. Two main alternatives
have been analyzed for providing the required mobility: a walking manipulator, and mobility within a rail. The lower
complexity of the second option makes it preferable at this stage of the study. Although no space applications are
known so far implementing this kind of technology, it is common on ground applications. On the other hand, different
walking manipulators have been implemented for space applications, including ERA and SSRMS.8 A proof of concept
of a walking manipulator for a space-based assembly application is currently under development in OG9-MOSAR.13

PERASPERA building blocks The autonomy requirements could be addressed by using the OG1 (ESROCOS) and
OG2 (ERGO) building blocks

• ESROCOS is a framework for developing robot control software applications. It includes a set of tools that
support different aspects of the development process, from architectural design to deployment and validation. In
addition, it provides a set of core functions that are often used in robotics or space applications.2

• ERGO is a framework for on-board autonomy systems17 based on two main components:

8
 AUTONOMOUS ASSEMBLY IN SPACE

– A functional layer which performs the requested actions by the executive layer. The functional layer will
use the ESROCOS framework to provide the interface with the hardware.
– An ERGO agent which controls the execution of the functional layer. This agent will enclose a set of
control, which can implement deliberative or reactive behaviours, and a central agent which ensures the
correct interaction among the different control loops.
The interface among different components of the ERGO agent will be based on goals (action or state
desired to be achieved) and observations (sensors data, or internal state deduced from the functional layers
information).
Adaptable perception, localization and mapping techniques are also required to guide the assembly process.
After the telescope is assembled, a metrology system needs to be employed for verifying the location and orientation
of each mirror tile, so that adjustments can be made to achieve the required accuracy and precision.
The OG3 (INFUSE) provides an open-source Common Data Fusion Framework (CDFF) by which data may be fused
in a modular fashion from multiple sensors.19
The OG4 (I3DS) relevant components for the PULSAR used-case are:
• The Instrument Control Unit (ICU) which provide "data concentrator" function. It is used as the central point
through which all of the OG4 Sensor Suite are connected. This allowed a single and standard interface between
the OBC and all sensors.
• The High Resolution (HR) camera is foreseen to perform visual servoing at short range.
• The pattern projector is also foreseen and will serve two purposes: first as an illumination device for the scene
and second coupled with the HR camera, it can provide 3D measurements.
• The stereo camera is proposed as a back-up: this can be used on the platform side to monitor the assembled
mirror.

3.4 AOCS Controller

The AOCS must be designed to support all mission’s needs from launch and early orbit phase (LEOP) to satellite
disposal. The satellite must implement at least the following operational phases: launch phase, transfer phase, deploy-
ment phase, mission phase, deorbit phase. Each phase is supported by one or more AOCS modes.14 In the context
of PULSAR, our main concern will be to design efficient controllers for the deployment phase and mission phases.
The deployment phase is normally entered when the satellite has finally reached the target orbit and all the satellite’s
appendages (i.e. solar arrays, antennas, instruments) are deployed. The mission phase begins after the deployment and
it is maintained up to the end of mission before satellite disposal.
The deployment phase is certainly the most challenging as far as the control design problem is concerned. During
this time period indeed, as already observed in the early work with ETS-VII,18 it is important to stabilize the attitude
with a reasonable accuracy to keep communication link despite the torque perturbations that are generated by the
robotic arm. Moreover, the robotic arm is used to build the primary mirror from tiles that are progressively deployed
from the main body. As a result, the inertia of the total satellite varies rather slowly but significantly during this
deployment phase. Many different strategies have been developed in the literature over the past thirty years to handle
attitude control problems in the presence of time varying inertia, for instance Adaptive Control Techniques for Linear
Time Varying systems,6, 15 linear parameter varying models,11 or robust control techniques that consider the variations
in the inertia matrix as time-varying uncertainties.22
Among possible approaches, one can mention Adaptive Control Techniques for Linear Time Varying systems
initially developed in15 and recently revisited in23 and.6 The central difficulty with adaptive control techniques is to ob-
tain a guaranteed performance level. This is why alternative LPV-based methods are often preferred.11 Moreover, when
the variations in the inertia matrix remains sufficiently small, the latter can be viewed as time-varying uncertainties and
robust control techniques become applicable.
In22 for example such a method, based on a smoothed sliding mode control strategy is applied to provide a
robust attitude controller using reaction wheels. Sliding mode control techniques are indeed very interesting since they
exhibit high robustness properties and well suited to nonlinear systems. However, they tend to generate aggressive
control inputs which often cannot be realized by limited reaction wheel systems. In this respect, a better compromise
is generally reached by robust control techniques mixing the LPV concept and the Hinfinity design framework.16
Based on results presented in,3 our approach to solve the problem will be based on a multi-model Hinfinity design
framework from which a robust and possibly parameter-varying attitude controller will be obtained. The perturbations
(typically those generated by the onboard manipulator) that cannot be rejected without a significant performance loss,

9
AUTONOMOUS ASSEMBLY IN SPACE

will be online estimated by a torque observer as shown in Figure 13. The latter can be designed either using the
Hinfinity or the LPV framework.

Figure 13: AOCS control structure

During the observation phase, the inertia matrix does not change significantly. However, pointing stability is the
driving requirement for high-quality imaging in a space telescope.4 The full primary mirror tends to generate badly
damped and rather low frequency torque perturbations. The main control design issue will then consist of enhanced
weighting functions tuning to optimize the compromise between a reasonable pointing accuracy and disturbance rejec-
tion. The general structure of the AOCS during deployment will be kept in order to facilitate control switching from
the deployment phase to the observation phase.

3.5 Modelling tools

Such control techniques need models highlighting physical parameters and the different dynamic couplings due to
torques induced by the robotic arm motions and different kinds of appendages such as flexible solar arrays, thermal
shield and the primary mirror of the telescope. These constraints lead to choose and couple two modelling tools :

• the first one is the Satellite Dynamics Toolbox (SDT - see https://personnel.isae-supaero.fr/daniel-
alazard/matlab-packages/satellite-dynamics-toolbox.html), based on a multi-body modelling ap-
proach. This toolbox allows to build generically a parametric linear model of a satellite composed of a rigid hub,
rigid appendages and/or flexible appendages. The key idea, detailed in,1 is to compute, for each substructure, its
inverse dynamic model: this transfer matrix has

– a input vector, composed of external forces and torques (wrench torque) applied on this substructure
– and a output vector, composed of the linear and angular accelerations (derivative of the twist vector) at a
point of the substructure.

A cantilevered connection of a -flexible or not- appendage on a satellite hub results in a simple feedback, since
the appendage is subjected to a force opposite to that which the hub undergoes. It is also at this connection,
that reference points and reference frame can be changed or that different kinds of connection (pivot,...) can
be considered. All couplings are then taken into account. Such an approach allows one to split the geometric
and dynamic parameters of each substructure and each link into specific blocks. Moreover, as each substructure
is separately modelled, the physical parameters are repeated minimally: it is also one of the advantages of the
approach.

• the second one is the open-source NPS-SRL/SPART (see https://spart.readthedocs.io/en/latest/),

which can derive easily the dynamics of the robotic manipulator. It is explicitly dedicated to kinematic trees
composed of rigid links and joints.

4. PULSAR: Prototype of an Ultra Large Structure Assembly

The PULSAR project aims at developing key technology listed in the previous technology review.
The retained method by PULSAR will involve two physical demonstrators (one focused on the assemble of a fully
functional section of a telescope mirror on earth conditions and the other on the assemble a very large structure in low
gravity - underwater - conditions) and one simulator (evaluation of the PULSAR technology in space conditions).

10
 AUTONOMOUS ASSEMBLY IN SPACE

4.1 Demonstrator of precise assembly of mirror tiles (dPAMT)

The demonstrator for precision assembly of mirror tiles will show the capabilities to autonomously assemble several
mirror tiles following specifications from a Master plan. This demonstrator will be implemented with a combination
of adaptable perception, integrated assembly and grasp planning, and compliant control of the manipulators. The
assembly demonstrator will rely on an assembly planner, which integrates a grasp planner and a motion planner, for
autonomously creating a master plan for the overall process starting with the specification of the desired assembly.
The system automatically decomposes a given assembly into a task sequence, which is then mapped to a sequence of
appropriate robotic skills. The skills exploit the capabilities of a lightweight and highly sensitive robotic manipulator,
the KUKA iiwa, for achieving compliant operations that guarantee successful execution of the robotic skills even in the
case of positional or sensorial uncertainties. Standard interfaces will be used both at the end point of the robotic arm
and at the mirror tiles, to facilitate the retrieval and repositioning of the SMTs. The main limitations of the demonstrator
will be gravity and the robotic arm’s payload limit, which restricts the achievable size of the assembled structure.
Visual servoing will be an important component for verifying the execution according to the nominal plan.
Additional external sensors are required to provide a ground truth measurement for robot positioning and motion, and
for measuring the success of the assembly process for the space telescope. An external measuring device will be used
to verify the pose of each individual mirror tile in order to validate the geometry and configuration of the primary
mirror, and to define the adjustments required to perform an optical alignment to a given focal point.

4.2 Demonstrator of large structure assembly in free floating environment (dLSAFFE)

To simulate on-orbit conditions, in particular the effects of micro-gravity, the autonomous assembly of a large seg-
mented mirror in underwater conditions will be demonstrated. This needs advanced mobility to overcome the limits
of robotic arm adaptability to the accumulated assembly errors, and an optimal Attitude and Orbit Control System
(AOCS) to stay in the required pose. An underwater platform endowed with a robotic manipulator will be used, and
thrusts in the platform will help to control the effects of impulsive forces created during the assembly operation. The
extended mobility of the arm will show the feasibility of assembly operations of a large structure. For this demonstra-
tor, all the technical sub-systems have to be adapted for underwater operation, including the connectors and the mirror
tiles.

4.3 Demonstrator of In-Space Assembly in Simulation (dISAS)

This last demonstrator will address the challenge of autonomously deploying a large structure in space while ensuring
the stability and safety of the spacecraft. To compensate the limitations of the fidelity of low-gravity facilities (such
as time-delay for the robotic platform and water-drag instead of neutral buoyancy), simulation means are retained as
the third demonstrator. This includes accurate physical models of spacecraft, robotic assembly system, and segmented
mirror tiles, to estimate torque disturbances involved in the deployment as well as robust controllers to manage them.
Software coming from the previous OGs will be embedded. The objective is to demonstrate that the deployment of
large structures and active tessellated mirror control can be carried out on-board a spacecraft respecting the AOCS
requirements.

5. Conclusion

This paper provided an overview of different technologies proposed for the assembly of a large structures in space.
The paper primarily focused on the mission analysis of telescope assembly in space and the maturity assessment of
the technology required. It was demonstrated that the complete assembly process requires, in particular, robotic arm
systems with advanced autonomous systems as well as stable control of the spacecraft.
The European project PULSAR was finally introduced, which aims to provide a first experimental verification for
low-level technologies that need to be further developed for in-space autonomous assembly of complex structures such
as telescopes. This goal will be achieved through three different demonstrators, based on a mobile robotic manipulator
(for testing autonomous assembly and optical verification of the telescope), an underwater platform (for testing assem-
bly in a low gravity environment), and a simulation-based approach for testing a full mission. The final demonstrations
will be performed in 2021.

11
AUTONOMOUS ASSEMBLY IN SPACE

6. Acknowledgments

We thank Aurelien Cuffolo, Sabrina Andiappane and Pablo Negro Lopez from Thales Alenia Space France for provid-
ing informations for the mission analysis.
The PULSAR project is funded under the European Commission’s Horizon 2020 Space Strategic Research
Cluster Operational Grants, grant number 821858

References

[1] Alazard Alazard, Christelle Cumer, and Khalid Tantawi. Linear dynamic modeling of spacecraft with various
flexible appendages and on-board angular momentums. In 7th International ESA Conference on Guidance, Nav-
igation Control Systems, 2008.
[2] M Muñoz Arancón, G Montano, M Wirkus, K Hoeflinger, D Silveira, N Tsiogkas, Jérôme Hugues, H Bruyninckx,
I Dragomir, and A Muhammad. Esrocos: a robotic operating system for space and terrestrial applications. In 14th
Symposium on Advanced Space Technologies in Robotics and Automation (ASTRA 2017), pages pp–1, 2017.
[3] J.M. Biannic, C. Roos, and J. Lesprier. Nonlinear Structured H∞ Controllers for Parameter-Dependent Uncertain
Systems with Application to Aircraft Landing. AerospaceLab Journal, (13):pages 1–15, November 2017.
[4] Lars Blackmore, Emmanuell Murray, Daniel P Scharf, Mimi Aung, David Bayard, Paul Brugarolas, Fred
Hadaegh, Allan Lee, Mark Milman, Sam Sirlin, et al. Instrument pointing capabilities: past, present, and fu-
ture. 2011.
[5] Matthew R Bolcar, Steve Aloezos, Vincent T Bly, Christine Collins, Julie Crooke, Courtney D Dressing, Lou Fan-
tano, Lee D Feinberg, Kevin France, Gene Gochar, et al. The large uv/optical/infrared surveyor (luvoir): Decadal
mission concept design update. In UV/Optical/IR Space Telescopes and Instruments: Innovative Technologies
and Concepts VIII, volume 10398, page 1039809. International Society for Optics and Photonics, 2017.
[6] Kaiwen Chen and Alessandro Astolfi. Adaptive control of linear systems with time-varying parameters. In 2018
Annual American Control Conference (ACC), pages 80–85. IEEE, 2018.
[7] Renee Cox, Steven D Creech, and David Hitt. Nasa’s space launch system: Unprecedented payload capabilities.
2018.
[8] V. Dubanchet. Modeling and control of a flexible space robot to capture a tumbling debris. In Doctoral disserta-
tion, École Polytechnique de Montréal, 2016.
[9] Jonathan P Gardner, John C Mather, Mark Clampin, Rene Doyon, Matthew A Greenhouse, Heidi B Hammel,
John B Hutchings, Peter Jakobsen, Simon J Lilly, Knox S Long, et al. The james webb space telescope. Space
Science Reviews, 123(4):485–606, 2006.
[10] Marko Jankovic, Wiebke Brinkmann, Sebastian Bartsch, Roberto Palazzetti, and Xiu Yan. Concepts of active
payload modules and end-effectors suitable for standard interface for robotic manipulation of payloads in future
space missions (sirom) interface. In 2018 IEEE Aerospace Conference, pages 1–15. IEEE, 2018.
[11] Rongyu Jin, Xueqin Chen, Yunhai Geng, and Zhili Hou. Lpv gain-scheduled attitude control for satellite with
time-varying inertia. Aerospace Science and Technology, 80:424–432, 2018.
[12] M Kortman, S Ruhl, Jana Weise, Joerg Kreisel, T Schervan, Hauke Schmidt, and Athanasios Dafnis. Building
block based iboss approach: fully modular systems with standard interface to enhance future satellites. In 66th
International Astronautical Congress. Israel, Jerusalem, pages 1–11, 2015.
[13] P. Letier, X. Yan, M. Deremetz, A. Bianco, G. Grunwald, M. Roa, R. Krenn, M. Munoz, P. Dissaux, J. Sanchez,
R. Ruiz, L. De Filippis, G. Porcelluzzi, M. Post, M. Walshe, and P. Perryman. MOSAR: Modular spacecraft
assembly and reconfiguration. In Symposium on Advanced Space Technologies in Robotics and Automation -
ASTRA, 2019.
[14] Leonardo Mazzini. Flexible Spacecraft Dynamics, Control and Guidance. Springer, 2015.
[15] Richard H Middleton and Graham C Goodwin. Adaptive control of time-varying linear systems. IEEE Transac-
tions on Automatic Control, 33(2):150–155, 1988.

12
 AUTONOMOUS ASSEMBLY IN SPACE

[16] Diego Navarro-Tapia, Andrés Marcos, Samir Bennani, Christophe Roux, and ELV SpA. Structured h-infinity
and linear parameter varying control design for the vega launch vehicle. In Proceedings of the 7th European
Conference for Aeronautics and Aerospace Sciences (EUCASS), 2017.
[17] J Ocón, FJ Colmenero, J Estremera, K Buckley, M Alonso, E Heredia, J Garciaa, AI Coles, AJ Coles, M Martínez,
et al. The ergo framework and its use in planetary/orbital scenarios. In International Astronautical Congress
(IAC), pages 1–5, 2018.
[18] M. Oda, K. Kibe, and F. Yamagata. ETS-VII, space robot in-orbit experiment satellite. pages 739–744, 1996.
[19] Mark Post, Romain Michalec, Alessandro Bianco, Xiu Yan, Andrea De Maio, Quentin Labourey, Simon Lacroix,
Jérémi Gancet, Shashank Govindaraj, Xavier Martinez-Gonzalez, et al. Infuse data fusion methodology for space
robotics, awareness and machine learning. In 69th International Astronautical Congress, 2018.
[20] M. A. Roa, K. Nottensteiner, A. Wedler, and G. Grunwald. Robotic technologies for in-space assembly operations.
In Symposium on Advanced Space Technologies in Robotics and Automation - ASTRA, 2017.
[21] Máximo A Roa, Korbinian Nottensteiner, Gerhard Grunwald, Lopez Negro, Pablo, Aurélien Cuffolo, Sabrine
Andiappane, Mathieu Rognant, Antoine Verhaeghe, and Vincent Bissonnette. In-space robotic assembly of large
telescopes. In Proc. 15th Symp. Adv. Space Technol. Robot. Autom., 2019.
[22] Lingling Shi, Jayantha Katupitiya, and Nathan Kinkaid. A robust attitude controller for a spacecraft equipped
with a robotic manipulator. In 2016 American Control Conference (ACC), pages 4966–4971. IEEE, 2016.
[23] Divya Thakur, Sukumar Srikant, and Maruthi R Akella. Adaptive attitude-tracking control of spacecraft with
uncertain time-varying inertia parameters. Journal of Guidance, Control, and Dynamics, 38(1):41–52, 2014.
[24] Robert M Warden. Cryogenic nano-actuator for jwst. In Proc. 38th Aerospace Mechanisms Symposium, pages
239–252, 2006.

13