Space Robotics I

Autonomous Space Systems

Space Robotics II

Introduction to Space Robotics

Planetary Robotics

Advanced robotic capabilities and operational autonomy are crucial enabling technologies for the envisioned progress in in-space servicing, assembly, and manufacturing and resource utilization. Intelligent robotic manipulation is deemed central to enabling and scaling up these applications, where capabilities like handling unknown objects in unstructured scenarios are crucial. In this context, this project proposes to develop a state-of-the-art robot vision pipeline and the underlying algorithms for space-borne vision-based grasping. The project's developments are geared toward testing, validation and use on Redwire's commercial light-weight robotic manipulation system. The development concerns synthesis of compute-efficient pipeline to accomplish the following tasks on an on-board processing unit: (1) accurate estimation of target object’s position and attitude and (2) subsequent estimation of grasp configuration that leads to a secure grasp on an unknown object in space.

Modular vISion for dynamic grasping of Unknown Resident Space Objects

Space debris is caused by millions of non-functional, human-made objects left in space that become a hazard for current and future space missions. Yet, no Active Debris Removal (ADR) systems currently exist. Many methods have been proposed, but most of them are still in early technological development stages. CASED concerns the design, prototyping, and testing of a SotA flexible capturing mechanism for ADR missions that can be placed in a CubeSat-based (CS) structure, targeting a range of uncooperative space debris, reliable and cost-effective. In addition, CASED will verify and validate (V&V) the developed capture system in the close-to-real facility of SnT that simulates orbital scenarios. Therein, the project will tackle the following objectives: OBJ-1: Design and development of robust, reliable, flexible CS capturing mechanisms for ADR missions. OBJ-2: Prototype and test these novel capturing mechanisms in relevant close-to-real orbital scenarios at SnT facilities. OBJ-3: V&V of the performance of the selected system for ADR missions.

Design of a Capturing, Absorbing, SEcuring system for active space Debris removal

This project investigates how model-based reinforcement learning can be applied to acquire manipulation skills for performing assembly tasks in different application domains of space robotics. The developed approach aims to facilitate self-supervised robot learning that is deployable to various real robot systems based on experience collected inside realistic simulation environments. Skill transfer among application domains will be accomplished by employing high-dimensional actions and observations that are invariant to the environment, task, robot kinematics and visual sensors. The applicability of the proposed approach will be evaluated on real robots under laboratory conditions of planetary and orbital scenarios, demonstrating the feasibility of sim-to-real transfer.

Space Assembly through Learning of Transferable Skills

The thesis is part of the HELEN project which aims to create a high-fidelity testing environment for the active space debris removal capturing system (FlexeS). The thesis is focused on developing real-time photorealistic in-orbit simulations (digital twin) for the capturing system, as well as designing and implementing a software-in-the-loop architecture for testing the latter.

Design and implementation of software in the loop architecture for active space debris removal high-fidelity scenarios

The goal of this project is development and performance evaluation of interaction control strategies for orbital robotic applications and ground based HIL simulations. To this effect, the focus of this work is two-fold. Firstly, development of application specific control algorithms for robotic manipulators in simulated microgravity environments. This involves a wide range of in-orbital robotic manipulation activities. Secondly, to use robot interaction control to simulate spacecraft interaction behaviour in the ground testing facilities.

Interaction Strategies for Robotic Manipulators in Space Applications

This PhD thesis projects aims to analyse the existing state-of-the-art visual Simultaneous Localisation And Mapping (vSLAM) software, verify its suitability for robotics exploration on the lunar surface, identify challenges and optimize the software for potential shortcomings as the algorithms have been optimized for terrestrial applications. The main challenges on the lunar surface consist of the extreme lighting conditions of extremely dark shadows and extremely bright sunlight that is not filtered through an atmosphere, and the feature-poor environment. To mitigate the shortcomings of the purely vSLAM algorithm, additional sensors such as (solid state) Light Detection And Ranging (LiDAR) sensors, wheel encoders and IMU can improve the robustness and accuracy of the mapping process, making it more adapted for the lunar surface.

LUNAR-SLAM

Using Machine Learning based techniques to improve the quality and resolution of remote sensing data of the Moon. Robotic task planning on lunar surface requires enormous amount of data, typically gathered with satellites. This project presents a direct solution to improve the already collected data, particularly images, enhancing the capabilities of the planning teams working in future lunar robotics missions. This work explores additionally an alternative to the launch of new satellites with better instrumentation into lunar orbits to obtain more precise data, with the corresponding economic and time saving.

Enhancing robotic mission planning on the Moon

FiReSpARX’ objective is to use distributed ledger technologies as well as intelligent agent technologies to develop space robots that follow a given governance (RegTech) and make autonomous economic decisions (FinTech) to jointly create collaborative MRS. This form of dynamic collaboration between space robots could lower market barriers in the space sector and thus transform the whole industry. This topic hence perfectly fits Luxembourg’s national research priorities “Autonomous and intelligent systems and robotics for earth and space“ as well as “Fintech/RegTech and transformative applications of distributed ledger technologies“.

DLT-based decision-making strategies of heterogeneous multi-robot system for planetary exploration

Nowadays, we are using more and more robots to replace the tedious, difficult, or even impossible tasks to be done by humans. This is especially true in the undergrounds, where the walls can be damaged by water, vibration, seismic waves or any aggression through time. The CAESAR-XR project allows, through all its dangers, to send entire fleets of robots to map, explore and discover points of interest in long-abandoned mines and caves. For such systems to work autonomously, they must 'see' and 'understand' the environment through different components. In the CAESAR-XR project, robots of all types can be used. Some sensors allow the robot to make a 3D map, while others use colour cameras or calculate the distance between the robot and the various obstacles in front of the robot. These fleets of intelligent robots will not always be connected with a human pilot, which is why they must communicate and make decisions by themselves when exploring the underground. CAESAR-XR aims to develop a simulated environment based on a randomised generation of underground environments to facilitate the validation and verification of the developed algorithms before accessing underground environments. Experts can then use the data collected by the robots to help identify potential weaknesses in the internal structure of the explored areas. The use of this information can be used in the mining industry as well as in space with potential explorations of lunar lava tubes that could serve as protection for a base against radiation and extreme temperatures.

CollAborative Exploration and Structural Analysis using heterogenous multi-Robot system in eXtreme enviRonments

MICRO5G focuses on industrial research on URRLC and MEC with the objective of acquiring new skills and knowledge related to the deployment and support of drone services in 5G.mURRLC will allow drone service providers to extend the flight time by reliably moving complicated processing tasks to the mobile edge. In parallel, MEC will support crowdsourcing data from multiple drones, which can be consolidated and utilized as a traffic management tool. The synergy out of combing these capabilities shall not only enable innovative services such as safe flight beyond visual line of sight (BVLOS) and dynamic guidance/geofencing, but also promote an interactive, rather than independent, integration of different 5G technologies.

Mobile Edge Computing for 5G DROne Systems

Advanced robotic capabilities and operational autonomy are crucial enabling technologies for the advent of industrialization in space. A key advantage of using robots in space is their capability of operating continuously in harsh environments. This enables applications that exceed the reach of possible human intervention. Redwire/Made in Space Europe is developing state-of-the-art light-weight low-cost robotic arms for space, to enable a commercial revolution in-orbit, on-surface, and beyond. To realize ambitious technology architectures for sustainable utilization, commercialization, and exploration of space, advanced in-space robotic capabilities are required. Operational and algorithmic autonomy is a core challenge in that direction. Together, SpaceR and Redwire/Made In Space Europe are working on frameworks, methods and algorithms for incrementally developing autonomy with focus on advanced robot vision and perception.

Modular Perception and Autonomy for Light-weight On-Orbit Robotic Manipulators

Space debris is caused by millions of non-functional, human-made objects left in space that become a hazard for current and future space missions. Yet, no Active Space Debris Removal (ASDR) systems currently exist. Many methods have been proposed, but most of them are still in the technology development stages. Thus, on-ground experimental facilities for the test, verification and validation of ASDR will be critical for pushing the technology to the operational stage. SpaceR and Spacety target to explore, within the HELEN project, the potential of the 2D micro-gravity facility (Zero-G lab) for validating FlexeS, a small Flexible Capture System for debris removal. Advanced computational methods will be developed combining HIL and SIL to recreate high-fidelity in orbit scenarios. The integration of virtual and physical systems will enable close-to real testing, speeding up the transition between the development and deployment stages of ASDR systems.

High-fidELity tEsting enviroNment for Active Space Debris Removal

Zero-G Lab is designed to allow students and researchers to test the movement of in-orbit robotics, satellites and other spacecraft in a micro-gravity environment – similar in concept to an air hockey platform. On Earth, we take gravity for granted, but moving in its absence presents a number of challenges for in-orbit operations. For instance, a small push between two orbiting systems could make one or both to tumble and get out of control. Seeing how spacecraft and orbital robotics can be controlled or perform with decoupled systems in this environment provides students the unique chance of understanding and forecasting their behaviour in space

Zero-G Lab - Multi-Purpose Zero Gravity Lab facility

The goal of this project is to develop a fully functioning VBN system prototype at technology readiness level (TRL 7) that would be representative of the future commercial off-the-shelf product for the space industry. This means the system should be completed and tested in a laboratory environment that is representative of the operational environment in orbit. The result of this project - the VBN prototype will be a standalone hardware system running custom navigation software algorithms, including AI and computer vision, to perform complex autonomous relative navigation maneuvers necessary to rendezvous and/or dock with an uncooperative target that does not have any active or passive aids for such maneuvers. In addition to that, skills and processes for product development and testing will be created during this project. This will be achieved by building necessary partnerships with academia and the industry for gaining access to state-of-the-art research, current technology and technological processes as well as the necessary testing facilities for validation. Algorithms will be coded and tested in the simulation environment using software-in-the-loop (SITL) approach. Once the algorithm performance matches or exceeds the performance in the requirements definition document, the algorithms will then be ported to a chosen embedded system and tested for efficiency using hardware-in-the-loop (HITL) simulations.

Vision-Based Navigation system (VBN) for autonomous satellite navigation in space

Descent and landing (D&L) are arguably among the most critical phases for the success of upcoming space exploratory missions. They can face difficult challenges due to extremely uncertain and variable physical environments of the target planet or asteroid bodies. These challenges require fully autonomous Guidance, Navigation and Control (GNC) algorithms to achieve mission success. Reinforcement Learning (RL) is a promising technique that has the advantage over traditional engineering GNC systems to adapt to unknown situations and circumstances that are hard to account for manually. The integration of RL with the GNC subsystem in a realistic infrastructure (such as LUXLANDER) could be a game-changer for upcoming space-exploration missions. Indeed, using RL-based algorithms in the real world it is crucial to develop a suit of the trained control policies for the given problem in a highly accurate simulation environment and to demonstrate their robustness with a realistic Verification& Validation (V&V) test-bench.

Recent Publications

Courses given recently

SpaceR Team