ESA GNC Conference Papers Repository
EROSS+ ground demonstrations from Model to Hardware in the Loop validation
The European H2020 project EROSS+ Phase A/B1 is standing for European Robotic Orbital Support Services, and has been led over 2021-2023 to mature the future robotic servicing missions with a highly-autonomous and coupled Guidance, Navigation and Control (GNC) architecture of a Servicer robotic spacecraft. This project is built upon the previous developments of the Operational Grants led by the Strategic Research Cluster in Space Robotics funded by the European Commission since 2016. More specifically, EROSS+ project aims at deriving a system design of a robotic Servicer approaching, capturing and servicing a Client satellite. It thus integrates and demonstrates the key European building blocks in space robotics by demonstrating their performances from the tuning with Model in the Loop (MIL) tests, to the end-to-end validation with Hardware in the Loop (HIL) experiments. The EROSS+ mission timeline is based on four main steps covering (1) the approach and inspection of the Client with an autonomous visual-based navigation using advanced processing and filtering techniques; (2) the robotic capture using state-of-the-art compliance control techniques to synchronize the robotic arm and its platform to smoothly grasp a mechanical target; (3) the mating of the two spacecraft through a dedicated interface for refuelling; and (4) the robotic exchange of a replacement payload designed with standard interfaces. This paper will focus on the validation process implemented to mature the autonomous software of the Servicer vehicle during the step (1) from numerical MIL validation with emulated relative sensors, to the end-to-end HIL experiments with real equipment. The validation scenario encompasses the safe approach of a Client spacecraft using relative sensors feeding the autonomous GNC software coupled with an advanced image processing from few kilometres to few meters from the Client. To reach the end-to-end validation of such a functional chain, a step-by step approach has been implemented to increase its overall Technology Readiness Level (TRL). The different software building blocks of GNC or image processing were initially at a prototype stage of TRL3/4, running on ground processing units (e.g., laptop) with ground operating systems (e.g., Linux). The core of this paper will be the presentation and feedbacks on the incremental validation used to move from TRL3/4 to TRL5 by implementing a four steps approach switching from (a) Model in the Loop (MIL), to (b) Software in the Loop (SIL), to (c) Processor in the Loop (PIL), and eventually to (4) Hardware in the Loop (HIL). This last step is qualified of end-to-end in the sense that it merges the key software hosted in a representative processing boards and running in a representative robotic test bench creating realistic datasets from the sensors equipment. The core architecture of this incremental validation is based on a complete numerical emulation of all the systems presents in the closed-loop in orbit. Its main elements can be cast into the Ground Segment for sending the telecommands and monitoring the telemetries, the servicer Space Segment with its On-Board Software (OBSW) and its GNC, the simulator of the Dynamics, Kinematics and Environment (DKE) of both the Servicer and Client, and eventually the hardware & data processing elements feeding the OBSW with either numerical/mathematical models or experimental sensors data. This architecture embeds the following elements: -Ground Segment: the commands from ground are sent to the Servicer OBSW by high-level commands of type GO/NO-GO to proceed with the next steps of the mission. In parallel the telemetries and video monitoring stream are used to have a full vision of the on-going mission, as real-time as possible. -On-Board Soft-Ware (OBSW): the software encompasses the Guidance, Navigation and Control (GNC) functions moving safely the Servicer around the Client, with the rest of the autonomy functions ranging from thermal control, to electrical power management, to sensors and actuators data processing. -Dynamics, Kinematics and Environment (DKE): this simulator is larger than the spacecraft dynamics in orbit as it also covers the sensors and actuators numerically emulated during the experiments. It covers both the orbital and rotational dynamics models of the Servicer and Client spacecraft; along with the sensors and actuators models fed by the dynamic models in measurements and feeding it on the actuation forces/torques applied; and eventually the Image Processing transforming the images (or any other data from relative sensors) into exploitable physical quantities at the GNC level. The main advantage of this modular architecture is to support the incremental validation by switching between: a)The numerical/mathematical models of the relative sensors and its processing for (a) MIL test, b)The processing software hosted on a standard laptop and running on emulated data of the relative sensor equipment fed by the dynamics simulator for (b) SIL tests, c)The processing software hosted on a space-representative processing board, and still running on emulated data of the relative sensor for (b) PIL tests, d)The processing software hosted on the space-representative processing board on realistic data from a prototype of the relative sensor in a robotic test bench for (b) HIL tests. The core elements of this paper will detail further the results obtained at each stage of this incremental validation until the end-to-end validation has been performed. The advantage of this approach is its inherent modularity allowing to update and re-validate some key elements of GNC or IP software by running the MIL tests first, and then jumping directly to the HIL architecture once the SIL & PIL processes have been clearly established with the autocoding and compilation chains. This has been especially true for the GNC development and validation as minor to large updates have been done throughout the project with a reduced integration time thanks to the SIL/PIL frozen workflow after the first prototype. EROSS+ project has been co-funded by European Unions Horizon 2020 research and innovation program under grant agreement N°101004346 and is part of the Strategic Research Cluster on Space Robotics Technologies as Operational Grant n°12. Thales Alenia Space has led this project in collaboration with DLR, GMV, SINTEF AS, and PIAP Space.