GMV Brain

Overview

What is GMV Brain?

GMV Brain is a pre-qualified high-performance software framework designed to empower autonomy of robotic systems -both space and terrestrial- through advanced autonomous decision-making capabilities.

Built upon the proven heritage of the ERGO project (EU Horizon 2019), GMV Brain represents a significant evolutionary leap, consolidating GMV’s state-of-the-art space robotics expertise into a robust and mature product ready for Category-B software qualification.

GMV Brain has been engineered as a flexible, configurable and extensible solution for covering ambitious space and terrestrial missions demanding high levels of autonomy.

For further information:

If you have a potential use case for GMV Brain and would like to discuss your requirements, CONOPS or request more information, please contact the team: [email protected]

+10

Projects that made the smart choice of selecting GMV Brain for implementing on-board and/or ground autonomy solutions, achieving successful V&V and final demonstration in the planetary/terrestrial scenarios (ADE, PROA-ACT, MIPTOOL, CoRob-X and OBPE-AI) and orbital scenario (MOSAR, EROSS, PERIOD, EROSS+, EROSS-IOD and currently in SCOPE and ASIMOV)

+30

GMV Brain has been successfully integrated, validated with more than 24 different partners (ADS, Thales, DLR, CNRS, DFKI, PIAP Space, SCISYS, SINTEF, SPACE APPLICATIONS SERVICES, SENER, SINTEF, MAGELLIUM,…) and 9 university research centres (LAS PALMAS DE GRAN CANARIA, MALAGA, SALENTO, BASEL, GRENOBLE ALPES, CRANFIEDL, LONDON, OXFORD AND STRATHCLYDE).

Highlights

GMV Brain framework has been designed, developed and validated to become the brain of the future space (planetary and orbital) and terrestrial missions demanding high-level of autonomy on-board and on-ground (EGSE, Digital Twins and Simulators).

GMV Brain design delivers fast, seamless customization to power on-board autonomy across the three pillars of ISAM (In-Space Servicing, Assembly, and Manufacturing).

In the planetary & terrestrial surface exploration scenarios GMV Brain can be used for implementing autonomy in operations like SLAM/VSLAM, planetary samples collection, Add-hoc Mesh Network, and manufacturing of planetary infrastructures.

More information

Main Features

How does GMV Brain work?

GMV Brain autonomy framework is designed as a layered architecture (executive, deliberative and functional) which allows the framework to be easily customized to any new mission.

The outcome of this customization is a GMV Brain software instance or orchestrator fully aligned with mission/subsystems CONOPS, specifications and I/Fs, and split into two major components:

Agent; an autonomous goal-oriented controller based on the well-established Sense-Plan-Act paradigm. It implements both reactive (executive) and deliberative (planning) behaviours on each (configurable) execution period (Tick).
The Agent orchestrates the mission or subsystems by periodically implementing the following processing sequence:

Sense: The Agent checks and processes new mission goals (from ground or other spacecraft/rovers) and periodic feedback statuses from the functional layer.
Plan: The Agent generates a plan to achieve and monitor the high-level mission goals while considering its environment. If unpredicted events are detected during the “Sense” phase, the Agent dynamically reacts by generating a new plan (re-planning) or by executing the specific FDIR solutions.
Act: The agent executes the nominal plan or recovery the actions by issuing low-level commands to the functional layer.

Functional Layer; responsible for the low-level interaction with all SW & HW subsystems and components (i.e. GNC, Thermal, Power, robotic arms, sensors, standard interconnects, refuelling units, etc.).

Key features of GMV Brain

Designed from the ground up to enable fast and straightforward customization, tailoring and extension when required.
Applicable to orbital, planetary and terrestrial missions.
Goal oriented, supporting E1-E4 ECSS autonomy levels.
On-board planning and replanning at E4 autonomy.
Applicable to single robot/satellite and multi-robots or constellations.
Available in different programming languages.
Well documented and compliant with Space product quality assurance (PA) standards.
Integrable with different ROCS, ROS and middleware.
Compatible with a wide range of HW platforms and Operating Systems.
Proven track record of V&V excellence across 10+ projects and counting.
Supporting OTA (Over The Air) updates.

Data sheet

Type	Highly customizable Autonomy Software Framework (ASF) for implementing Autonomy in space and terrestrial subsystems and applications.
Autonomy Levels	Supports all ECSS autonomy levels (E1-E4), from immediate and teleoperated TCs (E1) to goal-oriented TCs (E4).
Controller’s reactiveness	Configurable (1Hz by default).
Deliberative planning and replanning	Two options for smart decision taking: Deterministic PDDL planner based on symbolic AI algorithms. Non-deterministic optimizer based on AI constraint solver algorithms.
Dynamic replanning and FDIR	Yes, dynamic replanning supported following unexpected event occurrences, newly received goals and FDIR events handling.
Multi-Robot/Sat	Yes, supports collaboration and cooperation between different robots, different spacecraft and constellations. Applicable also to hybrid Robot-Human CONOPS.
Performance	Low CPU and memory consumption.
Communication’s Middleware	ECSS PUS (CCSDS Space Packets), NanoSat MO Framework (CCSDS Space Packets), ESROCOS, TASTE, ROS1 and ROS2.
Programming Language	C, C++ and Java.
SW Criticality Category	B
Operating Systems	Ubuntu, PetaLinux, Xilinx, RTEMS, PikeOS and Windows (terrestrial)
HW platforms	LEON, ARM, Xilinx and PC (terrestrial)

Applications

In Space (Orbital)

Mission/Payload Vehicle Manager (MVM/PVM).
Autonomous cooperative operations based on Inter-Satellite-Links (ISL).
Autonomous, semi-autonomous and hybrid approaches with human in the loop (robot-human).
ISAM/IOS (prepared/non-prepared and cooperative/non-cooperative space assets).
- Detection, inspection, rendezvous (CPO) and capture.
- Assembly, upgrade and repair of ORUs/SatApps.
- Refuelling.
- End of Life (EoL) relocation and Active Debris Removal (ADR).
- Manufacturing; space robotic factory for construction/assembly of space stations (hubs/hosts), and long solar panels.
Earth Observation
Communications
Exploration
Constellations

On Surface Planetary/Terrestrial exploration

Single or multi-robot/agent systems (rovers).
Autonomous, semi-autonomous and hybrid approaches with human in the loop (robot-human).
Surface exploration (Moon, Mars, Earth, mines, caves, nuclear and fuel plants, others).
Sample return missions (e.g. MSR).
Dingle/cooperative manufacturing; assembly and construction of future planetary ISRU (In-Situ Resource Utilization) and astronauts supporting infrastructures.
Planetary refuelling stations

On Ground