Objectives

The Intelligent System Study will define the path towards the achievement of the following expected benefits:

• No human involvement for selected use cases.
• Enabling handling much more complex systems
• Ability to react in real-time in case of unexpected events
• Advanced diagnostics in support to drastic downtime reduction

The Intelligent System Initiative is considered Instrumental to increase competitiveness in future Missions via:

• Reduction of ground infrastructure cost
• Increased availability (reduction of service downtime)
• Increasing orchestration capabilities

Overall it will enable and accelerate enhanced autonomy of future products, leveraging on selective use of  AI/ML Techniques.

Challenges

The main challenges identified in the system study are the following:

• The need for a Multidisciplinary and holistic assessment of the use cases and proposed solutions, including the views of:
  • Operators & service suppliers
  • Satellite integrators
  • Avionics & SW designers
  • Payload designers
  • AI-ML element providers
• Pursuit of Solution generality: to develop a reference in terms of architecture and workflow envisaged for applications of E2E autonomy for a potential multi-orbit & multi-application scenarios.
• Standardisation of onboard Telemetry and other on ground data sources in view of the concept generality and industrial adoption.

Data availability for model training in Machine Learning, impacting the quality and efficiency of the resulting models.

System Architecture

The study has defined a flexible reference frame (both for SC  designers/integrators/operators and SW library developers) that will enable European Industry to accelerate & demonstrate the development and early prototyping of:

• Intelligent system building blocks
• Autonomous E2E (space & ground) system

This reference frame will be based on:

• Distributed processing entities both within the Ground and onboard infrastructure
• Large and robust data storage solutions (volatile and non-volatile memories)
• Communication network, with high-speed and low-speed interfaces, to support large data volume exchanges
• Hypervisor-based SW architecture, and shared resources across multiple AI libraries running on the application layer
• Modular partitions, isolated from each other by the hypervisor

Plan

The project activity has been divided into the following main Tasks:

• Task 1: Identifying Satellite Communication Applications that benefit from Augmented Satellite Autonomy
• Task 2: Trade-off and identification of the most promising strategies and architecture concepts capable to enhance the degree of autonomy for system/satellite operations
• Task 4: Description of Reference Architecture and Building Blocks for AI and ML Adoption.
• Task 5: Assessment of industrial implementation aspects
• Task 6: IOE (In Orbit Experiment) Feasibility Assessment

During the study, two Industrial Consultations have been conducted, targeting to foster interactions across the supply chain, gather requirements on the use cases and sharing the high-level conclusions. This has strongly contributed to creating an ecosystem of active industrial players, paving the way to future interoperable solutions.

Current Status

• The System Study has quantitatively analysed the impact and benefit of increased autonomy applied to multiple functions and use cases.
• For the most promising applications a detailed methodology, workflow and general architecture trade-off has been conducted.
• Relevance of the use cases and proposed solutions have been iterated and validated with industry via Industrial consultation sessions.
• An architectural frame has been specified, outlining the necessary features (SW & HW) and key technical specifications for each of the use cases.
• Key aspects and considerations have been analysed to deploy a sustainable and competitive Industrial Ecosystem to enable and accelerate the adoption of the proposed solutions.
•  Potential roadmaps and opportunities for the deployment of an IOE have been explored.

Related Links

• AI-based On-board Reconfigurable FDIR and Lifetime Prediction for Constellations
• Intelligent System Initiative: Advancing Space-Based Systems

Companies

OHB System AG

https://www.ohb.de/

Germany

S.A.T.E. S.r.l.

http://www.sate-italy.com/

Italy

Kepler Communications

https://kepler.space/

Canada

Objectives

The Intelligent Platform for Constellations Study (IPFS) represents the initial phase of a strategic initiative aimed at developing a highly autonomous architecture for future satellite communication constellations. As satellite constellations become more complex to manage, the need for advanced platform autonomy and greater integration of Artificial Intelligence (AI) will grow significantly. The same is expected to happen to the whole system, therefore, even if this project focuses on the space segment, ground and general operational requirements are considered.

The first step is the identification of applications and relevant concept of operations by investigating different scenarios: standalone satellites, fleets and constellations. The most promising use cases are selected based on their potential benefits, evaluated up to the end-to-end system level. Further on, the baseline architecture and building blocks are defined, including solutions for AI integration, while considering industrial implementation aspects and operational impacts.

It is important for the proposed solutions to be widely supported by the entire European space sector and in harmony with its supply chain. Consequently, a consultation with key industrial stakeholders has been conducted in order to gather feedback and needs. The Study concludes with feasibility assessment of future in-orbit experiments (IOE) and preliminary implementation strategy for preselected autonomous functionalities.

Challenges

As the project strives for greater autonomy, it faces several challenges related to integration of ML-based software into spacecraft systems: lack of established standards in the space sector; the need for dedicated V&V process tailored for AI methodologies; and edge processing cybersecurity threats.

Moreover, there is a paradigm shift for ground operators and facilities, coupled with the absence of procedures for model re-training and fine-tuning. The implementation of these ambitious goals require at the present stage the use of Commercial Off-The-Shelf (COTS). High-end alternative solutions will be progressively scaled up as soon as the currently under development Rad-Hard option becomes available.

System Architecture

The Preliminary Reference Architecture for the intelligent platform represents a shift from traditional subsystem-based designs to an integrated solution that enhances efficiency and reduces complexity. Central to this architecture is the On-Board Computer (OBC), which consolidates functionalities typically distributed across separate subsystems, including transponder and GNSS capabilities.

For fleet and constellation applications, Inter-Satellite Links (ISLs) are crucial. The architecture offers two ISL options – an “RF omnidirectional link” and an “RF/optical link with a dedicated data router” – providing flexibility to meet mission requirements and ensuring reliable communication in dynamic environments.

To streamline the system, the design limits discrete lines by utilising modern serial links and distributing processing power across system edges. A high-performance COTS processing bench supports the OBC, enabling advanced AI and ML capabilities for autonomous decision-making.

A high-speed serial data platform, possibly using optical point-to-point links, serves as the communication backbone, leading to a dual-star topology with the OBC as the central node for redundancy.

Plan

The study has a duration of six months. The main milestones include the kick-off, an intermediate review and final review (final presentation). Progress meetings are organised monthly and as necessary.

Key events of the project are the industrial consultations: a first offline consultation midway through the project and a second consultation at the end of the project.

The scope of such a tight timeline (together with the ongoing parallel issue of the first batch of de-risking technology activities in the frame of the Intelligent System Initiative), is to drastically reduce the time to market for the targeted advanced autonomy.

Current Status

The project is currently ongoing. A thorough investigation of promising applications and use cases has been performed considering the standalone satellite system, the fleet and constellation. Evaluation metrics have been discussed and used for the selection of three use cases to be further extended in proper business cases.

A preliminary architecture for the Intelligent Platform has been outlined. However, refinements are envisaged according to the feedback received from the first industry consultation.

Related Links

• Opportunity: Intelligent Platform for Constellations System Study
• Intelligent System Initiative: Advancing Space-Based Systems

Companies

Thales Alenia Space Italia

https://www.thalesgroup.com/en/worldwide/space-italy

Italy

Airbus Defence and Space (Germany)

http://airbusdefenceandspace.com/

Germany

Objectives

The objectives of BOLERO were multi-fold:

• To identify potential satellite telecommunication (satcom) functionality and applications that could benefit from the use of continual learning methodologies in a fully transparent, quantifiable and reproducible way which will be reusable in future satcom and other missions for an informed selection of on-board machine learning (ML) models that could benefit from continual learning techniques.
• To explore, develop, and simulate different continual learning implementation techniques for the identified satcom applications. This project objective also aimed to explore the connection between offline and online learning and the current state-of-the-art methodologies that would allow models that have been pre-trained offline to be updated so they can be enhanced by online/continual learning.
• To identify and justify a suitable system architecture for on-board continual learning applications through performing the benchmarking process of all developed ML algorithms with continual learning techniques for all satcom applications and simulation scenarios, as well as through performing the theoretical trade-off analysis of the hardware and system architectures considered in BOLERO.

Challenges

The most important challenges of BOLERO related to:

• Availability of hardware architectures that could be used for benchmarking continual learning AI algorithms.
• Availability of datasets that could be used to verify and validate continual learning scenarios (therefore, we developed synthetic data simulators for all selected satellite communications applications).
• Building reproducible, unbiased and reproducible pipelines for the quantitative validation of AI algorithms.
• Developing an objective procedure for selecting satcom use-cases that would benefit from continual learning paradigms.

System Architecture

The technology developed in BOLERO is fully modular, and directly relates to the key product features, including:

• Synthetic data generators;
• Continual learning artificial intelligence algorithms developed for selected satcom applications;
• Assessment matrices for selecting a) appropriate satcom applications for on-board continual learning deployment and the b) best hardware architectures for such on-board implementation;
• Research and development roadmaps. All these components are stand-alone and self-contained entities that can be effectively used separately.

Plan

Project was planned and divided into specific Work Packages focusing on the following:

• WP100 SOTA: Review and analysis;
• WP200 Satcom applications: Identification and analysis;
• WP300 Algorithms: Continual learning for satcom;
• WP400 Hardware: Performance assessment;
• WP500 Programmatic and development gaps;
• WP600 Software;
• WP700 Management, outreach and dissemination.

Current Status

In BOLERO, we identified satcom functionalities that could benefit from on-board continual learning, selected them via a quantifiable process, and developed ML models accordingly.

We built data simulators to test various continual learning techniques, emphasising reproducibility and algorithm generalization in realistic settings. The project delivered an end-to-end pipeline for continual learning and online adaptation for satcom, validated in simulated scenarios. Finally, we benchmarked these methods across different hardware, including KP Labs’ Leopard and BrainChip’s Akida, providing comprehensive results for all algorithms, hardware, and applications explored within BOLERO.

Companies

KP Labs Sp. z o. o.

https://kplabs.space

Poland

OHB HELLAS

https://www.ohb-hellas.gr/

Greece

Eutelsat OneWeb (OW)

https://oneweb.net/

United Kingdom