Smart Orbits: Advancing AI for Space Exploration and Space Plasmas
Following the success of our first workshop, the ASAP (Automatics in Space Exploration) project invites you to our second workshop, dedicated to the next frontier of space technology: onboard Artificial Intelligence. As space missions venture farther and generate unprecedented volumes of data, the ability to process information and make decisions in real-time, directly on the spacecraft, is no longer a luxury but a necessity. This workshop will convene leaders from academia and industry to explore the innovations, challenges, and opportunities in creating truly autonomous systems for space.
The project has already made significant strides in laying the groundwork for robust, space-qualified AI, including the development of ML algorithms for in-situ analysis of Velocity Distribution Functions (VDFs), solar imagery, and Solar Energetic Particle (SEP) prediction concurrent with the setup of FPGA-based hardware testbeds.
Workshop Format and Important Dates
The ASAP Workshop will take place on Tuesday, 7 October 2025, from 09:00 to 17:30 CET. The format of this event is hybrid, featuring both face-to-face and online contributions.
Location of the venue for the face-to-face participants: KU Leuven Department of Mathematics, Celestijnenlaan 200B (room 05.100), Leuven (Heverlee), Belgium.
There will be three types of presentations:
- Keynote setting the scene: 40 min + 5 min Q&A
- Invited talks: 25 min + 5 min for Q&A
- Contributed talks: 10 min + 5 min for Q&A
Please note that there is no participation fee for this event.
Registration is closed.
To take part in the workshop, it is compulsory to complete the registration form.
For onsite participation the deadline to register is 22nd September 2025, EOB
For online participation the deadline to register is 1st October 2025, EOB
Workshop Themes & Topics of Interest
We are seeking contributions that address the full spectrum of on-board AI, from foundational technologies to mission-ready applications. Key themes include:
- Hardware for In-Space AI Acceleration: Exploring the design and implementation of power-efficient and radiation-tolerant hardware. This includes the use of FPGAs tailored for the harsh environment of space.
- AI for Predictive Space Weather Science: Presentations on the development and application of machine learning models for forecasting space weather events.
- On-Board Data Analysis and Clustering: Showcasing ML techniques for the autonomous, in-situ analysis of scientific data. This can include clustering methodologies for particle velocity distribution analysis, selective data downlink triggers, and real-time analysis of solar imagery.
- Software Frameworks and Virtual Validation: Use of advanced "digital twin" flight simulators that allow for the robust validation of AI algorithms in a virtual environment before deployment.
Contact
- George Miloshevich, george.miloshevich (at) kuleuven.be (Project Manager of ASAP)
- Anastasiya Boiko, anastasiya.boiko (at) vaboconsult.com (Project Support Manager)
Organizers
- George Miloshevich
- Ekaterina Dineva
- Panagiotis Gonidakis
- Anstasiya Boiko
Schedule
- 09:00 - Registration and arrival of participants
- 09:30 - Keynote: Enrico Camporeale, Queen Mary University of London
- (Onsite) "Machine Learning in Space Weather: recent progress and future challenges"
- 10:15 - Invited talk 1 Maria Federica Marcucci, National Institute for Astrophysics
- (Onsite) "Future perspective of space observations and artificial intelligence/machine learning advancements”
- 10:45 - Coffee break and posters discussion
- 11:15 - Invited talk 2: Simon Mackovjak, Slovak Academy of Sciences
- (Onsite) “From Orbit to Action: Using ML to Turn Satellite Data into Actionable Information”
- 11:45 - Contributed talk 1 Ivan Zaitsev, University of Helsinki
- (Online) “Applying the Hermite Transformation to Velocity Distributions in Simulations and Observations”
- 12:00 - Contributed talk 2 Soboh Al Qeeq, Laboratoire de Physique des Plasmas (LPP) - UMR7648
- (Online) “Automatic Classification of Plasma Regions and Detection of Plasma Jets in Near-Earth Space Using Supervised Machine Learning”
- 12:15 - Discussion
- 12:30 - Lunch, room 05.130
- 13:30 - Invited talk 3: Romain Maggiolo, Royal Belgian Institute for Space Aeronomy, BIRA/IASB
- (Onsite) “Magnetospheric region classification: insights from the manually labeled GRMB dataset and future directions”
- 14:00 - Invited talk 4: Vicki Toy-Edens, Johns Hopkins Applied Physics Laboratory
- (Onsite) “Towards Onboard AI: Designing from the Ground Up”
- 14:30 - Invited talk 5 Gianluca Giuffrida, IngeniArs
- (Online) “From ground to space: the evolution of hardware accelerators for AI and HPC applications in space domain”
- 15:00 - Coffee break and posters discussion
- 15:30 - Contributed talk 4: Pedro Antunes, KTH Royal Institute of Technology
- (Onsite) “Neural Network Models to Hardware in Space”
- 15:45 - Contributed talk 4: Luis Linan, KU Leuven
- (Onsite) “Modeling toroidal CME fields with Physics-Informed Neural Networks in EUHFORIA”
- 16:00 - Contributed talk 5: Panagiotis Gonidakis, KU Leuven
- (Onsite) “Segmentation of solar coronal structures and solar flare forecasting for SDO/AIA images”
- 16:15 - Contributed talk 4 Francesco Carella, KU Leuven & Turin University
- (Onsite) “Clustering Solar Wind Observations at 1 AU with Self-Organizing Maps”
- 16:30 - Discussion
- 17:30 - Closing words from the organisers and end of the workshop.
Venue
Onsite participation
The workshop will take place at KU Leuven Department of Mathematics, Celestijnenlaan 200B (Room 05.100), Leuven (Heverlee), Belgium.
Online participation
Online participation will be through Microsoft Teams. Link to the Teams meeting will be sent out to registered participants 1-2 business days before the workshop.
Address
KU Leuven Department of Mathematics, Celestijnenlaan 200B, Leuven (Heverlee), Belgium
Den här webbplatsen använder cookies.
Vi använder cookies för att analysera webbplatstrafik och optimera din webbplatsupplevelse. Genom att acceptera vår användning av cookies kommer dina data att aggregeras med alla andra användardata.