The research program involves three PhD students and a PostDoc within the Parallel Computing and Optimisation Group (PCOG), led by Prof. Dr. Pascal Bouvry and Dr. Grégoire Danoy (Deputy Head). Each of the PhD students, supervised by Dr. Mohammed Alswaitti, will conduct research on one of the sectors identified as fundamental for the economy in the national standardization strategy 2020-2030, namely: Information and Communication Technologies (ICT), aerospace and construction, with the ICT sector remaining a transversal axis for each of them.
- ICT project: “Satellite Images Data Marketplace” – Mr. Manuel Combarro Simon
This research project focuses on potential combinatorial problems when several images from different satellite data providers are needed, for example to obtain an image time series of a location or to form an image mosaic covering a large area. Especially in the latter case, the group of images must meet certain criteria to be merged.
“Our current research on this topic has two major goals: to develop a heuristic evaluation to recommend the lowest cost of image combination and to find gaps between the state-of-the-art research and existing standards for satellite image mosaicking. For example, are there guidelines for the characteristics that images must meet to be merged?”
- Construction project: “Building Information Modelling (BIM) and its integration with Artificial Intelligence (AI)” – Mrs. Hedieh Haddad
This research project aiming at leveraging machine learning models to learn from available data, via building information modeling (BIM) or even via on-demand sensors, offers some very promising perspectives. First, it would allow independent decision-making and finding the best available solution to optimize price, time and energy consumption during construction. Secondly, it would allow some anticipation on when certain actions need to be performed, such as planning, repair, or demolition time.
Taking advantage of the capabilities of Artificial Intelligence enables a variety of qualities in construction such as, but not limited to, faster decisions, cost and time forecasts, reducing the risk of human error and improving safety on the construction site. It allows to better organize information and to improve performance throughout the design, manufacturing and construction processes, which is one of the goals of standardization in construction.
- Aerospace Project: “Swarms of Nano-satellites” – Mrs. Maria Hartmann
This research project focuses on the application of federated learning to satellite swarm scenarios. It has interesting potential because it allows for online training on each machine and can be leveraged to train more robust intelligent swarms. Learning in a way that maintains an up-to-date model on each agent means that agents can still operate successfully if the connection to others is interrupted, and resume the learning process later. In addition, federated learning can potentially be performed on swarms where not all agents are connected to each other, for example in large-scale networks.
The federated learning approach could be extended to facilitate the fusion of different types of data, allowing for example the implementation of multiple models of satellites observing different aspects of their environment to make more accurate joint predictions about their environment.
Since satellite swarm autonomy is a new and rapidly developing research area, this research will likely lead to the identification of standardization needs and provide new perspectives on existing standards.
ILNAS is delighted with the prospects opened up by this research program, particularly in terms of standardization and education, since the research results will also be used to develop the MTECH Master “mastering smart ICT, standardization and digital trust for enabling next generation of ICT solutions”, jointly conducted by the University of Luxembourg and ILNAS, which will open its doors to a second class as early as February 2023 (registrations from September 2022), in collaboration with the Chamber of Employees (Chambre des salariés, CSL).