CRTES research
Most of our work is in the field of E/E architecture design and real-time communication networks, be it in the automotive or aerospace domains. Our work aims to further automate the design activities based on constraints and goals.
Areas of research
We have been investigating for several years an approach rooted in computational thinking where system designers break down the general multidimensional design problem into smaller problems that algorithmic tools can solve in a near optimal way. In the continuity of the work performed in 2020, progresses were made this year in the development of deep-learning models, based on Graph Neural Networks (GNN), that speed-up by several orders of magnitude the performance evaluation of Ethernet networks with respect to simulation or mathematical analysis.