Towards Certification of a Reduced Footprint ACAS-Xu System: a Hybrid ML-based Solution

Mathieu Damour , Florence de Grancey , Christophe Gabreau (1) , Adrien Gauffriau (1) , Jean-Brice Ginestet (2) , Alexandre Hervieu (2) , Thomas Huraux , Claire Pagetti (3) , Ludovic Ponsolle (4) , Arthur Clavière

1 Airbus [France]
2 Direction générale de l'Armement (DGA)
3 DTIS, ONERA, Université de Toulouse [Toulouse]
4 APSYS (Blagnac)

Approximating while compressing lookup tables (LUT) with a set of neural networks (NN) is an emerging trend in safety critical systems, such as control/command or navigation systems. Recently, as an example, many research papers have focused on the ACAS Xu LUT compression. In this work, we explore how to make such a compression while preserving the system safety and offering adequate means of certification.

Domaines: Intelligence artificielle [cs.AI]; Calcul formel [cs.SC]

Mathieu Damour, Florence de Grancey, Christophe Gabreau, Adrien Gauffriau, Jean-Brice Ginestet, et al.. Towards Certification of a Reduced Footprint ACAS-Xu System: a Hybrid ML-based Solution. SAFECOMP 2021: Computer Safety, Reliability, and Security, pp.34-48, 2021, 978-3-030-83903-1. ⟨10.1007/978-3-030-83903-1_3⟩. ⟨hal-03355299v1⟩