Today, Airbus A350 XWB is undoubtedly the most modern and efficient commercial jetliner in the world. Since January 2015, this aircraft is flying worldwide with a high performance robust model-based monitoring method on-boarded. This exceptional achievement has been resulted from an inspiring collaboration between my research team and Airbus on FDIR issues in aircraft systems, starting with mid-2000s. With current certification standards in civil aviation, extremely rigorous specifications are to be satisfied, not only in nominal flight regimes, but also in extreme, unusual, non-standard/off-nominal and unexpected flight conditions. Achieving flight-proven and certified model-based systems has been a tall order and required several years of research and innovation, V&V activities and maturation.

Five years after the first A350 commercial flight, this note posted on the IFAC Blog and IFAC social media looks back at some lessons learnt from this amazing success story. The note has been prepared by Dr. Philippe Goupil (eXpert FDIR – A/C Control Architecture & Functions, Airbus, Toulouse – France) and myself.

What next ? Today, A350 offers on-board cockpit technologies a pilot 2-3 decades ago could only dream about. But, what will next-generation aircraft and air transport look like in the next decades? The challenges tomorrow are far greater than those faced in the past as individual flight systems will have to operate with greater autonomy and intelligence within a cyber-physical environment. The next leap forward is Single Pilot Operations (reducing the commercial cockpit to a single pilot, 2030+). The move to SPO is mainly motivated by cost savings and the chronic problem of pilot shortage in the coming years. SPO will be a logical stepping stone on the way to self-learning and autonomous planes able to behave in a non-deterministic way (2050+). Today, autonomy is not a target in itself – the potential of autonomous technologies are explored for improving flight management operations and overall aircraft performance. From a scientific perspective, solving the puzzle of autonomy requires new concepts and cross-domain methods. An aircraft is a complex cyber-physical system and some research issues are for example: how to design efficient and scalable safety management methods on the way to operational autonomy, including high level discrete contingency monitoring and decision-making (which may be approached by robust IA techniques) – how to design autonomous cyber-physical systems that are actually safe enough to be deployed in real situations – how to find ways to validate safety strategies of such systems and to demonstrate that a sufficient level of safety can actually be achieved …

N.B. An exhaustive discussion can be found in: A. Zolghadri, On flight operational issues management: Past, present and future. Annual Reviews in Control. Vision article, 2018. pdf file.

On the 18th of December 2019, Airbus demonstrated first fully automatic vision-based take-off in an A350 configured for flight testing.