Flight Testing Algorithms for Graceful Performance Degradation after a Failure

摘要

Abstract. Due to Unmanned Aerial Vehicle (UAV) market expansion and their future integration in National Airspace System (NAS), failure management is becoming more and more crucial for the safety of UAVs and other vehicles. Failures can be managed at different levels of a UAV architecture and with different approaches, as proved by a rich scientific literature on this topic. The paper describes in detail a functional architecture and related algorithms for failure management, highlighting the crucial aspects related to this function and how the proposed implementation can guarantee a graceful performance degradation after a failure. Moreover, some results of a flight test campaign are included and demonstrate the effectiveness of the proposed architecture in managing both real and simulated non-nominal conditions following a system malfunction.