Prognostics for Autonomous Electric-Propulsion Aircraft

Abstract

An autonomous unmanned aerial system (UAS) needs, during the flight, accurate information about the current failure state of the aircraft and its capabilities in order to safely perform its mission and properly react to contingencies. The flight battery of an electric-propulsion aircraft is its most relevant resource. Model-based prognostics algorithms are used to obtain good estimates of its current state of charge and remaining capacity. However, these algorithms can have a large computational footprint. We present Prognostics-as-a-Service, a hybrid approach combining on-board computation with server-based prognostics on the ground. In this paper, we focus on the role, battery prognostics plays for the safe operation of a highly autonomous aircraft: prognostics for (1) continuous on-board safety monitoring, (2) for UAS operations, and (3) for contingency planning. We present the NASA Autonomous Operating System (AOS) and discuss how the autonomous components closely work together with on-board and server-based ground prognostics systems. We will illustrate the system with case studies on small NASA unmanned aircraft.

Keywords

• electric-propulsion aircraft
• prognostics