Flight mishap prevention for UAVs

Abstract

For Uninhabited Air Vehicles (UAVs) to achieve operational levels of flight safety comparable to those of piloted aircraft, it is important to replicate the critical skills and predictive capabilities of the pilot to ensure proper flight mode selection and prioritization of tasking. Each individual combination of modes must be able to accommodate both predetermined as well as unanticipated causes of uncertainty. The system must adapt to these changes based on sensory inputs, as well as be robust against uncertainties that have not been sensed or directly measured by the system. The purpose of Technologies for Reliable Autonomous Control (TRAC), a joint Lockheed Martin Aeronautics Company/General Electric/Jet Propulsion Laboratory mission management architecture, is to autonomously accomplish this complex mission and subsystem management task. To do this, it is desirable to provide "outer-loop" compensation for larger sources of disturbances that may require corrective maneuvers, mode switching, or changes to the command inputs to the flight control system. Flight Envelope Protection is one such feature. Based on our analysis of recent UAV data, emergency procedures account for about 26% of mishaps, while errors in operating procedures account for about 10% of mishaps. Thus, Flight Envelope Protection could prevent up to about 36% of mishaps. While there are many points of similarity between UAV Flight Envelope Protection (FEP) and FEP for manned aircraft, a number of new issues arise in the UAV case. Hence, this is not straightforward extension of existing techniques.