Files and links (1)
PDFOpen Access (License Unspecified), Open Access
Dissertation
A deep learning based failure detection, isolation, and recovery system for airplane elevators
Doctor of Philosophy (Ph.D.), Drexel University
Mar 2026
DOI:
https://doi.org/10.17918/00011320
Abstract
Aircraft loss of control (LOC), defined as the unintended departure of an aircraft from controlled flight, has been the largest contributor to fatal accidents over the past forty years. LOC events are complex and often have several, perhaps interrelated, contributing factors. This research is specifically focused on those that are attributed to actuator failures. In such an event, a pilot or flight control system must quickly identify the actuator failure and take the correct action to mitigate its effects. Failure to do so can have catastrophic results. This work presents a failure detection, isolation, and recovery (FDIR) framework designed to automatically protect aircraft against elevator failures. It is comprised of two components. The first component is a failure detection and isolation system. A deep neural network is used to estimate elevator, aileron, and rudder position based on sensed aircraft state. If the difference between the commanded elevator position and the estimated elevator position exceeds a certain threshold, the elevator is declared failed. The second component is a failure recovery system. Once an elevator is declared failed, it is no longer a control effector to be used to achieve commanded accelerations -- it is a disturbance that must be rejected. The failure detection and isolation system activates a backup set of control laws specifically designed to mitigate the effects of an elevator failure. It leverages the inherent control coupling of the aircraft and regulator theory to provide an alternate means of maneuvering the aircraft. The full FDIR system was developed and evaluated in a 12 degree of freedom nonlinear simulation. Time history simulation results are presented to demonstrate its effectiveness.
Metrics
1 Record Views
Details
- Title
- A deep learning based failure detection, isolation, and recovery system for airplane elevators
- Creators
- Joseph Masgai
- Contributors
- Bor-Chin Chang (Advisor)
- Awarding Institution
- Drexel University
- Degree Awarded
- Doctor of Philosophy (Ph.D.)
- Publisher
- Drexel University
- Number of pages
- xiii, 105 pages
- Resource Type
- Dissertation
- Language
- English
- Academic Unit
- College of Engineering (1970-2026); Mechanical Engineering (and Mechanics) (1970-2026); Drexel University
- Other Identifier
- 991022170151304721