Autonomous Drone Assisted Aircraft Inspections
The safety of the passengers, crew, and mechanics is of the utmost importance for any aircraft manufacturer or operator. Visual inspections of the exterior of aircraft are critical to their safe operation, as defects such as corrosion, dents, lightning strikes, or missing parts can compromise the st...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Published: |
Massachusetts Institute of Technology
2023
|
| Online Access: | https://hdl.handle.net/1721.1/151481 |
| _version_ | 1826205625666764800 |
|---|---|
| author | Mighty, Andrew |
| author2 | Barnett, Arnold |
| author_facet | Barnett, Arnold Mighty, Andrew |
| author_sort | Mighty, Andrew |
| collection | MIT |
| description | The safety of the passengers, crew, and mechanics is of the utmost importance for any aircraft manufacturer or operator. Visual inspections of the exterior of aircraft are critical to their safe operation, as defects such as corrosion, dents, lightning strikes, or missing parts can compromise the structural integrity of the whole aircraft. Currently, aircraft visual inspections are conducted by human mechanics in a process that is not only time consuming, but also puts the mechanics and the aircraft at risk, as mechanics must use lifts and cranes to inspect top portions of the aircraft, while at times even walking along the wings and spine. Throughout this process, paper records are maintained to document inspection findings, often without standard processes and dedicated equipment for capturing the current state of aircraft damage through imagery.
In an attempt to improve the safety, record management, and time required of this process, we developed an approach to the inspection process using autonomous small unmanned aerial systems (SUAS) to capture the required inspection imagery. This approach also implements the use of a computer vision model to process the inspection imagery, aiding the mechanic in the review of imagery and identification of inspection findings. During this process, we analyzed the effects of computer vision and machine bias on the human inspectors and inspection accuracy, recommending processes to mitigate these effects and maintain inspection accuracy equivalent to the current human-only process. |
| first_indexed | 2024-09-23T13:16:04Z |
| format | Thesis |
| id | mit-1721.1/151481 |
| institution | Massachusetts Institute of Technology |
| last_indexed | 2024-09-23T13:16:04Z |
| publishDate | 2023 |
| publisher | Massachusetts Institute of Technology |
| record_format | dspace |
| spelling | mit-1721.1/1514812023-08-01T03:38:50Z Autonomous Drone Assisted Aircraft Inspections Mighty, Andrew Barnett, Arnold Daniel, Luca Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Sloan School of Management The safety of the passengers, crew, and mechanics is of the utmost importance for any aircraft manufacturer or operator. Visual inspections of the exterior of aircraft are critical to their safe operation, as defects such as corrosion, dents, lightning strikes, or missing parts can compromise the structural integrity of the whole aircraft. Currently, aircraft visual inspections are conducted by human mechanics in a process that is not only time consuming, but also puts the mechanics and the aircraft at risk, as mechanics must use lifts and cranes to inspect top portions of the aircraft, while at times even walking along the wings and spine. Throughout this process, paper records are maintained to document inspection findings, often without standard processes and dedicated equipment for capturing the current state of aircraft damage through imagery. In an attempt to improve the safety, record management, and time required of this process, we developed an approach to the inspection process using autonomous small unmanned aerial systems (SUAS) to capture the required inspection imagery. This approach also implements the use of a computer vision model to process the inspection imagery, aiding the mechanic in the review of imagery and identification of inspection findings. During this process, we analyzed the effects of computer vision and machine bias on the human inspectors and inspection accuracy, recommending processes to mitigate these effects and maintain inspection accuracy equivalent to the current human-only process. S.M. M.B.A. 2023-07-31T19:43:08Z 2023-07-31T19:43:08Z 2023-06 2023-07-14T19:59:51.913Z Thesis https://hdl.handle.net/1721.1/151481 In Copyright - Educational Use Permitted Copyright retained by author(s) https://rightsstatements.org/page/InC-EDU/1.0/ application/pdf Massachusetts Institute of Technology |
| spellingShingle | Mighty, Andrew Autonomous Drone Assisted Aircraft Inspections |
| title | Autonomous Drone Assisted Aircraft Inspections |
| title_full | Autonomous Drone Assisted Aircraft Inspections |
| title_fullStr | Autonomous Drone Assisted Aircraft Inspections |
| title_full_unstemmed | Autonomous Drone Assisted Aircraft Inspections |
| title_short | Autonomous Drone Assisted Aircraft Inspections |
| title_sort | autonomous drone assisted aircraft inspections |
| url | https://hdl.handle.net/1721.1/151481 |
| work_keys_str_mv | AT mightyandrew autonomousdroneassistedaircraftinspections |