Classifying Aerial Objects from Trajectory Data
Washington State University
Master of Science (MS), Washington State University
2023
:
https://doi.org/10.7273/000005186
The recent availability of consumer-grade drones has dramatically increased the number of unmanned aerial systems piloted in the United States. Unfortunately, this has resulted in operators using drones with malicious intent, including smuggling contraband into federal prisons. Because of this, there have been wide-spread efforts from researchers to develop technologies which can detect and classify aerial objects, including drones. A key challenge of aerial object classification is differentiating between birds and drones, which is known as the bird-drone problem. Birds and drones are difficult to distinguish because of their similar size and velocities. Previous researchers have used a combination of image-based machine learning, radar cross sections, and acoustic methods to solve the bird-drone problem, with varying degrees of success. An alternative, less researched methodology considers classifying aerial objects from trajectory data, which exploits the fundamental differences between the flight patterns in birds and drones. This thesis is a collection of works which develop technology aiming to classify aerial objects from trajectory data.
82
21
- Classifying Aerial Objects from Trajectory Data
- Logan Dihel
- Sandip Roy (Advisor)Ali Saberi (Committee Member)Benjamin Belzer (Committee Member)Chester Dolph (Committee Member)
- Washington State University
- Electrical Engineering and Computer Science, School of
- Master of Science (MS), Washington State University
- Washington State University
- 95
- 99901019638201842
- English
- Thesis