Abstract

Unmanned aircraft systems (UAS) have become integral to various modern applications, including package delivery, security, medicine, and recreation. The autonomous operation of UAS significantly enhances their utility by improving performance and reducing operator workload, enabling collaborative missions such as boundary detection, border protection, and target tracking. To ensure safe integration of UAS, more work needs to be done in the areas of control, estimation, collision avoidance, path planning, hardware implementation, and autonomous practices in general. For this dissertation we chose to focus on the critical areas of control, collision avoidance, and remote autonomy. Our research is motivated by the need for robust systems to manage UAS in dense urban traffic, where reliable GPS information may be unavailable. Ground radar systems, as demonstrated in previous studies, can effectively track multiple aircraft and broadcast this information to enhance situational awareness. Building on this use case, we address some key challenges of how both participating and non-participating aircraft can utilize such information for safe and coordinated operations. In the area of control, we develop and analyze forward-propagating Riccati equations (FPRE) and their variations to ensure the stable control of highly nonlinear aircraft dynamics. For collision avoidance, we introduce the uncertainty-aware velocity obstacle (UVO) algorithm, which accounts for uncertainties in radar tracking to enable safe autonomous navigation. Finally, in remote autonomy, we propose a novel approach using monocular imagery and deep learning to remotely sense and control small UAS in our unreliable GPS and radar use-case. Through a combination of theoretical analysis, numerical simulations, and hardware experiments, this dissertation provides innovative solutions to facilitate the safe and efficient integration of UAS into urban airspaces. The contributions include new methods for stability in LTV systems, robust collision avoidance algorithms, and advanced remote control techniques, all aimed at enhancing the capabilities and safety of unmanned aircraft operations.

Degree

PhD

College and Department

Ira A. Fulton College of Engineering; Mechanical Engineering

Rights

https://lib.byu.edu/about/copyright/

Date Submitted

2024-11-26

Document Type

Dissertation

Keywords

linear time-varying control, Riccati equations, collision avoidance, estimation, deep learning, autonomy

Language

english

Download

Included in

Engineering Commons

Share

COinS