AN EXPERIMENTAL TESTBED FOR ROBUSTNESS ANALYSIS OF AUTONOMOUS FIXED-WING UNCREWED AERIAL VEHICLES

Uncrewed Aerial Vehicles (UAVs) have had a significant impact through their applications in the ever-changing world. Especially for autonomous fixed-wing UAVs, their advantages in terms of range and efficiency allow fixed-wing UAVs to be an effective tool for transportation schemes. As particular interest in these autonomous fixed-wing UAVs grows, ensuring the safety and the robustness of their autopilot control systems becomes critical. The difficulty of performing safety verification and robustness analysis through experimental testing in real-world operations involves exposure to multiple dynamic and unknown factors.

Implementing preliminary control algorithm development, however, may face safety concerns at full scale on real, outdoor fixed-wing UAVs, which may raise other complications on the system from environmental factors in the early testing stage. Major full-scale research verification and validation methods may face multiple cost, safety, environmental, or regulatory constraints. Therefore, a controlled, indoor experimental testbed is essential for verifying and validating these systems in a safe and repeatable environment before real-world deployment.

This thesis focuses on the development of an indoor, scaled, and controlled fixed-wing UAV experimental testbed that supports the experimental verification and validation of multiple research and development tasks on autonomous control designs. The testbed explores a scaled fixed-wing Uncrewed Aerial Vehicle (UAV) designed for indoor flights at the Purdue UAS Research and Test facility (PURT), including its control architecture and capabilities within a Mixed-Reality (MR) framework. This also supports the monitoring and controlling capabilities of the fixed-wing UAV from a remote ground control station. Moreover, a Software-in-The-Loop (SiTL) digital twin simulation is provided to enable preliminary communication protocol testing and tuning of algorithms prior to hardware-based deployment within the MR framework.

This work supports the robustness analysis of fixed-wing UAVs. The testbed enables the collection of reliable ground truth data and facilitates preliminary safety verification and validation of fixed-wing control systems. To demonstrate the capabilities of this proposed testbed, autonomous control systems for taxiing, flight, and landing, including flare and land abort phase, are discussed and deployed on a fixed-wing UAV within the developed testbed. The autonomous flight utilizes the Total Energy Control System (TECS), a commonly used autopilot control for fixed-wing vehicles that manages the total energy of the system and allocates it between potential and kinetic energy.

This thesis also provides a preliminary application of such robustness analysis by demonstrating Bounded-Input Bounded-Output (BIBO) stability for real-world flights using the fixed-wing UAV model supported by this developed testbed. A disturbed system model is utilized, and bounded output states are obtained based on the bounded external disturbance magnitude. The output bounded flow pipe is generated and compared to real flight data.