Coursey, Austin, Quinones-Grueiro, Marcos, & Biswas, Gautam. (2024). An experimental framework for evaluating the safety and robustness of UAV controllers. In Proceedings of the 2024 AIAA Aviation Forum and ASCEND, Las Vegas, NV, July 29 – August 2, 2024. https://doi.org/10.2514/6.2024-4548
This study focuses on improving the safety, robustness, and efficiency of unmanned aerial vehicles (UAVs) by developing a standardized framework to evaluate advanced control algorithms. While many algorithms have been proposed to enhance UAV performance, the lack of a consistent way to test their effectiveness has made it difficult to assess their capabilities. To address this gap, the authors created an experimental framework that evaluates UAV controllers based on customizable metrics, such as flight safety and robustness, across different conditions.
The framework includes four stages, allowing researchers to adjust flight conditions, trajectories, simulations, and evaluation metrics based on specific needs. The authors tested this framework using an octocopter UAV controlled by a PID algorithm, a common control method in UAVs. Through their experiments, they found that external factors like wind speed and motor faults significantly affected the UAV’s safety and mission success. The PID control scheme, under these challenging conditions, was unable to meet safety requirements.
To improve performance, the researchers developed a waypoint modification scheme, which reduced tracking errors and increased the likelihood of successfully completing missions. This work provides a valuable tool for systematically evaluating UAV control algorithms and demonstrates how safety and robustness can be enhanced by adjusting flight parameters and control strategies. The proposed framework could lead to safer and more efficient UAV operations in real-world scenarios.