Failure Detection and Recovery for Quadrotors in the Presence of Severe Rotor Failures With Multiple Model $\mathcal {L}_{1}$ Adaptive Controller
Chengyu Yang, Jiahao Yu, Sheng Cheng, Ziyin Han, Junjie Gao, Bihao Mo, Naira Hovakimyan
Abstract
Over the last few decades, quadrotors proved to be a viable platform for improving efficiency and achieving cost savings in a variety of industries, and yet the safety of flight remains a major challenge to be guaranteed, especially in the presence of rotor failures. In this paper, we revisit the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\mathcal {L}_{1}$</tex-math></inline-formula>Quad architecture and test its robustness for severe/complete rotor failure. To achieve fast stabilization we develop multiple model <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\mathcal {L}_{1}$</tex-math></inline-formula> adaptive controller, where different models are suited for different failures, accompanied with a failure detection scheme to quickly evaluate the failure mode and guide the selection of the suitable controller. The method has been validated in experiments, demonstrating fast detection within 0.045 seconds after the failure occurs. The detection robustness under disturbances and uncertainties is also demonstrated with 100% success rate when the quadrotor flies different trajectories in wind or with an added weight.
BibTeX
@inproceedings{ral2026_failuredetection,
title = {Failure Detection and Recovery for Quadrotors in the Presence of Severe Rotor Failures With Multiple Model $\mathcal {L}_{1}$ Adaptive Controller},
author = {Chengyu Yang and Jiahao Yu and Sheng Cheng and Ziyin Han and Junjie Gao and Bihao Mo and Naira Hovakimyan},
booktitle = {RA-L 2026},
year = {2026}
}