Modeling Asymmetric Blade Damage in Quadrotors Through System Identification Techniques

As quadrotors continue to become more popular for personal and commercial use, improving their safety is essential, especially in impaired operating states. With (asymmetric) blade damage(ABD) being a potentially dangerous type of impairment, it is beneficial to understand how it affects the dynamic behavior of a quadrotor. This research examines the effects of blade damage on the dynamic model of a quadrotor through system identification techniques. Time scaleseparation is used to split the low-frequency aerodynamic behavior and high-frequency (HF) dynamics. Aerodynamic models are identified using stepwise regression, and a novel approach for modeling HF dynamics –relying purely on on board sensors– using spectral analysis and simplex B-splines has been developed. A majority of the aerodynamic models surpass R ² values of 0.95, and the HF models exceed R ² values of 0.90. The findings provide new insights and implications for diagnosing ABD in quadrotors.