Modeling, Control and Implementation of Adaptive Reconfigurable ROtary Wings (ARROWs)

Inspired by the self-rotating motion of the descending maple seeds, we introduce a novel modular aerial robotic platform-ARROWs. With customized wing and control modules, ARROWs can be easily reassembled into different configurations. Unlike conventional multirotor aerial vehicles which rely on the direct thrust from propellers, ARROWs can generate more lift by their revolving wings. However, the complex dynamics causes difficulties in flight controller development. In this work, we first analyze the flight dynamics by considering a combination of effects from the propeller, aerodynamic force on the wing, as well as self-rotating motion. As a result, a cascaded flight controller with a unified framework is designed based on reduced flight dynamics and relaxed hovering conditions to achieve stable flights in all proposed configurations. In addition, a set of inertial measurement units is employed for each flight module to estimate the flight configuration to overcome the dynamic uncertainties caused by manual re-construction. Finally, our proposed platform and flight control strategy are validated using several flight experiments in 12 different configurations (include both centrosymmetric and centrally asymmetric cases). The results show an average position error of 8.9 cm with a deviation of 2.4 cm among all configurations in hovering tests.