An experiment study for unmanned aerial manipulator systems with L1 adaptive augmentation of geometric control

Abstract

Unmanned aerial manipulators, integrating multi-rotor unmanned aerial vehicles (UAVs) and multi-link manipulators, are extending the capabilities of UAVs to interact with the environment and demonstrating vast potential for diverse applications. However, the dynamic coupling between the UAV and the manipulator, along with the unmodeled nonlinear disturbances exerting on the aerial platform, pose challenges for high-performance control of the aerial manipulator. To this end, this paper proposes an $L_1$ adaptive augmentation control method for the aerial manipulator to compensate the disturbances brought by manipulator motion and unmodeled uncertainties. Specifically, a baseline geometric controller is constructed to directly compensate the measurable wrench disturbances exerted on the UAV by the manipulator motion, while an $L_1$ adaptive controller is designed to compensate the residual unknown wrench disturbances. By Lyapunov techniques, it is proven that the error signals of the closed-loop system are uniformly ultimately bounded, which demonstrates the theoretical effectiveness of the proposed method. Finally, the feasibility and robustness of the proposed control method are validated through three groups of hardware experiments.