DRL-Based Adaptive Control for Tethered UAV Under External Disturbances

As a unique UAV configuration, the tethered UAV typically operates in hovering states, where external disturbances significantly affect control performance. However, conventional control methods often lack sufficient robustness against such disturbances, resulting in degraded performance. In this paper, an intelligent control framework based on deep reinforcement learning (DRL) is proposed. A mathematical model of a tethered UAV is presented based on a discretized tether cable model and the Euler-Newton formulation. Focusing on the position subsystem of tethered UAV, an adaptive PD controller is implemented, in which the agent adjusts the PD parameters. Furthermore, to mitigate the influence of unmodeled dynamics and external disturbances, an additional compensation loop is introduced into the position control subsystem. Simulations results validate the effectiveness of the proposed intelligent control framework, demonstrating superior dynamic performance and robustness under wind and cable disturbances.