VSDRL: A robust and accurate unmanned aerial vehicle autonomous landing scheme

Abstract

Visually assisted unmanned aerial vehicle (UAV) autonomous landing has drawn a lot of interest as a vital technology with the quick development of UAV systems. From the perspective of robustness, a visual servoing disturbance rejection landing (VSDRL) scheme based on fractional-order linear active disturbance rejection control is novelly proposed to achieve the disturbance-resistant landing. The proposed VSDRL scheme is constructed by two modules: (i) A visual positioning algorithm combining YOLOv5 with Kalman filtering to solve the occlusion problem in the visual positioning module obtaining the relative position relationship; (ii) On the basis of linear active disturbance rejection control, fractional order is introduced to improve the antidisturbance ability and response speed. Theoretical analysis, computer simulations and real UAV experiments all verify the effectiveness and superiority of the proposed VSDRL scheme.