Fixed-Time Differentiator-based Adaptive Nonsingular Fast Terminal Image-based Visual Servoing for a Quadrotor UAV Subject to Turbulent Wind

This paper presents the design of an adaptive nonsingular fast terminal image-based visual servoing. The aim is to drive the position and heading of a quadrotor unmanned aerial vehicle to track a dynamic target without knowledge of target measurements. Additionally, the entire system is subject to turbulent wind disturbances of moderate and light intensity described by the Von Karman wind model. A tracking differentiator has been included to provide estimates of the image features-based state vector and its first derivative in a fixed time regardless of the initial estimation error. As a consequence, the design of the visual servoing is simplified since the control law of the proposed visual servoing is independent of velocity measurements from the rotorcraft and the target. Furthermore, a nonsingular fast terminal adaptive sliding mode control has been designed to provide robustness against bounded external disturbances and model uncertainties, practical finite-time state convergence, and chattering attenuation of the control input. A Lyapunov stability analysis is given to guarantee the total stability of the system. Finally, simulations in ROS/Gazebo framework show the advantages and feasibility of the proposed visual controller at a low-altitude operation.