Adaptive Output-Feedback Image-Based Visual Servoing for Quadrotor Unmanned Aerial Vehicles

Abstract

This brief presents an adaptive output feedback image-based visual servoing (IBVS) law for a quadrotor unmanned aerial vehicle. The control objective is to regulate the relative 3-D position and yaw of the vehicle to a planar horizontal visual target consisting of multiple points. The control is implemented using a minimal number of commonly available low-cost on-board sensors including a strapdown inertial measurement unit and a monocular camera. The IBVS method relies on moment image features which are defined using a virtual camera. Output feedback introduces a filter to the control which removes the common requirement for linear velocity measurement. The method is adaptive and compensates for a constant force disturbance appearing the translational dynamics and parameter uncertainty in thrust constant, desired feature depth, and mass. Exponential stability of the outer loop and combined inner-outer closed-loop error dynamics is proven. Flight tests demonstrate the proposed method's motion control performance and its ability to compensate parametric uncertainty and reject constant force disturbances.