Online Trajectory Correction and Tracking for Facade Inspection Using Autonomous UAV

Facade inspection is an emerging application of Unmanned Aerial Vehicle (UAV) due to its high efficiency and low risk. In this paper we propose an online trajectory correction and tracking algorithm that applies to planar facade inspection using autonomous UAV. A reference trajectory is firstly computed from a set of predefined waypoints based on a less accurate model of the target facade. During flight, an EKF-based estimator iteratively updates plane model estimation using real-time LiDAR sensor measurements. An updated trajectory is then computed which maintains a desired distance to the updated facade model. A Model Predictive Control (MPC) framework is applied to compute the control input for UAV at each sampling time. Theoretical analysis is carried out to prove the stability of trajectory tracking under changing tracking outputs. Simulations taking into account real-life challenges are conducted to validate the efficacy of the proposed algorithm.