ROLAND: Robust Landing of UAV on Moving Platform using Object Detection and UWB based Extended Kalman Filter

Abstract

Unmanned Aerial Vehicles (UAVs) have been intensively used in various fields thanks to their excellent maneuverability. However, the short operation time of the UAV is limiting the further utilization of the UAV. The limitation of the UAV can be overcome by landing on ground vehicles and charging the battery on them. Therefore, diverse researches have been done on robust landing on ground platforms. Unfortunately, the existing autonomous landing methods have such limitations that a) special tag or marker should be attached on the landing site; b) visibility of the landing site must be secured; c) platform should be static; In this paper, to robustly estimate the relative pose of the moving platform and successfully land on it, we propose a novel robust landing system of the UAV. A neural network based object detection is used to recognize the landing site without a special tag or marker, and an Ultra-wideband (UWB) sensor is adopted to compensate for the limited field of view of the camera. The Extended Kalman Filter (EKF) for estimating the relative position of the moving platform by fusing the information obtained from various sensors is developed. Additionally, a robust autonomous landing controller is also designed. The performance of the proposed method is verified by the simulation of the UAV and the moving platform.