Detection and Estimation of Moving Obstacles for a UAV

Abstract

In recent years, research interest in Unmanned Aerial Vehicles (UAVs) has been grown rapidly because of their potential use for a wide range of applications. In this paper, we proposed a vision-based detection and position/velocity estimation of moving obstacle for a UAV. The knowledge of a moving obstacles state, i.e., position, velocity, is essential to achieve better performance for an intelligent UAV system specially in autonomous navigation and landing tasks. The novelties are: (1) the design and implementation of a localization method using sensor fusion methodology which fuses Inertial Measurement Unit (IMU) signals and Pozyx signals; (2) The development of detection and estimation of moving obstacles method based on on-board vision system. Experimental results validate the effectiveness of the proposed approach.