Development of an Off-board Vision-Based Control for a Micro Aerial Vehicle

Abstract

Our research aims to develop an intelligent robot vehicle with multimodal locomotion capabilities for the forest monitoring purpose. Due to the complexity of the proposed research, in this paper, we proposed a visual-based control system; capable to detect fiducial markers and pose estimation even with limited capabilities and transmission losses, in order to hover, navigate and fly to the desired target location while remaining stable without adversely affecting the effective flight time due to additional sensors and computation. Based on the experimental results, the MAV was able to detect the fiducial markers with a success ratio of about 92.8% as well as to land on the desired location within a radius of about 10 cm from the center-point of the landing pad, with a reduction of the effective flight time of about 28%.