Localization and Mapping in Semi-Urban Environments with Unmanned Vehicles for Agriculture-Oriented Labours

The overcrowding access to unmanned aerial vehicles (UAVs) by companies and farmers has motivated their use in tasks such as monitoring and image capture. In order to increase the safety and autonomy of these vehicles inside semi-urban environments during monitoring or data collection tasks, this paper aims to address a strategy of Simultaneous Location and Mapping (SLAM) within the robotic operating system (ROS) for a terrestrial and aerial robotic platform to navigate in a three-dimensional space, with the ability to auto-locate and make real time decisions avoiding collisions with static or dynamic obstacles. Thus giving farmers a capable, modular and low-cost alternative for obstacle avoidance during the task in an a-priori unknown area. As a result, a SLAM algorithm is integrated into a ground wheeled robot of Mobile Robots, Pioneer-3DX, and within an aerial multirotor robot, DJI Matrice Pro 600, obtaining high functionality during repeated field tests, capturing environmental maps and recording possible obstacles to be avoided during execution of an autonomous trajectory.