Autonomous Tracked Agricultural UGV Configuration and Navigation Experimental Results

Nowadays agricultural techniques are increasingly focusing onto self-driving machines and smart robotic tools designed to support the farmer. This paper frames within this context, describing hardware and software configuration of the developed experimental robotic platform and reporting the results of practical tests about autonomous navigation within orchards. The robotic system consists in a completely-electrical tracked vehicle equipped with GPS receiver, 3D laser scanner and an inertial measurement unit, which provide to the robot the capability of navigating and localizing itself in the agricultural environment. To start the experiment the vehicle is placed before the beginning of an orchard row, then it has to detect the latter, navigate through it keeping a fixed lateral distance from the trees and stop the navigation once it has exited the row. In order to achieve this goal, algorithms for the detection of beginning, end and direction of rows are implemented, exploiting the semi-structured nature of orchards. These tools are seen as fundamental capabilities for smart navigation in agriculture.