DeepWay: a Deep Learning Estimator for Unmanned Ground Vehicle Global Path Planning

Abstract--Agriculture 3.0 and 4.0 have gradually introduced Successively, output waypoints are processed with a I. Finally, the global path is computed through Over the past years, several research activities related to the A* search algorithm. Extensive experimentation with the precision agriculture and smart farming have been published synthetic dataset and real satellite-derived images of different [1]-[5], sign of a new industrial revolution approaching the scenarios are used to validate the proposed methodology. Agriculture 4.0 brought a new concept of of our training and testing code and data are open source and agriculture based on the introduction of robotics, artificial publicly available Section 2 in order to increase production efficiency and to cut labour covers the synthetic dataset design and generation. In this regard, self-driving agricultural machinery plays 3, the proposed methodology is analyzed with a detailed a relevant role both in production efficiency, by providing explanation of the DeepWay architecture and the waypoint a 24/7 weather-independent working production system, and refinement and path generation processes. Finally, section 4 cost-cutting, since there is not the need of a paid driver presents the experimental results and discussion followed by when performing the required task anymore.