Line of Sight Detection in Industrial Environment: A Machine Learning Approach

Line of Sight condition is usually beneficial in wireless communication links, as it commonly corresponds to better quality of service and can also simplify the reliable execution of tasks like beamforming and localization. Existing models dealing with line-of-sight detection are limited to statistical assessment, whic consists of line-of-sight probability formulas. In this work, a machine learning-based tool for point-to-point assessment of the line of sight condition is presented. The model is tailored to the industrial environment, where wireless technologies have been gaining increasing importance in the development of nextgeneration smart factories. Machine learning is leveraged to get flexibility, i.e. to evaluate the presence of line of sight not only depending on the link distance but also on some general descriptive features of the industrial scenario, like machine size and density. Results show good performance and the overall physical soundness of the tool.