This activity, carried out in collaboration with the University of Padua, focused on the methodological design and experimental validation of a system aimed at analyzing and improving the efficiency of agricultural machinery using only GNSS (Global Navigation Satellite System) data. Specifically, the activity involved:

• the design and implementation of a model for classifying the operational states of tractors (working, turning, transport, idle) using machine learning algorithms (Random Forest) and spatial analysis techniques;
• the validation of the method on data collected at the experimental farm of the University of Padua, with particular attention to the system’s robustness under real field conditions;
• the integration of automatic field-boundary detection tools (DBSCAN and α-shape) and spatial-temporal post-processing steps to enhance the consistency and accuracy of classifications;
• the evaluation of operational efficiency indicators (Field and Time Efficiency), with the goal of providing metrics that can be easily applied even in agricultural contexts with limited resources.

The developed approach makes it possible to estimate the efficiency of agricultural operations using minimal data, accessible through simple GNSS receivers or mobile devices, without the need for proprietary sensors or telemetry platforms.