CANBUS Data for Site-Specific Tractor Performance Analysis and Prediction

Abstract

Tractors’ working performances can be affected by many environmental variables such as terrain slope, ambient temperature, terrain conditions. Such aspects interact with other performance parameters which are determined by agricultural activities, like power take-offs (PTOs) utilization, torque, and drive system. Since these parameters can be used to define the type of activities carried out on the field, determine machinery efficiency, and can also be employed to define machinery impact on operating costs and business outcomes, understanding the relationship among them can be a reliable source of information for performance analyses and fleet management systems (FMS).

This paper provides a solution for tractor performance analysis based on real-time machinery data that can be collected from tractors’ CANBUS and on external data sources such GPS and Digital Elevation Models (DEM) which help to better define the operating conditions of every activity that is being carried out in a vineyard during an entire year.

The result of data analysis generated a dataset containing information that allowed, with the use of clustering and machine-learning algorithms, to understand which agricultural activities have been carried out for the entire yearly dataset. From that point, comparisons could thus be made on tractor’s field capacity, fuel consumptions, torque at different moments on the same field, or at the same external conditions on close fields. Such information can be of great help for farming businesses, allowing finer cost analyses and the definition of the most suitable working set-ups for their agricultural machinery based on site-specific environmental and soil conditions.