Elektor Team

Self-learning driverless tractor

December 10, 2013 | 14:37
Self-learning driverless tractor
Self-learning driverless tractor

The department of Mechatronics, Biostatistics and Sensors (MeBioS) of Leuven University, Belgium, together with Flanders' Mechatronics Technology Centre (FTMC) successfully converted an existing mini tractor into an autonomous self-learning field robot.


The tractor learns to identify soil characteristics, and on the basis of results controls its speed and steering angle allowing it to follow a certain route very accurately, all without a driver. Precision in the control of agricultural machinery is an important feature for organic farming.


The self-learning tractor is equipped with linear drives for the accelerator pedal and steering control. In addition, the tractor carries a computer and various additional sensors including a GPS system. The autonomous control system acting as the driver is capable of examining and learning soil conditions, and makes an estimate of the expected wheel slip. Based on a model of the tractor, the optimal speed and steering angle get calculated for the soil to be worked. A GPS measures the actual position which is then compared to the expected position. On this basis, the system adjusts itself.


A prototype of the self-learning tractor will be shown at the 30th International Farming Tools Day, held in Oudenaarde (Bedlgium) on September 24 and 25.


Photo: Foto: LeCoPro

Loading comments...
related items