Neuro-Fuzzy Modelling and Stable PD Controller for Angular Position in Steering Systems

The precision agriculture and soil tillage are tasks which can be achieved by Automated Tractors (AT) through the integration of several servomechanisms. In order to reach a high autonomy under various work conditions of the AT, control laws’ design and tuning are paramount. The aim of this work is to develop an angle position controller for steering systems applied to an AT. We introduce an automatic electronic steering system by using a fuzzy model, obtained through an Adaptive Neuro-Fuzzy Inference System (ANFIS) algorithm training. A proportional derivative (PD) controller is also tuned through a Takagi-Sugeno fuzzy model (T-S). Furthermore, several closed-loop stability tests were carried out, in numerical simulation and real-time implementation. The feasibility of our methodology is illustrated through the tracking of several angles’ profiles in a real test scenario. For the tracking of a trajectory or several set-points, we have obtained mean errors about 0.6545 and 0.8651 degrees, respectively. Some mechatronic integrations to convert a conventional tractor into a basic Autonomous Agriculture Off-road Tractor (AAOT) are also shown.