Esterçamento automático e controle de velocidade de veículos agrícolas com simulações em tempo real

Abstract

In this work, an automatic steering and velocity control is designed for an agricultural tractor to improve agricultural efficiency due to the following reasons: to avoid missing or overlap between two adjacent tracks and to keep a constant velocity during operation. The vehicle is firstly modeled in Vortex Studio, which is a physics-based simulation software that uses multibody dynamics combined with several models, such as tire models, to compute the physical quantities. The Bekker soft ground tire model, which capture the non-linear friction behaviour and rolling resistance resulting from wheels sinking into the ground, is used to compute the tire forces. Then, a simplified mathematical model, based on the rigid body equations of motion, is developed. For the longitudinal dynamics the mathematical model is build based on system identification. Therefore, the mathematical model is compared with the Vortex Studio model and the compatibility is verified. Later, the simplified models are linearized and an analysis of the open-loop stability is performed. Based on this, the control requirements of performance are specified and a phase lead compensator with an integrator is designed for the automatic steering system. For the automatic velocity control a Proportional-Integral-Derivative (PID) compensator is used. The control system design is done based on a modern control technique, which determine the optimal feedback gains based on a linear quadratic formulation of the tracker problem. In the sequence, the performance requirements are verified based on the step response and the closed-loop stability is analyzed in terms of the disk margins for both control systems. Also, the Lookahead-Based Line of Sight Steering algorithm is implemented to generate the appropriate reference signal to the steering control system. Finally, Software-In-the-Loop (SIL) simulations are performed to verify the stability margins and to analyze the control system performance in the presence of noise, model uncertainties and terrain variations.