Trajectory tracking control of agricultural vehicles based on disturbance test

Abstract

To improve the trajectory tracking robust stability of agricultural vehicles, a path tracking control method combined with the characteristics of agricultural vehicles and nonlinear model predictive control was presented. Through the proposed method, the path tracking problem can be divided into two problems with speed and steering angle constraints: the trajectory planning problem, and the trajectory tracking optimization problem. Firstly, the nonlinear kinematics model of the agricultural vehicle was discretized, then the derived model was inferred and regarded as the prediction function plant for the designed controller. Second, the objective function characterizing the tracking performance was put forward based on system variables and control inputs. Therefore, the objective function optimization problem, based on the proposed prediction equation plant, can be regarded as the nonlinear constrained optimization problem. What’s more, to enhance the robust stability of the system, a real-time feedback and rolling adjustment strategy was adopted to achieve optimal control. To validate the theoretical analysis before, the Matlab simulation was performed to investigate the path tracking performance. The simulation results show that the controller can realize effective trajectory tracking and possesses good robust stability. Meanwhile, the corresponding experiments were conducted. When the test vehicle tracked the reference track with a speed of 3 m/s, the maximum lateral deviation was 13.36 cm, and the maximum longitudinal deviation was 34.61 cm. When the added horizontal deviation disturbance Yr was less than 1.5 m, the controller could adjust the vehicle quickly to make the test car return to the reference track and continue to drive. Finally, to better highlight the controller proposed in this paper, a comparison experiment with a linear model predictive controller was performed. Compared to the conventional linear model predictive controller, the horizontal off-track distance reduced by 36.8% and the longitudinal deviation reduced by 32.98% when performing circular path tracking at a speed of 3 m/s.