Model Predictive Controller Design for Precision Agricultural Robot

Abstract

The basis of human existence, agriculture is essential to the continued existence of humanity. Accurate and efficient agricultural control technology ensure the welfare of people in the world. There are still difficult and urgent agricultural issues, and traditional methods of controlling agricultural output are labor and time-intensive. Agricultural systems can be challenging to control using traditional control technology due to their complexity, variability, and unpredictability. Model predictive control (MPC) techniques enable rolling optimization in a constrained time domain, which increases precision, and can provide very accurate control actions with moderate complexity. The differential drive robot trajectory tracking to the minimum error using a mathematical model that controls kinematics and dynamics without coordinate transformation is the subject of this paper. A single state-space linear model is derived by augmenting the dynamic and kinematic models. Constraints on manipulated and controlled variables are considered in the control scheme. To show the performance of the controller, time domain specifications and error evaluation were implemented using MATLAB software and findings are examined critically. The tuning parametrs used in model predictive controller (MPC) design were prediction horizon P=10, control horizon m=5 and a sample time of 0.1 s. The study demonstrates that, for the system under investigation, the suggested MPC controller is more consistent and powerful to changes in internal parameters as well as external disturbances. Keywords: MPC, Differential drive, Kinematics, Dynamics