Implementation of the trajectory controller based on model predictive controlfor unmanned ground vehicle

Abstract

This paper is devoted to one implementation of a trajectory controller based on model predictive control for an un-manned ground vehicle (UGV). The statement of the problem of autonomous movement along a given trajectory, the assumptions and restrictions adopted are presented. The features of the application of model predictive control for solving the problem are described. A feature of the proposed implementation of the trajectory controller is the use of a linearized chassis dynamics model to predict the position of the robot after a given period of time and control the movement without solving the optimal control problem. To determine the parameters of the dynamic model of the robotic complex, taking into account the drive level, the identification problem was solved, as a result of which two linear models with delay were obtained, describing the dynamics of the longitudinal movement and rotation of the robot. The experiment showed that the obtained models provide high compliance with experimental data, while the robot moves on a surface with a constant coefficient of friction. An algorithm for calculating the predicted position, a block diagram of the trajectory controller and its software implementation are described. The implemented trajectory controller provides autonomous movement of an UGV in an industrial-urban environment or in slightly rugged terrain.