Nonlinear Model Predictive Control for Pose Regulation Robot and Obstacle Avoidance

The present work addressed the problem of autonomous navigation of wheeled mobile robots, specifically the Differential Driving Robot (DDR). The DDR kinematic model is nonlinear, which requires adequate automatic control strategies for good performance in autonomous navigation. The variables of the DDR mathematical model are the robot’s position expressed in a global Cartesian reference frame, its orientation, its linear speed, and its angular speed. It was proposed that the navigation problem of a DDR be solved by following a reference trajectory using model-based nonlinear predictive control. In addition, a potential field algorithm was added for obstacle avoidance. Experiments were carried out in a dynamic simulator. Several simulations provided convincing evidence of the feasibility of implementing a real robot using the proposed approach.