Development of a New Robust Controller With Velocity Estimator for Docked Mobile Robots: Theory and Experiments

The tracking control problem of docked mobile robot systems is challenging due to their nonlinear and underactuated system dynamics as well as limited access to the required states of robots. The majority of the previously developed controllers in the literature are not robust to model uncertainties and are based on the assumption that full states are accessible. In this paper, we develop a new robust tracking controller for a docked nonholonomic mobile robotic system with online velocity estimation. Our proposed controller, composed of sliding mode and robust saturation controllers, is developed to be robust to external disturbances, unmodeled dynamics, and parameter uncertainties. To provide the required states for the controller, a model-aided particle filter estimator is developed to estimate the translational and rotational velocities. We performed extensive experiments to verify the effectiveness of our proposed control and estimation methodologies as well as the integrated system. We also compared our results with some conventional controllers, including the well-known robust sliding mode controller, and demonstrated its superior performance in terms of model uncertainties over all the controllers. The results showed that, compared with sliding mode control, our approach improves the steady-state tracking performance up to 8.3% and 11% for unmodeled dynamics and parametric uncertainties, respectively. Our proposed integrated (controller–estimator) method can be used in uncertain systems with good tracking performance, where accessing velocity directly is not possible.