EPSAC Predictive Control Applied to Path Tracking of Wheeled Mobile Robots

Abstract

This paper presents the application of the Extended Prediction Self-Adaptive Control (EPSAC) to the path tracking problem of wheeled mobile robots (WMRs). The main complexity of the control problem is that a WMR is a MIMO (multiple-input and multiple-output) system, being also nonsquare, nonlinear, and with input constraints. EPSAC, which belongs to the class of Model-based Predictive Controllers (MPC), is an interesting option to overcome such difficulties because its main properties are: (i) the control performance can be improved when the future reference trajectory is previously known as in the case of WMR, (ii) it has the ability to deal with constraints during the calculation of the control law in a fairly straightforward way, and (iii) with proper tuning in nonlinear models, a sub-optimal solution closer to nonlinear MPC can be obtained if compared to the conventional linear MPC. Simulations and experiments on a real robot have shown improved performance regarding trajectory tracking associated with reduced computational cost if compared to other MPCs applied to WMR proposed in literature.