In this paper, we have developed a novel visual servo-based model predictive control method to steer a wheeled mobile robot (WMR) moving in a polar coordinate toward a desired target. The proposed control scheme has been realized at both kinematics and dynamics levels. The kinematics predictive steering controller generates command of desired velocities that are achieved by employing a low-level motion controller, while the dynamics predictive controller directly generates torques used to steer the WMR to the target. In the presence of both kinematics and dynamics constraints, the control design is carried out using quadratic programming (QP) for optimal performance. The neurodynamic optimization technique, particularly the primal-dual neural network, is employed to solve the QP problems. Theoretical analysis has been first performed to show that the desired velocities can be achieved with the guaranteed stability, as well as with the global convergence to the optimal solutions of formulated convex programming problems. Experiments have then been carried out to validate the effectiveness of the proposed control scheme and illustrate its advantage over the conventional methods.
IndexTerms-Model predictive control (MPC), neurodynamics, nonholonomic mobile robots (NMRs), quadratic programming (QP), visual servo steering. since 2014. His current research interests include robotics, control, and human-robot interaction.Chun-Yi Su (SM'99) received the Ph.D. degree in control engineering from the His current research interests include mobile robot, model predictive control, neural network control, and optimization.Weidong Zhang received the B.S., M.S., and Ph.D.