Path following of wheeled mobile robots using online-optimization-based guidance vector field

This article studies a path-following problem for a wheeled mobile robot with nonholonomic constraints. The path-following task is represented by a guidance vector field (GVF), for which an online optimization procedure is adopted to estimate the path error. By exploiting a matrix-measure-based contraction principle, the convergence property of the designed GVF with respect to the task path is theoretically guaranteed. Then, a nonlinear controller is developed to track the defined GVF such that the target path is followed by the controlled mobile robot in the presence of unknown disturbances, including the unmodeled dynamics and the surface friction. Robustness properties of the closed-loop system are analyzed, and it is shown that the path error eventually converges to a residual set, which can be reduced by increasing control gains. Experiments are provided to validate the effectiveness of the desired GVF and the proposed control design.